DESIGN IMPROVEMENTS OF DISTRIBUTION TRANSFORMERS How to improve conditions of transportation in Vietnam
DESIGNFÖRBÄTTRINGAR AV DISTRIBUTIONSTRANSFORMATORER Hur man kan förbättra transportförhållanden i Vietnam
Lisa A. Magnusson Bachelor Degree Project in Product Design Engineering G2E, 30 ECTS Spring term 2014
Industry Supervisor: Brad Havlin, ABB Vietnam Supervisors: Erik Brolin and Tomas Walander, University of Skövde Examiner: Dan Högberg, University of Skövde 1
Certificate of Originality Submitted on December 22, 2014 by Lisa Anna Magnusson to the University of Skövde, Sweden as a Bachelor Degree Project in Integrated Product Development/Product Design Engineering at the School of Engineering Science. Hereby, I certify that all material in this report, which is not my own work has been identified and that no material is included for which a degree has previously been conferred on me.
Abstract Nowadays international companies wish to relocate their production to developing countries in Asia. The product design is often developed in Europe and is, therefore, not always adapted to the production (transport and manufacture conditions) present in the country of manufacture. The consequence is less qualitative products e.g. distribution transformers of ABB Vietnam get damaged during transport; a problem none existing in Europe. The purpose of the project was to develop the design of distribution transformers for ABB Vietnam, so they better withstand the transport conditions. This Bachelor Degree Project in Integrated Product Development follow a general design process typically for product development and is divided into four main phases. The four faces are; Exploration (of the ill-defined problem space), Generation (of concepts), Evaluation (of the design proposals), and Communication (of the final design solution). Methodologies as literature studies and empirical investigations (observations and interviews) performed at the ABB Vietnam factory in Hanoi provided information about the problem space and a brief overview of the production of ABB transformers. In the study it emerged that the underlying problem was due to both the structure of the product and the infrastructure of the country e.g., extreme road conditions, lower standard of transportation means etc. Design methodologies were implemented to systematically create ideas, generate and evaluate a number of concepts. The final design concept chosen was simulated to behaviour adequately of transport in real life to verify that the concept manage its purpose. The concept can be further improved and optimised. The outcome of the study revealed the possibilities of improvement, designers can contribute to increased quality of products by understanding the country of manufacture. The thesis is intended to enlighten people working with product development what they shall consider when designing for transportation, and may be used as a practical example of similar problems. The final design concept is an external support frame attached to the transformer to enhance support and contribute to improved quality – especially under extreme road conditions.
Sammanfattning Internationella företag vill idag flytta sin produktion till utvecklingsländer i Asien. Deras produkter vars design har utvecklats i Europa är därmed inte alltid anpassade till transport- och tillverkningsförhållanden som finns i tillverkningslandet. Konsekvensen blir mindre kvalitativa produkter. Ett exempel på detta problem är distributionstransformatorer från ABB Vietnam som skadas under vägtransporter pga. dåliga vägar, ett problem som är omärkbart i Europa. Syftet med projektet var därför att analysera och utveckla designen av distributionstransformatorer, för ABB Vietnam, så att de bättre klarar av de svåra vägtransporterna. Detta examensarbete inom Integrerad Produktutveckling följer den allmänna designprocessen typiskt för produktutveckling och består av fyra faser som inkluderar Utforskning (av det vagt definierade problemområdet), Generering (av koncept), Utvärdering (av designförslag) och Kommunicering (av det slutgiltiga designförslaget). Metoder som litteraturstudier och empiriska undersökningar (observationer och intervjuer) utfördes på ABB Vietnam, i fabriken som ligger i Hanoi. Därifrån uppdagades information om problemområdet och en översikt över produktionen av ABB transformatorer. I studien framkom att det underliggande problemet berodde på en kombination av produktens struktur och landets infrastruktur t.ex. extrema vägförhållanden, lägre standard för transportmedel, etc. För att systematiskt skapa nya idéer implementerades diverse designmetoder, för att utveckla och värdera ett antal koncept. Det slutliga designkonceptet som valdes simulerades med avseende på verkliga transportförhållanden för att verifiera att konceptet kommer att motstå dessa påkänningar. Konceptet kan förbättras och optimeras ytterligare. Resultatet av studien visade att designers kan bidra till förbättringar och öka kvaliteten på produkter genom att förstå förhållanden i tillverkningslandet. Studien syftar till att upplysa vad de som arbetar med produktutveckling ska tänka på när de designar för transport och den kan användas som ett praktiskt exempel för liknande problem. Det slutliga designkonceptet är en extra stödram fäst på transformatorn för att öka transformatorns hållbarhet och förbättra kvaliteten - särskilt under extrema vägförhållanden.
List of Abbreviations ABB ANSI ASEAN ASTM CAD DTR FEM G2/ Hz HV Hz ISO ISTA kVA LV MPa MVA N NA PDS PNI PSD PTR RQ
Asea Brown Boveri American National Standards Institute The Association of Southeast Asian Nations American Society for Testing and Materials Computer Aided Design (program) Distribution Transformers Finite Element Method Acceleration versus Frequency High Voltage Hertz International Organization for Standardization International Safe Transit Association kilo Volt-Amperes Low Voltage Mega Pascal Mega Volt-Amperes Newton Not Applicable Product Design Specification Positive, Negative, Interesting Power Spectral Density Power Transformers Research Question
Table of Contents CERTIFICATE OF ORIGINALITY ............................................................................................................... IIII ABSTRACT ........................................................................................................................................ III SAMMANFATTNING ............................................................................................................................ V LIST OF ABBREVIATIONS....................................................................................................................... V TABLE OF FIGURES AND TABLES.......................................................................................................... VIIII
BACKGROUND .................................................................................................................... 1
THE PURPOSE ..................................................................................................................... 6
DESIGN PROCESS ................................................................................................................. 9
ABB ................................................................................................................................. 2 VIETNAM ........................................................................................................................... 3 PRESENTATION OF THE PROBLEM ........................................................................................... 4 ABOUT TRANSFORMERS ....................................................................................................... 4
Exploration ............................................................................................................................. 10 Generation ............................................................................................................................. 11 Evaluation .............................................................................................................................. 12
ORGANISATIONAL STRUCTURE ............................................................................................. 13
STATISTICS OF THE LEAKING PROBLEM ................................................................................... 14
TRANSPORT CONDITIONS IN ASIA ......................................................................................... 20
STRESS ANALYSIS OF WELD ................................................................................................. 23
................................................................................................................ 25 ACTIONS TOWARD SUSTAINABLE TRANSFORMERS .................................................................... 25
HOW TO PREVENT FAILURE ................................................................................................. 29
ESTABLISH REQUIREMENT ................................................................................................... 35
THE PRODUCTION PROCESS ................................................................................................. 16
CONCEPT GENERATION ....................................................................................................... 38
FURTHER CONCEPT GENERATION ......................................................................................... 43
EVALUATION AND CONCEPT SELECTION ................................................................................. 47
DETAIL DESIGN ................................................................................................................. 53
TEST AND VERIFICATION ..................................................................................................... 56
PRODUCT DESCRIPTION ...................................................................................................... 68
DISCUSSION ..................................................................................................................... 70
PRODUCT DESIGN SPECIFICATION ......................................................................................... 69
FURTHER RECOMMENDATIONS ............................................................................................ 72 SOURCES OF ERRORS .......................................................................................................... 74
REFERENCE LIST OF FIGURES ............................................................................................................... 78
Table of Figures and Tables Figures Figure 1‐1 Medium distribution transformer, a corrugated tank with fin walls (ABB, 2014). .................. 3 Figure 1‐2 A simplified demonstration of the power grid (Lisa Magnusson, 2014). ................................. 5 Figure 1‐3 Distribution transformer with the active part on the inside visible (Power Energo, 2011). .... 6 Figure 2‐1 A simple four‐stage model of the design process by Nigel Cross (1994). ................................ 9 Figure 2‐2 The product development process by Ulrich and Eppinger (2008) ........................................ 10 Figure 2‐3 Based on the Five‐step‐concept‐generation method by (Ulrich and Eppinger, 2008). ........ 12 Figure 2‐4 Based on the compositions of product development team (…)(Ulrich and Eppinger, 2008). 13 Figure 3‐1 a) The most common leakage is located in the corrugated weld, b) (…) (© ABB). ................ 15 Figure 3‐2 An illustration of the general steps in the manufacturing process of DTR transformers. ..... 16 Figure 3‐3, a) Assembly of the steel core, b) copper winding machine, c) the yoke (© ABB) ................. 17 Figure 3‐4 a) Fin wall folding machine, b) welding a corrugated tank, c) weld along (…) (© ABB). ...... 17 Figure 3‐5 a) Wooden breaks between the transformer and the truck wall. (…). (© ABB). ................... 19 Figure 3‐6 Examples of transformers tightened with steel wires to the container floor (© ABB). ......... 19 Figure 3‐7 The textile lashing ratchet belts placed between fins and belt goes under (…) (© ABB) ....... 19 Figure 3‐8 a) A transformer with sensitive terminals (unwrapped), b) (…).(© ABB). ............................. 20 Figure 3‐9 A few example of road conditions in Thailand and India. ..................................................... 21 Figure 3‐10 Demonstration of how the hydrostatic pressure affects the weld (…) ................................ 23 Figure 3‐11 a) The load on the inside of the transformer due to hydrostatic pressure, b) (…) ............... 24 Figure 3‐12 Round steel bar (© ABB) ..................................................................................................... 27 Figure 3‐13 Steel rod squared ©ABB ...................................................................................................... 28 Figure 3‐14 Bracket by Hung (2013) (© ABB) ......................................................................................... 28 Figure 3‐15 Vertical placed stiffeners ©ABB .......................................................................................... 29 Figure 3‐16 Angled flat iron in the tanks corners (©ABB) ...................................................................... 29 Figure 3‐17 a) Double weld vs. outside weld and b) an illustration of how the inside could be (…) ....... 31 Figure 3‐18 Two examples of poor quality welds with air bubbles (porosity) (©ABB). .......................... 32 Figure 3‐19 a and b) The fin to the left is welded wrong and can affect the weld quality (…)(©ABB). .. 32 Figure 3‐20 An example of bad material quality of some of the locally supplied rods (…)(©ABB). ........ 32 Figure 3‐21 A solution strategy of how to prevent oil leak of distribution transformers of ABB ............ 33 Figure 4‐1 a) Two alternative(…)(IACS, 2005), b) (…)(Alibaba.com, 2014), c) (…) (Amazon.com, 2014).40 Figure 4‐‐2 The six concepts are based on the six idea solutions no. 2‐4, 7, 9, and no. 10 .................... 43 Figure 4‐3 Top view of the transformer with concept B inside. The beams could be (…) ....................... 44 Figure 4‐4 Concept C with the 45° degree cut‐out of the fin wall edges (Lisa Magnusson, 2014). ........ 45 Figure 4‐5 a) Beam is cut off to overlap (…), b) a U‐beam placed vertical on the inside, c) (…) ............. 45 Figure 4‐6 Details of Concept E a) the flat bar vertically placed outside fins (…), b) shorter flat bar(…).46 Figure 4‐7 Different combinations of the console attached to different flat bar solutions .................... 47 Figure 4‐8 a and b) The result of the concept ranking survey was summarised in two tables. .............. 50
Figure 4‐9 The black rectangle represents the transformer tank, and the blue part is the active (…) .... 52 Figure 5‐1 The support solution is attached to the bottom of the fin wall sections and (…) .................. 54 Figure 5‐2 The support frame seen from a top perspective, four iron flat bars are joined as a (…) ....... 54 Figure 5‐3 a and b) The console on the short side (bottom) is longer than the original console (…) ...... 55 Figure 5‐4 A simplified shorter section of the original CAD‐model with three main parts. .................... 57 Figure 5‐5 Stress distribution of the CAD‐model subjected with a) Load case 1, b) Load case 2, c) (…)..58 Figure 5‐6 The support frame was added to the simplified CAD‐model and the result of stress (…) ...... 59 Figure 5‐7 a) The left figure shows the stress result “before reinforcement”, b) figure to the right(…) . 60 Figure 5‐8 The high stress in the attachment of the console are critical. The blue colour (…). .............. 60 Figure 5‐9 The examined factors of the console were the lengths L1‐L4, number of consoles Q1 (…) ... 61 Figure 5‐10 The 16 different console shapes. (…) ................................................................................... 62 Figure 5‐11 The effect of each factor and interactions. .......................................................................... 64 Figure 5‐12 a, b, and c) represent the console shapes no. 7, no. 11 and no. 13. .................................... 65 Figure 5‐13 a) The first chamfer is a 45 ° cut‐out of the edge and in b) is the second chamfer (…) ....... 66 Figure 5‐14 a) CAD‐model of one 20 mm thick console, b) two 10 mm thick consoles and c) (…) ......... 66 Figure 6‐1 A transformer with the support solution of steel attached underneath the fin wall (…)…....68 Tables Table 3‐1 Wishes ranked in order of importance.................................................................................... 36 Table 3‐2 The PDS represents the requirements of a design solution for distribution transformers(…) . 37 Table 4‐1 Ten generated solution ideas. ................................................................................................. 41 Table 4‐2 A positive, negative and interesting aspects of each solution idea. ....................................... 42 Table 4‐3 The concept screening matrix. ................................................................................................ 49 Table 4‐4 Table of manufacturing possibilities of each concept. ............................................................ 51 Table 5‐1 Load cases with respective constrains and loads .................................................................. 57 Table 5‐2 The levels of the factors. ......................................................................................................... 62 Table 5‐3 The design matrix of the full factorial design experiment. The stress (…) .............................. 63 Table 5‐4 Extract from Table 5‐3 The design matrix of the full factorial design experiment. ................ 65
Introduction This Bachelor Degree thesis was conducted during spring of 2014 by Lisa Magnusson, a Product Design Engineer student at the University of Skövde. The thesis is a product development project performed in collaboration with ABB Vietnam, and took place in Hanoi, Vietnam. From personal experiences and according to RC PERRET (2013) the infrastructure of Vietnam such as roads, railway system, ports, electric supply and airports are not efficient. The situation in Vietnam has been central to the project and is further explained in Chapter 1.3. At the start of the project the purpose was to develop a “design for transportation” methodology, how design engineers shall design for transportation conditions present in Vietnam and Asia. Simultaneously, a case study was planned to evaluate the methodology when developing a product. However, the focus of the thesis was later to be changed because a methodology was not applicable to the product design development process of ABB Vietnam. The primary focus and the main purpose of the project then shifted to develop the design of an existing product to withstand the transportation conditions present in Vietnam (further explained in Chapter 1.6). The following chapters give an introduction to the background (Chapter 1.1) and the initial problem of the thesis (Chapter 1.4). The introduction chapter also reviews brief information about the company ABB (Chapter 1.2) and the products of ABB, transformers (Chapter 1.5).
Background Today companies in many European countries move their production to Asia. When an international high-tech company wishes to relocate their manufacture to a developing country, it often causes a positive expansion. However, the company might not take into account that 1
the way things work in Europe is not necessarily the way things work in Asia. The international companies need to gain a clearer understanding of the developing country’s culture, norms, and environment to contribute to increased success and prosperous businesses. It was believed that there is a need for deeper knowledge on how people working with product development shall adapt the design of the products to the conditions of the country of manufacture. The foundation for this belief is based on the problem typical for ABB Vietnam, where the products (with a Europe centred design used globally) are exposed to risk of damage and reduced quality after transportation. A problem non-existing in Europe because transport conditions in Europe are relatively stable compared to the extreme road conditions of Asian roads. The thesis intended to prove that the product would have improved outcomes, a higher standard and sustainability by adapting the product development to the road conditions of the country of manufacture.
ABB This chapter introduces facts about ABB and information about ABB Vietnam, the company that the thesis was aimed for. ABB is described as a global leader in power and automation technologies that enable utility and industry customers to improve performance while lowering environmental impact (ABB, 2004). The ABB group of companies operates in around 100 countries and employs about 150 000 people (ABB, 2014a). The ABB Power businesses focus on power transmission, distribution and power-plant automation and serve electric, gas and water utilities, as well as industrial, commercial, and government customers. (ABB, 2014b). ABB is also among the world’s leading supplier of transformers offering a complete range of power and distribution transformers (liquid/dry), designed to grant the reliability, durability, and efficiency required. (ABB, 2014c). in fact, ABB deliver over 1,500 power transformers and over 400,000 distribution transformers annually from its 57 production facilities worldwide (ABB, 2010). ABB Vietnam, have invested in a factory in Hanoi where they produce small and medium liquid-filled distribution transformers (Figure 1-1) and power transformers. The design of the distribution transformers produced in Vietnam are basically the same as the global design for smaller distribution transformers of ABB that is developed at the Technology Lead Centre in Poland.
Figure 1‐1 Medium distribution transformer, a corrugated tank with fin walls (ABB, 2014d).
Vietnam Vietnam is a developing country which is transitioning towards a middle-income market economy (IMF, 2011). Vietnam is a country about the same size as Norway (Embassy of Sweden, 2012) with a population of 90 million people and borders to Cambodia, China, and Laos (NationMaster.com, 2014a). Larger cities are Ho Chi Minh City (south), Hai Phong (northeast), Da Nang (southeast) and the capital Hanoi (north). The growing urban industrialization and population migration are rapidly degrading the environment in Hanoi and Ho Chi Minh City (IndexMundi.com, 2014a). Vietnam has a long history of war and was not until the second of July 1976 united with the socialist Republic of Vietnam (Wikipedia.org, 2014a). The French started to colonize Vietnam 1856 and during the Second World War Japan occupied Vietnam. In 1961, the Americans started sending troops and took an active part in the conflict between North and South Vietnam in the so-called American War (also called the Vietnam War in West). The American War continued to 1975 (Wikipedia.org, 2014a). Vietnam decided in 1986 to open up for reforms and in the middle of the nineties they became a member of the ASEAN countries. Consequently, the economic development during end of the nineties in Vietnam was one of the biggest in the world (Embassy of Sweden, 2012). Sweden was the first western country which opened up for diplomatic relations with North Vietnam in 1969 during the American War and supported Vietnam with development aid afterward. There are about 80 Swedish companies active in Vietnam today, e.g. Ericsson, ABB, Tetra Pak, Electrolux, Atlas Copco, SKF, SAAB, and IKEA. (Embassy of Sweden, 2012). The main export products of Vietnam are textile, foot ware, rice, electronic parts (mobile phone, printers, computers, etc.) and coffee and fish products (General statistics office of 3
Vietnam, 2007). A global distribution of export products takes place every day on the 206.633 km long roads and 2.632 km long railway existing in Vietnam (NationMaster.com, 2014b). Paved roads in Vietnam were 47.60 % of the total roads as of 2007 (Index Mundi.com, 2014b). The condition of the roads are not efficient, e.g. to travel by car from the capital Hanoi to the main port city Hai Phong, a distance of 100 km, takes nearly 3 hrs. The max allowable speed is 80 km/hrs, but due to overloaded trucks many roads are damaged with bumpy surfaces that partly reduce the speed less than 80km/hrs (personal experience). Road safety is one of the biggest issues in the country (RC PERRET, 2013). According to International Monetary Fund (IMF) Vietnams “development to date has been nothing but remarkable, and its potential to continue to deliver rising living standards for the Vietnamese people in the future is high” (IMF, 2011).
Presentation of the Problem Transport of transformers is not an easy task. The big issue is today related to the design of the so-called corrugated tank construction of medium distribution transformers which is sensitive to impacts such as vibrations and harmonics created during transportation on roads with severe road conditions. The problem with less qualitative products after transport is caused when applying the standard design of European ABB into Asian ABB supply. The design of the products is not adapted to the transportation conditions present in the country of manufacture, and the consequence is that the product does not withstand the transport. What happens when some transformers are exposed to transportation, is that failures of the constructions arises and the mineral oil inside the transformers leak out, which is both dangerous and harmful to the environment. The leakage is located near the weld around the sealed outlets. The leakage problem is discovered during the transportation or after the transformer arrives at the site of the customer and the situation to repair the leakage afterward is consequently hampered and can become very costly. ABB Vietnam aims to be able to guarantee that their products do not get damaged during transport and therefore actions to prevent leakage are highly prioritised. Therefore, there is a need for deeper knowledge of how to adapt the design of the distribution transformers so that the products can withstand transportation without any reductions in quality.
About Transformers Transformers exist all over the world in cities and communities supplying electric power to infrastructural buildings and machines such as the hotel lift, the underground railway, amusement parks, and every kind of factory (ABB, 2004). The main requirement of a transformer is according to the Distribution transformer handbook “(…) that it shall transfer a certain amount of electric power at a constant frequency while the voltage is being changed from one level to another with a minimum of power losses” (ABB, 2003, p. 26). 4
Transformers are electrical devices in a major world-wide network, in the combined electrical power transmissions and distribution system, known as the power grid, see an illustration of the power grid in Figure 1-2.
