Multi-attribute decision making green electrical discharge machining

Chia sẻ: Đăng Xuân Phương | Ngày: | Loại File: PDF | Số trang:5

Thêm vào BST

Báo xấu

54
lượt xem 3
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

(BQ) This paper aims to develop a combination of Taguchi and fuzzy TOPSIS methods to solve multi-response parameter optimization problems in green manufacturing.

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Multi-attribute decision making green electrical discharge machining

Expert Systems with Applications 38 (2011) 8370–8374 Contents lists available at ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa Multi-attribute decision making for green electrical discharge machining S.P. Sivapirakasam a,⇑, Jose Mathew a, M. Surianarayanan b a Department of Mechanical Engineering, National Institute of Technology, Tiruchirappalli, India b CISRA, Central Leather Research Institute, Chennai, India a r t i c l e i n f o Keywords: EDM TOPSIS Green manufacturing Multi-attribute decision making A high carbon high chromium tool steel a b s t r a c t This paper aims to develop a combination of Taguchi and fuzzy TOPSIS methods to solve multi-response parameter optimization problems in green manufacturing. Electrical Discharge Machining (EDM), a commonly used non-traditional manufacturing process was considered in this study. A decision making model for the selection of process parameters in order to achieve green EDM was developed. An experimental investigation was carried out based on Taguchi L9 orthogonal array to analyze the sensitivity of green manufacturing attributes to the variations in process parameters such as peak current, pulse duration, dielectric level and ﬂushing pressure. Weighing factors for the output responses were determined using triangular fuzzy numbers and the most desirable factor level combinations were selected based on TOPSIS technique. The model developed in this study can be used as a systematic framework for parameter optimization in environmentally conscious manufacturing processes. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Manufacturing processes generate large amounts of various solid, liquid, and gaseous wastes. Apart from the generation of waste, the manufacturing process is considered to be an energy intensive activity, which also indirectly affects the environment. Implementation of stringent government regulations and growing public awareness made the environmental issues in the processes one of the most important topics in strategic manufacturing decisions (Sheng & Srinivasan, 1995). Green manufacturing is an advanced manufacturing mode, aiming at improving the efﬁciency of the process as well as minimization of environmental impact and resource consumption during the manufacturing process (Tan, Liu, Cao, & Zhang, 2002). Die sinking Electrical Discharge Machining (EDM) is one of the most popular non-traditional manufacturing processes suitable for machining very hard and brittle materials. Recent advances in the EDM technology made it a valuable and viable process in the manufacturing of critical parts such as aerospace and aeronautical components. Despite its advantages the EDM is considered as a hazardous process in which large amounts of toxic solid and liquid wastes and exhaust gas are discharged, resulting in serious occupational and environmental problems (Tonshoff, Egger, & Klocke, 1996). High discharge energies of this process can lead to the arising of a number of reaction-products of the dielectric, which can emit from its surface as aerosols or gases. Apart from the air emissions, hazardous substances can also concentrate in the slurry and ⇑ Corresponding author. Tel.: +91 431 2503408; fax: +91 431 2503402. E-mail address: spshivam@nitt.edu (S.P. Sivapirakasam). 