Architecture of intelligent wear prediction system

Abstract

To address the shortcomings of traditional offline monitoring for rotating machinery friction pairs, such as strong lag, high deployment costs, and dependency on external networks, a friction condition intelligent prediction system based on a three-layer IoT architecture using the ESP32 controller was designed. The system adopts a local offline deployment model, capable of real-time acquisition of multi-source parameters of friction pairs, wear feature extraction, and remaining life prediction. It features lightweight design, easy scalability, and high reliability, meeting the online monitoring requirements of industrial friction pairs.