Fault Diagnosis System For Hydraulic Shovel Based On Rough Set And Support Vector Machine

Hydraulic shovel plays an important role in engineering machinery field, which is widely used in construction working conditions. The poor and high intensive working conditions increase the probability of mechanical failure. Randomness and abruptness of fault diagnosis system make it very urgent to develop an intelligent one for hydraulic shovel. The accuracy of fault diagnosis system mainly relies on engineer’s experience, which has been mostly compromised in today’s hydraulic shovel field combined with electromechanical technology. Diagnosis system based on body sensing and experiences couldn’t meet the requirement to detect mechanical failure. Aimed to this problem, an intelligent fault diagnosis system with combination of rough set, and particle swarm optimization algorithm based on SVM was proposed for hydraulic shovel.This paper firstly put forward diagnosis methods and went through feasibility analysis for it according to the construction of EC230 hydraulic shovel and backgrounds of triggering mechanical failure. Rough set theory was utilized to extract features from complicated mechanical failure samples in order to eliminate the redundant features, which resulted in improved effectiveness of those samples. The fault diagnosis method based on SVM had been reviewed. After building model on it, parameters optimization of the model had been conducted to improve the accuracy of diagnosis and generalization ability of fault diagnosis model. Ultimately, the fault diagnosis system had been developed by VC++6.0, which lead to high performance and increased the benefits of EC230 hydraulic shovel.