Sistema de inspeção de bateria de lítio em placas eletrônicas: uma proposta a partir da integração de CLP e aprendizado de máquina

Abstract

The objective of this work is to develop a lithium battery inspection system based on the integration of Programmable Logic Controllers and Machine Learning, identifying advantages, disadvantages, and challenges associated with meeting the needs of industrial facilities. The project seeks to use low-cost materials and technologies, ensuring its economic viability and ease of construction. The innovation of the work lies in the adaptation of organizational needs, using Industry 4.0 tools, providing a practical, efficient, and assertive solution. Thus, the proposed system aims to address these aspects, offering an accessible solution for detecting assembly, optimization, and quality failures. The system inspection is based on a supervised learning that detects the battery via webcam and then compares the captured image with the model created from the OK and NG datasets. Therefore, the difference between the two results in an approved or rejected result for the part. The developed system performed well in its inspections during testing and validation. In addition, the work designated the observation of this integration in other segments, instigating opportunities and improvements for effective advancement. Therefore, the study concluded that, despite the significant challenges to implementing AM in PLCs, the integration shows promise for optimizing operational efficiency, flexibility, data transmission, cost reduction, and decision-making support in companies with limited resources and a need for modernization.