use style: paper title

International Journal of Applied Engineering Research, ISSN 0973-4562 Vol. 10 No.55 (2015) © Research India Publications; httpwww.ripublication.comijaer.htm

Optimization of Removal Rate in Electrochemical Machining for EN 19 Using Fuzzy logic 1

Abhishek Tiwari, 2Amitava Mandal, 3Kaushik Kumar* Research Scholar, Department of Mechanical Engineering, BIT, Mesra, Ranchi, 835215, India 2 Professor, Department of Manufacturing Engineering, NIFFT, Ranchi, 835215, India 3 Associate Professor, Department of Mechanical Engineering, BIT, Mesra, Ranchi, 835215, India 1

Abstract

This paper is concerned with the design of fuzzy logic based model to simulate the material removal rate (MRR) in electrochemical machining (ECM) of EN 19 tool steel. The main objective to develop fuzzy model is its practical application, such that a design engineer can use it in process planning time to time. The efficiency of the model is measured by comparing the predicted with the practical results. Outcomes shows that fuzzy model is an effective and an alternative tool that can be effectively applied to determine the responses for any input combination with high accuracy. Keywords: Optimization; Fuzzy logic;EN 19 tool steel; Electro chemical machining; MRR.

vibrations for higher accuracy using the RSM. Many other researchers have done a lot of work on ECM, like De Silva et al. [10] (on process monitoring of electrochemical machining), Siebertz et al. [11] (on Design of Experiments), Z. Liu et al. [12] (on electrochemical slurry jet micro-machining), B. Bhattacharyya et al. [13] (mathematical modelling of ECM using the RSM) etc. 2. Electrochemical machining ECM is opposite of deposition process [14]. Thus ECM can be thought of a controlled anodic dissolution at atomic level due to flow of high current through an electrolyte. The variations in process parameters greatly affect the performance. Therefore, proper selection of the machining parameters can result in better removal rate in the ECM process.

1. Introduction

Inevitable requirement of any industrial operation is higher MRR, increase in MRR has some technical hitches; like high tool wear and poor roughness features. During any conventional process there is tool and work-piece interface occur, also scrap flow along the tool. ECM process is an ideal machining process where higher surface quality needed with higher MRR, since there is no contact between tool and workpiece. Therefore; Number of research work have been focused on ECM to heighten the machining process; since the introduction of experimental ECM process by W. Gussef in 1929. Newer machining processes like ECM, EDM have advantage, material is removed by the combination of electrical and chemical energy [1-3]. Jagdev et al. [4] presented fuzzy model to simulate the MRR in ultrasonic drilling for porcelain ceramic. Milan Kumar Das et al. [5] have presented the grey-taguchi optimization technique on EN-31 for Surface Roughness and MRR in ECM. Prabhu et al. [6] predicted the surface roughness for with and without CNT based cutting fluid using fuzzy logic approach to explore the behaviour of control factors toward response. Jo˜ao Cirilo da Silva Neto et al. [7] also conducted a testing to predict the dominant variables in electrochemical machining. The material removal rate (MRR), roughness and over-cut were taken as response for study. Xiaolong Fang et al. [8] study the effects and behaviour of pulsating electrolyte flow, experiments were conducted to corroborate the possibility of the proposed method and its effects on MRR. S.J. Ebeid et al. [9] address an hybridized ECM process with low-frequency

3683

2.1. Setting the Levels of Welding Parameters Experiments were conducted on METATECH machining system. For experimental work EN 19 tool steel is selected as sample work-piece of diameter 25 mm and height 15 mm, its chemical composition (wt %) shown in Table.1. Table 1: EN 19 chemical composition (wt. %) Carbo Mang Chro Molyb Silico n anese mium denum n 0.350.45%

0.500.80%

0.901.50%

0.200.40%

0.100.35%

Phosp horou s 0.03%

Sulph ur 0.050 %

Potassium chloride is used as an electrolyte. In the present research work four dominated control factor were selected, shown in table 2 with their levels. Table 2: Design parameters and their levels Process parameter Level Symbol (unit) 1 Electrolyte conc. (%) A 10 Voltage (V) B 8 Feed rate (mm/min) C 0.1 Electrode gap, mm D 0.2 3.

Responses and measurement

Level 2 15 10 0.21 0.25

Level 3 20 12 0.32 0.3


Material Removal Rate is taken as the response variable in the present study.

= weight of work piece before machining = weight of work piece after machining = Material density T = Time of machining Fig. 2: Membership functions for Concentration.

