Development of Hybrid ANN–Genetic Algorithm Model for Optimization of Surface Roughness in CNC Turning of AISI 1040 Steel

Surface roughness is a critical indicator of machining quality that directly governs the functional performance, fatigue life, tribological behaviour and aesthetic acceptability of turned components. This paper presents a hybrid Artificial Neural Network (ANN) and Genetic Algorithm (GA) model for the prediction and optimization of surface roughness (Ra) in Computer Numerical Control (CNC) turning of AISI 1040 medium-carbon steel. A set of machining experiments was designed using the Taguchi L27 orthogonal array, considering cutting speed, feed rate and depth of cut as the controllable input parameters. A feed-forward back-propagation neural network was trained on the experimental data to establish a non-linear mapping between the cutting parameters and the resulting surface roughness. The trained ANN was subsequently embedded as the fitness function of a genetic algorithm, which searched the continuous parameter space to identify the combination of cutting conditions that minimizes Ra. The hybrid ANN–GA model achieved a prediction accuracy in excess of 97% on unseen test data, with a mean absolute percentage error markedly lower than that of conventional regression models. The optimized parameter set recommended by the GA reduced the predicted surface roughness by approximately 28% relative to the average experimental value. The results confirm that the hybrid intelligent approach offers a robust, accurate and computationally efficient framework for parameter optimization in intelligent manufacturing, and that it can be deployed for real-time decision support in modern CNC turning operations.