Using machine learning techniques to predict the surface roughness of titanium alloy

Production engineering focuses on designing, optimizing, and managing manufacturing processes to produce goods efficiently. Turning is a machining process where a cutting tool removes material from a rotating workpiece to create cylindrical shapes. Key parameters include cutting speed, feed rate, and depth of cut. Surface roughness is a key challenge in turning, impacting product quality. Achieving the desired finish is crucial for tight tolerance and performance. Engineers use optimization techniques to minimize roughness. Advancements in tool materials and technology help address roughness challenges for improved efficiency in turning. Predictive models for surface roughness are vital for optimizing machining processes, ensuring quality, and enhancing performance. They guide decision-making, improve efficiency, and drive innovation in manufacturing. In this paper, 25 machine learning models have been used and optimized to accurately predict the surface roughness of the turning process of five previous studies on Titanium alloy. Resulting in the best performance with the lowest MSE was for the artificial neural network with 3 hidden layers: the first has 5 neurons, the second has 10 neurons and the last one has 5 neurons. The MSEs are 0.053072, 0.555763, 0.059667, 0.051867, and 0.554829 for the five studies, respectively.