Application of fuzzy logic-based Eco-QFD for a disconnecting switch

Abstract

The importance of human life and the continuing endeavours to enhance its quality have sparked great interest in the improvement of design and manufacture of engineering materials and components in order to be eco-sustainable. The need for environment-friendly manufacturing technologies is essential and vital in today's global state of affairs. But such eco-friendly products do not sustain in the markets though they are environmentally conscious. Customers' aspirations and cost aspects are not prioritised while developing such products. In this context, this article reports a study in which Eco-quality function deployment (Eco-QFD) has been applied for a 95 KV disconnecting switch used as a part of electrostatic precipitator in power plants. Fuzzy method is applied to Eco-QFD to overcome the vagueness and uncertainty and to attain optimal balance between customer satisfaction and environmental concerns. The practical benefits reaped from the study have been presented.

Keywords:

• eco-sustainability
• fuzzy logic
• quality function deployment
• Eco-QFD
• disconnecting switch

1. Introduction

Sustainability in its various forms has, will and continue to be the governing criteria in industries. Needs promote changes and the need for introducing environmental considerations in product design and development is the need of the hour. Resource optimisation and environmental impact have become the focus of government agencies. The ever-growing need for environmental certifications and demand for eco-label products necessitate the setting up of environmentally friendlier technologies and assessment of existing technologies. Several environmental impact assessment tools (Rapoza et al. 1996) and models (Hui et al. 2002) have been developed to assess the environmental impact associated with products. Some of them include Health Hazard Assessment and Scoring (Barzilai and Golany 1994), Sustainable Process Indices (Sage 1993), Toxic Release Indices (Horvath et al. 1995), Life Cycle Analysis and Design for Environment. However, the question remains on the application of environmental criteria in the product design. Design modifications for environmental purpose should not compromise on the traditional design requirements such as function, cost and quality (Fiksel 1996). Also the constraints associated with the inclusion of environmental criteria have to be tackled and related costs have to be reduced. This is very important in view of the participation of companies in manufacturing eco-designed products in sustaining profits. The receptivity of such eco-designed products has not been favourable in the markets. This may be due to the fact that they are more focused on Environmental Impact Analysis rather than customer requirements (CRs). The key issue for a successful eco-designed product is not only to meet environmental objectives but also to take into account cost effectiveness, market demand and multi-functionality requirements (Lee et al. 2001). Quality function deployment (QFD) is used to incorporate customer needs into product design concepts, product planning and process planning to achieve total customer satisfaction (Akao 1990, Clausing 1994, Chan and Wu 2002–2003). QFD, in general, is a means to translate CRs into appropriate engineering characteristics during each stage of product design and development. Traditional QFD consists of product planning, parts deployment, process planning and production planning. QFD has been applied in design engineering, engineering management, teamwork, production costing, etc. Few researchers have contributed towards the introduction of environment aspects into QFD: Green quality function development (Green QFD) (Cristofari et al. 1996), (GQFD II; Zhang et al. 2000) and QFD for environment (QFDE; JEMAI 2001, Masui et al. 2001). QFDE is focused on the development of environmentally friendlier products, which consists of four major phases (Masui et al. 2003). Phases 1 and 2 allow the user to identify environmentally significant components of the products. Phases 3 and 4 allow the user to select the most environment-friendly decision from several alternatives. The unstructured nature of QFD structure possesses certain drawbacks. The weighting factors are described using vague linguistics. Customer information is gathered in a subjective manner. Majority of the efforts are spent on capturing the customer voice. In order to overcome the vagueness associated with linguistic assessments, fuzzy methods are used. Fuzzy approach to QFD (Kuo et al. 2009) has been used in this article. The novelty of the study is that it reports an application of fuzzy logic-based QFD for a disconnecting switch to achieve balance between customer satisfaction and environmental acceptability. Eco-designed product development problem has been formulated as a fuzzy multi-objective model. The introduction of environmental considerations into QFD model leads to optimal solutions. Using this approach, an optimal balance between overall customer satisfaction and environmental acceptability has been obtained. The approach has been exemplified with an industrial case study of a disconnecting switch used in an electrostatic precipitator (ESP). This approach has been selected after a discussion with the industry experts.