Figure 1‐2 A simplified demonstration of the power grid (Lisa Magnusson, 2014).
The power grid is generally divided into two parts. The first part, the Transmission system includes power transmission along transmission lines over long distances with extreme high voltage electricity from generating power plants to electrical substations. The second part, the Distribution system, distributes the electric power from so-called substations to the customer with the help of distribution transformers (Wikipedia.org, 2014b).
Classification Transformers can be classified by the power range which is expressed in kilovolt-amperes (kVA) or megavolts-amperes (MVA). The power range indicates the amount of power that can be transferred through the transformer (McDonald, Wojszczyk, Flynn, and Voloh, 2013). Transformers are divided into two categories, power transformer (PTR) and distribution transformer (DTR). Traditionally, transformers that transform the voltage down to the domestic consumer voltage level (less than 400 V) are called distribution transformer. ABB includes transformers with a power range up to 10 MVA in the category distribution transformers (ABB, 2003). In this project, the distribution transformers exposed to the problem have different power ranges between 1.5 MVA – 5 MVA.
Major Components of Transformers Distribution transformers share certain fundamental components, see Figure 1-3 (Power Energo, 2011. TERMINAL PARTS
Figure 1‐3 Distribution transformer with the active part on the inside visible (Power Energo, 2011).
The active part on the inside is where the electromagnetic induction occurs between windings and the core (CBSA, 2012). Transformers use electromagnetic induction between circuits to increase/decrease the output voltage levels being transmitted (CBSA, 2012). The active part has a housing in the shape of a tank. The corrugated tank (a tank with fin walls), typical for distribution transformers, works as a cooling system together with mineral oil as the cooling media (CBSA, 2012). Terminal parts as bushings are HV- and LV-inputs/outputs.
The Purpose As mentioned the thesis main focus and purpose shifted to develop the design of an existing product to withstand the transportation conditions present in Vietnam. It was not until very late in the starting phase of the project revealed that there was not an official design process, research or development department existing at ABB in Vietnam where the thesis took place. Which is the reason why the focus was changed. There is a mechanical design department but the current design of the distribution transformer at ABB is a global design that is designed, developed or updated by the ABB’s Technology Lead Centre located in Europe. Thereby, there was no design process to analyse or get in depth with at ABB Vietnam, but it also meant that finding information about how to conduct the methodology was considerably complicated. The possibility to perform a case study and to develop a product did still exist. Hence, even if there is a global construction design of the transformer there are some of the factories around the world with varying transformer designs because e.g. different types of suppliers are available. As an example, ABB in Vietnam has a fin wall supplier in Thailand while ABB in 6
Poland produces their own fin walls so that is why their corrugated tank design can differ a bit from the one at ABB Vietnam, even though the function is the same. This is the reason why one ABB factory can have individual problems with the construction design. The problem must then be solved locally and requires cooperation between e.g. factory workers and mechanical design engineers to solve the design problem. ABB Vietnam was exposed to this type of local design problem that became a foundation to base the case study on. It would not be a case study where a methodology was evaluated. However, it would still be possible to use this case study as an example of how a design problem at ABB Vietnam can be tackled with the help of design methodologies from a general design process. The design process of the case study and the methodologies used in this thesis are further presented in the method chapter (Chapter 2).
Research Objectives and Questions The primary research objective of this thesis was to: Understand the characteristics of the transport conditions as well as the manufacturing conditions in Vietnam and consider what possibilities there are to improve the quality of delivered ABB transformers. In addition to the primary objective following corollary research questions (RQ) were formulated: RQ1:
What are the considerations that need to be taken into account when designing transformers for transportation on Asian roads?
What are the characteristics of the transport conditions in Asia? This includes road characteristics as vibrations/harmonics that need to be considered for simulations undertaken and current packaging features.
What is the root-cause of the transport problem and is it possible to find out when the problem occurs?
What can be done or changed to prevent the leakage problem? Should the focus of design solutions be toward packaging or develop the original transformer design to stabilise the product during transport?
How should a simulation model be designed to model the behaviour adequately of transport in real life situations?
How can it be proven with physical testing that any proposals are effective?
Goals of Thesis The goals of the thesis were to deliver:
A literature review of relevant transportation information of developing countries, including the collection of specific data related to logistics and transportation in Vietnam
Results from empirical investigations of the root-causes of the transport problem, where the focus is to find insight why the problems occur and what can be improved.
Potential solution strategies, a summary of how to improve transportation conditions, specifically adapted for ABB, which can be set as general leading examples for companies with transport problems of goods produced in Asia.
A product development case study, based on the corresponding solution strategy, where an existing transformer’s design is adapted to conditions in Vietnam.
Method The product development project for this thesis can be categorised as incremental improvements to an existing product described by Ulrich and Eppinger in the book Product design and development (2008). This means that modifications or features are added to the existing product to remedy flaws to keep the product competitive. The methodologies, the activities as well as the design methods used in this product development project are presented below in following chapters.
Design Process A product development project is based on a problem or a need that is wished to be solved or satisfied, it is an opportunity to develop a product, method or service. To satisfy the mentioned need or problem, a designer or product development team can methodologically solve the problem with the help of the steps in a design process. A simple four-stage model in the book Engineering design methods by Nigel Cross (1994) describes the overall design process used in this project, see Figure 2-1.
Figure 2‐1 A simple four‐stage model of the design process by Nigel Cross (2008, p. 30).
The four-stage model illustrates how a project starts with an initial exploration of the ill-defined problem space. A proposal arises from the generation of concepts by the designer. The design proposal is subjected to evaluation against the goals and the final design solution is communicated at the end of the process and prepared for manufacture (Cross, 1994). The arrow represents the iterative phase of the process, hence the process may not be like a straight line, in fact, evaluation can lead to further generation (Cross, 1994). The product development process described by Ulrich and Eppinger (2008) is a different variation of the design process described by Cross (1994), see the product development process in Figure 2-2 (Ulrich and Eppinger, 2008). The product development process is especially adapted for the development of industrial products
Figure 2‐2 The product development process by Ulrich and Eppinger (2012, p. 9)
Phase 0, Phase 1 and Phase 3 were especially applicable to this project. It was necessary to perform Phase 4 to evaluate the outcome of the three earlier phases, but Phase 5 extends beyond the project range. The planning phase was performed at the beginning of the project with the help of a Gantt chart to keep track of the various project activities and deadlines. The detail design phase was performed straight after the concept development phase because there was no need for the system-level design phase (Phase 2). The transformer has no sub-systems to develop e.g. as a computer or machines have. The four-stage model by Cross (1994) and the activities of the design process of the project are explained further in the chapters below. As well as the activities performed in the concept development phase, phase 1, by Ulrich and Eppinger (2008).
Exploration The exploration phase of the project was all about understanding, as well as exploring the problem space, and to reach a basic understanding of the objectives earlier presented in the introduction chapter (Chapter 1.6.1). A problem-solving methodology was adapted to the exploration phase; methods such as literature studies and empirical investigations were used (Cross, 1994). The literature studies in the exploration phase reviewed the characteristics of the transport conditions in Asia and about transformers in general. The empirical investigation was based on observations and interviews (Cross, 1994). The empirical investigation was performed in Vietnam and made it possible to
collect data e.g. about oil leakage failures of transformer, and to collect information about relevant work such as existing solutions that earlier had been implemented. Direct observations in Vietnam, was necessary to gain a good insight of the production process to be able to analyse and describe the production and the packaging of ABB Vietnam in Hanoi. The observations had a “case study” approach which is a flexible approach that can continually be adapted, questions and the focus can be changed during the study (Cross, 1994). Supplementary work was performed e.g. strength calculations to study the leakage problem and a type of benchmarking of the existing solutions at ABB where their pros and cons were evaluated. The most valuable information was collected through the “open-ended” face-to-face interviews with the workers at ABB Vietnam (Cross, 1994). To get a variety of information such as opinions and knowledge about the problem, different representatives from the company with various occupational background were interviewed along the project’s process. From the exploration phase, a deeper understanding of the problem was initiated, and the customer’s needs of ABB Vietnam were developed when the root-causes of the leakage problem were identified. If the customer’s needs according to Ulrich and Eppinger (2008) are identified early in the development process then it is easier to establish the product design specification (PDS). The PDS is according to Hurst (1999) the most critical phase of the design process. A product design specification with metrics and values is a description of what a product should be able to perform. The PDS function in the early design process works as an aim, a target, of what to achieve with the product. But not until the final concept has been chosen it is possible to establish the final specification of the product (Ulrich and Eppinger, 2008). An initiation of the PDS (see Chapter 3.7.) was established at the end of the exploration phase and during the generation phase the aim of the generated ideas and concepts was to fulfil the necessary demands and wishes of the PDS.
Generation In the generation phase, a number of concepts were generated and at the end of the phase only one final proposal remained. Before the development of industrial design concepts was performed the activities suggested by Cross (1994) to identify lead users and identify competitive products were helpful. Design methodologies, typical of product development were implemented to systematically generate creative ideas that would solve the initial project problem. For the generation phase the five-step concept generation method (Ulrich and Eppinger, 2008) and various brainstorming methods were used in different group sessions to generate ideas. Examples of brainstorming methods are the gallery method described in Chapter 4.1.3 and the 3-6-5 method described in Chapter 4.1.2. The five-step concept generation method presented in Figure 2-3 is a method used to break down a complex problem into more simple sub-problems and identify concepts from internally and externally search procedures (Ulrich and Eppinger, 2008). 11
2. Search Externally 1. Clarify Problem
3. Search Internally
Consult expert Benchmarking
Figure 2‐3 Based on the Five‐step‐concept‐generation method by (Ulrich and Eppinger, 2012, p. 120).
The steps in the method are performed step by step to solve the problem gradually. Step 2 and 3 could conveniently be performed parallel. Step 2 was about finding inspiration for ideas, by searching externally for existing solutions or technologies used in products that solve similar problems. The collection of external information was done by consulting experts and benchmarking competitive solutions (Chapter 3.5.2). Step 3 was about searching for ideas internally, inside the company e.g. with group sessions together with ABB workers as well as from personal creativity and knowledge (Ulrich and Eppinger, 2008).
Evaluation The evaluation phase includes general activities such as investigate feasibilities of the product concepts, build and test experimental prototypes, estimate the manufacturing cost, and assess production feasibility (Ulrich and Eppinger, 2008). However, in this project there was no prototype due to lack of time. According to Ulrich and Eppinger (2008) the evaluation of the design concepts can be done e.g. by intuition, by considering pros and cons, by testing or with decision matrices, where the concepts are weighted against the specification. However, it is preferred if a choice is made rationally than by intuition according to Cross (1994). The evaluation methods used in the project were e.g. the PNI-method where pros and cons of the concepts were evaluated (Chapter 4.1.4.), concept screening which rates the concepts after criteria (Chapter 4.3.2.), and the manufacturing possibilities of the concepts were evaluated (Chapter 4.3.4.). After the concepts were evaluated one concept was selected for further development and detail design. To be able to evaluate if the final solution met the requirements, a digital mockup of the final concept solution was constructed in a computer aided design program (CAD) where the CAD-model was tested in a finite element method (FEM) analysis (Chapter 5.2.1). To optimize the final design solution, a one-factor-at-a-time experiment and a full factorial design experiment described by Bergman and Klefsjö (2010) were implemented with the help of the FEM-analysis (Chapter 5.2.3). The design experiments for optimisation is about testing and analysing the result of stress when changing dimensions of the solutions e.g. increase the length or the thickness and compare the best alternative shape. 12
Organisational Structure Even though there was no specific product development team at ABB Vietnam, the work was done through interaction and collaboration with workers of ABB. The role of the auhtor was similar to the role of a team leader as illustrated in Figure 2-4, based on the compositions of product development team for an electromechanical product of modest complexity by Ulrich and Eppinger (2008).
Figure 2‐4 Based on the compositions of product development team for an electromechanical product of modest complexity (Ulrich and Eppinger, 2012, p. 4). Illustration by Lisa Magnusson (2014).
At ABB Vietnam, the factory workers work close with Manufacturing Engineers and Quality Engineers inside the factory. The task of the Mechanical Design Engineers and the Electrical Design Engineers is to compose and customise distribution transformers to the performance and needs of the customer, by combining present technology and construction parts from the Technology Lead Centre of ABB. Key people to the project included Quality Engineers that were an excellent source of information about the leakage problem and statistics, and the Mechanical Engineers contributed with considerable knowledge about the transformers tank design. Other important people were people working directly with the production line such as the Production Manager and the Welders. The workers of ABB Vietnam were very supportive when it came to evaluate which ideas would endure, considered the manufacturing possibilities, etc.
Exploration As mentioned in the method chapter, the exploration phase was about exploring the problem space, thus answering the research questions defined at the beginning of the project. This meant to find out when and why the problem occurred, and what caused the damage of the product. This also meant to explore what the characteristics of the transport conditions were in Asia (RQ2) and to understand the considerations that need to be taken into account when designing transformers for transportation on Asian roads (RQ1). The exploration chapter presents essential information needed for the development project such as information about the leakage problem (Chapter 3.1), a brief presentation of the production and packaging (Chapter 3.2). One of the goals with the thesis was to provide a literature review of transportation information typical for developing countries in Asia and therefore three studies concerning road vibrations were analysed in Chapter 3.3. Other essential inputs for the development of concepts was previous actions toward more sustainable transformers such as existing solutions (Chapter 3.5.2) and different root-cause analysis (Chapter 3.5.1). The main reasons for decreased quality (RQ3) of transformers are summarised in Chapter 3.6.1, and a solution strategy on how to prevent failure (RQ4) is illustrated in Chapter 3.6.2. The information gathered during the exploration phase resulted in a product design specification that describes the needs of what the product must fulfil and is presented in Chapter 3.7.
Statistics of the Leaking Problem It is believed that the weld around the outlets that seals the tank is subjected to hydrostatic pressure which is created inside the transformer when the transformer tank is filled with oil. If 14
the weld gets hurt e.g. during transport and a small crack is generated and the risk of leakage has been increased. The location of oil leakage may differ due to the fact that the transformer is a sealed tank with several sealed outlets, such as drain valve, LV/HV bushings, radiators, pressure relief, oil pipe, filling pipe etc. ABB has brought statistics since 2005 of the number of DTR units that had leaks after they reached the final destination. The statistics provided by Mr. Nguyen P. Minh, Quality Engineer at ABB Vietnam (personal contact, 24 February 2014) indicated that year 2013, the ABB factory in Hanoi delivered over a thousand distribution transformers to customers globally. The percentage of all transformers, where oil leak arisen at site, was less than 3 % per year 2013. The statistical data also revealed that most (20%) of the transformers with oil leakage problems, had the location of leakage at the corrugated weld where the fin wall section is welded together with the tank-bottom, see Figure 3-1a and b. The priority of ABB Vietnam was therefore to prevent the leakages in the corrugated welds along the fin walls.
Figure 3‐1 a) The most common leakage is located in the corrugated weld, b) a damaged transformer, where the leakage occurred at the corrugated weld on three places and mineral oil have dripped along the wall of the tank bottom (© ABB).
In general, the corrugated welds along the longer fin wall sections have the most number of leaks compared to the weld on the short sides of the transformer (Mr. Bradley Havlin, DTR Product Group Manager at ABB Vietnam personal contact, 18 February 2014). Therefore, the focus of this project was to prevent the leakage problem in the weld along the long fin wall sections. Observe most leakage problems are not so severe that the tank needs to be replaced, on the contrary, it is in general possible to repair the problem on site. Possible actions to repair the damages are re-welding, re-tighten bushing nuts or replace a new set of gasket, depending on the location of leakage.
The Production Process The empirical investigations, observations and interviews performed at ABB Vietnam factory provided a better understanding of the production process of ABB transformers. A brief summation of the production process is described in this chapter. Knowledge of the manufacturing process of distribution transformers was crucial for further development of the design. It was also useful to observe the preparations of the transformers, e.g. how it was handled and packaged before shipping.
The Manufacturing Process The factory of ABB Vietnam in Hanoi has the capacity to produce transformers from the size of 10kVA to 12.5 MVA. Most steps in the process of making the transformer as winding the windings, cutting the steel core and folding the fin walls are performed with the help of machines. The general steps of the manufacturing process of distribution transformers are illustrated in Figure 3-2. A: CORRUGATED TANK PROCESS A1
C: TRANSFORMER PROCESS C1
B: ACTIVE PART PROCESS B1
B3 A: A1: fin folding machine, A2: fin welding machine, A3: corrugated tank assembly, A4: tank leaking test, A5: galvanising and painting. B: B1: core cutting machine, B2: winding, B3: core assembly, B4: active part assembly, B5: dry active part assembly in oven. C: C1: seal tank and fill with oil (vacuum chamber), C2: ISO standard test (electricity), C3: prepare transport.
Figure 3‐2 An illustration of the general steps in the manufacturing process of DTR transformers.
As Figure 3-2 illustrates there are two processes which take place simultaneously. The two processes manufacture the parts of the transformer, the transformer tank and the active part located on the inside of the transformer. The active part and the corrugated tank pass numerous steps before the active part and the tank are united, sealed and filled with mineral oil etc. A quality control of each transformer is performed before delivery. Where each is routine tested which includes separate tests for No Load Loss testing and Load Loss testing. This is two types of test, one for current and one for voltage. 16
188.8.131.52. The Active Part Process The production process of the yoke (the active part) involves the steps of cutting the steel sheets (automated) and assembling (manually) the core by apply the steel sheets in layers, see Figure 3-3a. The copper windings are winded with different types of machines for different sizes, see Figure 3-3b. The windings are placed around the steel core to form the yoke, see Figure 3-3c.
Figure 3‐3, a) Assembly of the steel core, b) copper winding machine, c) the yoke (© ABB).
184.108.40.206. The Corrugated Tank Process The most relevant process for this project was the production process of the corrugated tank. The process may differ depending on the capacity of the factory. For example, ABB in Poland manufacture the fin wall sections with a fin folding machine (Figure 3-4a) which fold a 1.5 mm thick steel sheet to one fin wall section, whereas ABB in Hanoi pre-order their fin walls from a manufacturer in Thailand. A local company produces the tank-bottom and tank-top parts of in Vietnam. The different parts of the corrugated tank are assembled in a welding workshop. Figure 3-4b shows a welder in Poland assembling a corrugated tank. The fin wall sections are welded to the tank-bottom and the tank-top parts; the weld both attaches the fin wall (see the enlarged picture in Figure 3-4c) to the tank assembly and seals it. Every assembled tank is leakage tested in an air pressure test of 0.2-0.3 Bar for 20-30 min before they are painted.
Figure 3‐4 a) Fin wall folding machine, b) welding a corrugated tank, c) weld along the fin wall section (© ABB).
To be able to implement realistic concepts to the manufacturing it is important to take the process into account when developing concepts. For example, it was first supposed that the amount of damaged products could be reduced if the transformers are transported empty with no oil, meaning there would be much less weight during the transport affecting the structure. However, due to the complicated process to fill the transformer it is not possible to transport the transformers empty and fill them at the site of the customer. Hence, it was revealed through the process that it is of great importance that there is no moisture inside the active part, and the active part is placed in an oven to remove all moisture. Additional, to assure there is no moisture inside the transformer the tank is filled (with approximately 90% of its volume) with mineral oil inside a vacuum chamber to ensure as much air (moisture) is removed.
Preparations for Transport A demonstration and instructions for loading, fixing and tightening for transformers within DTR have been provided by Pham Minh Nguyen (Quality Engineer at ABB Vietnam, personal contact 24 February 2014). Instructions were implemented at the end of year 2012 to ensure that the transformer is loaded and tightened in the correct way on a truck, lorry or container before delivery. The transformer is loaded onto the truck with the help of a crane. It is not allowed to lift the transformer from the bottom of the fin walls. Therefore the lifting chains are attached to lifting lugs generally located on the top cover of the transformer tank. A 5-10 mm thick rubber mat is placed underneath the transformer to protect it from scratches (see the mat in Figure 3-5a). Wooden breaks are used to fix the transformer with additional 100 mm long nails to nail the break onto the truck bed made out of wood (Figure 3-5b). If there is a steel floor as in a container, the breaks are only inserted between the roller beams and the wall to fix the transformer. The transformer is lashed with steel wire rods or textile lashing belts to strap the transformer through lifting lugs on truck or container wall (Figure 3-6). The lashing belts can also be placed on the cover as long as the belt is placed between the delicate fins (Figure 3-7). Generally transformers are packed together if there is enough space on the truck. It is important to assure that there is sufficient space between the transformers and the walls of the truck so that the unprotected fin walls do not collide. The observations at ABB Vietnam showed that the distribution transformers have no special packaging, the only part wrapped before transport are the sensitive parts on top of the transformer, see Figure 3-8a and b.
