Research Of Fault Diagnosis On The Temperature Control System Of Coal Spontaneous Combustion Laboratory Terrace

With the development of science and technology, the rising of the industry automation, more and more computer control systems are applied in the area such as railway, coal mineral, iron, petrochemical. Many complex systems with high technology and all- round function come into being, hose systems are large-scale and valuable. Once they have the faults, the result is probably disastrous. So those systems must be high secure and well reliable. Under the requirement , fault diagnosis emerge with the tide of the times. It gives many important helps to enhance the security of the control system.In the area of coal mine, coal spontaneous combustion is very hazardous, it can bring on vast economic loss and human being injury or decease. So it is significant to study the rule of coal spontaneous combustion. And if we carry out coal spontaneous combustion experimentation, we can study the reason of coal spontaneous combustion and the rule in the course of spontaneous combustion raising temperature. So it is very important to undertake normal working of the coal spontaneous combustion experimentation system. The research content of this programer is to carry out fault diagnosis of the coal spontaneous combustion experimentation temperature control system and predict its working condition tendency. The programmer carries out the research of fault diagnosis on the temperature control system by means of grey system theory and unnatural neural network. Aiming for an experimentation of coal spontaneous combustion, we apply both methodologies. With the grey predicting method, we predict the behaving condition of the equipment, and carry out simple fault diagnosis. At the same time, we expound the enlarged information net is founded by neural network, resting on the fault diagnosis strategy we conclude that the fault occurs or not. If there is the fault, we must figure out where it is and recover the signal of fault. We take advantage of BP network to learn, and put MATLAB 6.1 to use to software achievement, finally we get the conclusion.The result of instance emulation demonstrates the both methodologies possess feasibility, the result of fault diagnosis comes up to the fact, the accuracy is high.