2. Literature review

The literature has been reviewed from the perspectives of QFD and fuzzy QFD application in enabling environmental conscious design and manufacture. Dr Yoji Akao is regarded as the father of QFD and he has contributed a widely used definition of QFD (Besterfield et al. 2004). QFD provides the need for translating consumers' demands to appropriate technical requirements during each stage of a product/process development (Sun et al. 2003). It enables the development of customer-friendly and high-quality products. QFD emphasises quality in the design process to prevent the likelihood of defects at the early stages, thereby reducing cost and improving productivity (Chan and Wu 2002, Akao and Mazur 2003). Other benefits of QFD include reduced design changes, increased market share and improved market quality (Chan and Wu 2002, Akao and Mazur 2003).

Some researchers have been working on incorporating environmental aspects into QFD. Pun (2006) has presented the determinants of environmentally responsible operations and suggested that Green QFD as one of the tools for environmentally responsible operations. Cristofari et al. (1996) introduced the concept of Green QFD by integrating QFD with a life cycle approach to product development. This is useful for evaluating different product concepts and deploys environmental requirements throughout the development process. Zhang et al. (2000) developed a method called GQFD-II, which includes the integration of life cycle analysis (LCA) and life cycle costing (LCC) into QFD. They integrated LCC into QFD matrices and suggested the deployment of quality, environmental and cost requirements throughout the entire product development process to evaluate different product concepts. Masui et al. (2003) presented a concept called QFDE in which QFD has been applied to environmentally conscious design. Sakao (2009) has presented a QFD-centred design methodology for environmentally conscious product design. The author has combined LCA, QFDE and theory of inventive problem solving (TRIZ) and applied the combination to a hair dryer to effectively support the product planning and conceptual design stages.

In a traditional QFD exercise, the correlation between CRs and technical requirements as well as the importance for each CR is determined by the members of a design team using linguistic expressions (e.g. weak, average and strong). These linguistic terms are then scaled into crisp values for ranking of each alternative. This crisp assessment for correlation evaluation in QFD analysis has difficulty in coping with uncertainty among design team members (Khoo and Ho 1996). The major problem is that the assignment of crisp values cannot reflect the imprecision or vagueness inherent in these types of assessments. Accordingly, the inconsistent ranking result could be generated due to the sensitivity of crisp evaluation, where no imprecision or approximate concept is allowed (Lin and Chen 2004).

The practical approach facilitating the balance between customer satisfaction and environmental acceptability through customer voice translation incorporated with fuzzy method gains vital importance. In this context, fuzzy logic-based Eco-QFD proposed by Kuo et al. (2009) has been selected as the technique for our study.

3. Methodology

3.1 Identify CRs and technical attributes in QFD

The process of congregating and scrutinising customer needs is of vital importance in QFD for design and development of customer-oriented products. The customer needs are identified from a wide variety of sources, such as surveys, focus groups, interviews, trade shows, complaints and even expert opinions (Griffin and Hauser 1993, Gryna 2001). The needs are then translated into technical attributes (TAs). In reality, customer needs are satisfied by completing those required TAs. To integrate the environmental concerns (ECs) with the QFD, the philosophy generated by the Japan Environmental Management Association for Industry (JEMAI 2001) is used. That is, the voice of the customer consists of traditional and environmental customer needs. The marketing needs need to be gathered from the customers' perspective. But more often than not, customers find themselves unable to describe their needs precisely and clearly due to lack of information and experience. Their needs tend to be vague, subjective and qualitative. In such scenarios, expert views are preferred as they are more technical, objective and easily understandable. The voice of ECs is identified to consider the environmental requirements. Figure 1 shows the model QFD structure constructed using customer and environmental requirements.

Figure 1 QFD structure.