Figure 3‐5 a) Wooden breaks between the transformer and the truck wall. Underneath the transformer is the rubber mat placed (© ABB), b) wooden breaks are nailed to the wooden truck bed (© ABB).
Figure 3‐6 Examples of transformers tightened with steel wires to the container floor (© ABB).
Figure 3‐7 The textile lashing ratchet belts placed between fins and belt goes under the steel rods (© ABB).
Figure 3‐8 a) A transformer with sensitive terminals (unwrapped), b) transformer with the sensitive parts wrapped (© ABB).
Transport Conditions in Asia In an attempt to answer RQ2, available articles with vibration analysis of truck shipment in Asian countries were studied. Valuable information about the current road surfaces and how they affect packaged products was found. A better understanding of the characteristics of transport conditions present in Asia was developed, and useful data for future testing became available. The vibration analysis and their measurements are presented in the chapter below.
Performed Tests of Road Vibrations Damages to transported products can in most cases be associated with various vibration forces that originate from the transportation means (Jarimopas, Singh and Saengnil, 2005). It is important to understand the types and levels the forces have, to be able to reduce damage when designing packages (Jarimopas et al., 2005). Three studies about these kinds of road vibrations were found during the literature studies. The titles of the three analysed studies are; Measurement and Analysis of Truck Transport Vibration Levels and Damage to Packaged Tangerines during Transit (Jarimopas, Singh and Saengnil, 2005), Measurement and Analysis of Truck and Rail Shipping Environment in India (Singh, Sandhu, Singh, and Joneson, 2007) and Measurement and Analysis of Truck and Rail Vibration Levels in Thailand (Chonhenchob, Singh, P., Singh, J., Sittipod, Swasdee, 2010). The data presented from the studies could assist product and package designers to reduce damage in transportation. These countries have infrastructure and road conditions comparable to Vietnam, and the data could therefore be useful to ABB Vietnam when simulating vibration tests. Test methods from the American Society for Testing and Materials (ASTM), and the International Safe Transit Association (ISTA) makes it possible for the users to develop customized test plans for test of packages for global distribution (Singh, Sandhu, Singh, and Joneson, 2007). ASTM and ISTA already provides various references of previously measured vibration data (mainly from conditions in North America) to be used in the test methods ASTM 20
D 4169 or the new standard Project 4AB of ISTA. Therefore, the purpose of the studies performed in Thailand and India was to fill the data void and contribute to updated data for the most common distribution means in Thailand and India (Singh, Sandhu, Singh, and Joneson, 2007). Vibration levels observed from various segments of transportation are presented in the form of power spectral density (PSD) (Chonhenchob et al. 2010). A PSD function displays at which frequencies variation (energy) is high and at which frequencies variation is weak (power density [g2 /Hz] versus frequency [Hz]) (Singh et al., 2007). It is possible to combine the power density data to determine a composite spectrum that can be used to develop vibration tests. A summation of the result from these studies, of quantified levels of vibration for particular trucks and road conditions are: Jarimopas, Singh and Saengnil (2005) present a study with the purpose to measure vibration levels in commercial truck shipments in Thailand and to observe the effects on packed fruit. In the study, the vibration levels in the two of the most commonly used truck types to ship packed goods were measured as a function of road condition and vehicle speed. The trailers studied had leaf-spring suspensions. Singh et al. (2007) present a study with the purposes to establish extensive awareness of measurement of the truck and rail vibration environment among distribution hubs in India and to compare the vibrations levels in trucks versus those in railcars. Chonhenchob et al. (2010) presents a study that provides a broad understanding of the vibration levels measured in the truck and rail shipments along major distributions routes in Thailand. The collected vibration data from the studies were analysed and presented in the form of PSD graphs. The vibration levels were measured for various road conditions, some examples are shown in Figure 3-9, (Chonhenchob et al., 2010, Jarimopas, et.al., 2005, Singh et al., 2007).
Figure 3‐9 A few example of road conditions in Thailand and India. From left: (top and bottom) Various road conditions of Thailand (Chonhenchob et al., 2010), (middle), a two‐lane highway in India with bad shoulders (Singh et al., 2007), (top and bottom) laterite road surface and concrete highway surface in Thailand (Jarimopas, et.al., 2005).
A scoop of the most important conclusion from the studies was:
The vibration levels and its severity increased with the speed of the vehicle (Jarimopas et al., 2005).
The kind of road surface that contributed to most severe vibrations was the laterite roads (unpaved roads) followed by concrete highways. The lowest vibration damage occurred on asphalt roads (Jarimopas et al., 2005).
The measured vertical vibrations were more severe and had higher intensity vibration levels in India than levels used in existing test methods from North America provided by ASTM and ISTA (Singh et al., 2007).
The vertical vibration levels had the most intense levels (measured of both truck and rail shipments), followed by lateral and longitudinal levels (Chonhenchob et al., 2010).
The new vibrations PSD, was recommended to be used to simulate the vibrations conditions in truck and rail shipments for shorter bed trucks in Thailand (Chonhenchob et al., 2010).
Conclusions The studies provides a basis of data for proper test methods e.g. ASTM 4169 and ISTA 4AB, that contributes to better simulations of road conditions of Thailand and India, more correctly than the levels provided by ASTM and ISTA themselves (Singh et al., 2007). Hence, the road conditions in India and Thailand are similar to those in Vietnam the data from the studies are believed to be helpful also for simulation of Vietnam road conditions. An interesting result of the studies showed that the vibration level in the vertical direction is severe and had the most intense levels followed by lateral and longitudinal levels (Chonhenchob et al., 2010). This means that the vertical impact and vibrations should be counted in test simulations. Another thing that was interesting was that roads typical for Vietnam such as unpaved roads and concrete highways are the type of road surfaces that provokes the worst roads conditions possible (Jarimopas et al., 2005). Also, the conclusion that the severity of vibration levels increases with the vehicle speed suggests that a transformer shipped overseas to a country like Australia, with higher speed limits than in Vietnam but not necessarily better road conditions, is exposed to higher vibration levels than in Vietnam. Another notable fact was that vibration levels can be improved with the use of air-ride suspensions instead of steel leaf-spring suspensions systems (Jarimopas et al., 2005). Singh and Marcondes confirmed that air-ride suspensions had significant improvement for the ride quality and reduced both acceleration levels and damages (1992 referred by Jarimopas et al. 2005). 22
Countries of advanced vehicle technology have adopted air suspensions system widely for both passenger cars and trucks but in developing countries there are only some luxury passenger cars and small part heavy trucks that are provided with air suspensions systems (Yonglitai, 2012). Thereby, it is more likely that trucks available in the distributions system of Vietnam are equipped with coil springs instead of air springs. The studies provided relevant transportation information of developing countries, they also gave an idea of the transport condition in countries with similar infrastructure as Vietnam and provided specific data to better perform vibrations test of possible products. From the studies it was also suggested what needed to be done to improve the transportation e.g. the road vibrations effecting the products can be reduced if the trucks were equipped with air-springs, which may be one explanation of why the leakage problem of transformers are not present in Europe. Hence, even though there are occasionally extreme road conditions in Europe the vehicle of the distribution systems are better equipped and can thereby more efficiently suppress the vibrations from the roads.
Stress Analysis of Weld
The number of leaks is higher at the bottom weld than at the weld above the fin wall. A general hypothesis was that this is because the bottom welds are subjected to higher pressure from the oil inside the transformer. If the stress concentrations in the structure are initially high due to hydrostatic pressure this also means that additional exposure to road impacts from transport increases the risk of leakage. Therefore, simple hand calculation was done in order to estimate the stress concentrations in the weld caused by the hydrostatic pressure. The hypothesis of the believed stress situation that occurs when a distribution transformer is filled with oil is illustrated in Figure 3-10.
FIN Inside tank
Figure 3‐10 Demonstration of how the hydrostatic pressure affects the weld on the outside in a negative way (Lisa Magnusson, 2014).
Hypothesis: The volume and the density of the mineral oil creates an internal pressure (hydrostatic pressure) on the inside of the transformer. The pressure becomes larger closer to the bottom, and this is most likely why the number of leaks is higher at the bottom weld than at the weld above the fin wall. The pressure on the fin wall section subjects the weld with a lot of
stress. Because, there are nothing that stiffens or carries the fin wall in the middle so the fin wall wants to fluctuate (observe the figure is excessive). To calculate the amount of stress in the weld due to hydrostatic pressure the stress situation was simplified to a load case of a simply supported beam with linearly distributed load, with minimum load q1 and maximum load q2 (see Figure 3-11). The load cases with respective equations were taken from a general formula sheet (e.g. Lindell, 2005). SIMPLY SUPPORTED BEAM
Weld thickness = the thickness of the steel sheet
Figure 3‐11 a) The load on the inside of the transformer due to hydrostatic pressure, b) a simply supported beam with linearly distributed load (Lindell, 2005).
The equations used to calculate the average stress in the weld are presented below. See the full calculations with the values included in Appendix A. Calculations of a simply supported beam with linearly distributed load: Designations (from Figure 3-11) are: L: Beam length 1500 mm is same as the fin wall height of a 4 MVA DTR q : Minimum load case q : Maximum load case R : Reaction force in support A R : Reaction force in support B
The stress in the bottom weld ( the use of Equation 1:
) caused by the hydrostatic pressure was calculated by
Designations from Equation 1 are: : Stress in bottom weld at point B : Weld area
The reaction force ( ) affecting the bottom weld of the tank and the weld area ( calculated with the help of Equation 2 and Equation 3:
/6 2 ∙
Designations from Equation 2 and 3 are: b: The width of surface exposed to pressure and the length of the weld along one fin (25 mm) t: is the thickness of the steel sheet of the fin (1.5 mm) and
are calculated with the help of Equation 4 and Equation 5:
Designations from Equation 4 and 5 are: : Minimum internal pressure due to the density of oil fulfilment in transformer tank (0.00099 MPa) : Maximum internal pressure is set to 0.15 MPa (values taken from AESC, 2013)
Conclusions The stress at the bottom of the support B ( ) has the approximate value of 50 MPa, which is the same as a force of 50 N over each mm2 affecting the weld. These rough hand calculations gave only an estimation of the average stress caused by the oil inside the transformer tank, affecting the weld on the bottom of the fin wall section. The average stress does (fortunately) not exceed the yield stress of the material (250 MPa). Suggesting that the leakage is not only due to the pressure from the oil inside the transformer. Unfortunately, the peak stresses that will occur cannot be defined with these hand calculations. It can only be concluded that if the average stress is significant from the hydrostatic pressure, then the peak stress will be even more hazardous for the construction. The analytical result was later compared with the stress result from the FEM-analysis results in Appendix D.
Actions toward Sustainable Transformers The exploration phase gave a deeper understanding of the initial project problem; to understand what to consider when designing transformers for transportation (RQ1), what others have considered before and what can be done to prevent the leakage (RQ4). To realise how to prevent the problem requires not only good knowledge about the problem but also a deeper understanding of the cause of the problem (RQ3). 25
Previous actions that have been done about the leakage problem with purpose to improve the transport conditions of transformers are presented in the chapters below. A root-cause analysis of the failure where the focus was to find insight why the problems occur is summarised in Chapter 3.5.1, and existing solutions found during the empirical investigation are benchmarked in Chapter 3.5.2.
Previous Root-Cause Analysis Before the start of this project earlier actions were done in attempt to solve the leakage problem, including two FEM-analysis of distribution transformer where the aim was to find the rootcause of the leakage and develop a design solution that prevents the leakage. Two cases of analysis are presented, one static analysis and one vibration analysis.
220.127.116.11. FEM-analysis from AES and AESC Mr. Hung (personal contact, 5 March 2014) from the company AES Vietnam Co. Ltd, in corporation with AESC Co. Ltd, performed in 2013 a Finite Element Method (FEM) analysis of an ABB transformer, model 2 MVA, to find out the root cause of oil leakage. The analysis document is confidential and proprietary of AES and AESC. Nevertheless, the analysis have been helpful for this project and provided ABB Vietnam with one of the root-causes. The analyses was a kind of linear static analysis where the load cases as self-weight (gravity load) and internal pressure (hydrostatic pressure due to the oil in the transformer) was applied, as well as inertial load (acceleration of 2G in lateral and longitudinal directions). The conclusions from the analysis of AES and AESC (Hung, 2013) confirmed that the load as self-weight is not harmful to the transformer. However, when the self-weight was combined with hydrostatic pressure, maximum stress (peak stress) in the weld along the bottom fin exceeded the yield stress of the material (Hung, 2013). The stress levels became naturally worse when an acceleration of 2G was added. The peak stress area in the analysis agreed accordingly with the actual leak area in reality, in the weld along the bottom of the fin wall section. Thus, the hydrostatic pressure was believed to be the root-cause of the leakage problem. Meaning, the weld does not withstand the stress formed by a combination of the self- weight and the internal pressure of the transformer (Mr L. Hung, Engineer at AES Vietnam Co. Ltd personal contact, 5 March 2014). The outcome of the analysis was a modified solution, a bracket solution that releases the stress around the bottom fin wall area to avoid leakage, by increasing the attachment of the fin walls. The bracket reduces the stress in the weld that join the fin wall to the tank wall. The bracket was applied to the production of transformers to test its efficiency during a 6 month period simultaneously as this thesis was performed. The results to date clearly indicate that the bracket is working. A summary of the pros and cons of the bracket is presented in Chapter 3.5.2 with existing solutions, and the bracket can be seen in Figure 3-14.
18.104.22.168. FEM-analysis by University of Lodz Another study was performed in 2008 by Kołakowski, Kubiak, and Mania from the Department of strength of materials and structures, University of Lodz. They made strength test of a TNOSCTRU 1600-1610 PNS transformer. The aim of the study was to measure the types of damage that arise during transportation on roads by subjecting the transformer to vibrations. They tested what frequency contributed to natural frequency of the transformer. It was found that natural frequency is reached, as for this transformer at 22.3 Hz respective 19.7 Hz. When this occurs there are dominant bending deformations in the outer vertical edge of the fin wall section, forcing the fin wall to deflect in the middle of its height (Kołakowski, Kubiak, and Mania, 2008). Stress levels were also found that exceed the yield stress of the material in the area where the weld joins the fin wall to the tank-bottom. The outcome of the study of Kolakowski et al. (2008) resulted in an support solution used today for transformers, which is a round steel rod at the bottom and the top edge of the fin walls, see Figure 3-12 in Chapter 3.5.2. The rod was believed to stabilise the fins and prevent them to collide during transport when vibration occurs. This rod was later proven by Mr L Hung to reduce up to 50 % of the stress in the weld (Mr. L. Hung, Engineer at AES Vietnam Co. Ltd personal contact, 5 March 2014).
Existing Solutions In this chapter, the current solutions applied to the original transformer design in order to prevent the leak problem are presented. The design solutions have been listed with both positive and negative aspects. Earlier solutions were studied to help answer RQ4 (what can be done to prevent the leakage) but were also an excellent source of inspiration for the generation phase. Solution 1 Name: Round Steel Bar (8 mm Ø) Location: At bottom and top edge of fin wall (see Figure 3-12) Positive aspects: According to Hung (personal contact, 5 March 2014) can the peak stress be reduced up to 50% when the bar is applied. The bar goes all the way around the transformer to stabilise the fins during transport. Negative aspects: The bar is attached to the transformer with several spot welds, preferable would be continuous weld for more strength. Figure 3‐12 Round steel bar (© ABB)
Solution 2 Name: Steel Rod (10x10 mm) Location: 250mm from bottom and top edge of fin wall (see Figure 3-13). Positive aspects: The rod goes all the way around the transformer to keep it stable during transport and joining the transformers short side with its long side Negative aspects: The rod is hard to weld and attach, only attached with spot welds, preferable would be continuous weld for more strength. The 10x10mm rod is not standard size in Vietnam so the material quality can be poor and can break e.g. when it is bent around the structure. The 10x10mm rods are too weak, and more robust rods are needed to stabilise the structure.
Solution 3 Name: Steel Bracket Location: Below bottom of fin wall (see Figure 3-14). Positive aspects: According to Hung (2013) the brackets reduces the stress levels in the weld on the transformer to nonhazardous. A long test period clearly indicate that the bracket is working. Negative aspects: The bracket is hard to weld and attach, attached with spot welds and only welded on the front side of the bracket. The space behind the bracket is narrow causing difficulties to repair the weld behind the bracket if necessary.
Figure 3‐13 Steel rod squared ©ABB
Figure 3‐14 Bracket by Hung (2013) (© ABB)
Solution 4 Name: Stiffeners Location: Two stiffeners vertical attached on the long side of the fin wall, attached to the bottom and tank-top part (see Figure 315). Positive aspects: The stiffeners connect the tank top and bottom, and thereby stabilise the transformer during lift. Negative aspects: There is no weld joining the stiffener and fin wall section because there is too little welding space. Figure 3‐15 Vertical placed stiffeners ©ABB
Solution 5 Name: Angled Flat Iron Location: Four angled flat iron inside the tank in each corner (see Figure 3-16). Positive aspects: The stiffeners connect the tank top and bottom, and thereby stabilise the transformer during lift. Negative aspects: The angled flat iron is hard to weld and attach due to the deep and narrow space inside the transformer. To transformer with ABB VN design the angled flat iron would impair the oil flow to the fins on the sides.
Figure 3‐16 Angled flat iron in the tanks corners (©ABB)
How to Prevent Failure The exploration phase gave a deeper understanding of the failure (the leakage problem) and gave a better insight about the problem space with the decreased quality of transformers. One believed root-cause causing the product failure has already been presented in Chapter 3.5.1, with the FEM-analysis performed by Hung (2013). Nonetheless, from the exploration phase it was observed that the decrease in quality is caused by a number of influencing factors. The main reasons, the so-called root-causes of the transformer’s failure are presented in Chapter 3.6.1. An illustration of how to prevent failure, a type of solution strategy is presented in Chapter 3.6.2.
Main Reasons for Decreased Quality This chapter is an attempt to answer RQ3, to explain the root-cause of the transport problem. The believed main root-causes of the leaking problem, distinguished from the exploration phase were:
The design of the transformer and the forces it is subjected to.
The quality of material
The design of the transformer and the forces it is subjected to: The construction of the transformer was believed to be subjected to high stress along the bottom fin wall section on the long side as a consequence of the necessary oil filling. A combination of the high stress and the impacts and vibrations from the roads surface causing the weld to failure and leak (described more thorough in Chapter 3.5.1). The product is thereby not designed to withstand its requirements. This hypothesis is based on results from previously performed FEM-analysis (Hung, 2013) and logic. Europe-centric design: The design is not optimal as it originates from Europe. ABB transformers have a Europe-centric design, developed and engineered from and adapted for Europe where the leakage problem is insignificantly small. Extreme road conditions typical for Asia (illustrated in Chapter 3.3) have not been taken into consideration when designing. Henceforth, the construction requires reinforcement afterwards. Infrastructure: Vietnam is a developing country and so is the traffic that is contrary to the secure traffic found in Europe. The infrastructure is poor due to the long war history of Vietnam, and consequently are the road conditions not efficient. For example, there is only one main highway (Highway 1) that runs through the country from north to south. It is impossible to drive faster than 80 km/h (average speed is more like 40 km/h), and the trucks are often overloaded which causes bumpy road surfaces.
Transportation means: In Vietnam the transportation means do not have as good standard as transportation means in Europe. For example, trucks and trolleys with air suspensions are standard means in Europe while in Vietnam it is more common with coil springs or leaf springs as suspensions. Air suspensions would dampen vibrations and impacts from the road more. Packaging: It was revealed from the observations of the shipping of transformers at ABB that there was barely any packaging of the transformer itself. There was no packaging surrounding the transformer tank with the purpose to suppress the vibration from the road surface. Instead the distribution transformer was placed directly onto the truck bed (describe more thoroughly in Chapter 3.2.2). Weld techniques: It was revealed from observation of the manufacturing of the tank assembly that there were some transformers without fully welded insides. Even though the recommendation is to always fully weld the inside of the tank. Overlapping sections should be welded with double weld, see Figure 3-17a. A continued or spot weld on the inside (Figure 3-17b) of the transformer could unload some of the stress on the outside weld.
Figure 3‐17a) Double weld vs. outside weld and b) an illustration of how the inside could be welded (Lisa Magnusson, 2014).
Some important actions have been taken toward improving the weld techniques and the quality of the weld at ABB Vietnam, but they need to continue to follow the improving action plan by ABB Vietnam. Continued visual inspections are required to assure there are no parts welded wrongly or porosity in the weld, see an example of poor weld in Figure 3-18.
Figure 3‐18 Two examples of poor quality welds with air bubbles (porosity) (©ABB).
The visual inspection is extra important to perform on transformers parts supplied by a different company e.g. in Figure 3-19a the triangular opening of the fin wall section has been welded wrongly. Figure 3-19b shows how it should be done.
Figure 3‐19 a and b) The fin to the left is welded wrong and can affect the weld quality negatively, the example to the right is done correctly (©ABB).
