Research On Conditional Monitor And Fault Diagnosis Based On Internet For Large-scale Hoisting System

Mine hoist is very important in mining industry, and its fault would result severe accident. But mining technology dropped far behind. The paper put forward an internet-based condition monitoring and fault diagnosis system.First, the paper introduced the-state-of art of hoist monitoring items and methods, analyzed two kinds of internet-based monitoring and fault diagnosis models, Browse and Server(B/S) model and Mobile Agent model. Based on the real working condition of mine hoist and repository-updating needs of fault diagnosis, a model is put forward.Next, according to the fault sort and severe degree, a fault tree is set up. The paper emphasized on the research of hoist braking system, established 3-layer Back propagation(BP) Artificial neural net(ANN) on the basis of Load Pressure Method (LPM).Considering the real require of fault diagnosis, the number of nerve cells for each layer has been confirmed , sample collecting and training has been accomplished. The paper introduced component and function of hoist hardware system. After analyzing the requirement and building models, according to the thoughtway of software engineering, application software compiled by Visual Basic has been accomplished and debugged.Finally, the condition monitoring and fault diagnosis test has been made ,The result shows that the software can work well in hoist monitoring and diagnosis, the ANN training program can be used to train other object's samples ,the ANN diagnostic model can find the fault of other objects without any changes, the model can be used in general, and the software system can update by itself.