Machine learning hyperparameters optimization of control strategies for emission reduction in sustainable leather cutting using semiconductor laser diode

Leather valued for its durability and versatility is widely utilized in industries such as footwear, automotive and handicrafts. Despite technological advancements conventional leather cutting techniques remain labour-intensive, inefficient and produce substantial material waste impacting both productivity and environmental sustainability. This study explores the optimization of leather cutting using semiconductor laser diode technology emphasizing emission reduction through Random Forest hyperparameter optimization of various control strategies. The experimental investigation explores fixed and adaptive control approaches using 20W semiconductor laser diode with Standoff Distance (SOD) control, Pulse Width Modulation (PWM) control and a combined adaptive SOD and PWM control. The implementation of adaptive control approaches contributed to optimal performance by reducing internal emissions during leather cutting. Real time monitoring and analysis of carbon based emissions were conducted using the SCD30 sensor ensuring compliance with environmental and health standards. The development of advanced control approaches with emission control unit reduces hazardous emission during leather cutting and promotes sustainable manufacturing in line with UN SDG-7 and SDG-13. The implementation of machine learning optimization in leather cutting contributes towards cleaner leather processing. Future research should focus on refining this technology for broader industrial applications contributing to economic diversification and growth and further optimizing emission reduction techniques.