The quality of material: The quality of the material in Vietnam is not always as good as the material available in Europe. Only because it happens to be an easily accessible standard material in Europe does not mean that it is available in Vietnam, which is not always considered by designers. One example, is the 10x10 mm steel rod developed in Europe to reinforce the transformer. 10x10 mm rods may be standards material in Europe but the quality of the local supplied 10x10 mm rods in Hanoi sometimes break during assembly (poor quality) as it is not standard material (Figure 3-20).
Figure 3‐20 An example of bad material quality of some of the locally supplied rods to ABB Vietnam (©ABB).
Solution Strategies One goal of the report was to review potential solution strategies and thereby answer RQ4. A general solution strategy that revises possible alternatives for ABB Vietnam how to solve the oil leakage is presented in this chapter, see the solution strategy illustrated in Figure 3-21. The solution strategy was developed from observations at the factory in Vietnam and has the aim to divide the problem into sub-problems (problem areas). In long terms, ABB Vietnam can tackle these sub-problems one by one. This solution strategy could be used as a practical example and approach for similar problems with transportation of goods.
PREVENTION OF OIL LEAK
Better vehicles with air suspension
Better lashing equipment
Temporary attached support structure
Damping pallet etc
New/ modified design ‐without stresses
Ad external solution
Enhance support/ increase weld surface
Figure 3‐21 A solution strategy of how to prevent oil leak of distribution transformers of ABB Vietnam.
As the chart in Figure 3-21 illustrates, there were two main alternatives to prevent the oil leakage, either ABB Vietnam could improve transportation or reduce stress concentrations of the transformer. A further explaination and discussion regarding the two main alternatives of the solution strategy is presented in chapter 22.214.171.124 and 126.96.36.199. The conclusion of the solution strategy is presented in chapter 188.8.131.52.
184.108.40.206. Improve Transportation The alternative “improve transport” is divided into three problem areas suggesting that improvements are necessary when it comes to transport means, packaging and road conditions. Ideal would be to transport transformers on roads with better road conditions, although, this factor is dependent on the infrastructure of Vietnam and nothing ABB Vietnam can affect directly. It would also be preferred if the transportation means were better equipped in the cargo shipments of Vietnam, with better standards of trucks, with more advanced vehicle technology and better equipment for lashing. 33
ABB Vietnam has the potential to improve the equipment for lashing e.g. provide with thicker rubber mats and more advanced wooden breaks to fix the transformer on the truck bed. However, ABB Vietnam hires independent shipping companies to distribute their products to the site without supervision of ABB. Therefore, there is no guarantee that more advanced lashing equipment (provided by ABB) on the truck leaving the factory of ABB Vietnam will be transferred to the next truck used in the transportation chain. Consequently, improved equipment will not guarantee improved transport conditions of ABB Vietnam. The problem area, where ABB Vietnam could have a direct influence on, is the packaging of distribution transformers. From the observations how ABB Vietnam prepare transformers for transport (Chapter 3.2.2.) it was revealed that there is no specific packaging protecting distribution transformers. However, from the observation of the power transformers production process at ABB Vietnam it was revealed that the power transformers are shipped with a type of packaging, with simple packaging like pallets and boxes made out of wood. They are packed in boxes because they are significantly larger than distribution transformers and cannot be distributed as a completely assembled product. The distribution transformers could also benefit from being shipped away with corresponding wood pallets. Wooden pallets would most likely damp the vibration from the truck bed. Also, there are pallets equipped especially with damping effects available on the market. An alternative to packaging is temporary customised support structures. Temporary attached support structures could be used when it is not possible to adapt the original structure of the product to withstand transportation. Temporary support structure would be attached to the transformer to enhance support only during transport that generally is one time during the product life for transformers. Instead of adding extra material and cost on permanent support solutions, a temporary solution could be re-used for transport of other transformers.
220.127.116.11. Reduce Stress Concentrations The leakage problem as explained earlier is due to high stress concentration of the tank structure.Transport aggravates the situation, which is why attempts to reduce the stress of the structure has been of high priority for ABB Vietnam. The possibilities for ABB Vietnam to reduce stress can be done according to the chart in Figure 3-21 by changing the transformer tank design, or by adding an external solution (permanent). The purpose of changing the original design of the transformer or modifying the shape of the fins is to reduce the initial stress concentrations in the structure, and in the weld. It is possible to reduce the initial stress concentration in the structure e.g. by removing the sharp corners between the fin wall sections and the tank-bottom, or by shortening the length of the fin walls sections due to the problem is typical for the long sides of the transformer. If the stress is reduced then the structure better withstands the impact of transportation. An earlier priority of ABB Vietnam has been to support the structure with external support solutions. The purpose of an external solution added to the transformer is to enhance the attachment of the fin wall, by providing more weld surface and to attach the fin wall better onto the tank. If the structure and the fin walls can be better supported then the initial stress 34
concentrations of the structure can be better distributed and the carrying weld would be less affected by the pressure caused by the oil inside the transformer. The alternative to replacing the established design of the distribution transformer with an improved design may be a good idea in the long run. Hence the current product design clearly does not manage its purpose without extra added support structures. However, a new design may be harder and more expensive to implement in the manufacturing process of ABB.
18.104.22.168. Conclusions There are several possibilities for optimisation of the manufacturing process as well as the transport process of distribution transformers at ABB Vietnam that possibly could result in improved quality of the transformers after transport. As mentioned, ABB Vietnam could e.g. improve the packaging of distribution transformers and suppress the impacts and vibrations during transport, or change the design of the transformer to reduce the stress concentrations in the structure. As mentioned, the initial stress concentrations can be reduced either by changing the original transformer design or by attaching an external solution to support the transformer. To change the original design of the transformer is more complicated compared to adding an external support solution regarding the manufacturing process. When the solution strategy had been established, with ABB Vietnam it was discussed which of the two main alternatives were suitable to base the product development on. It was decided that the project was to focus on developing the product design of the transformer and not a type of packaging.
Establish Requirement As mentioned in the chapter of methods (Chapter 2), a Product Design Specification (PDS) is an establishment of requirements, with metrics and values, and explains what a product should be able to do (Ulrich and Eppinger, 2008). Results from the empirical investigations, identified needs of ABB and a better understanding of the conditions of transformers provided the foundation to the PDS of the solution. The PDS was used to represent the target that was sought throughout the developmental process and signified the ambitions of the project. The PDS has two types of requirements; necessary (N) demands and wishes (W). The necessary demands describe what the final product must perform or fulfil to be considered as a good working product or solution (Olsson, 1995). Wishes do not have the same importance as necessary demands (Olsson, 1995). Although, if the wishes are fulfilled (as well as the necessary demands), then the product will be competitive and differentiate from the existing products on the market. The rank of wishes is presented in Table 3-1, and the PDS is presented in Table 3-2.
Table 3‐1 Wishes ranked in order of importance.
RANK OF WISHES 1 Not significant increase the weight, even if it has no great significance for the use. 2 Simple to manufacture 3 Give enough access to weld, to paint and sandblast and access to repair welds around fin wall edges. 4 Withstand extreme transport conditions in a physical test. 4 Should be possible to attach/weld onto the transformer used today with specific fin thicknesses and tank thicknesses. 5 Should not have any projecting parts that can get damaged during lifting. 6 Express robustness 7 Have a product lifespan of 30 years or more. 8 Be as cheap and effective as possible. 8 Be designed to simplify transport of the transformer. 9 As many components as possible should be locally produced and assembled in Vietnam. 10 The material used should be recyclable. 11 Minimize incorrect designs and constantly increase the quality. 12 The solution shall not change the overall picture of the original transformer.
Table 3‐2 The PDS represents the requirements of a design solution for distribution transformers with a corrugated tank (fin wall sections). Criteria category No N/W
Functions : 1 2
Prevent oil leakage of transformer tank.
Distribute the peak stresses in the fin‐wall section and reduce x < 250 the stress to x MPa (the yield stress level of the material). MPa
YES x < 100 MPa
Hinder the fin to expand in width from 8‐26 mm.
Be able to operate in an outdoor environment. Humidity range 0
N N N
Take little of space, total size (width and height) of the width + width + 400mm 0mm transformer shall increase as little as possible. No Have any projecting parts that can get damaged during lifting. YES Be designed to simplify transport of the transformer. YES Any modification shall meet the ABB standard and be ISO9001/ 14001 certified. Change the overall picture of the original transformer. NO ‐
The amendment must be reliable.
17 Weight: 18 Manufacturing 19 process: 20
Increase weight, even if it has no great significance for the use. Simple to manufacture.
All components should be locally produced in Vietnam.
YES Be implemented to the regular manufacturing process of ABB factory in Hanoi. Should be possible to attach/weld onto fin thickness w mm, w = 1.2, x = 1.5, z = 10 y = 6 and x mm, tank thickness z mm, and y mm. YES Steel, same/similar material used to the transformer tank.
11 12 13
Withstand an impact load of x times the Gravity force in all x < 1.5 G x < 2 G directions (lateral, longitudinal and vertical). Withstand vibrations levels in all directions (lateral, ‐ longitudinal and vertical). Be suitable adapted to ABB distribution transformers of size YES between 1.5‐5 MVA. Hinder the cooling system (the fin‐walls) to cool down the oil. NO Air should pass trough without problems. Hinder the volume to be expanded 5% or more during use. NO
N W W N
YES YES NO YES YES
Materials: 23 Cost: 24
Have a low production cost of x USD.
Have a product lifespan of x years.
Have enough access to weld, paint, sandblast etc., to repair welds around fin‐wall edges. Use hot‐dip‐zinc‐coating for outdoor applications.
Use the suitable coatings (post processing’s).
Inspections: 30 Testing: 31
x mm layer of paint is required on all exposable surfaces.
Withstand the leak control of each product. After the transformer tank is assembled it will be subjected to pressure test of x Bar. Withstand extreme transport conditions in a physical test.
The material used should be recyclable.
32 Disposals: 33
x < 60000$ x < 1500$ x > 20
x > 30
YES x >2 mm
0.2< x <0.23 Bar
Generation One goal of the project was to deliver a product development case study. This was done by following the solution strategy of how to prevent oil leakage in ABB transformers (presented in Chapter 3.6.2.). Thereby, the case study suggests an answer to RQ4 (what can be done to prevent the leakage problem), which was to focus on the development of the original transformer design. The generation phase (described in the method chapter, Chapter 2.1.2) is the creative part of the design process (Cross, 1994). In the generation phase, creative methods are used to generate ideas and to create many concepts and suggestions to a final solution. However, the focus of this project was not to come up with as many concepts as possible, the focus was instead directed to generate complete solutions and realistic ideas from start that was technically and economically realizable. The goal was to generate ideas that helped reduce the stress on the leakage area by changing the original transformer design or add an external solution to the transformer. From the concept generation eventually six concepts were selected, improved and evaluated until a final concept was chosen with the help of experts’ knowledge, group sessions, and observed manufacturing possibilities from the empirical investigation. The chosen concept’s final design was further improved and examined (Chapter 5). The methods in the generation process, the ideas from the concept generation is presented and evaluated below.
Concept Generation Which existing solutions can be adapted for the application and what new concepts might satisfy the established needs are two important questions to consider during the concept 38
generation according to Ulrich and Eppinger (2008). These questions were particularly possible to apply in this product development project. For example, ABB had already some existing solutions which could be analysed and improved in a function analysis (see Chapter 4.1.1.) and new concept ideas could be generated with creative methods (see Chapter 4.1.2.).
Function Analysis As mentioned in the method chapter (Chapter 2.1.2.), concepts can be identified from internally and externally search procedures according to the 5-step-concept generation method (Ulrich and Eppinger, 2008). Inspiration for ideas was searched both externally and internally of the company. Step 2 in the 5-step-concept-generation method was about searching for existing solutions which was summarised in Chapter 3.5.2., and from there a type of function analysis was done to see what different parts and functionalities the existing solutions had. Pros and cons of the different solutions were established, and this became useful in the concept generation. For example;
It can be wise to have a solution that is joined on the long side and the short side of the transformer to achieve more stability.
It is preferred to have a continued weld rather than spot welds for better attachment.
One purpose of stiffeners connecting the tank-top part together with the tank-bottom part is to stabilise the transformer tank structure vertically.
These conclusions were considered when developing new solutions.
Idea Generation Step 3 in the 5-step-concept-generation method was all about searching for ideas internally, inside the company e.g. from personal creativity and knowledge, with help from group sessions together with ABB workers (Ulrich and Eppinger, 2008). Together with an expert from the ship construction business (Mr. Dan Magnusson, Service Manager at TTS Marine VN personal contact, 9 March 2014) a brainstorming session was performed to generate realistic idea solutions of how to reduce the stress inspired by existing solutions. Inspiration for the brainstorming was also taken from existing techniques used in hull design (Figure 4-1a), and packaging as wooden boxes which is a robust structure that endure shipping (Figure 4-1b), and flange brackets for shelves (Figure 4-1c). The generation session was based on the 6-3-5 method described by Wright (1998) where each participants (preferably 6) generates 3 ideas in 5 minutes and then forward the idea to the
next person for further idea generation. The purpose with the 6-3-5 method is to develop each other's ideas.
Figure 4‐1 a) Two alternatives how to repair the stringer deck of container ships with fracture (IACS, 2005), b) fumigation wooden box (Alibaba.com, 2014), c) a flange bracket that carries shelves (Amazon.com, 2014).
The Gallery Method With the help of workers from ABB Vietnam, the generated ideas from the earlier brainstorming session were evaluated and further generated during a group session. For the group session, the gallery method was used described by Ulrich and Eppinger (2008) where the ideas was presented one by one before any discussion was allowed. The participants were given the opportunity to take a closer look at the ideas as if they “walked into an art gallery” . Spontaneous comments and feedback were welcomed as well as further idea generations. The workers were experts from each area of the manufacturing process at ABB, such as welders, quality controllers, mechanical design engineers, and the production manager. The ideas were discussed from a production perspective. The ten generated ideas can be seen in Table 4-1 and is further described in Appendix B.
Table 4‐1 Ten generated solution ideas (Lisa Magnusson, 2014). No. 1
The Positive, Negative, and Interesting Method The positive, negative, and interesting (PNI) features with the ideas from Table 4-1 were decided with the help of the PNI-method from Mind Tools (2014). The PNI-method was used to evaluate the ideas, but also to understand what negativities needs further improvement and what interesting ideas could be useful. Table 4-2 present the result from the PNI-method.
Table 4‐2 A positive, negative and interesting aspects of each solution idea.
PNI No. 1
5 6 7
POSITIVE (P): + stronger than the 8 mm bar + better quality + beam goes all around tank + beam is stiffer
– will probably rust –do not provide support for tank during
* a rod on the bottom edge resist ~50% of the stress in the weld so maybe an angled iron would resist even more stress!
vertical lift –spot weld to attach –different size of tank need different length of angled iron – transformer size will increase the circumference – do not provide support for tank during vertical lift – different size of tank need different length of the beam
+ prevent fluctuation * do not need to be + stronger than the permanently attached, 10x10mm rod could be mounted with + better quality nuts and bolts instead + beam goes all around the of weld tank *different shape of the + beam is stiffer beam is possible + will not increase – this solutions may require angled iron on the * if angled iron is used the circumference inside (more material) it provides support + an angle iron during vertical lift – spot weld to attach inside provides * can an angled iron be – do not provide support for tank during vertical lift attached to the inner more weld surface than no. 2 – different sized tanks = different beam lengths corners? * can the volume still expand as needed, + reduces stress by soften – material goes to waste the material stiffness or is it necessary to compensate for the – extra work in the removed volume? + simple solution production process * in theory, if this work properly, no + no extra material needed for fin wall‐supplier other added solutions is necessary + a known technique [same as no.4} [same as no.4] [same as no.4] +better appearance than no.4 – hard to cut, special tool required [same as no.4]
[same as no.4]
+ less waste of material than no.4 and no. 5
– hard to cut, special tool required – support on the inside may take
* by cut‐out + original fin wall design but shorter are the U‐beam used too much space + by unite bottom and tank‐top is the whole – large space between fin‐sections it is possible to put more tank stabilised, not only the fin wall is necessary to be able to weld + the sturdy structure provides support for weld surface – different size of tank need on the inside transformer during vertical lift different length of the beam + increases attachment and weld surfaces [same as no.7] – large space between fin‐sections is necessary to be able * if the flat bar + won’t take any to weld placed on outside extra space – different size of tank need different length of the beam it can contribute to the cooling * utilise + by unite bottom and tank‐top is the – vacuum inside box is needed whole tank stabilised, not only the fin wall – box is hard to seal without letting air the fin as support + utilise the existing fins as support inside material. + have a continuous weld + the cut‐out on top and bottom (chamfer) – if oil leaks into the air‐tight box there will be hard to notice if the tanks is is good, aloud space of weld edge damaged and repair the leak inside box * how + leaves space enough to reach the weld to paint and – bracket on short side thick re‐weld differs to the one on long need the + cross‐section of bracket is robust in vertical direction side bracket + easy solution – maybe different size of to be? +easy to manufacture tank need different length + increases attachment and weld surfaces of the bracket + no spot weld used, only continuous weld – need enough space + only a few brackets needed between brackets to be able + little material used to weld
Further Concept Generation From the PNI-method and consultation with the ABB workers it was decided that six out of the ten solution ideas were chosen for further development, these were solution ideas no. 2, no. 3, no. 4, no. 7, no. 9, and no. 10. Consultations with experts and teachers and cooperation with mechanical engineers at ABB Vietnam were essential for further concept development. New sketches and ideas were generated about the features and details. The result was six thoroughly developed concepts (Figure 4-2). The concept was then evaluated in Chapter 4.3. The six developed concepts are described and presented in Chapters 4.2.1 – 4.2.6. Concept A
Figure 4‐2 The six concepts are based on the six idea solutions no. 2‐4, 7, 9, and no. 10 (Lisa Magnusson, 2014).
Concept A The 10x10 mm rods attached to the transformer today, have in Concept A been replaced with a rectangular beam frame, attached around the transformer with spot welds (see Concept A in Figure 4-2). It is simple to add this external support solution. It is believed that beams compared to rods, are stronger, and will limit any fluctuation of the fins, and make the construction more robust. Consequently the beams will reduce the stress concentrations in the fin wall sections. The large distribution transformers as a 4 MVA in the figure could have two beam frames while a smaller DTR may only need one. It is possible to use any shape of beams. In the picture Ubeams are used but ideal is to use angled iron beams so that rain water better runs off. 43
Concept B Concept B has a rectangular beam frame attached along the inside of the four fin walls to prevent any fluctuation of the walls from hydrostatic pressure (see Concept B in Figure 4-2). In a strength point of view, Concept B should have the same robustness as Concept A. The original solution idea was no. 3 and included angled bars in the corners on the inside, which connecting the tank-bottom and tank-top parts. This would increase the attachment of the beams and create rigidity. However, this was not applicable to DTRs of ABB Vietnam hence the flow of oil in the fins located at the sides would be negatively affected with angled bars in the corners. The four U-beams must instead be welded onto respective fin wall section before the tanks walls are assembled and afterward join the beams with welds or smaller flat bars as Figure 4-3 shows.
Figure 4‐3 Top view of the transformer with concept B inside. The beams could be joined after the walls are assembled and joined with flat bars or weld (Lisa Magnusson, 2014).
Concept C Concept C is a modified design of the original transformer. It looks robust and less bulky. The original design of the solution idea no. 4 was further modified to make the concept more adaptable for manufacturing. The chamfered edge has now a slope of 45° degrees, which is a standard angle to cut and the cut-out should be enough to reduce the material stiffness, see Figure 4-4. The cut-out of material in Concept C adds a step in the manufacturing process, which can be performed before the metal sheet is folded in the fin wall folding machine. Furthermore, new tools or techniques may be required for the modified fin wall design. The manufacturability needs to be overlooked by the fin wall supplier.
Figure 4‐4 Concept C with the 45° degree cut‐out of the fin wall edges (Lisa Magnusson, 2014).
Concept D Concept D aims to reduce the stress concentrations in the weld by divide the fin wall section on the long side into two shorter sections. A U-beam is attached vertically in between the sections and between the tank-bottom and tank-top (Figure 4-5b). The stiff U-beam makes the tank more stable, and the fin walls are no longer the only carrying section of the structure. The U-beams are cut off and chamfered in the ends (Figure 4-5a). The cut-outs make it possible to overlap and to weld the beam on the inside wall of the tank-bottom and tank-top, to increase the weld surface. To be able to weld the two sections onto the beam enough space is required in-between the two sections (Figure 4-5c). Therefore, the long side of the tank in Concept D has fewer fins than the original design would have, which could affect the volume negatively. The beams were placed on the long side hence the leak problem occurs most times on the long sides rather than on the short sides. Nonetheless, Concept D could be implemented on the short side as well and divide the short side into two sections if necessary.
