فهرست نشریات علمی دسترسی آزاد ایران

In this study, the prediction capabilities of the adaptive neuro-fuzzy inference system (ANFIS) and the particle swarm optimization (PSO)-based ANFIS were compared. First, the material removal rate (MRR), surface roughness (Ra), and tool wear ratio (TWR) were modeled using the ANFIS technique during the ultrasonic-assisted electrical discharge machining (US/EDM) process. The ANFIS model was developed to predict these output parameters, and subsequently, its parameters were optimized using PSO to reduce prediction error. The pulse-on time (Ton) and current (I) were selected as input factors. The models were trained, tested, and validated with experimental data, and statistical analyses were conducted to evaluate the effectiveness of both ANFIS and ANFIS–PSO approaches. The greatest reduction in the average prediction error percentage was observed for MRR, decreasing from 10.87% in the ANFIS model to 6.19% in the ANFIS–PSO model. Overall, the experimental results demonstrated that the ANFIS–PSO algorithm provided superior performance in estimating machining parameters compared with the conventional ANFIS model.