| Coal mine main ventilator is one of the main safety equipment for coal mine, the ventilator run safely and effectively is the guarantee for the underground workers and coal mine economic benefit. Take real-time monitoring for ventilator combined the feature of the coal mine ventilator, and fault self diagnosis for ventilator, such as rotor imbalance, axis misalignment, rotating stall, oil whip, design of real-time monitoring and fault diagnosis system.This thesis briefly analyzes the mechanism of common faults, select vibration signal of ventilator, oil temperature of bearing, revolving speed of ventilator, exit pressure of ventilator and flow of ventilator as the monitor parameters of this system. This coal mine ventilator real-time monitoring and default diagnosis is based on multi-sensor information fusion, and select characteristic layer fusion and decision-making layer fusion for data fusion. Adopt the improved BP network as the characteristic layer fusion algorithm, and the D-S fusion rule as the decision-making layer fusion algorithm.This thesis designed two4-9-5construction BP networks as the characteristic layer fusion algorithms, select the front and back bearing bush as the collection points of the vibration signal, pick up the RMS1, RMS2and GLP1, GLP2of these signals as the input of the first BP network, obtained an estimate of the run status of ventilator; select the oil temperature of bearing, revolving speed of ventilator, exit pressure of ventilator as the input of the second BP network, obtained another estimate of the run status of ventilator. From simulation result, the performance of the improved BP network has promoted. From the two output can judge the run status of the ventilator, but error is too big, the effect can not reached the expected. Through the D-S fusion rule to fusion the two outputs of BP network into one output, experimental result, the output after D-S fusion is better than the two output of the BP network. |
PDF Full Text Request