0957-4174/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2011.01.026 dielectric itself. These toxic substances can enter the body of operating personnel through ingestion, inhalation and skin contact. The performance characteristics of this process and the amount of waste generated from it are strongly inﬂuenced by the process parameters. Thus optimization of process parameters is an essential requirement to achieve green EDM. The selection of optimum process parameters to achieve green manufacturing involves contradictory criteria which necessitates the implementation of sophisticated Multi-Attribute Decision Making (MADM) methods. Analytic Hierarchic Process (AHP) (Saaty, 1980), Technique for Order Preference by Simulation of Ideal Solution (TOPSIS) (Hwang & Yoon, 1981), VIKOR (Tong, Chen, & Wang, 2007) and gray relational analysis (Deng, 1989) are the MADM techniques normally employed in solving engineering problems. Several researchers used these techniques in environmental impact assessment and green manufacturing (Kuo, Chang, & Huang, 2006; Tesfamariam & Sadiq, 2006). Yeo and New (1999) used a prioritization matrix for dielectric selection in a die sinking EDM process. However, no reported literature on the optimization of process parameters for green EDM is available. Among the MADM methods TOPSIS, which can handle multi-response problems with both continuous and discrete data, is the most suitable technique in manufacturing applications (Tong & Su, 1997). The basic philosophy of this method is to select the best alternative that has the shortest distance from the positive ideal solution and the farthest distance from the negative ideal solution. Real life multi-criteria decision-making problems usually involve uncertain, imprecise and subjective data, which make the decision-making process more complex and challenging (Aiello, Enea, Galante, & Scalia, 2009). In evaluating such data, decision makers S.P. Sivapirakasam et al. / Expert Systems with Applications 38 (2011) 8370–8374 generally view risk in terms of linguistic variables like low, high, very high, very low, etc. Fuzzy set theory (Zadeh, 1965; Bortolan & Degami, 1985) deals effectively with this type of uncertainty (vagueness), thus allowing linguistic variables to be used for approximate reasoning. Generally triangular and trapezoidal fuzzy numbers are used for representing linguistic variables. The main objective of this work was to select optimum process parameters for die sinking EDM which best reﬂect the manufacturing priority between environmental and machining factors. The experiments were designed using Taguchi L9 orthogonal array. An analytical model was developed for optimizing the process parameters. Peak current, pulse duration, dielectric level and ﬂushing pressure were the process parameters considered in this study. Relative importance of output parameters were designated by triangular fuzzy numbers (TFN). TOPSIS was used to evaluate the overall performance index values corresponding to each experimental run and optimum factor level combinations are identiﬁed based on the same. 2. Decision making model for green EDM In die sinking EDM process, an electric arc struck between two conductive electrodes produces the energy required for the material removal. The process is carried out in a dielectric medium. The main cause of the material removal is the high temperature reached by the surface of both the electrodes due to the thermal energy generated in the discharge channel. Heat generated in the channel causes some of the work material to melt and even evaporate. As the spark collapses, the evaporated metal and part of the molten metal are carried away by the dielectric ﬂuid which is ﬂushed using pressure (Abbas, Solomon, & Bahari, 2007). In order to assess the relationship between the process parameters and output responses, a modiﬁed form of the input-process output model (Fig. 1) proposed by Choi, Kaebernick, and Lai (1997) was employed. The inputs of the process include the process parameters (peak current, voltage, pulse duration, ﬂushing pressure, etc.), materials (work, tool and dielectric) and electrical energy. The outputs are the material removal rate, tool wear rate, air emissions, dielectric wasted in the form of liquid, eroded work and tool materials, heat and noise. Organizations implementing green manufacturing for the EDM process should consider two types of attributes: manufacturing and environmental. Each attribute is related to several output parameters. Fig. 2 illustrates the proposed analytical decision making model for parameter optimization in green EDM. In this model two responses were considered under manufacturing aspects and three responses were considered under environmental aspects. The following discussion is concerning the output responses assessed in this work. 8371 Fig. 2. Decision making model for green EDM. time and cost of operation. The time required for unit material removal was considered as a factor in the present work. 2.2. Relative tool wear ratio During the electric discharge, some of the discharge energy applied to the tool produces a crater in the tool material. This electrode wear inﬂuences the cost of operation as well as the amount of waste generated. The amount of erosion suffered by the tool compared with that of the work piece was referred to as the relative tool wear ratio. 