Figure 4‐5 a) Beam is cut off to overlap tank bottom and increase the weld surface, b) a U‐beam placed vertical on the inside, c) space between the two fin sections is necessary to weld them onto the beam (Lisa Magnusson, 2014).
Concept E Concept E is an external support solution consisting of three parts, one flat bar attached vertically along the outside of two adjacent fins (Figure 4-6a), and two shorter flat bars attached to the top and bottom fins (Figure 4-6b). The flat bars are 10 mm thick, and the edges are chamfered to optimize the weld possibilities. The fins and the flat bars together create a sealed “box” which acts like a stiffener that makes the surrounding fins stiff. The aim with stiffeners is to prevent any fluctuation in the fin walls and make the tank more stable. For larger distribution transformers, it is better to use more than one “box” on the long side (see Concept E in Figure 4-2). The box needs to be airtight to prevent any corrosion.
REINFORCED FLAT BAR OUTSIDE FINS
FLAT BAR BELOW
Figure 4‐6 Details of Concept E a) the flat bar vertically placed outside fins must be reinforce to be stiff enough, b) shorter flat bar are attached below the fins and above the fins (Lisa Magnusson, 2014).
Concept F Concept F is a steel frame made out of flat bars and consoles (see Concept F in Figure 4-2) and is an all-around support to the fin walls, similar to how consoles support a shelf. The idea with Concept F is to increase the attachment of the fin wall sections onto the transformer tank bottom and support the fin walls especially during vertical impacts from bumpy roads. The original idea no.10 was to weld three consoles to three respective fins to make these and the surrounding fins stiffer. However, the original idea requires a continued weld along the thin fin edge which could cause material melting. Therefore, a flat bar was implemented in the design and attached to the bottom side of the fins (with spot welds). The consoles are instead welded directly (with continuous welds) to the flat bar. A transformer with an attached flat bar frame around the whole structure becomes stiffer compared to a few consoles attached to single fins. The flat bar also helps to distribute the stress from all the fins to the connected consoles. The advantage with Concept F is that the consoles are thin but have a robust vertical cross section. The concoles are attached parallel along the fins, which partly gives better access to 46
the weld under the fin wall to paint or to re-weld if necessary. The cross section of the console is specially adapted to make the transformer endure vertical forces better. Different design constellations of consoles and flat bars were considered, see some of the combinations presented in Figure 4-7. The final design selected for Concept F was the original L-shaped console design from the solution idea no. 10 (see top left corner of Figure 4-7) with a flat bar between the console and the fin. It was believed to be a durable, cheap, alternative to manufacture and the most suitable solution to attach along the fin wall.
Figure 4‐7 Different combinations of the console attached to different flat bar solutions (Lisa Magnusson, 2014).
Evaluation and Concept Selection More or less systematic and structured approaches like intuition, "knowledge", list of pros vs. cons, testing, and decision matrices can be used when evaluating concepts according to Johannesson, Persson, and Pettersson (2013, p 180). The already performed PNI-method was equivalent to the list of pros vs. cons. To perform testing of each concept was not an alternative at this stage of the project. Therefore, the concepts were evaluated with the remaining methods of the approach by Johannesson et al. (2013), with intuition and knowledge of ABB workers and different kind of decision matrices. Decision matrices is a powerful systematic evaluation tool to compare the solutions alternatives total value/quality because decisions are based on the criteria from the specification (Johannesson et al., 2013, p 181). The purpose of the PDS is to measure if the final solution fulfils the expectations but it can also be used to evaluate developed concepts during the evaluation phase (Ulrich and Eppinger, 2008).
The evaluation processes used in the project are:
clearing out the concepts alternatives that do not qualify with the help of an Elimination matrix,
a relative comparisons of the different solutions in a Concept screening matrix,
a concept ranking survey among the ABB workers,
evaluation of the manufacturing possibilities of each concept with the manufacturing workers at the factory.
The matrices from the evaluation process and result tables from the survey presented below can be found in Appendix C.
Elimination Matrix The first step of the evaluation process was to clear out or “eliminate” the bad solutions (Pahl and Beitz, 1995). In the Elimination matrix the concepts are examined to clarify whether they solve the main problem, meet the requirements or if they can be implemented and within the framework of cost thereto (Pahl and Beitz, 1995). The elimination matrix that was made can be found in Appendix C. The matrix did not lead to any direct elimination. However some concerns were encountered and further investigation and information was necessary before any decision could be made. The only flawless concept was Concept F that agreed well with all the statements in the matrix.
Concept Screening The purpose of Concept screening is to reduce the number of concepts, and clear out the worst ones. In Concept screening the selection is based on relative comparisons between different concepts (Johannesson et al., 2013, p 183). The selection criteria in the matrix were taken from the PDS, and all the concepts were compared with a reference solution, a datum. According to Johannesson et al. (2013, p 183) the reference can be an existing solution. Therefore, the bracket solution of ABB transformers were used as a datum and was given a net score of zero. The concepts were one by one evaluated if they met the selection criteria better than (+), as well (0) or worse (-) then the referred datum. A net value was then computed, and a rank of order became available. The result from the Concept screening matrix according to Pugh’s method (1995 referred by Johannesson, 2013, p 184) can be seen in Table 4-3. From the matrix, it was revealed that Concept B should be eliminated hence it was rated as worst. It was rated worst because the concept would be complicated to manufacture, and it might not even withstand vertical impacts during transportation and therefore it was not 48
considered to be a realistic solution. Concept A was rated second worst. The datum solution shared the fourth place with Concept E. Concept F was best rated followed by Concept C on second place and Concept D on third place. The credibility of Concept A and E were questioned because Concept E was believed to be too complex to manufacture and Concept A was not believed to reduce enough stress nor endure vertical impact loads. The two worst concept were kept in the evaluation process until the manufacturability had been looked over for all concepts. Concept F was considered to be a competitive solution adapted for distribution transformers of ABB. The screening and the result were based on the judgment and perception of the author and therefore a second opinion was desired. Consequently a survey was carried out. Table 4‐3 The concept screening matrix. Criteria Reduce/ distributes stress Withstand impact load (vertical) Low production cost Realistic solution Take little of space, do not increase the total circumference Low weight Simple to manufacture Weld is in reach for welder, paint reach every inch Total amount of + Total amount of 0 Total amount of ‐ Net Value Ranking Further development
Concepts B C 0 + ‐ 0 0 ‐ ‐ +
DATUM 0 0 0 0 0
A ‐ ‐ 0 0 ‐
0 0 0
‐ + +
D + + ‐ +
E 0 0 0 0
F 0 + + +
+ + 2 1 5 4 1 4 2x0 3x0 1x0 2x0 6x0 4x0 ‐4 ‐4 ‐2 ‐2 ‐1 0 ‐2 ‐3 3 2 0 4 5 6 2 3 4 1 ? No Yes Yes ? Yes
Concept Ranking Survey The purpose of the survey was to examine which concept that the members of the ABB workers believed was the best. The workers were asked to rate the concepts qualities using a scale from 1 – 5 points (see Figure 4-8a). The qualities were presented as statements for example “ the concept has a low production cost” or “the concept expresses robustness”. 5 points meant that the concept agreed very well with the statement. 1 point meant that the concept disagreed badly with the statement e.g. the concept was given 1 point if it was believed to become costly or was perceived as a weak structure. The concept with the highest total score was interpreted as the most popular among the staff at ABB. The survey revealed which qualities of each concept were most appreciated e.g. even if two concepts received similar scores they were appreciated for different types of qualities. A table of all the numbers of “highest score votes” that the concepts were given for every statement is presented in Figure 4-8b. 49
Concept A Concept B Concept C Concept D Concept E Concept F Highest scored concept :
Participant A Express robustness
Simple to manufacture
Low production cost
Take little of space
Participant B Express robustness
Simple to manufacture
Low production cost
Take little of space
Figure 4‐8 a and b) The result of the concept ranking survey was summarised in two tables.
The result from the survey was summarised; Concept F was the concept given the highest score from 3 out of 4 participants and therefore had the highest total score of 93 points. Concept F also received the highest number of votes, with the highest score. The other concepts had comparable scores, but e.g. Concept C (79 points) was mainly believed to reduce the stress. Concept B (79 points) expresses robustness, and Concept D (85 points) reduces stress and takes little space. Concept E (80 points) is also believed to take little space and was believed to be simple and inexpensive to manufacture. While Concept A (81 points) was believed to have a low production cost and be simple to implement and manufacture. The conclusions from the survey were that Concept F was believed to be the most appropriate concept solution for ABB, mutually by the designer and the participating ABB workers. Concept A and E, that earlier had been questioned by the designer after the concept screening were in the other hand according to the survey highly rated among the ABB workers. Concepts C that was rated high earlier in the concept screening shared the lowest rank together with Concept B in the survey. Therefore, the evaluation of the manufacturing possibilities for each concept would play a significant role in the final selection. The original sizes of the tables in Figure 4-8 can be found in Appendix C.
Evaluation of Manufacturing Possibilities A concept becomes useless, no matter how good it seems to be if it is not possible to manufacture or implement the concept in the production process of ABB Vietnam. The evaluation of manufacturing possibilities was carried out at the weld-workshop of ABB Vietnam where the welders were delegated to decide whether it would be possible to produce the concept solution ideas or not. Possible actions to enable and enhance the weld performances were discussed. It was very rewarding to turn to experts who had knowledge of the welding techniques and had good insights in the assembly procedure of ABB. There and then it became apparent which of the concepts were possible to manufacture, see Table 4-4. 50
Table 4‐4 Table of manufacturing possibilities of each concept. Concept
Possible to assemble/weld
Comments: Spot weld used
Spot weld used
Contact supplier and ask if possible
Weld onto the fin walls before assembly. The space required between the fin wall sections depends on the fin depth e.g. depth 330 mm minimum 100 mm space.
A continual vertical weld along the fin edges is required. Weld cause heat, risk of deformation. Would be difficult to paint inside after the box is sealed. Spot weld used
Concept A, B, D and F are possible to manufacture and assemble. Concept B however is a bit more complicated due to the narrow space inside the transformer e.g. it requires that the beam on the inside is attached to the back of the fins before the assembly of the tank. Concept D is possible to assemble but requires more finesse. It requires that the beam on the inside first is attached to the tank-bottom and tank-top parts before the fin wall sections are assembled onto the tank. Concept E is not possible to manufacture. Hence, the external support solution is not possible to weld onto the transformer without causing material meltdown that would deform the fins. The concept that could not be manufactured became automatically eliminated. Concept C requires changes to the original fin wall design which is entirely up to the fin wall supplier to determine if this is possible. Unfortunately, it was not possible to enquire the fin wall supplier, all attempts of communication was in vain, and therefore the manufacturing possibilities for Concept C were not possible to examine.
22.214.171.124. Further Essential Aspects In addition to the information from the weld-workshop, further information was revealed during the evaluation process about some of the concepts. The information was given by Mr. Lam, Design Manager Distribution Transformers at ABB Vietnam (personal contact, 13 April 2014). Even though Concept B and D were possible to weld and manufacture at the factory, Mr. Lam observed that the support solutions consisting of U-beams placed on the inside of the tank wall unfortunately affect the safety distance negatively. The safety distance is the distance between the tank wall and the active part in the tank, see Figure 4-9. Consequently, it is only possible to keep the distance and manufacture Concept B and D if the tank volume is increased. An increase of the tank volume means an increase of material and oil volume, which would be very expensive to implement (Mr. Lam, Design Manager Distribution Transformers at ABB Vietnam personal contact, 13 April 2014). The 45° degree cut-out of the fin design of Concept C adds multiple steps in the manufacturing process. The extra steps would require the supplier to change both tools and maybe even machines to be able to change their standard fin wall design according to Mr. Lam. It is highly unlikely that the fin wall supplier would, or could change their production process customised for ABB. Regardless; it would be very costly for ABB. 51
TANK WALL x mm x mm
Figure 4‐9 The black rectangle represents the transformer tank, and the blue part is the active part of the transformer seen from above, x mm represents the safety distance between the tank wall and the active part.
Conclusions The concept selection required many methods as decision matrices (Johannesson et al., 2013, p 180) mixed with intuition and knowledge to eliminate the number of concepts gradually. A combination of these methods made it easier to see the bigger picture and step by step all but one concept were eliminated. Through the elimination matrix, all concepts were in need of further investigation and information except Concept F that agreed well with all the statements. Through the concept screening matrix, all concept ideas were valued to various criteria and given a net value that was compared with a datum (an already existing solution). To get a second opinion a survey among the workers at ABB was performed to rate all the concepts after criteria. The most important moment of the concept selection was to revise the manufacturability of each concept which was necessary to complete the evaluation and make a selection. It was concluded that Concept E could not be manufactured, and Concept C, B, and D would be too expensive to manufacture. Only two concepts were left to choose between; Concept A and Concept F. The final decision making can be summarised as: The favoured Concept F did meet the specification very well. Concept F was first ranked in the survey as well as in the concept screening and passed the elimination matrix flawlessly. The credibility of Concept A had been questioned, and the concept was never ranked highly. Concept A also clashed with both the need (N) “to take little space” (it increase the circumference) and the wish (W) “to not have any projecting parts that can get hurt during lifts”. Concept F had a vertical support and therefore was believed to withstand vertical impacts better than Concept A that mainly strengthens the structure during horizontal impact. Concept F had not only proven to be the best concept but was also considered as the most popular and trusted of them all and was therefore selected for further development.
FINAL DESIGN CONCEPT The final design concept, concept F, had been evaluated and selected for further development. It was believed to be robust, to reduce stress concentrations, and was supposed to be simple to manufacture. The design was adapted to function with the appearances and production of distribution transformers at ABB. The structural strength was tested in various simulations with finite element method (FEM) to verify if the new design solution worked as expected.
Detail Design The final design concept is an external support frame welded to the transformer’s tank bottom and its fin wall sections, see Figure 5-1. The solution is aimed to give the transformer support and stabilise the structure when subjected to road impacts during transport. The solution increases the attachment of the fin wall onto the tank, which should generally make the structure more stiff and robust. The solution is attached to the transformer before the transformer is corrosion treated with paint. The intentionally designed space between the fin wall and the console gives access to the weld below the fin wall section during painting and welding, or if re-weld is required.
Figure 5‐1 The support solution is attached to the bottom of the fin wall sections and onto the tank of a 4 MVA distributions transformer of ABB (Lisa Magnusson, 2014).
The Frame Design The steel frame (Figure 5-2) consists of four steel flat bars joined as a rectangular frame, and L-shaped consoles are welded to the flat bars that are spot welded to the bottom of the fin walls.
Figure 5‐2 The support frame seen from a top perspective, four iron flat bars are joined as a rectangle and four consoles are placed on each long side and a minimum of two consoles on the short sides (Lisa Magnusson, 2014).
The four flat bars used in the frame are 10x100 mm and have two lengths (one size intended for the long side and one for the short side of the transformer) customized for each size of DTRs. The console is a type of L-shaped steel flat bar, see Figure 5-3a. Ideally, would be if one standard size of the console is suitable for all the varieties of distribution transformers in need of support (transformers between 1.5 MVA – 5 MVA). However, a console developed intended for a large transformer may be “too big” for smaller transformers as the fin depth varies. The efficiency of the consoles ability to reduce stress may also vary depending on the position of attachment along the fin wall. The optimal dimension of the console for a 4 MVA transformer was examined with the help of an FEM-analysis in Chapter 5.2.3.
Adaptable Dimensions Both the positions and the dimensions of the consoles can be adapted for transformers with an unusual appearance. For instance, under the fin walls on the short side of a 4 MVA DTR lashing lugs in shape of consoles are placed (Figure 5-3b) which differs from the regular design of smaller DTRs. Consequently, the support consoles could not be placed on the short side between the lashing lugs. Instead, the consoles were positioned on the edge of the long side to not interfere with the lashing lugs. Consequently, the length dimensions of the console on the short side must be increased as the longer example of the console in Figure 5-3a.
Figure 5‐3 a and b) The console on the short side (bottom) is longer than the original console design (top), hence a 4 MVA DTR have lashing lugs on the short side (Lisa Magnusson, 2014).
The solution is attached to the transformer before the transformer is corrosion treated with paint. The intentionally designed space between the fin wall and the console gives better access to the weld during painting and welding, and if re-welding becomes necessary.
The Intention of the Final Design Concept The transformer was believed to be subjected to high stress concentrations, which are aggregated by extreme road conditions (vibrations and impacts) significant for Asia. The high stress concentrations are believed to be harmful to the sealing weld along the long side of the transformer. The high stress concentrations were believed to be due to a combination of the self-weight of the transformer and the hydrostatic pressure on the inside of the tank caused by the oil (Hung, 2013). The intention of the support frame was to both reduce the mentioned stress concentrations by increasing the attachment (help the structure to carry itself) but also to distribute the stress concentrations away from the weld. The stresses in the knots along the weld would then be carried from the fins to the flat bar and distributed among the consoles. The efficiency of the support frame can be confirmed by strength calculations and preferably some physical tests that simulates and subjects the transformer or a CAD-model of a transformer to conditions similar to transport conditions in Vietnam.
Test and Verification This far into the development process the new solution was believed to work, based on theory and engineers’ credence. Therefore attempts with various analyses were performed with FEMsimulations to be able to verify and quantify the efficiency of the new design solution further. With the help of the FEM-analysis, it was proved that the attachment of the fin wall is too weak. Hence, when the console was attached to the structure, and the attachment of the fin wall was increased, the critical stress concentration along the weld became reduced. However, the outcome also showed that the attachment of the bracket to the tank wall instead was exposed to high stress concentrations. The stress on the attachment of the console in the tank wall can also be reduced by developing the console further and change the shape of the L-shaped console. A summary of the performed FEM-analysis is described in these chapters. The first simulation of the FEM-analysis was conducted in order to confirm that the FEMsimulation resembled stress concentrations similar to reality, see Chapter 5.2.1. To prove that the new design solution would reduce stress concentrations in the weld along the fin wall further FEM-analysis were performed, see Chapter 5.2.2. An attempt to optimise the design was performed, see Chapter 5.2.3. The aim of the optimisation was to investigate which design of the L-shaped console was the best by varying different factors like the design dimensions of the console. A second attempt to optimise the console design was performed (see Chapter 5.2.4) hence the stress concentration in the attachment of the optimised console remained critical. Factors as thickness, quantity, and chamfers were added to the console design to confirm the possibilities to reduce the stress concentration in both the weld along the fin wall and the attachment of the console.
Verify with FEM-analysis In this chapter, a summary and the conclusions of the FEM-analysis are presented. The procedure and results can be found in Appendix D. Three simulations in the first FEM-analysis were conducted in order to confirm that the FEM-simulation resembled stress concentrations similar to reality. For the simulation a simplified shorter section of the original CAD-model of a 4 MVA ABB transformer was built with solid elements, see Figure 5-4.
x z y
Figure 5‐4 A simplified shorter section of the original CAD‐model with three main parts.
The shorter and the simplified CAD-model was then exposed to three different types of load cases, three different forces representing loads that the transformer is subjected to in reality, to hydrostatic pressure, impact load and gravity. The first two load cases were typical linear static analyses, where Load case 1 subjected the model with self-weight (gravity load), and Load case 2 subjected the model to both self-weight and pressure. The pressure symbolises the hydrostatic pressure due to oil in the transformer. Load case 3 was a simplified dynamic analysis where the transformer section was subjected to inertia load of 2 G (2 times gravity load) in the travel directions, see Table 5-1. Table 5‐1 Load cases with respective constraints and loads Load case 1: Load case 2:
Load case 3:
Self‐weight (1G) in vertical direction (Y‐direction) + Constraint Self‐weight (1G) in vertical direction (Y‐direction) + Pressure (Hydrostatic Pressure) + Constraint Self‐weight (1G) in vertical direction (Y‐direction) + Pressure (Hydrostatic Pressure) + Inertial load in Z‐travel direction (2G) + Constraint
126.96.36.199. Conclusions from the FEM-analysis in Appendix D The outcome of the first FEM-analysis was: The simulation with Load case 1 (Figure 5-5a), indicated that the self-weight of the transformer alone did not affect the transformer with any high stress levels. The stress result from this simulation had comparable levels to the results of previously simulations made by Hung (2013),
which is positive. When the stress result behaved similar, it showed that the simplified CADmodel worked and received stresses like the original CAD-model of the transformer would. The simulation with Load case 2 (Figure 5-5b) attained stress result on the bottom notches of the fins along the weld that exceeded the yield stress in the material, which corresponds to the leakage of the transformer tank in reality. It seems like, the weld is subjected to high stress concentrations from a combination of the self-weight and the internal pressure. This assumption agrees with the conclusion of the earlier stress analysis performed by Hung (2013). However, the stress results with Load case 2 in this FEM-analysis (and the earlier stress analysis by Hung (2013)) were too severe. They implies for example that an oil filled transformer’s weld would get damaged due to the hydrostatic pressure of the oil combined with the self-weight of the transformer even when the transformer stands still. Which does not correspond to reality. The conclusion was, therefore, that the FEM-simulations (and the simulation of Hung (2013)) give too excessive results. The simulation with Load case 3 (Figure 5-5c) proved that a transformer tank which initial stress concentrations in the weld (from Load case 2) would more or less crack if it was further exposed to extreme transport conditions. Meaning, if inertial load in the travel direction was added to the loads of self-weight and hydrostatic pressure). The inertial load of 2 G that was used is a very high amount and is not very likely to occur, but the damages if so happen would be disastrous for the transformer. It was decided to continue to use the criteria of 2 G as a dynamic load because it was used in the stress-analysis of Hung (2013).