4. Optimization methodology Present work comprise four input and one output (Fig. 1). A fuzzy logic unit comprises a fuzzifier, membership functions, a fuzzy rule base, an inference engine, and a defuzzifier. First, the fuzzifier uses membership functions to fuzzify the input values. And the inference engine performs fuzzy reasoning based on fuzzy rules to generate a fuzzy value. Finally, defuzzification converts the fuzzy quantity into a non-fuzzy value. Deffuzifying method determines the accuracy of the developed model [15]. In this study, centroid defuzzification method, which estimated centroid of the area under the membership function, was used. It gives more accurate result compared to the other deffuzifying methods.

Fig. 3: Membership functions for Voltage.

Fig. 1: Structure of the four-input-one-output fuzzy logic unit. The fuzzy rule base consists of a group of if–then control rules. Nine fuzzy rules (Table 6) are directly derived based on effect of input variable on the removal rate. In the paper, three fuzzy subsets are assigned in the four inputs (Figs. 2-5). Five fuzzy subsets are assigned in the output (Fig. 6).

Fig. 4: Membership functions for Feed-rate.

Low Medium large

Electrolyte concentration (%) Low Medium Large Very Small Small small Small Large Large Mediu Large Very m large Low Medium Large Inter-electrode gap (mm)

Low Medium large

Feed -rate

Voltage (V)

Table 3: Fuzzy rule table

Fig.5: Membership functions for Gap.

3684


Fig. 6: Membership functions for Metal Removal Rate

5.

Result and discussion Table 4: Experiment and predicted MRR using fuzzy model result for response parameter Exp. No. A B C D MRR S/N Predicted (MRR) value 1 1 1 1 1 21.6561 26.7116 22.8 2 1 1 2 2 17.2399 24.7306 14.6 3 1 1 3 3 19.7452 25.9092 22.1 4 1 2 1 2 14.1189 22.9960 16.8 5 1 2 2 3 14.837 23.4269 13.5 6 1 2 3 1 16.1359 24.1558 18.6 7 1 3 1 3 19.7707 25.9204 23.2 8 1 3 2 1 24.2038 27.6776 24.8 9 1 3 3 2 23.0828 27.2657 25.8 10 2 1 1 2 13.2484 22.4432 16.5 11 2 1 2 3 16.8153 24.5140 18.5 12 2 1 3 1 16.4968 24.3479 17.1 13 2 2 1 3 16.1783 24.1786 14.3 14 2 2 2 1 25.4777 28.1232 21.4 15 2 2 3 2 26.9427 28.6088 28.4 16 2 3 1 1 24.1401 27.6547 26.5 17 2 3 2 2 25.3715 28.0869 28.4 18 2 3 3 3 22.8238 27.1677 20.6 19 3 1 1 3 23.0998 27.2721 26.4 20 3 1 2 1 21.8259 26.7794 23.5 21 3 1 3 2 23.5669 27.4468 24.5 22 3 2 1 1 22.1806 26.9194 19.6 23 3 2 2 2 25.2529 28.0462 21.5 24 3 2 3 3 32.3931 30.2090 32.9 25 3 3 1 2 22.293 26.9633 19.8 26 3 3 2 3 24.7134 27.8586 25.4 27 3 3 3 1 37.707 31.5284 35.2

Absolute error 1.1439 2.6399 2.3548 2.6811 1.337 2.4641 3.4293 0.5962 2.7172 3.2516 1.6847 0.6032 1.8783 4.0777 1.4573 2.3599 3.0285 2.2238 3.3002 1.6741 0.9331 2.5806 3.7529 0.5069 2.493 0.6866 2.507

modelling were compared with experimental values and results were plotted as shown in Fig. 8.

5.1. The evaluation of optimal setting

Optimal combination of process variable estimated to enhance the performance. Optimal parameter setting is measured on the basis of mean parameter. In this regard residual values for response has been plotted (Fig. 9), it shows at test run 24 (concentration 20%, voltage 10V, feed-rate 0.32mm/min, and inter-electrode gap 0.3) has lowest residual value respectively 0.5069.

Scatter plots for the experimental and predicted values with fuzzy modelling is depicted in Fig. 7. The efficiency of the developed fuzzy models was evaluated by R2 values. In fuzzy models, R2 values for MRR was found as 96.5%, which meant the high correlation that existed between the experimental and predicted values. The predicted values achieved by fuzzy logic

3685


5.2. Confirmation test In order to ascertain the extent of the improvement as indicated by the optimum parameters analysis, a validation test was required to be undertaken. The predicted MRR was tallied with the experimental ones respectively (Table 5). The midlevel combination of process parameters is assumed as the initial condition. It can be easily seen that the residual error for A3B2C3D3 is less as compared to the A2B2C2D2 From the table, it give the improvement of MRR at the optimal condition is about 2.5149 dB .

