An Acquisition and Calculation Platform for Predictive Maintenance in Coal Mine

Abstract

The fault diagnosis system for monitoring the traditional coal mine mechanical and electrical equipment are applied to collecting various types of data underground, and it transmitted the data up to the ground by ring network, processing data by Industrial Personal Computer (IPC) or cloud computing. There are some disadvantages in this traditional method: diagnostic or control delay, limited amount of raw data, higher cost, poor flexibility and so on. According to these problems, a platform based on STM32F4 with ARM Cortex-M4 core, integrated FPU floating-point arithmetic unit and DSP instruction set is designed to predicting maintenance, data acquisition and calculation. Meanwhile, the platform processes some sensor data acquisition units with an 8-channel parallel port 16bit A/D sampling module, which maximum sampling frequency is 200Ksps. The results indicated that the data collection platform can effectively acquire various types of monitoring data, and the peak-to-peak error of the collected vibration parameters is less than 4%. Moreover, the platform is not only competent for data frequency spectrum analysis, but also envelope spectrum analysis, the demand of edge computing, real-time analysis, as well as processes flexible layout.