A forecasting model for optimizing abrasive water jet machining (AWJM) parameters based on the adaptive neuro-fuzzy inference system and meta-heuristic algorithms

Abstract

Innovative materials have been steadily produced in the machining process, and it has stimulated substantial attention among researchers to assess the ideal parameters for the processing of these materials. In the manufacturing industry, conventional and modern or unconventional machining retains their imprint in delivering quality products. One of the most frequently used contemporary machining processes is abrasive water jet machining (AWJM) for cutting materials. The optimization of parameters, however, is essential to manufacturing high-quality products quickly and affordably. This study focused on optimizing the machining parameters of AWJM using the Adaptive Neuro-Fuzzy Inference System (ANFIS) by encrypting the evolutionary algorithms, including the reptile search algorithm (RSA), beetle antennae search (BAS), and particle swarm optimization (PSO). As a result of the developed methods, the surface roughness of titanium alloy (Ti–6Al–4V) during AWJM machining is predicted by considering inputs such as standoff distance, abrasive flow rate, and traverse speed, which affect the surface roughness. Based on a comparison of the prediction accuracy of the proposed methods, it was found that the combined ANFIS and RSA outperforms the others. In the simulation results, ANFIS-RSA (Adaptive Neuro-Fuzzy Inference System and Reptile Search Algorithm) yielded the most accurate estimates of surface roughness based on the RMSE (0.1241), MAE (0.0961), MAPE (0.0201), and R² (0.9781) values. In particular, the analysis findings demonstrated that the combination of ANFIS and RSA was the most effective approach for predicting surface roughness on titanium alloy during AWJ machining.