Development On The Mine's Main Ventilator Condition Monitoring And Fault Pre-warning System

The underground environment of the coal mine is complex and harsh,and there are various toxic,flammable and explosive harmful gases,which threatens the safety of the underground workers.As the core equipment of the coal mine ventilation system,the mine's mine ventilator can effectively discharge harmful gases and supply fresh air to ensure the working environment of underground personnel,which is an important guarantee for coal mine safety production.The health status of the mine's main ventilator is related to the working environment of the mine.Once a fault occurs during the operation of the mine's main ventilator,it will seriously affect the safe production of the coal mine and even cause major accidents.Therefore,this paper studies the operation status monitoring,diagnosis and pre-warning of the mine's main ventilator.By analyzing the common fault mechanism and characteristics of the mine's main ventilator,the operation status monitoring program is formulated,and the vibration signal-based fault feature extraction,intelligent fault diagnosis and trend pre-warning method are proposed,and an integrated system of the mine's main ventilator condition monitoring,diagnosis and pre-warning to ensure its safe operation.Aiming at the problem that the non-stationarity vibration signal of the mine's main ventilator makes it difficult to accurately extract the fault features,an idea of feature extraction based on the working conditions of the speed variation is proposed.Non-stationary operating conditions are divided into two types:speed fluctuation and variable speed.For the speed fluctuation conditions,the wavelet packet decomposition is used to extract the energy of the vibration signal frequency band under different faults as the fault characteristics and for the variable speed conditions,the order analysis is used to convert the non-stationary signal in the time domain to the pseudo-stationary signal in the angular domain,and then compares the angular domain characteristics of different fault signals to obtain the fault characteristic parameters.Through the combination of the two methods,the vibration failure characteristics of the mine's main ventilator can be accurately extracted under non-stationary conditions.In order to solve the non-linear problem of the mine's main ventilator fault diagnosis based on data,the BPNN method is introduced,and the extracted feature are input into BPNN for classification training and recognition.The experimental results show that BPNN has the disadvantages of slow convergence speed and easy to fall into a local minimum.In this paper,the particle swarm optimization(PSO)algorithm is introduced to optimize BPNN.The research results show that the proposed method has significantly improved the accuracy of fault recognition and the speed of algorithm convergence.In view of the time series non-linear fitting and solving problems of the mine's main ventilator failure pre-warning,based on the study of BPNN,the output layer activation function is improved to enhance the network dynamic tracking performance.A method for pre-warning combined with dynamic time series and PSO-BPNN is proposed,and the correctness of this method is verified by the data of laboratory and ventilator field operation.Finally,the condition monitoring and fault pre-warning system of the mine's main ventilator is developed,including the hardware platform and software.The hardware platform uses an ordinary PC as the host computer,and consists of sensors,acquisition cards,and conditioning circuits to form a complete signal transmission path.The software uses Lab VIEW and MATLAB joint programming to realize the condition monitoring,vibration feature extraction,fault diagnosis and pre-warning,data management of the mine's main ventilator.The function verification of the system each module was completed in the coal mine production site.The results show that the proposed feature extraction,fault diagnosis and pre-warning methods all achieve the expected functions and meet the needs of field applications.