Fig 7: Scatter plot for experimental and predicted MRR

Fig. 9: Variation of residual with training pattern

Fig. 8: Variation of Experimental and predicted MRR Table 5: Confirmation runs of optimal parameters for MRR Parameter Mean parameter for EN 19

S/N RATIO MRR

Mean A2B2C2D2 27.6941

Optimal parameter for EN 19 Predicted

27.851 32.9 Improvement in S/N ratio = 2.5149 dB

3686

Experimental A3B2C3D3 30.2090 32.3931


Fig. 10: 3D surface plot shows the effect of (a) voltage and concentration (b) concentration and gap (c) feed-rate and concentration (d) voltage and feed-rate on MRR The 3D plot for MRR from the fuzzy logic model was constructed as shown in Figs. 10 (a-d). As per Fig. 10a, MRR is optimum at the higher concentration (20%) and significantly raised by increasing the voltage. In fig 10b, as the gap decreases MRR increase. Since in small inter-electrode gap current density increase which increase the anodic dissolution rate. From Fig. 10c, it is clearly seen that as the feed rate increase MRR, but increase in removal rate quit less initially. When feed rate close to mid-level MRR increases sharply. Similarly in fig 10d, shows the voltage and feed rate plot, as the voltage increases current density increases between the gap which increase the MRR, higher feed-rate also help the MRR.

6. Conclusion In this paper, fuzzy model has been devolved for ECM and metal removal rate is predicted for electrochemical machining of EN19 tool steel. The experiment study illustrate that error for the MRR have minimum value 0.5096 for A3B2C3D3 input combination (concentration 15%, voltage 10V, feed rate 0.32 mm/sec, and 0.25 gap mm). So with reference to the midlevel response characteristics is greatly improved and supported by fuzzy based model. Also it can be concluded that a fuzzy model can tackle high degree of non-linearity that exist between the output and input variable.

3687


References 1. V.K. Jain, Advanced Machining Processes, Allied Publishers, Mumbai, India, 2004. 2. J.A. McGeough, Principles of Electrochemical Machining, Chapman & Hall, New York, 1974. 3. Denkena B, Kindermann R, Hollmann C. High Performance Cutting of Aerospace-Materials. Production Engineering, 2004; 11:15–20. 4. Jagdev Singh and Simranpreet Singh Gill, 2009. Fuzzy Modeling and Simulation of Ultrasonic Drilling ofPorcelain Ceramic with Hollow Stainless Steel Tools, , Materials and Manufacturing Processes, 24: 468–475, 2009, DOI: 10.1080/10426910802714407. 5. Milan Kumar Das, Kaushik Kumar, Tapan Kr. Barman and Prasant Sahoo, 2014. Optimization of Surface Roughness and MRR in Electrochemical Machining of EN31 Tool Steel using Grey-Taguchi Approach. Procedia Materials Science 6, 729 – 740. 6. S. Prabhu, M. Uma, B. K. Vinayagam, 2015 , Surface roughness prediction using Taguchi-fuzzy logic-neural network analysis for CNT nanofluids based grinding process., Neural Computation & Application, 26:41–55, DOI 10.1007/s00521-014-1696-8. 7. Jo˜ao Cirilo da Silva Neto, Evaldo Malaquias da Silva, Marcio Bacci da Silva, 2006. Intervening variables in electrochemical machining. Journal of Materials Processing Technology 179, 92–96. 8. Xiaolong Fang, Ningsong Qu, Yudong Zhang, Zhengyang Xu, Di Zhu, 2013. Effects of pulsating electrolyte flow in electrochemical machining. Journal of Materials Processing Technology 214, 36– 43. 9. S.J. Ebeid, M.S. Hewidy, T.A. El-Taweel, A.H. Youssef, 2004. Towards higher accuracy for ECM hybridized with low-frequency vibrations using the response surface methodology. Journal of Materials Processing Technology 149, 432–438. 10. De Silva, A., McGeough, J.A., 1998. Process monitoring of electrochemical machining, J. Mater. Process. Technol. 76, p.165-169 11. Siebertz, L., Van Bebber, D., Hochkirchen, T., 2010. Design of Experiments (DoE), Springer Heidelberg Dordrecht London New York. 12. Z. Liu, H. Nouraei, M. Papinic, J.K. Spelt, 2014. Abrasive enhanced electrochemical slurry jet micro-machining: Comparative experiments and synergistic effects, Journal of Materials Processing Technology 214, 1886–1894. 13. B. Bhattacharyya, S.K. Sorkhel, 1999. Investigation for controlled electrochemical machining through response surface methodology-based approach, Journal of Materials Processing Technology 86, 200–207. 14. P.C. Pandey, H.S. e Shan, Modern Machining Processes, Tata McGraw Hill, New Delhi, 1980. 15. Kovac, P., Rodic, D., Pucovsky, V., Savkovic, B., & Gostimirovic, M. (2013). Application of fuzzy logic and regression analysis for modelling surface roughness in face milling. Journal of Intelligent Manufacturing, 24, 755–762.

3688