AI-Enhanced Laser Drilling of Alumina Ceramics

Laser drilling of hard and brittle materials is challenging due to defects like hole taper, heat-affected zones, and microcracking, requiring careful process parameter optimisation for their mitigation. Conventional approaches involve time consuming and often costly experimental trials to identify the suitable process parameters. Artificial Intelligence (AI) has proven successful across various domains, enhancing task performance and efficacy. However, based on a comprehensive literature review, its apparent that AI application in laser-based manufacturing remains limited compared to other areas. This paper investigates the use of AI techniques, specifically machine learning and predictive modelling, to predict optimal laser drilling parameters, including laser power, pulse frequency, and drilling speed. The study uses secondary data such as surface characteristics and geometric outputs to facilitate AI training and validation. By analysing extensive process data, AI models can predict defect formation, such as cracks and taper, thereby improving drilling quality. Furthermore, AI-driven optimisation strategies can increase drilling speed while maintaining desired dimensional accuracy, thus reducing cycle times and operational costs. An optimal method for building a predictive model was determined by comparing and analysing different machine learning algorithms. The mean squared error of the experimental data, and the model predicted results for hole taper angle were found to be 0.5 degrees for the Random Forest algorithm. Integrating AI into laser-based manufacturing processes presents a significant advancement by identifying optimal process parameters, predicting and quantifying defects, and addressing challenges associated with laser drilling of alumina ceramic. This integration is expected to advance precision manufacturing and smart automation.