These requirements are then translated into appropriate technical and environmental attributes. Life cycle analysis applied from the raw material, manufacturing, distribution, use and recycling stages is employed to effectively identify the attributes. The next step involves the evaluation of the relationship between the requirements and attributes. Typically, the relationship is determined qualitatively, such as strong, medium, weak and no relation with appropriate weights by a cross-functional team (or a group of experts).

Customer needs and ECs are complicated and often imprecise. So fuzzy approaches such as fuzzy sets, fuzzy arithmetic and fuzzy defuzzification methods can be applied (Zhou 1998, Wang 1999, Vanegas and Labib 2001). Fuzzy methods can be implemented to determine relationship between needs and attributes objectively (Fung et al. 1998).

3.2 Fuzzy sets, fuzzy operations and fuzzy preference relation

Define a universe of discourse X, as a collection of objects all having the same characteristics. The individual elements in the universe X will be denoted as x. As per Zadeh (1965), a fuzzy set (subset) A in X is characterised by a membership function f _a (x) which associates with each point in X, a real number in the interval [0,1] with the value of f _a (x) representing the ‘grade of membership’ of x in A. The larger f _a (x) represents the stronger degree of membership of x in A. In classical sets, the transition for an element in the universe between membership and non-membership is abrupt and well defined or ‘crisp’. For an element in a universe that contains fuzzy sets, this transition can be gradual. This transition among various degrees of membership can be thought of as conforming to the fact that the boundaries of the fuzzy sets are vague and ambiguous. Hence, membership of an element from the universe in this set is measured using a function that attempts to describe vagueness and ambiguity. Thus, a fuzzy set is the one containing the elements that have varying degree of membership in contrast to a crisp set, where membership is assigned a value 1 (full) else 0 (Ross 1995). The height of the fuzzy set means the maximum degree of membership. If there is at least one element with the height of one, then the fuzzy set is normalised.

A fuzzy number is a convex, normalised fuzzy set A belonging to R whose membership function is at least segmentally continuous and has the functional value μ _A (x) = 1 at precisely one element. In this study, the triangular fuzzy numbers are used. A fuzzy number A in the real line R is a triangular fuzzy number, denoted by A = (a, b, c) and presented in Figure 2, if its membership function f _a (x): R → [0, 1] is equal to

Let A and B be two normal, convex fuzzy subsets of R with piecewise continuous membership function F _A (x) and F _B (y), ∀ x, y ∈ R, respectively.

Figure 2 Triangular membership function.

3.3 Operations of triangular fuzzy numbers

The fuzzy arithmetic operations of triangular fuzzy numbers are described as follows. If two triangular fuzzy numbers are A _i  = (L _i ,M _i ,U _i ) and A _j  = (L _j ,M _j ,U _j ), (A _i >0 and A _j >0), then those operators are defined as

Multiplication:

•
•	(U i U j  ≥ 0).

Trapezoidal fuzzy numbers can be employed as well. Since our study comprises the views of three experts and the fact that the triangular membership functions are the simplest, we have chosen triangular fuzzy numbers. The study does not involve precise calculations that are involved in operations of fuzzy controllers, where the type of membership function may prove to be important in reducing errors. Even in such cases, triangular fuzzy numbers have proved to be highly accurate (Pedrycz 1994).

3.4 Linguistic variables

Since a vast amount of the information in human communication involves natural language terms that, by their very nature, are often vague, imprecise, ambiguous and fuzzy, the use of fuzzy sets gains importance. Linguistic variables are the variables whose values are words or fundamental terms or phrases of our natural language, which can be expressed as very low, low, medium, high and very high. Basic variables can be defined as [0,1] in order to evaluate suitable situation of the criteria (attributes), where ‘1’ can be regarded as the optimal suitable situation and ‘0’ can be regarded as the poorest situation. Since words, in general, are less precise than numbers, the concept of a linguistic variable serves the purpose of providing a means of approximate characterisation of phenomena which are too complex or too ill-defined to be amenable to description in conventional quantitative terms (Ross 1995).