2.3. Process energy The electric energy consumed during the EDM operations indirectly affects the environment as more waste is produced in order to generate more electricity. This energy is determined by the gap voltage during discharge, the discharge current and the length of time that the current ﬂows. In the present study the energy consumed for unit material removal was considered as the response variable. 2.4. Breathing zone concentration of aerosol Occupational exposure to toxic aerosols is an important hazard potential of the process particularly when hydrocarbon dielectric ﬂuid is used. The aerosol generated from the process may consist of metallic particles and reaction products of the dielectric material. Mass concentration of respirable particulates in the breathing zone of the operator was considered as a factor in the study. Generally this value is used as a measure of risk in similar processes. 2.1. Process time 2.5. Dielectric consumption Material removal rate is the most important machining parameter in the EDM process. This parameter determines the machining During the EDM process, the dielectric ﬂuid is generally wasted through three paths (Yeo, Tan, & New, 1998): (i) Coating of the dielectric ﬂuid on the work piece. (ii) Coating of the dielectric ﬂuid on materials removed from both the work piece and tool. (iii) Vapor of the dielectric diffused into the surrounding environment. Fig. 1. Input-process-output diagram of EDM process. The dielectric consumed during the process has economical and environmental impacts. The wasted dielectric in the form of gas and liquid may cause problems to the operators as well as the environment. In the present investigation the mass of dielectric consumed per unit material removal was considered as a factor. 8372 S.P. Sivapirakasam et al. / Expert Systems with Applications 38 (2011) 8370–8374 3. Materials and methods 3.1. Determination of output parameters The experiments were conducted on a conventional die sinking electric discharge machine manufactured by Victory Electromech. A high carbon high chromium tool steel plate of size 4 cm Â 4 cm Â 1.5 cm was used as the work piece and a copper rod of diameter 25 mm was used as the tool. Commercially available kerosene was used as the dielectric ﬂuid and side ﬂushing was opted. Kerosene is a blend of hydrocarbons (C12–C15) which is widely used in EDM because of its high ﬂash point, good dielectric strength, low viscosity and low speciﬁc gravity. The open gap voltage was kept constant at 100 V. Duty factor was kept at 0.5. During the machining, the tool ‘jumped’ periodically to a height of 1.25 mm to clean the gap between the work piece and tool. The machining was interrupted at each jump. The jump duty cycle (ratio of machining time to the total time) was kept at 0.75. Aerosol in the process location was sampled using a Universal Air Sampler (SKC model No. 224-PCXR8) with cyclone attachment. The cyclone attachment was used to separate particles of size 5 lm and above which are not respirable. A PVC ﬁlter of diameter 37 mm was used as the sampling medium. Samples of the aerosol were taken at a sampling point of 200 mm vertical distance and 200 mm horizontal distance from the dielectric surface above the process location. This location was assumed to correspond to an operators breathing zone in the worst case. The velocity of the sampler was kept at 2.5 l/min and the sampling was done for a duration of 8 h. The sampler was calibrated before and after sampling using a soap bubble meter, with the sampling medium in line. The weight of the ﬁlter paper was measured before and after sampling using a sensitive balance (accuracy ±0.01 mg). The concentration of aerosols (CA) in the work atmosphere was calculated from the following equation: ðW fb À W fa Þ Â 1000 CA ¼ ts Â v ð1Þ where Wfa and Wfb are the weights of ﬁlter paper in mg, before and after sampling, ts is the sampling duration in minutes and v is sampling speed in