Load Case 1
Load Case 2
Load Case 3
Figure 5‐5 Stress distribution of the CAD‐model subjected with a) Load case 1, b) Load case 2, c) Load case 3.
It was concluded that the FEM-analysis did not give actual levels of stress because the analysis was simplified analysis. However, the stress distribution was believed to behave correctly or as expected e.g. the high peak stress occur in the bottom notches along the fin wall that agreed accordingly with the actual leak area in reality. Therefore, it is possible to add new concept solutions to the CAD-model and examine and compare if the stress levels are increased or decreased by performing the same FEM-analysis. Load case 3 was chosen to be the load case used for further stress analysis, hence of all simulations it was evident that Load case 3 provoked the highest amount of peak stress. 58
Analysis of the New Design Solution This chapter presents a summary and the conclusions from the performed analysis of the new design solution. The full procedure and results can be found in Appendix E. The aim was, with the help of the FEM-simulations, to verify the efficiency of the new design solution. The analysis of the new design solution was performed in the same way as earlier FEM-analysis, with the same constraints and loads as in Load case 3. The CAD-model had a section of the support solution attached (see Figure 5-6). The stress was measured along the fin wall weld (green line in Figure 5-6) and in the attachment weld below the console (red line) in Figure 5-6. The simulation results were then compared to earlier simulation result from Appendix D, to see if any stress was reduced in the case when the console was added, see Figure 5-7.
WELD OF FIN WALL
WELD OF CONSOLE
Figure 5‐6 The support frame was added to the simplified CAD‐model and the result of stress was measured along the green line and the red line.
188.8.131.52. Conclusions from the FEM-analysis in Appendix E The stress concentrations in the weld along the fin wall was both reduced and redistributed when a console was added to the construction. The maximum stress concentration found in the weld along the fin wall was now believed to be harmless, see Figure 5-7b.
Figure 5‐7 a) The left figure shows the stress result “before reinforcement”, b) figure to the right shows the stress result “after reinforcement”. Blue coloured areas have zero stress levels and red colours have stress levels of 250 MPa or more.
The simulation confirms the theory that it is possible to analyse how any reinforcement would influence the stress levels, positive or negative, which was the main goal to achieve with the simulation. Even though the stress level in the weld along the fin wall was harmless it was believed that the area, where the console is attached to the tank wall, was instead exposed to a high amount of stress. This must be considered a risk, see Figure 5-8.
HIGH STRESS AREA
Figure 5‐8 The high stress in the attachment of the console are critical. The blue colour represent zero stress levels and the red colour represent stress levels above the yield stress of the material.
Figure 5-8 shows that the support frame redistributes a lot of the stress from the fin wall to the attachment of the console. Meaning, there is a risk of cracks in the tank-bottom wall where the consoles are attached. The final conclusion was that further design development of the support solution was needed and the aim should be to reduce the stress furthermore even at the console weld.
Optimising with Full Factorial Design Experiment This chapter presents the first attempt to optimise the console design. The goal was to optimise the console design, in a systematic way, and obtain a robust console design, which used as little material as possible. The aim was to reduce stress concentrations in the weld along the fin wall section to levels under the yield stress of the material. There was a need for the optimisation because the first design dimensions of the console had been randomly chosen. It was believed other variations of the design dimensions of the consoles e.g. a longer or shorter console could provide a better result with further reduced stress concentrations in the transformer. Two experiments, one-factor-at-a-time experiment, and a full factorial design experiment were performed accordingly to the design experiments from the book Quality, from customer needs to customer satisfaction by Bergman and Klefsjö (2010). The experiment was to change the shape of the console design systematically and then test the shape in a reality test or as in this case with FEM-analysis simulating the reality. Bergman and Klefsjö (2010) talk about the importance of the use of experiments early in product development that can provide knowledge of relevant design parameters, and they describe design experiments as means of obtaining robust products. The full factorial design experiment is useful when there are two or more design factors interacting e.g. two length dimensions of the console. When two factors are interacting the level of one factor can influence the effect of changing the other factor in a so-called one-factor-ata-time experiment, but in a full factorial design experiment all the combinations of the factors are considered (Bergman and Klefsjö, 2010). The one-factor-at-a-time experiment is described in Chapter 5.2.4. An explanation of the full factorial design experiment of the console design is presented below. In the full factorial design experiment four length factors of the console’s dimensions were examined (Factor L1, L2, L3, and L4), see all examined factors in Figure 59.
(T1= Thickness) (Q1= Quantity)
Figure 5‐9 The examined factors of the console were the lengths L1‐L4, number of consoles Q1 and the thickness of the console T1.
The console thickness (Factor T1) and the console quantity (Factor Q1) were tested separately in the one-factor-at-a-time experiment to provide a rough estimation if any stresses could be reduced by increasing the consoles thickness or by increasing the quantity of consoles, see Chapter 5.2.4. Two possible values, a high (+) level, and a low (-) level were chosen for each factor, see Table 5-2. Table 5‐2 The levels of the factors.
4 consoles /long fin wall side 8 consoles/ long fin wall side
All possible combinations of the four length factors L1, L2, L3, and L4 were systematically considered, and 16 different console shapes were developed, see Figure 5-10. If all length factors (L1, L2, L3 and L4) are kept at a low level then the console dimensions looks like consoles no. 1 in Figure 5-10. Another example is console no. 16 in Figure 5-10 where all the factors are kept at a high level.
Figure 5‐10 The 16 different console shapes. Console no. 1 keep a low level of all factors and is the smallest of them all meanwhile no.16 keep a high level of all factors and is the largest one.
There were 16 so-called runs (simulations) performed to test the performance of each console shape from Figure 5-10. The simulations of the 16 different L-shaped consoles were performed, in the same way, as the earlier FEM-analysis of the console in Appendix E, using the same constraints and loads as in Load case 3. The stress was also measured, in the same way, as in earlier FEM-analysis, along the fin wall weld and the weld under the console (green line and red line in Figure 5-6).
184.108.40.206. Stress Results from the Full Factorial Design Experiment The different console shapes received various results of stress concentrations in the weld along the corrugated tank and the attachment of the consoles. The stress result from each run is summarised in Table 5-3. The results of the experiment with the 16 different shapes of consoles were also compared to the initial stress results from the original console design, see run no. 0 in the same table (Table 5-3). The level of the four length factors in the original design in run no. 0 were set to zero hence the initial dimension values are between the high and the low levels.
Table 5‐3 The design matrix of the full factorial design experiment. The stress results from the simulations for each run is in the right columns. Run no. 0 (Original design) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
0 + + + + + + + +
0 + + + + 0*, (-) + + + +
0 + 0*, (+) + + + + + +
0 + + + + + + + +
Peak stress results in weld [MPa] 168 252 275 155 140 275 390 105 197 120 170 110 160 105 122 179 110
Peak stress results in weld console [MPa] 1150 397 1470 999 900 1028 981 908 912 853 792 968 740 713 600 650 750
* Run 6 and 10 with zero (0) change in factor L3 and L2 retained the original design dimensions, hence it was not possible to change all factors at the same times and still keep the L‐shaped console design. The symbol inside the brackets shows what the supposed level should have been.
Observe! The amount of stress from the simulations should not be seen as actual because the CAD-model and the FEM-simulations are simplifications of the reality. The CAD-model used in the performed simulation was only a simplified section of the original CAD-model of an ABB transformer. Nevertheless, the occurred stress in the simulation was believed to behave in a realistic way even if it may not have been realistic stress values. Henceforth, it was a possible to analyse if the stress levels increased or decreased with the different console shapes. The result from the design matrix table shows that the high factor levels in run no. 16 are more preferred than the low factor levels in run no. 1, hence console no. 16 contributes to a much 63
lower level of stress along the fin wall weld than compared to console no.1. However there are other runs with low stress results, which also use less material. Further calculations were necessary to be able to conclude systematically which shape out of the 16 consoles was the most preferred. The stress result from the 16 different consoles was examined in a systematic way by calculating and comparing the estimated effect of each factor that is further described in the chapter below.
220.127.116.11. Estimation of the Interaction Effect According to Bergman and Klefsjö (2010) factors as the console’s length dimensions affect the stress result with different effects. It is possible to estimate the average effect by raising the value from a low to a high level for each factor. It is also possible to estimate the interaction effect between factors (which is designated e.g. L1xL2 or L1xL2xL3), i.e. the effect of increasing one factor is affected by the level of the other one (Bergman and Klefsjö, 2010). The effect was calculated accordingly to the book Quality, from customer needs to customer satisfaction by Bergman and Klefsjö (2010). The calculated effects are shown in the chart of Figure 5-11.
Figure 5‐11 The effect of each factor and interactions.
By calculating the effects of each factor, it was possible to interpret which factor affected the stress result most, and thereby it was possible to better understand what the best console shape should look like. The conclusions about the estimated effects are: The dominating factors and interactions of the console dimensions are in following order factor L2, the interaction L2xL3 and L1xL2, as well as the factor L3. This means that factor L2 has the most dominant effect on the stress results in the weld; when factor L2 is kept at a high level (+) it contributed to lower stress concentrations in the weld. By analysing the effect it became apparent that the lowest stress result arises along the fin wall weld when the console’s factor L2 and factor L4 are kept at a high level (+). Factor L1 should 64
be kept at the opposite level to factor L3, as the console design no. 7 (Figure 5-12a) and no. 13 (Figure 5-12c), see Table 5-4. Table 5‐4 Extract from Table 5‐3 The design matrix of the full factorial design experiment. Run no. 7 11 13
Factor L1 +
Factor L2 + + +
Factor L3 + -
Factor L4 + + +
Peak stress results in weld [MPa] * 105 110 105
Peak stress results in weld console [MPa]* 908 968 713
*Observe! The amount of stress from the simulations should not be seen as actual because the CAD‐ model and the FEM‐simulations are simplifications of the reality.
Console no. 7 from run no. 7 utilises less material than console no.13 in run no. 13. However, console no. 13 had the lowest peak stress result in the weld below the console, which is also highly prioritised. If saving material is a higher priority, then run no. 11 (Figure 5-12b) is the best alternative.
Figure 5‐12 a, b, and c) represent the console shapes no. 7, no. 11 and no. 13.
The three console designs in Figure 5-12 were believed to be the best ones out of the 16 different shapes (Figure 5-10). Hence they required a low amount of material and reduced the stress concentrations in the weld along the fin wall to stress levels below the yield stress of the material. However, there was still harmful stress concentrations created in the console. Therefore additional analysis were performed to see if further stress could be reduce by just adding thickness to the console, increasing the quantity of consoles, or chamfering the edges.
Further Optimisation In this chapter, the second performed attempt to optimise the console design is presented. The aim was to optimise the console in a way that the stress, both in the weld of the console and in the tank, was reduced. Hence, if the high stress concentrations in the attachment of the consoles was not reduced then cracks and leakages could take place instead. 65
The remaining design factors of the console, factors as thickness, quantity, and chamfers were added to the console design. Chamfers (cut edges) were also added see Figure 5-13a and b. The console thickness (factor T1) see Figure 5-14a, and the number of consoles used (factor Q1) see Figure 5-14b were tested in a so-called one-factor-at-a-time experiment by Bergman and Klefsjö (2010). In a one-factor-at-a-time experiment one factor is changed and then the result is analysed (with FEM-analysis) to examine if the stress have been reduced or not. The factors were also changed simultaneously e.g. see Figure 5-14c. The simulations of the remaining factors used Load case 3 and were performed as the earlier FEM-analysis of the console in Appendix E. The stress was also measured in the same way as earlier FEM-analysis, along the fin wall weld and along the weld under the console (see example in Figure 5-6). The initial console design was attached to the CAD-model, and the mentioned factors were applied to the initial console design.
Figure 5‐13 a) The first chamfer is a 45 ° cut‐out of the edge and in b) is the second chamfer added and some material removed.
As mentioned, a chamfer was added to the edge of the console in Figure 5-13a, and a second chamfer was added close to the console weld in Figure 5-13b. Material of the console was removed in Figure 5-13b as an attempt to reduce the stress in the material by lowering the material stiffness.
Figure 5‐14 a) CAD‐model of one 20 mm thick console, b) two 10 mm thick consoles and c) two 20 mm thick consoles
In Figure 5-14a factor T1 was also increased from low level (10 mm thick) to high level (20 mm thick). In Figure 5-14b factor Q1 is kept at a high level, the quantity of number of consoles 66
are increased to the double amount (to 8 consoles/ long side of the transformer). In Figure 514c both factor T1 and Q1was kept at a high level, which means both the thickness and the number of consoles were increased. The result from the simulations of the remaining factors is presented more thoroughly in Appendix F.
18.104.22.168. Conclusions The outcome of the simulations is presented more thoroughly in Appendix F and summarised below: The outcome of the FEM-analysis of the remaining factors proved that it is possible to reduce the stress in the fin wall weld further by adding thickness to the console and increasing the number of consoles. However, the stress reduction was not enough and the changes, which mean extra cost and material use, cannot be considered sufficient. As mentioned earlier, it is not possible to state the actual amount of stress due to the simplification of the simulation, but the stress in the console weld was alarmingly high compared to the stress to the fin wall weld. This could mean that the console’s bottom edge would most likely give the tank-bottom wall cracks due to the high stress levels, even though both the thickness and the quantity of the console were increased. The most important outcome of the simulation with added chamfers (Figure 5-13a and b) was that it is possible to reduce the stress that is generated in the console weld by changing the shape of the console from the original L-shape to a shape that does not generate excessive stress in the weld at the consoles. Further discussions concerning the high stress in the consoles can be found in Chapter 7.2, with Further Recommendations.
Product Description The developed product is a support frame made for distribution transformers between 1.5–5.0 MVA, and its main purpose is to prevent the transformer to get damaged during transportation and avoid that any oil leakage occur. See the support frame positioned below the fin wall sections and connected to the tank-bottom in Figure 6-1.
Figure 6‐1 A transformer with the support solution of steel attached underneath the fin wall sections (Lisa Magnusson, 2014).
The solution is aimed to support and stabilise the transformer structure when subjected to various road impacts. The solution increases the attachment of the fin wall onto the tank that is believed to decrease and distribute the stress concentrations in the welds, which carry and attach the fin wall sections to the transformer tank. The details of the solution have earlier been described in Chapter 5.1. A requirement of the product was to be suitable for various sizes of transformers between 1.5 – 5 MVA. Thereby, the dimensions of the frame and its consoles can easily be adjusted and positioned for different transformer sizes or transformers with special appearances. The advantages that differs this support solution from earlier support brackets or bars are that it is primarily developed to resist vertical impacts, which is the most critical direction of impacts during transport (Singh et al., 2007). The earlier bracket solution was sometimes complicated to mount on the tank-bottom part of larger transformers, hence objects like lashing lugs etc. were placed in the way of the bracket. This problem will not occur with the welded support frame due to its flexible design and thin consoles that take little space. Compare to the bracket the design of the consoles intentionally contributes to better access to the weld below the fin wall section even after the support solution is attached e.g. if painting or re-welding is required.
Product Design Specification The final design concept was compared to the established requirements continuously during the design process and was chosen partly because of how well it agreed with the needs. Thus, the final design concept agrees with most of the established needs in the PDS in Chapter 3.7. However, some of the needs have not been completed or confirmed in this thesis but are possible to achieve in the future. For example, it has not been tested if the product withstands extreme transport conditions in a physical test because there have not been any prototype available. The same applies to the need to withstand the leak control test that has not been tested, and the length of the product life is not possible to confirm before any physical test is performed. Neither has it been confirmed if the solution meets the ISO90001/14001 certificate. The most critical need to fulfil was to reduce the oil leakage and withstand impact loads and vibrations comparable to extreme road conditions. Observe, the FEM-analysis performed in Chapter 5.2 were not advanced enough to be able to confirm if the solution meet the set values of impact load (of 1.5 gravity force in all directions) but is believed to do so. Nonetheless, the FEM-analysis did confirm that the support frame decreases the stress as expected by distributing the stress among the consoles. Vibrations tests were not included in the FEManalysis because more details are required about the set value (the exact amount of vibrations and the expected amplitudes). Further details are also needed for the set value of impact loads because the set value was believed to be unrealistically high.
Discussion About the procedure of the project The main purpose of the project was to improve the conditions of transportation of distribution transformers in Vietnam. The final result was a design improvement of distribution transformers, aimed to withstand transport conditions and make the product quality more sustainable. The project studies the characteristics of the transport conditions (Chapter 3.3.) as well as the manufacturing conditions (Chapter 3.1) in Vietnam (to partly answer RQ2). The final result could not have been performed without in depth studies of the problem areas. For example, it was necessary to investigate what to consider when designing transformers for transportation (to answer RQ1), and what the possible root-causes that affected the product failure were (to partly answer RQ3). By allocating the different root-causes and defining a solution strategy (Chapter 3.6.), it was possible to understand what can be changed to improve the problem of decreased quality of delivered distribution transformers. A decision was then made to adjust the direction of the project and to focus more particular on improving the product design and developing an external support solution (to answers RQ4). This was done with the purpose to narrow down the problem of the project, but it also excluded the possibility to focus on improving the packaging of the product. In that moment, the decision was necessary but afterwards it was possible to conclude that the packaging direction should not have been excluded so early in the developing process. There could have been interesting solution ideas combining the areas of packaging and product design. 70
Methodologies used in the project's generation phase to create creative ideas could have been better. The quantity of ideas could have been increased with further brainstorming sessions e.g. with independent people outside the company as other Product Design Engineer students. There was a great need to have access to more people who are used to creative thinking and creative methods to ask for help during the project. Hence, one can easily fixate on one's own ideas if working alone. The final step in the project’s evaluation phase was to examine the manufacturability of each concept, but this should have been examined much earlier in the project and could have saved both time and effort. Other methods used during the evaluation phase were the concept screening method and a survey among the workers. In the concept screening method all concept ideas were valued to various criteria and compared with a reference. In the survey was workers at ABB Vietnam rated all the concepts after criteria. The survey only highlighted the good qualities of each concept while the concept screening considered both pros and cons which were believed to give a more accurate and fairer result than the survey did. This project would not have been possible without the knowledge and information given from the workers at ABB Vietnam, which was the most crucial input to the thesis. This project would neither have been possible without the expertise and knowledge of the supervisors from the University of Skövde; their guidance was especially useful in the FEM-analysis section. About the criteria of the moment of damage In this thesis, the existing problem areas are defined and the transport conditions are fairly identified, but the conditions that contribute to the moment when the damage of the transformers occurs are unidentified. For instance, failure can arise when the transformer is exposed to a certain amount of impact or vibrations of a certain frequency, but the exact level or if it is a combination of both was not possible to establish during the project time. The speed of the truck as well as the type of road surfaces can also have a certain influence. It was necessary to know the criteria of the moment of damage to be able to establish what the transformer must withstand, to set the requirement, and to be able to develop a design solution as efficient as possible by testing the product against the criteria. However this was not possible to verify. Instead, the criteria were estimated, which consequently affected the possibility to prove the efficiency of the final result. Meaning that the final design concept is believed to work in theory but its efficiency have not been assured and will not be until it has been tested against proper input criteria of the moment of damage. In the report, the FEM-analysis was an attempt to illustrate how to model the behaviour of transportation adequately in real life situations (and answer RQ5). However, only a simplified simulation of impacts was performed where vibrations were excluded. The actual level of stress could not be calculated due to the simplification but also as mention due to the absence of the criteria of when the damage occurs. Nonetheless, the result from the FEM-analysis showed that even with less advanced FEM equipment like Creo Parametric 2.0 the outcome could become useful. With the help of the simulations, comparisons could be made if the final concept could reduce stress levels or not.
