The Embedded Fault Diagnosis System Of Shearer Based On Neural Network Optimization

China is a country of abundant coal resources, with the highest coal production in theworld. Coal winning machine, as one of the key equipments in fully mechanized coal mining,is necessary to improve the mechanization of coal mining. However, due to the poor workingcondition, the complicated structure, variability of coal layer as well as the frequent overload,the coal mining machine is often damaged and even fails to work, which may causes hugeeconomic loss and accident. Therefore, it is necessary to diagnose the coal mining machineonline; find and handle the fault timely, to reduce the unnecessary economic losses andprotect the safety of staff.With the development of technology, embedded system is well developed and widelyused in many field, especially the digital signal processing chip (DSP), which has a widerange of application in aerospace, automotive navigation, portable devices, industrial controland other fields. The digital signal processing chip TMS320F28335, developed by TI, iscommonly used in industrial control and fault monitoring.This thesis researches a embedded fault diagnosis system of high-power electric haulagecoal mining machine, with TMS320F28335as the core chip, and also builds a reasonableperipheral circuit, making it a complete hardware system. In terms of software system, itadopts the combination of CCS and MATLAB, reducing the developing time and realizing thefunction of A/D sampling, FIR filter, FFT transform, SD card storage, eigenvalue extractionas well as rough set—neural network fault identification and diagnosis. Finally, theRS232-RS485converter circuit and relevant software are designed to realize the serialcommunication between DSP and PLC and ultimately the fault diagnosis system based onembedded online.