4. Case study

4.1 About the case company

The case study has been carried out at Sajas Electricals (hereafter referred to as SE), a company specialising in the design and the manufacture of process monitoring and control and instrumentation systems and related components for power plants, chemical and sugar industries located in Tiruchirappalli, Tamil Nadu, India. It is an ISO9001:2000 certified by M/s TUV SUD organisation. SE aspires to enhance environmentally friendliness in their product design and development practices. It has been realised that the integration of Eco-quality function deployment (Eco-QFD) and LCA would enable sustainability improvement at SE. In this context, this case study has been motivated to be conducted at SE.

4.2 Case study

The disconnecting switch is depicted in Figure 3.

Figure 3 Disconnecting switch.

An industrial case study of a disconnecting switch used as a part of ESP circuit has been considered. The switch is used in precipitator circuits in thermal power plants, maintenance application in isolation and grounding during maintenance of the high-voltage circuit of the ESP. An ESP is a particulate collection device that removes particles from a flowing gas (such as air) using the force of an induced electrostatic charge. ESPs are highly efficient filtration devices that minimally impede the flow of gases through the device and can easily remove fine particulate matter such as dust and smoke from the air stream. The switch is used to disconnect the field from high voltage in off condition to earth. So the onus is on making the switch eco-sustainable by improving the design and manufacturing processes.

The first step involves identifying the various technical and environmental requirements for the design and manufacture of the switch. This suffices all customer needs. These needs are weighed by experts whose weights are in the form of linguistic variables. These variables are then defined as triangular fuzzy numbers. Let us say the weight for each CR is determined by P experts. We have as the weight for the jth CR. Finally, aggregate weights through all experts by means of fuzzy addition and scalar multiplication for the jth CR to form an average value triangular fuzzy number are determined by the following Equation (Kuo et al. 2009):

Based on the LCA, the product stages of raw material, manufacturing, distribution, use and disposal processes have been evaluated and analysed. The experts have identified 20 CRs by the company's sales network and marketing surveys as shown in Table 1.

Table 1 Fuzzy result for weighing importance of each CR.

			Expert 1		Expert 2		Expert 3
Customer requirements			Crisp	Fuzzy	Crisp	Fuzzy	Crisp	Fuzzy	Weight
Price	CR1	Yield-rate reduction	7	(6,7,8)	6	(6,7,8)	6	(3,4,5)	(5,6,7)
	CR2	Cost reduction	7	(7,8,9)	9	(8,9,9)	9	(8,9,9)	(7.7,8.7,9)
Function	CR3	Corrosion resistant	8	(7,8,9)	8	(8,9,9)	8	(8,9,9)	(7.7,8.7,9)
	CR4	Withstanding impulse test voltage	8	(8,9,9)	8	(7,8,9)	8	(8,9,9)	(7.7,8.7,9)
	CR5	Withstanding power frequency test	8	(8,9,9)	8	(8,9,9)	8	(8,9,9)	(8,9,9)
	CR6	Withstanding torsion test load	8	(6,7,8)	7	(7,8,9)	8	(8,9,9)	(7,8,8.7)
	CR7	Withstanding tensile test load	8	(7,8,9)	8	(7,8,9)	8	(6,7,8)	(6.7,7.7,8.7)
	CR8	Knife contact durability	8	(8,9,9)	8	(8,9,9)	8	(8,9,9)	(8,9,9)
	CR9	Withstanding operational high voltage	8	(8,9,9)	8	(8,9,9)	8	(8,9,9)	(8,9,9)
	CR10	Ingress protection	6	(6,7,8)	9	(8,9,9)	9	(8,9,9)	(7.3,8.3,8.7)
	CR11	Pre-treatment and paint quality	5	(5,6,7)	6	(6,7,8)	7	(6,7,8)	(5.7,6.7,7.7)
	CR12	Adherence to critical dimensions	8	(8,9,9)	8	(7,8,9)	8	(8,9,9)	(7.7,8.7,9)
Appearance	CR13	Constructional aesthetics	3	(3,4,5)	5	(4,5,6)	6	(5,6,7)	(4,5,6)
	CR14	Painting aesthetics	4	(3,4,5)	4	(4,5,6)	6	(5,6,7)	(4,5,6)
Delivery	CR15	Delivery on time	8	(8,9,9)	8	(8,9,9)	8	(8,9,9)	(8,9,9)
Environment	CR16	Recycle material usage	8	(6,7,8)	8	(7,8,9)	8	(8,9,9)	(7,8,8.7)
	CR17	Toxic material reduction	4	(5,6,7)	5	(5,6,7)	6	(5,6,7)	(5,6,7)
	CR18	Waste reduction	8	(6,7,8)	6	(5,6,7)	5	(5,6,7)	(5.3,6.3,7.3)
	CR19	Waste maintenance	6	(5,6,7)	6	(5,6,7)	5	(4,5,6)	(4.7,5.7,6.7)
	CR20	Life cycle extension	4	(3,4,5)	5	(4,5,6)	5	(4,5,6)	(3.7,4.7,5.7)