l/min. The material removal rate (MRR) and the tool wear rate (TWR) were calculated by taking the weights of the work piece and tool before and after the experiment MRR ¼ ðW Wa À W Wb Þ tm ð2Þ where WWa and WWb are the weights of workpiece in mg, before and after machining and tm is the machining time TWR ¼ ðW Ta À W Tb Þ tm ð3Þ were WTa and WTb are the weights of tool in mg, before and after machining and tm is the machining time. Process time (T) is the time taken in seconds to remove 1 mg of material. This factor was calculated using the equation shown below T¼ 60 MRR ð4Þ where fd, fi, V, and I are the duty factor, jump factor, gap voltage and peak current, respectively. The volume of dielectric in the dielectric sump was measured before and after machining and the dielectric consumed for removing 1 mg of material (Dc) was calculated as follows: Dc ¼ Da À Db tm Â MRR ð7Þ where Da and Db are the volume of dielectric in the sump (cm3), before and after machining and tm is the machining time. The peak current (I), pulse duration (tp), dielectric level (l) and ﬂushing pressure (p) were the independent variables studied to analyze the following responses; process time, relative tool wear ratio, process energy, breathing zone concentration of the particulates and dielectric consumption. In this study L9 (34) orthogonal array was selected to conduct experimental runs. The process variables and their levels for the design used in this study are shown in Table 1. The design of experiment matrix and experimental results are presented in Table 2. 3.2. Optimization using fuzzy TOPSIS In this paper important weights of the output responses were assigned linguistic variables as shown in Table 3. The linguistic variables were described using triangular fuzzy numbers. Four experts form a committee to act as decision makers. Each decision maker rated each attributes weight with respect to linguistic term. The results are presented in Table 4 and aggregated fuzzy weight of each output parameters are shown in Table 5. The optimization was done using the normalized performance matrix. The normalized performance matrix was obtained using the following transformation equation: xij rij ¼ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ P9 2 i¼1 xij ð8Þ where xij represents the actual value of ith attribute of jth experimental run and rij represents the corresponding normalized value. The normalized performance matrix was then multiplied by weights associated with each attribute thus yielding the weighted performance matrix S. The positive ideal value set, S+, and the negative ideal value set, SÀ, were then determined as follows: Sþ ¼ f½maxðsij Þjj 2 J or ½minðsij Þjj 2 J 0 ; i ¼ 1; 2; . . . ; 9g ¼ fsþ ; sþ ; . . . sþ g 1 2 5 ð9Þ SÀ ¼ f½minðsij Þjj 2 J or ½maxðsij Þjj 2 J 0 ; i ¼ 1; 2; . . . ; 9g ¼ fsÀ ; sÀ ; . . . sÀ g 1 2 5 ð10Þ where J = {j = 1, 2, . . . , 5|j}: Associated with higher the better performance parameters. J0 = {j = 1, 2, . . . , 5|j}: Associated with lower the better performance parameters. All the performance parameters considered in this study were of lower the better type. According to the weighted normalized Relative tool wear ratio is the ratio of TWR and MRR RTWR ¼ TWR MRR ð5Þ Parameters Process energy (E) is the energy in Watts used to remove 1 mg of material. This was calculated using the following equation: E ¼ fd Â fj Â V Â I Â T Table 1 Input parameters and their levels. ð6Þ Unit Level 1 Level 2 Level 3 Current Pulse duration Dielectric level Flushing pressure A 2 2 40 0.3 4.5 261 60 0.5 7 520 80 0.7 ls mm kg/cm2 8373 S.P. Sivapirakasam et al. / Expert Systems with Applications 38 (2011) 8370–8374 Table 2 Experimental results. Sl. No. Input parameters Peak current (A) Pulse duration (ls) Dielectric level (mm) Flushing pressure (kg/cm2) Process time (s) REWR Process energy (W) Conc. of aerosol (mg/m3) Dielectric consumption (cm3) 1 2 3 4 5 6 7 8 9 2 2 2 4.5 4.5 4.5 7 7 7 2 261 520 2 261 520 2 261 520 40 60 80 60 80 40 80 40 60 0.3 0.5 0.7 0.7 0.3 0.5 0.5 0.7 0.3 0.7258 1.5357 1.6393 0.4705 0.3415 0.3942 0.4062 0.2381 0.2646 0.3899 0.0055 0.0051 0.3496 0.0041 0.0049 0.3452 0.0065 0.0076 54.433 115.178 122.951 79.389 57.620 66.516 106.632 62.4884 69.469 0.82 0.77 0.64 1.22 2.13 1.98 2.4 4.12 5.05 0.0665 0.0981 0.0865 0.051 0.0332 0.0394 0.0497 0.0351 0.0434 Output parameters Table 3 Linguistic variables for the