Research On Fault Diagnosis Of Explosion-proof Electric Rubber Wheel Vehicle Based On Neural Network

As auxiliary transportation equipment in mines,explosion-proof electric trackless rubber-wheeled vehicles have significant advantages such as large carrying capacity,superior dynamic performance,and easy operation.However,their electrical structure is complex and the working environment is harsh.Once the vehicle fails,it is difficult to quickly locate and find out the cause of the failure.In order to improve the safety production and transportation efficiency of coal mines,it is urgent to strengthen the fault diagnosis and monitoring of trackless rubber-tired vehicles.Aiming at the fault problem of explosion-proof electric trackless rubber-tired vehicles in coal mines,this paper proposes a fault diagnosis mode based on neural network diagnosis and supplemented by fault tree analysis.A fault diagnosis expert system is developed by combining the BP neural network optimized by the algorithm(Sparrow Search Algorithm,SSA),which helps the field maintenance personnel to deal with the field faults in a timely manner and quickly eliminate the faults,thereby improving the mine explosion-proof electric trackless rubber wheels.The fault diagnosis efficiency of the vehicle does not affect the production and transportation of coal mines.In this paper,the expert system inference engine is divided into a fault tree analysis module and a neural network inference module.The two methods complement each other and verify each other.First,by analyzing the fault characteristics of the explosion-proof electric trackless rubbertired vehicle,the common faults of each system are summarized,and the fault tree of each system of the explosion-proof electric trackless rubbertired vehicle is established as the reasoning basis for the fault tree reasoning module of the expert system;Each system of the vehicle can find fault signal collection points,collect fault vibration signals,extract the eigenvectors of the system under different fault states through wavelet packet decomposition as the input vector of the neural network,train and simulate the neural network,and use the stable network as the expert system neural network The reasoning basis of the reasoning module.In this paper,MATLAB neural network toolbox,C# programming language and SQL Server database are used to realize the software design of fault diagnosis expert system in Visual Studio program development platform.Finally,it is verified by an example that the results have high accuracy and reliability,and meet the design requirements.