These 20 CRs are further categorised into five levels with cost, function, appearance, delivery and environment. To evaluate the weight for each CR, a scale of 1–3–5–7–9 is used, where 1 means the weakest and 9 means the strongest relation. Each expert provided a crisp value (linguistic variable is converted into crisp value) for each CR. Then, the crisp value for each CR given by each expert is fuzzified. For example, the computation of weight of CR₁₀ is as follows:

This gives us the weight for each CR.

Table 1 summarises the fuzzy results of determining the weight for each CR. A total of 20 CRs were identified by the experts. The experts proceeded to value these CRs on a scale of 1–9 as explained earlier. Table 1 includes both crisp and fuzzy values. The former have been included to differentiate them from the latter. The equation at the end of Step 1 has been used to find the weights of each CR. These weights will be used in Tables 2-4.

Table 2 Eco-QFD result for the experts' ratings.

Table 3 Eco-QFD table.

Table 4 Ranking difference with exclusion and inclusion of environmental factors.

The second step involves the evaluation of the relationship between CRs and TAs. The relationship may be strong, medium or weak, represented by numbers (9, 5 and 1), while the blank shows no relationship. However, as the dimension of customers' requirements increases, the number of potentially related design variables becomes much larger. One of the main problems associated with a large relationship matrix is the difficulty in laying down priorities and importance using an accurate number. This is due to the fact that the value-setting process is typically vague or imprecise in practice. Second, the data available for product design are often limited, inaccurate or vague at best particularly when developing an entirely new product. We find ourselves with a similar problem here. To solve these problems, the fuzzy logic technique is typically used (Kuo et al. 2009). So let S be the relationship matrix with an order of n × r between CRs and TAs assuming that there are r TAs (here 19) based on life cycle analysis. The element represents the relationship between the ith CR and jth TA. The relationship between CRs and TAs, , is evaluated by a group of P experts and can be fuzzified and transformed by the fuzzy inference techniques with , where are the lower bound, average value and the higher bound, respectively (Kuo et al. 2009):

Table 2 depicts the Eco-QFD result based on the expert's ratings.

A total of 19 TAs that affect the product's performance are shown in Table 2. The relative importance of each TA with respect to each CR was valued. As in the previous case, three experts valued this inter-relationship. Only the final weight (after considering all three values) has been presented for a particular CR vs. TA relation. The calculations were done using the equations at the end of Step 2. These results are presented in Tables 3 and 4.

The third step involves evaluating the importance of each TA. This is done by calculating the rank of each TA. The ranking value between ith CR and jth TA would be determined by multiplying with . Therefore, the ranking value, denoted by A _j , for the jth TA is , where is a crisp value associated with by centre of gravity (COG) defuzzification. Aggregate fuzzy performance ratings with fuzzy weights by means of fuzzy multiplication to form an important weighted, comprehensive decision matrix E are given by (Kuo et al. 2009):

These three steps help us prioritise the various TAs. With this, the target attributes have been identified and are analysed for improvement.

Table 3 depicts the important attributes that control the design and manufacture of the switch, the ones that can be improved so as to make it as eco-sustainable as possible. Ranks give us the relative importance of the TAs.