importance weight of each output criterion. Importance Abbreviation Fuzzy weight Extremely low Very low Low Medium High Very high Extremely high EL VL L M H VH EH The proximity of a particular experimental run to the ideal solution is expressed using the closeness coefﬁcient (CCi) which was calculated as (0, 0, 0.1) (0, 0.1, 0.3) (0.1, 0.3, 0.5) (0.3, 0.5, 0.7) (0.5, 0.7, 0.9) (0.7, 0.9, 1) (0.9, 1, 1) À CC i ¼ Decision maker DM1 Process time REWR Process energy Concentration of aerosol Dielectric consumption DM2 DM3 DM4 H VL M VH H H L H EH H VH L M VH M H VL L VH H Table 5 Fuzzy weights. Output parameter Fuzzy weight Process time REWR Process energy Concentration of aerosol Dielectric consumption 0.55, 0.05, 0.35, 0.75, 0.45, 0.75, 0.925 0.2, 0.4 0.5, 0.7 0.925, 1 0.65, 0.85 þ 5 X ð14Þ The closeness coefﬁcients for each experiment of the L9 orthogonal array were calculated as discussed in the previous section (Table 6). According to the performed experiment design, it could be clearly observed from Table 6 that the EDM parameters setting of experiment No. 5 yielded the highest closeness coefﬁcient. Therefore, experiment No. 5 had the optimal machining parameters setting for the desirable output responses simultaneously (i.e. the best multi-performance characteristics) among the nine experiments. The response table for the Taguchi method was used to calculate the closeness coefﬁcient for each level of the input parameters. The procedure is: (i) group the closeness coefﬁcients by factor level for each column in the orthogonal array and (ii) take the average of them. The closeness coefﬁcient values for each level of process parameters are shown in Table 7. Regardless of the category of performance characteristics, a greater closeness coefﬁcient value corresponds to better performance. Therefore, the optimal level of the machining Table 6 Closeness coefﬁcients. Sl. No. fuzzy decision matrix, the elements sij were normalized positive triangular fuzzy numbers and their ranges belong to the closed interval [0; 1]. The distance of each experimental result from positive and negative ideal solutions were calculated using the following equations: di ¼ di À þ di 4. Results and discussion Table 4 Importance of output responses. Output response þ di dðsij ; sþ Þ; j i ¼ 1; 2; . . . ; 9 ð11Þ dðsij ; sÀ Þ; j i ¼ 1; 2; . . . ; 9 ð12Þ Current (A) Pulse duration (ls) Dielectric level (mm) Flushing pressure (kg/ cm2) Closeness coef. 1 2 3 4 5 6 7 8 9 2 2 2 4.5 4.5 4.5 7 7 7 2 261 520 2 261 520 2 261 520 40 60 80 60 80 40 80 40 60 0.3 0.5 0.7 0.7 0.3 0.5 0.5 0.7 0.3 0.694042 0.47655 0.484248 0.7304 0.857285 0.830036 0.617257 0.709666 0.600028 i¼1 À di ¼ 5 X Table 7 Response table for closeness coefﬁcient. i¼1 where d(x, y) is the distance measurement between two fuzzy numbers. This distance between two triangular fuzzy numbers was calculated using the following equation: rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 ½ðx1 À y1 Þ2 þ ðx2 À y2 Þ2 þ ðx3 À y3 Þ2 dðx; yÞ ¼ 3 ð13Þ Input parameter Average closeness coefﬁcient Level 1 Peak current Pulse duration Dielectric level Flushing pressure Level 2 0.5516 0.6806 0.7446 0.7171 0.8059 0.6812 0.6023 0.7236 0.6423 0.6381 0.6529 0.6414 Max–min Level 3 0.2543 0.0431 0.1423 0.0822 8374 S.P. Sivapirakasam et al. / Expert Systems with Applications 38 (2011) 8370–8374 parameters was the level with the greatest closeness coefﬁcient value. Based on the closeness coefﬁcient values given in Table 7, the optimal machining performance for the green EDM was obtained for 4.5A peak current (level 2), 261 lm pulse duration (level 2), 40 mm dielectric level (level 1) and 0.5 kg/cm2 ﬂushing pressure (level 2). As listed in Table 7, the difference between the maximum and the minimum value of the closeness coefﬁcient of the EDM parameters was as follow: 0.2543 for peak current, 0.0431 for pulse duration, 0.1423 for dielectric level and 0.0822 for ﬂushing pressure. The most effective factor affecting performance characteristics was determined by comparing these values. This comparison demonstrated the level of signiﬁcance of the input parameters over the multi-performance characteristics. The most effective controllable factor was the maximum of these values. Here, the maximum value was 0.2543. This value indicated that the peak current had the strongest effect on the multi-performance characteristics among the input parameters. The order of importance of the controllable factors to the multi-performance characteristics in the EDM process, in sequence can be listed as follows: peak current, dielectric level, ﬂushing pressure and pulse duration. 