Implementación de herramientas de aprendizaje automático (IA) para la predicción de rugosidad de superficie en torneado

Abstract

The implementation of new technologies in engineering is at the forefront, thanks to the continuous advancements in computer science, given the ability to transmit and analyze information automatically. This is the case with the incorporation of machine learning methods in machining processes through algorithms, codes, and artificial intelligence, which optimize part production. This undergraduate monograph proposes the implementation of machine learning tools, primarily artificial intelligence, for predicting surface roughness in turning, based on the calculation of cutting parameters relevant to the manufacturing process. To this end, we propose the validation of machine learning from emerging artificial intelligences (ChatGPT, Meta, and Gemini) through collaborative interaction and information provision. This includes the development of a code for calculating cutting parameters in Python and experimental data from turning processes (extracted from databases) to train these tools. The respective validation of the feasibility of incorporating these machine learning models for surface roughness prediction will be carried out through statistical evaluation and comparison. It is expected that, through proper training of artificial intelligence, results will be obtained that generate precision in surface roughness prediction, while limiting errors due to misinterpretation and ambiguity by AIs, given that they are Large Language Models (LLM) that tend to generate errors in matters related to computation and correlation of numerical information.