Each TA was evaluated as shown in Table 3. The calculations were done using the equations involved in Step 3. The summarisation presents the total impact of each TA. Relative importance of each TA has been calculated from which the rank of each TA was determined as shown in Table 3. The ranking results are used in Table 4.

5. Results

From Table 3, we infer that certain attributes could be improved for overall improvement of the switch, which enable it to be economically and environmentally sustainable. Although the method only covers the part of identifying and prioritising the attributes, it is, however, an important step in implementing changes in design and manufacture. This is followed by the assessment of each attribute in an order of their ranks for effective improvements. In the present study, changes were made in materials constituting the switch and in the manufacturing stages with improved overall sustainability of the switch. For example, from Table 3, changes were made in the manufacture of sheet steel components, which reduced material wastage and recycling scrap. This method is best suited for products consisting of 10–20 components.

From Table 4, it could be inferred that certain attributes have a lesser rank with EC. These attributes have a considerable impact on the environment and have to be targeted for improvement. Such an improvement helps to minimise resource utilisation, resource wastage, increase resource recycling and minimise environmental impact as a whole. Although these methods have been adopted in certain industries in the country, there remains still a large portion of small- and medium-scale industries, in which the impact of minimisation is still low priority. These industries still use a major portion of the resources and methods have to be adopted to reduce environmental impact as a whole. The modifications made in sheet steel and copper components are presented as follows.

Sheet steel

•	Computerised, optimised CNC machinery.
•	Production of sub-components as a downstream process of main component production.
•	Produced scrap in shapes, sizes suitable for rerolling mills.

Copper components

•	Machined out of blanks. Forgings produce less scrap and recycling is trimmed down.
•	Punching holes instead of drilling holes. Punching produces buttons, which can be sub-components or recycled. Drilling produces boring scrap, which escapes to environment and pollutes.

Table 4 depicts the ranking difference with exclusion and inclusion of environmental factors.

The relative importance of each TA without considering environment considerations was calculated and the rankings are presented in Table 4. Then, rankings with and without the inclusion of environmental factors were compared and studied.

6. Industrial impact

The conduct of the case study enables the introduction of environmental awareness to CRs, assessment of environmental performance as a design objective and evaluation of the potential of product for re-use and recycling. The organisation has reaped the following benefits:

•	Better adherence to statutory regulations: the key attributes were identified and onus was on improving them with better material selection and better equipments.
•	Productivity has been improved to an extent of 5%.
•	Reduced cost of product at better quality by means of implementation of ISO14000 standard.
•	Cheaper downstream processes of effluent treatment.
•	Reduced scrap production is achieved using value stream mapping. The scrap percentage got reduced from 27 to 22%.
•	Workspace utilisation has been improved by means of ensuring proper flow of operations.
•	Improvement in international customer base has been achieved with enforced standards and adherence to regulations. The industry gains licence to produce sophisticated machinery for a number of industries by conforming to environmental certifications like ISO14000 Environmental Management Systems and ROHS compliance.
•	Better customer satisfaction.

7. Conclusion

This paper applies Eco-QFD to provide a framework for designing eco-friendly products by integrating the LCA with QFD throughout the entire product development process. The Eco-QFD framework is a useful tool to integrate not only the ECs but also the quality, cost and customer needs to improve the product design process, thus making the product economically worthwhile. By applying fuzzy theory in Eco-QFD, the product development team can overcome the vagueness and uncertainty faced in group decision-making processes. On conduct of the case study, the framework enabled the product development team to prioritise TAs and ECs so that the team can focus on utilising their limited resources on vital issues to develop customer-oriented eco-friendly products. The study has also lead to many practical benefits. This is very important in the contemporary industrial scenario for the development of sustainable products.

8. Scope for future research

The inferences derived from the conduct of the study are applicable to similar manufacturing organisations. The case study could be carried out using several tools/techniques of sustainable engineering for different manufacturing organisations across varied sectors in different countries. The methodology can be programmed and software tools could be created that help in achieving results in very quick time. The creation of such software tools will pave the way for easy, effective and quick processes.