Clasificación de piezas mecánicas mediante redes neuronales artificiales

Abstract

A spare parts warehouse will be exposed to acquire defective mechanical parts that is why an automated system is required to perform quality control of parts purchased in a spare parts warehouse and reduce the distribution of defective mechanical parts. Therefore, the research objective was to develop a mechanical parts classification program using artificial neural networks to achieve a distribution of quality parts to customers. The methodology had a quantitative approach that based on data analysis and tabulation allowed choosing a mechanical part for the research. A descriptive level that specifies characteristics of the element and correlational that quantifies, analyzes, and establishes links. An experimental and transversal design was applied to relate the variables together with their incidence, being essential a documentary study which allowed the development of artificial neural networks and a field study that allowed the application of them through MATLAB software. By this methodology it was possible to apply artificial neural networks in the classification of mechanical parts subsequently selected. In addition, the productivity of an automotive parts warehouse was achieved. In this research, it is concluded that the classification of mechanical parts by means of artificial neural networks is successful in a spare parts warehouse and leads to improve its productivity, this provides confidence to acquire mechanical parts for the company.