About RQ6 – Physical testing The final design concept has not been tested in a physical test because there was no time to develop a prototype. Thereby the last research question (RQ6) about how to prove the efficiency of the product with physical testing remained unanswered. If the criteria of the moment when the damage occurs had been available from the start of the project there would have been more time to put aside, and a higher priority, to develop and test a prototype. A contact person with expertise about road conditions, and how these affect transportation means would have been very helpful and would add value to the project. For example, there must be people in the car industry working daily with exposing various car chassis to vibration and impact tests. Test inputs and etc. from those kinds of tests could have been crucial for the project. About the console design The report suggests how to optimise the design of a product in a strategically way with the full factorial design experiment but as mentioned the console could been further improved and optimised. Due to lack of time the optimisation of the consoles was not performed to its final stage. The full factorial design experiment was helpful to develop the best L-shaped consoles, but too late in the project it was realised that an L-shaped console was probably not the best shape of consoles. It is not believed to be the optimal shape of console because the L-shaped console is attached perpendicular to the tank wall creating a 90 degree angle between the console and the wall that creates too much stress concentration in the tank wall material below the console.
Further Recommendations Recommendations for simulation and testing An identification of those conditions that contributes to the moment when the transformers get damaged would be of interest for further investigations. They are essential inputs to perform realistic FEM-analysis or physical tests of the support solution. Perform further FEM-analysis that are more reliable than the ones performed in this report, by model a full transformer instead of one simplified section. Use the criteria from the moment when the damage occurs as an input and more reliable stress values would most likely appear. If this criteria is not available, perform simulations as the simulation process of Hung (2013) suggests and subject the transformer model to self-weight, hydrostatic pressure and add dynamics (especially in the vertical direction). Prefferably, also include vibrations in the simulation to get a better understanding of the behaviour of the transformer. Further recommendations are to perform physical tests of the final concept. Firstly, cooperate with people with expertise about how to perform test adequately to road conditions and secondly, subject the transformer in a test lab under controlled circumstances to conditions that would in theory damage the transformer. Analyse if the final concept withstands or not.
The earlier studies mentioned in Chapter 3.3 about performed test of road vibration in Thailand and India provides input values that could be used in advanced vibration tests. They could be useful even for ABB Vietnam as those road characteristics examined in the studies are comparable to roads in Vietnam. Another physical test that is possible to perform is to implement the support solution in the production of transformers that are shipped and distributed during e.g. a six month period and analyse the statistics of failure. Then the final concept solution would be subjected to the transformers transport environment. After the test period, ABB should have convincing statistics if the percentage of transformer with leakage failure was reduced or not. Recommendation about the final concept Further development of the console shape is recommended. So that the final console design reduces the stress concentrations to levels under the yield stress in both the weld along the fin wall and under the console. The solution was developed and adapted to prevent leakages at the weld located below the fin wall that is the most common leakage location. The support solution could as well be attached on top of the fin wall, to prevent leakage along the weld above the fin wall section if necessary. The consoles would then be attached to the tank-top part and on top of the fin wall. Recommendations about how to increase quality of transformer One of the recommendations of how to prevent the decreased quality in Chapter 3.6.1 was to always fully weld the inside of a tank and have a double weld for overlapping sections which is not always accomplished at ABB Vietnam. This was recommended because an overlapping weld could reduce the stresses on the weld on the outside of the fin wall section. When talking to welders at ABB Vietnam, it was only possible in theory to weld the inside, hence in practice more equipment is needed at the workshop. To weld the inside of transformers is hard work and could also be unsafe due to the narrow space inside the tank. It is unsafe because the fin walls have a greasy layer of oil that can cause flames. The worker would need to stand inside the transformer tank to weld the inside and therefore better equipment such as air hoses are needed to provide air draft to these narrow spaces to prevent fires. Another option is to hang the transformer tank upside-down on the lifting crane at the workshop and weld the inside, this would be much safer. However, it takes time to weld so there would be a need for another lifting crane at the workshop to lift and move other transformers at the same time. It is therefore suggested that ABB Vietnam investigates the possibilities to provide the equipment needed to perform higher quality work on the transformers.
Sources of Errors
A lot of the information in the report is based on the information from interviews with workers at ABB Vietnam and it could, therefore, be room for misleading information about ABB.
The FEM-analysis in the report was based on the previous simulations performed by Hung (2013) to compare the result but, of course, there is a small chance that their simulation is incorrect.
The values for the pressure in the FEM-analysis and the initial stress analysis, were taken from the simulation by Hung (2013), who tested a 2 MVA transformer while beam length and the CAD-model in the FEM-analysis was a simplification of a 4 MVA. This means that the pressure levels applied in both the initial stress analysis with hand calculations and in the FEM-analysis were too low and possibly affected negativly the stress results. This was first noticed after the FEM-analysis was performed.
References ABB. (2014a) About ABB. [online]. Available at: http://new.abb.com/about [Accessed 16 December 2014]. ABB. (2014b) ABB for Government Power and Energy Solutions. [pdf]. Available at: http://www05.abb.com/global/scot/scot299.nsf/veritydisplay/b1a4638fd3aa516a852 57ce800698152/$file/government%202%20pager%20(2).pdf [Accessed 16 December 2014]. ABB.
(2014c) ABB Liquid-filled - Transformers. [online]. Available at: http://www.abb.com/product/fi/9aac30300162.aspx [Accessed 16 December 2014].
ABB. (2010) ABB supports the smooth and efficient operation of Shanghai World Expo. [online]. Available at: http://www.abb.com/cawp/seitp202/d5280d096b3cc471482576b20028170b.aspx [Accessed 16 December 2014]. ABB. (2004) Transformer handbook. ABB Power Technologies Division. Ident. no. 1LAC000010 – 04. [online]. Available at: http://dotorresg.files.wordpress.com/2011/12/abbtransformerhandbook.pdf [Accessed 18 November 2014]. ABB. (2003) Distribution transformer handbook. ABB Power Technologies Division. Ident. no. 1LAC000003. [online]. Available at: http://data.biitbook.com/111118798-ABBDistribution-Transformer-Handbook.pdf [Accessed 18 November 2014]. Bergman, B., and Klefsjö, B. (2010) Quality from customer needs to customer satisfaction (3rd). Lund: Studentlitteratur AB. CBSA. (2012) Statement of reasons: concerning the making of a final determination with respect to the dumping of certain liquid dielectric transformers originating in or exported from the republic of Korea. Ottawa. [online]. Available at: http://cbsaasfc.gc.ca/sima-lmsi/i-e/ad1395/ad1395-i12-fd-eng.pdf_br [Accessed 4 December 2014]. Chonhenchob, V., Singh, S. P., Singh, J. J., Sittipod, S., Swasdee, D., and Pratheepthinthong, S. (2010) Measurement and analysis of truck and rail vibration levels in Thailand. Packag. Technol. Sci., 23: 91–100. DOI: 10.1002/pts.881. [online]. Available through: Wiley InterScience website www.interscience.wiley.com [Accessed 17 March 2014].
Cross, N. (1994) Engineering design methods (2nd). Chichester: John Wiley and sons Ltd. Embassy of Sweden. (2012) Om Vietnam (About Vietnam). SwedenAbroad.com [online]. Available at: http://www.swedenabroad.com/svSE/Ambassader/Hanoi/Landfakta/Om-Vietnam/ [Accessed 2 December 2014]. Hung. (2013) Find out root cause of oil leakage and bring out a modified design to avoid oil leakage- MODEL: 2000 KVA. [pdf]. AES Vietnam Co., Ltd. (unpublished). Hurst, K. S. (1999) Engineering design principles. New York: John Wiley and Sons Inc. Jarimopas, B., Singh, S. P., and Saengnil, W. (2005) Measurement and analysis of truck transport vibration levels and damage to packaged tangerines during transit. Packag. Technol. Sci., 18: 179–188. DOI:10.1002/pts.687. [online]. Available through: Wiley InterScience website www.interscience.wiley.com [Accessed 17 March 2014]. IMF.
(2011) Consultative group meeting for Vietnam. [online]. Available at: http://www.imf.org/external/np/dm/2011/120611.htm [Accessed 4 March 2014]. IndexMundi.com. (2014a) Vietnam geography profile 2014. [online]. Available at: http://www.indexmundi.com/vietnam/geography_profile.html [Accessed 2 December 2014]. IndexMundi.com. (2014b) Vietnam – Roads. [online]. Available at: http://www.indexmundi.com/facts/vietnam/roads [Accessed 2 December 2014]. Johannesson, H., Persson, J.G., and Pettersson, D. (2013) Produktutveckling –Effektiva metoder för konstruktion och design (Product development - Effective methods of construction and design). Poland: Liber AB. Kołakowski, Z., Kubiak, T., and Mania, R., (2008) Badanie wytrzymałości kadzi TRAF0 1600 – analiza modalna (Strength test TRAF0 1600 - modal analysis). Opracowanie K-12 PŁ. Nr I-112/231/08 NB. Łódź. [pdf]. (unpublished). Lindell, K. (2005) Ingenjörshandbok (Engineer handbook). LTH Ingenjörshögskolan vid Campus Helsingborg. [pdf]. Available at: http://lup.lub.lu.se/luur/download?func=downloadFile&recordOId=1329009&fileOI d=1329010 [Accessed 24 November 2014]. McDonald, J.D., Wojszczyk, B., Flynn, B., and Voloh, I. (2013) Distribution systems, substations, and integration of distributed generation. In: M.M. Begovic (ed.), Electrical Transmission Systems and Smart Grids: Selected Entries from the Encyclopedia of Sustainability Science and Technology (p. 7-68). DOI 10.1007/978-1-4614-5830-2_2. New York: Springer Science+Business Media. 76
[pdf]. Available on Internet: http://www.springer.com [Accessed 18 November 2014]. Mind Tools. (2014) Plus, minus, interesting - Weighing the pros and cons of a decision. [online]. Available from: http://www.mindtools.com/pages/article/newTED_05.htm [Accessed 16 April 2014]. NationMaster.com. (2014a) Vietnam people stats. [online]. Available at: http://www.nationmaster.com/country-info/profiles/Vietnam/People [Accessed 2 December 2014]. NationMaster.com. (2014b) Vietnam transport stats. [online]. Available at: http://www.nationmaster.com/country-info/profiles/Vietnam/Transport#2013 [Accessed 2 December 2014]. Olsson, F. (1995) Principkonstruktion (Principle design). Lund: Lunds Tekniska Högskola. Institutionen för Maskinkonstruktion. Pahl, G., and Beitz W. (1995) Engineering design: A systematic approach (2nd). London: Springer-Verlag. RC PERRET. (2013) The run of civets: CIVETS countries transport infrastructure overview. [pdf]. Available at: http://rcperret.com/netcat_files/121/71/The_Run_of_Civets.pdf [Accessed 4 March 2014]. Singh, S. P., Sandhu, A. P. S., Singh, J., and Joneson, E. (2007) Measurement and analysis of truck and rail shipping environment in India. Packag. Technol. Sci., 20: 381–392. DOI: 10.1002/pts.764. [online]. Available through: Wiley InterScience website www.interscience.wiley.com [Accessed 17 March 2014]. Ulrich, K., and Eppinger, S. (2008) Product design and development (4th). New York: McGraw-Hill Publishing Company. Wikipedia.org. (2014a) History of Vietnam. [online]. (Last modified on 24 November at 15.33, 2014). Available at: http://en.wikipedia.org/wiki/History_of_Vietnam [Accessed 4 December 2014]. Wikipedia.org. (2014b) Electrical substation. [online]. (Last modified on 8 October at 08:49, 2014). Available at: http://en.wikipedia.org/wiki/Electrical_substation [Accessed 10 October 2014]. Wright, I. C. (1998) Design methods in engineering and product design. Berkshire: McGrawHill Publishing Company.
Yonglitai. (2012) Application and development trend of air suspension. Foshan: Yonglitai Axle Co., Ltd. [online]. Available at: http://www.ltcmc.com/blog/article/509.html#.UzdZWCKCgI [Accessed 28 August 2014].
Reference List of Figures © ABB. Internal archive, all photos copyright to ABB. ABB. (2014d) Medium distribution transformers (316-2,499 kVA) - Distribution Transformers [image]. Available at: http://new.abb.com/products/transformers/distribution/medium [Accessed 27 November 2014]. Alibaba.com. (2014) CD1 Wire rope electric hoist/ hoist machine, view electric hoist, bdslo product details. [image]. Baoding Sanlong Crane and Hoist Machinery Co., Ltd. Available at: http://hbsanlong.en.alibaba.com/product/608048471219346150/CD1_Wire_Rope_Electric_Hoist_Hoist_Machine.html [Accessed 30 November 2014]. Amazon.com. (2014) Knape & Vogt 183L ANO 10 183 series 10" heavy-duty left flange bracket, anochrome. [image]. Available at: http://www.amazon.com/Knape-VogtHeavy-Duty-Bracket-Anochrome/dp/B0006FKOC8 [Accessed 30 November 2014]. Chonhenchob, V., Singh, S. P., Singh, J. J., Sittipod, S., Swasdee, D., and Pratheepthinthong, S. (2010) Measurement and analysis of truck and rail vibration levels in Thailand. [photograph]. Packag. Technol. Sci., 23: 91–100. DOI: 10.1002/pts.881. [online]. Available through: Wiley InterScience website www.interscience.wiley.com [Accessed 17 March 2014]. Cross, N. (2008) Engineering design methods (2nd). [illus.]. Chichester: John Wiley and sons Ltd. p. 30. General statistics office of Vietnam. (2007) Press release: socio-economic statistical data, 1st Quarter in 2007. [online]. Available at: http://www.gso.gov.vn/default_en.aspx?tabid=487&idmid=&ItemID=5842 [Accessed 2 December 2014]. IACS, (2005) Container ships: Guidelines for surveys, assessment and repair of hull structures. [illus.]. International Association of Classification Societies. London: Wtherbys Publishing. p.73. [online]. Available at: http://www.iacs.org.uk/document/public/Publications/Guidelines_and_recommendat ions/PDF/REC_84_pdf233.pdf [Accessed 2 December 2014]. 78
Jarimopas, B., Singh, S. P., and Saengnil, W. (2005) Measurement and analysis of truck transport vibration levels and damage to packaged tangerines during transit [photograph]. Packag. Technol. Sci., 18: 179–188. DOI:10.1002/pts.687. [online]. Available through: Wiley InterScience website www.interscience.wiley.com [Accessed 17 March 2014]. Kołakowski, Z., Kubiak, T., and Mania, R.. (2008) Badanie wytrzymałości kadzi TRAF0 1600 – analiza modalna (Strength test TRAF0 1600 - modal analysis). [images]. Opracowanie K-12 PŁ. Nr I-112/231/08 NB. Lodz. [pdf]. (unpublished). Lindell, K. (2005) Ingenjörshandbok (Engineer handbook) [illus.]. LTH Ingenjörshögskolan vid Campus Helsingborg. [pdf]. Available at: http://lup.lub.lu.se/luur/download?func=downloadFile&recordOId=1329009&fileOI d=1329010 [Accessed 24 November 2014]. Lisa Magnusson. (2014) All photos copyright to Lisa Magnusson. Power Energo. (2011) Řada amorphous - abb - produkty (Many amorphous - abb - products) [image online]. Power-energo s.r.o. Available at: http://www.powerenergo.cz/produkty/rada-amorphous-abb.html [Accessed 27 November 2014]. Singh, S. P., Sandhu, A. P. S., Singh, J., and Joneson, E. (2007) Measurement and analysis of truck and rail shipping environment in India. [photograph]. Packag. Technol. Sci., 20: 381–392. DOI: 10.1002/pts.764. [online]. Available through: Wiley InterScience website www.interscience.wiley.com [Accessed 17 March 2014]. Ulrich, K., and Eppinger, S. (2012) Product design and development (5th). [illus.]. New York: McGraw-Hill Publishing Company.
Appendix A – Stress Analysis of Weld This appendix presents complete calculations with real values applied in the equations from the performed stress analysis presented in Chapter 3.4. Load case: Simply supported beam with linearly distributed load, with maximum minimum :
SIMPLY SUPPORTED BEAM
Weld thickness = the thickness of the steel sheet
Figure A‐1a) An illustration of the load on the inside of the transformer due to hydrostatic pressure, b) a similar load case as at the inside of the transformer is the simply supported beam with linearly distributed load (Lindell, 2005).
Designations: : beam length : Minimum load case : Maximum load case : Reaction force in support A : Reaction force in support B : Weld area : Thickness of the steel sheet of the fin b: length of the weld along one fin
Constants: 0.00099MPa (value based on Hung (2013)) 0.15MPa (value based on Hung (2013)) 25 1500 1.5
Conditions: 1 1 /
Equations: ∙ ∙
The reaction force ( following equation:
) affecting the bottom weld of the tank was calculated with the help of
2 0.00099 ∙ 25 The weld area ( ∙t
2 ∙ ∙ 7,5
2 ∙ 0.15 ∙ 25
was calculated with the help of following equation:
25mm ∙ 1.5mm
The sought stress in the bottom weld ( calculated as: , .
) caused by the hydrostatic pressure was
The answer is; the amount of stress in the weld due to hydrostatic pressure is 50 MPa.
Appendix B – The Initial Ideas Solution idea no. 1. Solution idea no. 1 in Figure B-1 could replace the existing round 8 mm bar on the bottom on the edges of the transformer. Instead of the rod a stronger bar such as an angled iron steel bar could be around the transformer. An angel iron bar in the bottom instead of the rod would stiffen and stabilise the transformer more. The beams could only be mounted with spot welds on the fins like the round rods is attached today, which is not the most preferred weld technique because it is less durable. However the angle bar maybe collect water in the corner when it is raining and may rust.
Solution idea no. 2 Idea no. 2 is a modification of the 10x10 mm rod solution used today around the transformer. It is believed that the 10x10mm rod is weak, even hand force could bend it. To add rigidity a stronger steel beam (Figure B-2) in any shape should replace the rod. A beam is stronger, stiffer, and it will prevent any fluctuation and keep the fin wall in a fixed position. If the beam takes too much space the possibility to only attach it during transport with nuts and bolts could be revised. Figure B‐2
Solution idea no. 3 A beam on the outside of the transformer as in solution idea no.2 increases the circumference. A beam (preferred U-shaped) placed on the inside would preferably not affect the transformer size. Idea no. 3 is a beam attached with spot weld along the inside of the fin wall (Figure B-3). Optionally, if an angled iron bracket could be used in the corners so the U-beam can be attached to these then the weld surface and attachment would be increased and more rigid. The transformer would be stronger vertically with the angled iron. Hence it joins the tank-bottom and tank-top part further.
Figure B‐3 a), inside view of idea no. 3, b) top view, c) side view (cross section y‐y).
Solution idea no. 4, no. 5 and no. 6 The purpose with idea no.4, no.5, and no. 6 is to modify the design of the fin wall and reduce the stress by reducing the material. If the material is reduced in the edge of the fin, then the material stiffness becomes less stiff and the stress in the weld is reduced. This technique is used for reparation of hull structures (Mr. Dan Magnusson, Service Manager at TTS Marine VN personal contact, 9 March 2014). The edged of the fin could have less amount of stress if the fin wall material is reduced 1525mm from the weld see Figure B-4. Idea no. 4 (Figure B-4a) is a simple and quick industrial solution with straight cuts. A cut with a radius as in idea no.5 (Figure B-4b) is more preferable. The transformer needs to be able to expand and increase the volume capacity with 5%. With the cut-out too much volume is reduced and then compensation is necessary e.g. by lengthen the height or the depth of the fin wall. To avoid to reduce too much volume the idea no. 6 have a cut-out semi-circle where as much volume of the fin wall is kept (Figure B-4c). To change the design of the fin wall section the manufacturing process has to be considered. The cutting must occur either before or after the metal sheet is folded in the fin wall folding machine. Furthermore, the welding process may require other tools or techniques for edges with radius. 83
The additional cut-out on the new design adds at least one new step or change in the production process, which is very difficult but at this moment of the project the ideas were allowed to be theoretical. Theoretically, with idea no. 4, no. 5 or no.6 there is no need to reinforce the transformer with any brackets or rods afterwards hence no high stress concentrations exists.
Figure B‐4 Modifications of the original fin design.
Solution idea no. 7 and 8 Idea no. 7 (Figure B-5a) and no. 8 (Figure B-5b) are modifications of the original transformer design. These ideas are based on the theory that the stress level is higher in the middle of the fin wall section on the long side. This means that the weld have higher stress concentrations in the middle of the section than on the sides. To reduce the stress in the middle, the fin wall section on the long side is divided into two shorter sections. A U-beam (as in idea no.7) or a flat bar with reinforcement (as in idea no. 8) is attached between the sections. The bar or beam unite the tank-bottom and tank-top together which make the tank more stable, and the fin walls are no longer the only carrying section of the structure. The welding area of the fin wall sections is also increased. In idea no. 8, the reinforcement of the flat bar stands out from the transformer tank in the same directions as the fins and could contribute to cooling the oil. In idea no.7 is the U-beams reinforcement towards the inside. To be able to weld the edges of the sections onto the flat bar or the beam a large space between the separated sections is required to reach with a weld. A large gap means downscaling of the number of fins, and removing too many fins can in turn affect the entire tank expansion capability, and compensations of the volume may be required.
