Thermal inspection System Prediction for Circuit Breaker on Load Panel With Image Classification Technique By Tensorflow : A Case Study Textile Industry

This research seeks to design and create a thermal monitoring system for load panels utilizing a thermal scanner camera and TensorFlow's image classification method. The system utilizes Arduino Portenta H7 and MLX90640 infrared thermal imaging camera as sensors to detect and measure the heat radiation from the Load Panels. TensorFlow is employed to classify and analyze the thermal images, enabling the prediction and identification of preliminary issues in the Load Panels based on the heat patterns. When high temperatures are detected, the system sends notifications through the Node-RED system via MQTT Broker, and the data is stored in a database for accuracy checking and maintenance planning. The results show that the thermal image classification achieves an accuracy of 95.1% in detecting and identifying issues in the Load Panels. However, for real-world applications, it is necessary to increase the training data and improve the system's intelligence.