5. Conclusions The present work proposed a combination of Taguchi method and TOPSIS to solve the multi-response parameter optimization problem in green electrical discharge machining. An analytical structure was developed to perform multi-criteria decision making. The responses were ranked based on the scores obtained by the summarization of ﬁnal global preference weights. Triangular fuzzy numbers were used to assign preference values to the output responses. The optimum factor level combinations were identiﬁed based on the closeness coefﬁcient values. The optimal machining performance for the green EDM was obtained for 4.5 A peak current (level 2), 261 lm pulse duration (level 2), 40 mm dielectric level (level 1) and 0.5 kg/cm2 ﬂushing pressure (level 2). From analysis of the closeness coefﬁcients, it was identiﬁed that the peak current was the most inﬂuential parameter in multi-performance characteristics used in this study. The computational and experimental effort needed to optimize these parameters was rather small. It was illustrated that the method was efﬁcient and effective for multi-attribute decision making problems in green manufacturing. Acknowledgements The authors are grateful to the Director, NIT Tiruchirappalli for his continuing encouragement and support for this work. Part of the work is supported by the Ministry of Environment and Forests, Government of India (F.No. 19/102/2008-RE). References Abbas, N. M., Solomon, D. G., & Bahari, M. F. (2007). A review on current research trends in electrical discharge machining (EDM). International Journal of Machine Tools & Manufacture, 47, 1214–1228. Aiello, G., Enea, M., Galante, G., & Scalia, G. La. (2009). Clean agent selection approached by fuzzy TOPSIS decision-making method. Fire Technology, 45, 405–418. Bortolan, G., & Degami, R. (1985). A review of some methods for ranking fuzzy subset. Fuzzy Sets and Systems, 15(1), 1–19. Choi, A. C. K., Kaebernick, H., & Lai, W. H. (1997). Manufacturing process modeling for environmental impact assessment. Journal of Material Processing Technology, 70, 231–238. Deng, J. L. (1989). Introduction to grey system. Journal of Grey Systems, 1(1), 1–24. Hwang, C. L., & Yoon, K. (1981). Multiple attribute decision making, a state of the art survey. New York: Springer-Verlag. Kuo, T. C., Chang, S. H., & Huang, S. H. (2006). Environmentally conscious design by using fuzzy multi attribute decision making. International Journal of Advanced Manufacturing Technology, 29, 419–425. Saaty, T. L. (1980). The analytic hierarchy process. New York: McGraw-Hill. Sheng, P., & Srinivasan, M. (1995). Multi-objective process planning in environmentally conscious manufacturing: A feature-based approach. Annals of the CIRP, 44(1), 433–437. Tan, X. C., Liu, F., Cao, H. J., & Zhang, H. (2002). A decision making framework model of cutting ﬂuid selection for green manufacturing and a case study. Journal of Materials Processing Technology, 129, 467–470. Tesfamariam, S., & Sadiq, R. (2006). Risk based environmental decision-making using fuzzy analytic hierarchy process (F-AHP). Stochastic Environmental Research and Risk Assessment, 21, 35–50. Tong, L. I., Chen, C. C., & Wang, C. H. (2007). Optimization of multi-response processes using the VIKOR method. International Journal of Advanced Manufacturing Technology, 31, 1049–1057. Tong, L. I., & Su, C. T. (1997). Optimizing multi-response problems in the Taguchi method by fuzzy multiple attribute decision making. Quality and Reliability Engineering International, 13, 25–34. Tonshoff, H. K., Egger, R., & Klocke, F. (1996). Environmental and safety aspects of electrophysical and electrochemical processes. Annals of CIRP: Manufacturing Technology, 45(2), 553–568. Yeo, S. H., & New, A. K. (1999). A method for green process planning in EDM. International Journal of Advanced Manufacturing Technology, 15, 287–291. Yeo, S. H., Tan, H. C., & New, A. K. (1998). Assessment of waste streams in electric discharge machining for environmental impact analysis. Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, 212, 393–401. Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8(3), 338–353.