Figure B‐5a) solution idea no. 7, b) the fin wall is divided into two sections, c) solution idea no.8.
Solution idea no. 9 Idea no. 9 (Figure B-6) is an external support solution attached to the transformer by weld flat bars on two adjacent fins as an enclosed “box”. The so-called “box” reinforces the structure, make the structure stiffer in the middle of the fin wall that reduce the stress concentrations and prevent the fin wall to bulge out. The box also unites the tank-bottom and tank-top together. The side view demonstrates how there is room for the weld between the flat bar and the fins if the flat bar is chamfered. Top view demonstrates how the flat bar on the outside of the fin wall can be further reinforced with another flat bar perpendicular attached (Figure B-6b). The box must be airtight so no humidity impacts the inside, to prevent rust. To make the enclosed space airtight can be done with different methods, e.g. throw a burning material inside the box and then quickly seal the box or by useing a certain powder that removes the air and creates vacuum in an enclosed area (Mr. Dan Magnusson, Service Manager at TTS Marine VN personal contact, 9 March 2014). As Figure B-6c suggest, there can be more than one box along the wall.
Figure B‐6a) Solution idea no. 9, b) (top) top view and (bottom) side view, c) two “boxes” on one fin wall side.
Solution idea no. 10 Solution idea no. 10 is an external support solution and inspired by the currently used bracket on ABB transformer, developed by AESC and AES. The idea is to support the fin wall, especially during vertical impacts, and increase the attachment of the fin wall onto the tank by increasing the amount of weld area. The bracket used today along the fin wall bottom overlapping the fins and is only attached with spot welded. The idea with solution no.10 is to use a more continuous weld along the fin edge and the tank bottom. A continuous weld contributes to a more durable attachment. A console attached to a fin increase the stiffness of the wall similar to vertical stiffeners. Three consoles or more are needed on the long side (Figure B-7a). The steel console must be sufficiently rigid and preferably have a radius in the corners (Figure B-7b). A support attached parallel with the fins gives much better access to the weld for painters and welders (much more than the current bracket).
Figure B‐7 a) A few (about 3) consoles would be needed close to the middle of the fin wall to enhance attachment, b) the size of console may differ depending on different sizes..
The dimensions and the shape of the console can be optimised further with the help of advanced strength calculation or FEM-analysis. See examples of different consoles shapes in Figure B8).
Appendix C – Evaluation Table C‐1 The Elimination matrix from the evaluation and concept selection chapter (Chapter 4.3.1.).
Solves main problem
Within the cost range
Easy to manufacture
Suitable for ABB
Elimination matrix Elimination criteria [+] Yes [‐] No [?] More information needed [!] Check with specification Decision [+] Continue with / Keep [‐] Eliminate solution [?] Find more information [!] Check with specification
+ total width
+ to much space of the inside volume?
C D E F
+ + + +
! + ! +
? ? ? +
‐ + + +
? + ? +
+ the fin‐wall supplier + Is it possible to assembly + Is it possible to seal the box without air? +
Maybe take to much space of transformer
Is it possible to manufacture. Does it take
Manufacturing possibilities depends on
!,? ? !, ? +
Table C‐2 The concept screening matrix from the evaluation and concept selection chapter (Chapter 4.3.2.).
Criteria Reduce/ distributes stress Withstand impact load (vertical) Low production cost Realistic solution Take little of space, do not increase the total circumference Low weight Simple to manufacture Weld is in reach for welder, paint reach every inch Total amount of + Total amount of 0 Total amount of ‐ Net Value Ranking Further development
Concepts B C 0 + ‐ 0 0 ‐ ‐ +
DATUM 0 0 0 0 0
A ‐ ‐ 0 0 ‐
0 0 0
‐ + +
0 4 87
D + + ‐ +
E 0 0 0 0
F 0 + + +
+ + 2 1 5 4 1 4 2x0 3x0 1x0 2x0 6x0 4x0 ‐4 ‐4 ‐2 ‐2 ‐1 0 ‐2 ‐3 3 2 0 4 5 6 2 3 4 1 ? No Yes Yes ? Yes
Table C‐3 Result from the concept ranking survey performed by four workers at ABB Vietnam (Chapter 4.3.3.).
Highest scored concept :
Statement : Express robustness Reduce stresses Simple to manufacture Low production cost Realistic solution
4 4 5 5 4
5 5 3 3 4
3 5 2 3 4
4 5 3 3 3
4 4 3 4 4
5 5 5 4 4
Take little of space
Express robustness Reduce stresses Simple to manufacture Low production cost
5 4 2 3
Participant B 5 2 4 2 1 5 3 1
5 5 4 3
2 2 2 1
3 2 2 3
Take little of space
3 3 5 4 2
Participant C 4 2 4 5 3 2 3 1 3 3
Express robustness Reduce stresses Simple to manufacture Low production cost Realistic solution
4 4 3 1 4
3 4 4 4 3
4 5 4 5 5
Take little of space
Express robustness Reduce stresses Simple to manufacture Low production cost Realistic solution Take little of space
4 4 3 3 3 3
Participant D 4 5 5 5 3 2 3 2 2 4 2 5
4 4 3 3 2 2
4 3 4 4 2 5
4 4 4 4 3 5
Table C‐4 The “Highest score votes” table (Chapter 4.3.3.). Highest ranked qualities: Number of highest score votes Concept A
Low production cost
Simple to manufacture
Total Score : 81
4 times highest score
Low production cost
Total Score: 79
5 times highest score
Simple to manufacture
Take little of space
Total Score: 79
10 times highest score
Low production cost
Take little of space
Total Score: 85
8 times highest score
Concept E Simple to manufacture
Low production cost
Take little of space
Total Score: 80
4 times highest score
Simple to manufacture
Low production cost
Take little of space
Total Score: 93
11 times highest score
Appendix D – FEM-analysis To be able to verify and quantify any efficiency of the new solution, there was a need for a FEM analysis comparable to the earlier mentioned analysis provided by the company AES and AESC (Hung, 2013) in Vietnam. Therefore, the simulations presented were conducted in order to confirm that the FEM-analysis resembled stress concentrations similar to reality. A thorough demonstration of the FEM-analysis, and the method and the results are explained below.
Method First the FEM model was subjected to a linear static analysis, to self-weight (see Load case 1 in Table D-1). Then pressure was added to the FEM model to symbolise the hydrostatic pressure from the oil inside the transformer (see Load case 2 in Table D-1). Furthermore, inertia loads in the travel directions were added to represent a simplified dynamic analysis (see Load case 3 in Table D-1). The stress result was then analysed and compared to the result of the analysis provided by Hung (2013) to gain perception of the result. Model The computer aided program (CAD) used for the FEM analysis was Creo Parametric 2.0. The CAD-model used in the simulations for this report was a simplified section of the original CADmodel of a 4 MVA ABB transformer with a corrugated tank (fin wall) built with solid elements, see Figure D-1a.
Figure D‐1 a) The CAD‐model of the transformer tank section used in the FEM‐analysis b) front view, c) back side, d) side view.
The CAD-model used a high quality “ triangular” (tetra) shaped element mesh with 24504 elements, see Figure D-2. 90
Figure D‐2 A high quality triangular shaped element mesh.
Material properties The material of the transformer tank are painted carbon steel grade SS 400 (JIS), see the material properties in Table D- 1. Properties like elastic module and poisons ratio were set up for the steel material in material settings in the CAD program.
Table D‐ 1. Material properties of SS 400
Description Values Elastic modulus E [GPa] Poisons ratio Yield stress [MPa] Fatigue limit [MPa] Ultimate tensile stress: UTS [MPa]
210 0.3 250 227 400
Constrains The tank wall section is rigidly locked in the directions of the blue arrows in Figure D-3. The edge where the arrows are located in the bottom is very stiff relative to the tank wall section. Hence, the bottom edge of the CAD-model is aligned to the tank bottom. The edge of the tank-top part in the CAD-model is in reality aligned to a corner of the tank lock, so the edge is stiff relative to the side of the plate. The fin wall section is attached to the two tank wall plates similar to how a weld would join them.
z x y
Figure D‐3 A demonstration of constraints of the CAD‐model and how they are applied.
The three different load cases used for the simulations are presented in Table D- 2. Table D‐ 2 Load cases with respective constraints and loads
Load case 1: Load case 2: Load case 3:
Self‐weight (1G) in vertical direction (Y‐direction) + Constraint Self‐weight (1G) in vertical direction (Y‐direction) + Pressure (Hydrostatic Pressure) + Constraint Self‐weight (1G) in vertical direction (Y‐direction) + Pressure (Hydrostatic Pressure) + Inertial load in Z‐travel direction (2G) + Constraint
The pressure load in Load case 2 and 3 was copied from the analysis by Hung (2013) which was calculated from the density of the oil and the volume of a 2000 KVA transformer tank. This means that the pressure inside the 4 MVA transformer should be slightly higher, but the simulation is a rough approximation and hence a precise stress value was not sought, and this amount of pressure was accepted. The hydrostatic pressure is 1.5 kPa on the top and increases linearly with y-axis along the height L m. In the bottom is the pressure 150 kPa, see Figure D4. In Load case 3 the inertial load of 2G is a simplification (and possibly an exaggeration) of a travelling dynamic case in z-direction (the travel direction) e.g. if a transformer stands on a 92
bed of a moving truck and the truck breaks hard then the inertial load of the steel structure. Then the oil inside the transformer will continue to have dynamic motion in the traveling direction. The loads in Load case 3, vertical gravity load and the inertial load in the travel direction and the hydrostatic pressure are displayed in Figure D-4.
Figure D‐4 Load arrows are placed on the backside of the model to represent the situation of the hydrostatic pressure on the inside of the tank wall, also the load of 1G in gravity direction and of 2G in traveling direction.
Results Observe, all stress mentioned in this report is based on von Mises stress theory. The stress result at the leakage zone in the weld was measured along the curve (the green line) in Figure D-5a, and the stress amount was then showed as a chart as in Figure D-5b.
Stress von Mises
Figure D‐5a The stress is measured along the green line (the weld), b) the result of stress is shown in a Length vs. Stress chart.
There was three different load cases (see Table D-2) used in the performed simulations of the FEM-analysis and the three stress results of Load case 1, Load case 2, and load case 3 are presented below. Result of Load case 1: Self-weight (1G) in vertical direction (Y-direction) + Constraint Stress distribution result due to self-weight is shown in Figure D-6.
Figure D‐6 The stress distribution from the FEM analysis with Load case 1.
The maximum stress result in the CAD-model due to self-weight (Load case 1) was 103 MPa (harmless). The dark blue colour represents zero stress in the material. The magnified image shows stress result on leakage zone. The highest value (peak) read from the chart of the stress amount along the weld was at most 55 MPa. The value of the stress between the peaks (the fin wall notches) varied between 15-25 MPa. Result of Load case 2: Self-weight (1G) in vertical direction (Y-direction) + Pressure (Hydrostatic Pressure) + Constraint Stress distribution result due to self-weight and pressure is shown in Figure D-7.
Figure D‐7 The stress distribution from the FEM analysis with Load case 2.
Maximum stress result in the CAD-model due to Load case 2 was 664 MPa. The stress result of the transformer tank exceeded the yield stress of the material. The dark blue colour represents zero stress in the material, and the red colour represents a stress level of 250 MPa. The highest value read from the chart of the stress amount along the weld was at most 300 MPa. The value of the stress between the peaks (the fin wall notches) varied between 70-110 MPa. Result of Load case 3: Self-weight (1G) in vertical direction (Y-direction) + Pressure (Hydrostatic Pressure) + Inertial load in Z-travel direction (2G) + Constraint Stress distribution result due to self-weight, pressure and inertial load is shown in Figure D-8 8.
Figure D‐8 The stress distribution from the FEM analysis with Load case 3.
Maximum stress result in the CAD-model due to Load case 3 was 1434 MPa. The stress result of the transformer tank exceeded the yield stress of the material. The dark blue colour represents zero stress in the material, and the red colour represents a stress level of 250 MPa. The highest value read from the chart of the stress amount along the weld was at most 750 MPa. The value of the stress between the peaks (the fin wall notches) varied between 200-225 MPa. The peak stress exceeded the ultimate tensile stress of the material, but this is not possible because the FEM program only consider a linear elastic range of material. In reality the material would reach plasticity after 400 MPa.
Conclusions Load case 1 (Figure D-6) The simulation with Load case 1 indicated that the self-weight of the transformer, alone, does not affect the transformer with any high stress levels (see Figure E-6). The stress result of the simulation with self-weigh got similar levels as the results in the previously simulation with self-weight made by Hung (2013) although their simulation was more advanced and precise. This is an affirming result. When the simplified CAD-model received similar stress result to the analysis of Hung (2013), it was looked at as proof that the FEM-analysis worked and was performed correctly.
Load case 2 (Figure D-7) In the simulation with Load case 2, the peak stress result reached above the yield stress level (see Figure D-7). The areas of stress concentrations in the CAD-model agreed to the leakage areas of the transformer in reality. The stress results from Load case 2 were compared to the performed analytical stress analysis presented in Chapter 3.4. The rough hand calculations gave a stress amount of 50 MPa while the stress levels in the simulation (located between the peak stresses) were between 70110 MPa. The analytical result was expected to be less, hence, only the hydrostatic pressure was considered as a load. However, if the stress results from Load case 1 (15-50 MPa) was theoretically added to the hand caluclations result of 50 MPa then the combined level of stress would be approximatley 50+50 MPa, which is consistent with the stress levels located between the peak stresses in Load case 2. With the hand calculations only the average stress can be defined and not the peak stress, so if the peak stress levels in load Case 2 were actual levels could not be determined. In theory, if the simulations peak stresses are neglected temporarily then it seemed like the average stress created from self-weigh and hydrostatic pressure would be harmless to the weld along the fin wall. If the peak stress levels in Load case 2 instead are considered as actual, then it seems like a load combination of self-weight and hydrostatic pressure gives the transformer harmful stress concentrations in the weld. The later assumption agrees with the stress result established in the previous stress analysis by Hung (2013). However, the peak stress results due to Load case 2 implies that an oil filled transformer weld would recive damages when the transformer just stands still, which does not correspond to how it is in reality. Therefore, the final conclusions made from the simulation with Load case 2 were that the FEM-simulations in this report (as well as the simulations by Hung (2013)) gives too excessive peak stress results. Load case 3 (Figure D-8) The simulation with Load case 3 (see Figure D-8) proved that a transformer tank that initially had existing stress concentrations in the weld (from Load case 2), would definitely broke if it was exposed to extreme transport conditions as well (if inertial load in the travelling direction was combined with self-weight and hydrostatic pressure). Observe, the inertial load of 2 G that was used in the simulation is a very severe amount of inertial load, and it is believed that the damages would be disastrous for the transformer if it is subjected to this amount of inertial load in the reality. It was decided to use the criteria of 2 G as a dynamic load because it was used in the stress-analysis of Hung (2013). The level of amount of stress in Load case 3 cannot be confirmed as true. It is believed, as mentioned that the FEM-simulations gives too excessive peak stress results. Also the simulation of Hung (2013) are doubted. The FEM-analysis did not give actual stress result levels, but the stress distribution is believed to behave correctly or as expected e.g. the high peak stress occur in the bottom notches along the fin wall. Therefore is it possible to add the concept solutions to the CAD-model and examine and compare if the stress levels are increased or decreased by performing the same FEM-analysis.
Appendix E – Analysis of the new design solution The simulation runs in Appendix D were conducted in order to confirm that the simulations in the FEM-analysis resembled stress concentrations similar to reality. The simulation runs in this chapter were prioritised to confirm if the added support solution reduced the stress concentrations in the weld or not. Also to analyse if it the stress was enough reduced or if further design development was necessary. A thorough demonstration of the simulation, where the method and the results are explained below.
Method After the first simulations in Appendix D it was possible to test if the new design solution, the added support frame (with the flat bars and consoles), could lower the stress in the weld to a level that was below the yield strength. A section of the support frame was added to the simplified CAD-model (see Figure E-1). 80 mm
Original Console Design
Figure E‐1 The new design solution (left), with the original console shape with respective dimensions to the right, the thickness of the flat bar and the console is 10 mm.
The analysis of the solution was performed in the same way with same constraints and loads as in Load case 3 (self-weight + hydrostatic pressure + inertial load in Z-travel direction (2G) + constraints). The stress result was measured along the fin wall weld (green line) and in the attachment weld below the console (red line) in Figure E-2.
WELD OF FIN WALL
WELD OF CONSOLE
Figure E‐2 The result of stress measured along the green (the fin wall weld) and the red line (the console weld).
The amount of stress was read from charts as Figure E-3.
Stress von Mises
Weld length of console 0 mm
Figure E‐3 the measured stress along the 10 mm long console weld (red line in Figure E‐2) was shown in a Length vs. Stress chart
Result The stress distribution after the support solutions was added, due to self-weight, pressure and inertial load on the weld, and the console weld can be seen in Figure E-4.
Figure E‐4 Stress distribution of FEM analysis of new design solution and load case 3.
The stress at the weld along the fin wall got reduced to harmless levels and the fin wall bottom notches closest to and above the console had lowest amount of stress (see dark blue area) compared to the notches furthest away from the console (see magnified image in Figure F-4). This proves that the flat bar and the console distribute the stress from the fin wall notches in an effective way to the tank wall instead. However, the maximum stress result was now located at the weld at the bottom edge of the console, it had a peak stress of 1150MPa. The red area in Figure E-4 has stress levels above the yield stress of the material and is likely to damage it. The stress distribution before and after a console was added to the transformer section was compared, see Figure E-5.
Figure E‐5 The left figure shows the stress result “before reinforcement” and figure to the right shows the stress result “after reinforcement”. Blue coloured areas have zero stress levels and red colours have stress levels of 250 MPa or more.
The peak stress in the weld was reduced from 750 MPa (left) to 168 MPa (right) in Figure E-5. However, the amount of stress from the simulations cannot be taken as accurate because the CAD-model and the FEM-simulations are simplifications of the reality. If the stress amount exceeds the yield stress of the material can, therefore, not be confirmed because the CAD-model and the FEM-simulations are simplifications of the reality. For instance, the amount of applied load representing pressure and its distribution was only an approximation. However, the stress distribution was believed to behave correctly or as expected. 99
Appendix F – Simulation with Remaining Factors This appendix presents the result from the simulations of the remaining factors more thoroughly, see Table F-1. Observe, the presented amount of stress levels in Table F-1 cannot be accounted as precise only comparison can be made.
Table F‐1 The result from the simulations with the remaining factors T1, Q1 and chamfers are presented as Run A, B, C, D, E and F. RUN A: Factor involved + Level: T1 Low, 10 mm console Q1 Low, 4 consoles/side Peak stress in weld; 168 MPa Peak stress in console; 1150 MPa CAD-model:
RUN B: Factor involved + Level: T1 High, 20 mm console Q1 Low Peak stress in weld; 180 MPa Peak stress in console; 737 MPa CAD-model:
RUN C: Factor involved + Level:
RUN D: Factor involved + Level:
T1 Low Q1 High: double amount/ side
Peak stress in weld; 150 MPa Peak stress in console; 1107 MPa CAD-model:
T1 High Q1 High
Peak stress in weld; 130 MPa Peak stress in console; 667 MPa CAD-model:
First Chamfer added
Peak stress in weld; 160 MPa Peak stress in console; 990 MPa CAD-model:
Second “Chamfer” added and some of the material is removed
Peak stress in weld; 170 MPa Peak stress in console; 721 MPa CAD-model:
In Run B factor T1 kept a high level and in run C kept factor Q1 a high level, see Table F-1. Run A represents the initial simulation of the support solution with low levels of both Factor T1 and Q1. In Run D both factor T1 and Q1 increased to a high level which means both the thickness and the number of consoles was increased. In run E chamfer was added to the edge of the console and in Run Fa second chamfer was added close to the console weld, material was also removed in an attempt to reduce the stress in the material by lowering the material stiffness. The outcome of the simulations is: The stress result in the fin wall weld from the simulation in run B was similar to the stress levels in run A, even though factor T1 was changed from low to high values (from 10 mm thick console to 20 mm thick console). The stress in the console weld however was reduced but still critically high. The same applies about the simulation in run C, the stress reduction was insignificant when increasing the amount of consoles to the double amount. The cost, time and material of extra consoles would thereby not be worth the change. The stress concentrations were reduced at the most in run D, where both the consoles thickness increased to 20 mm and the console quantity was doubled, but the material cost and manufacturing cost increased as well. Attempts to reduce the high stress in the weld of the console were performed by adding chamfers to the console. The console in run E had a 45° degree cut-out of the material in the end of the console but unfortunately not any significant stress was reduce. In run F an extra chamfer was added and the stress result in the weld along the console was reduced but it was still critical, see Figure F-1.
Figure F‐1 FEM‐analysis with Load case 3 and two added chamfers, some material have been removed from the original console design.