A combined approach of artificial neural network, multi-objective genetic algorithm, and response surface methodology for enhanced PMMA micro-channeling in low power fiber laser beam machining

The most cost-effective approach for generating micro-channels on polymethyl methacrylate (PMMA), a polymer highly sought after in the biomedical industry, is by employing laser technology. The study aims to evaluate the feasibility of utilizing a low-power fiber laser for the purpose of generating micro-channels on PMMA, as well as enhance the data analysis and forecasting of results for implementation in the biomedical sector. This study utilizes response surface methodology (RSM) to create micro-channels on PMMA using low-power fiber lasers. Furthermore, a multi-objective genetic algorithm (MOGA) is employed to optimize the responses, namely the cut width and depth, within the predetermined range of process parameters. Moreover, the artificial neural network (ANN) technique is utilized to predict the results of MOGA and RSM data. The study encompasses process parameters such as laser power, pulse frequency, scan speed, and number of passes. The ANN is trained using input-output patterns derived from RSM. Three methods are integrated not only to optimize but also to forecast the data set in order to reduce material waste, costs, and increase efficiency. The experimental research shows that the artificial neural network technique has higher prediction accuracy compared to the other two methods. All three strategies utilized in this study have been well aligned with the empirical results for predicting the dimensions of the micro-channels.