An Adaptive Neuro-Fuzzy Inference System (ANFIS) for Wire-EDM of Ballistic Grade Aluminium Alloy

Abstract

Intricacy and complexity of ballistic missile and aerospace parts makes WEDM an essential machining process. The current study aims to formulate an ANFIS model for Wire-EDM of ballistic grade aluminium alloy. The experimentation has been conducted with four input variables namely pulse on time (T on), pulse off time (T off), peak current (I p), and servo voltage (V s). Material removal rate (MRR) is employed as process performance evaluator. The values predicted by the developed model are found closer to experimental outcome and thus ensures the model suitability for prediction purpose and intelligent manufacturing. Machined surfaces are also examined by the scanning electron microscope (SEM) to obtain better insight of the process.

Verification of Fuzzy Inference System for Cutting Speed while WEDM for the Abrasion-Resistant Steel Creusabro by Conventional Statistical Methods

“…The predicted maximum cutting speed by linear regression will be 6.3471 mm•min −1 and the optimum position (except for the insignificant wire feed) is consistent with the result obtained from the fuzzy inference system. The results of the regression analysis "confirm the correctness" of the selected explanatory variables for cutting speed and were the same as in the study by Singh [27], which dealt with aluminum alloy machining. In this study the authors did not perform a similar statistical analysis and the effect of these variables on the response The optimal setup of the parameters for Creusabro steel machining according to the validation experiment is gap voltage = 60 V, pulse on time = 10 µs, pulse off time = 30 µs, discharge current = 35 A, and the wire feed parameter arbitrarily (according to validation experiment it has no significant statistical influence on the response).…”

“…The predicted maximum cutting speed by linear regression will be 6.3471 mm•min −1 and the optimum position (except for the insignificant wire feed) is consistent with the result obtained from the fuzzy inference system. The results of the regression analysis "confirm the correctness" of the selected explanatory variables for cutting speed and were the same as in the study by Singh [27], which dealt with aluminum alloy machining. In this study the authors did not perform a similar statistical analysis and the effect of these variables on the response was taken as given.…”

“…In contrast, a regression model using the same criterion using five regressors describes 90.69% of the observed cutting speed variability. It should be emphasized that "worse" results of the "global" FIS prediction response (as compared to Çaydaş [9] and Singh [27] studies) were achieved using 18 a priori propositions, instead of defining neural network on experimentally determined data. It is undisputed that FIS-based methods have many uses in WEDM for modelling other responses that were studied in Çaydaş [9], Soepangkat [17], Rupajati [16], Yan [13] articles, or multi-criteria optimization of multiple output variables at the same time [12,15].…”

A study of wire tool surface topography and optimization of wire electro-spark machined UNS N06690 using the federated mode of RSM-ANN

“…, 2018; Shandilya et al. , 2010; Singh, 2018). WESM was probably first introduced in the late 70s to replace the use of form-tool for metal removal in industries.…”

Performance Analysis on Eco-friendly Machining of Ti6Al4V using Powder Mixed with Different Dielectrics in WEDM

“…Performance efficiency of WEDM process is also strongly dependent on the material characterisation properties based on their thermal, electrical and physical properties. The comparative study was made for modelling and optimisation on different materials using Buckingham pi, RSM, Genetic algorithm and other methodology to model the WEDM responses such as MRR and SR [6][7][8][9][10][11]. The selection of wire electrodes in the WEDM process also plays a vital role in enhancing the outer response characterisation [12].…”