Research On Fault Diagnosis System Of Hoisting Rope Tension Based On Intelligent Diagnosis

The hoist is the key equipment for normal operation of mine hoisting system.The rope is a flexible body connecting the hoist with other components.And it stands some various loads during the working process,accompanied by vibration and shock.The running state of the hoist can be estimated by monitoring the tension in real time.The existing fault diagnosis system based on rope tension is monitored by setting threshold and formula expression for overloading,uncompleted unloading and unbalance of tension and so on.The state of monitoring is less and the degree of intelligence is poor.Therefore,it is of great significance to study a fault diagnosis system of the tension based on intelligent diagnosis for improving the fault diagnosis system of tension and the safety production.The core between intelligent diagnosis systems is the differentiation of diagnosis algorithm,and data is the foundation of all algorithms.Therefore,in order to facilitate the generalization of the intelligent diagnosis system in this paper and to obtain the dynamic tension data under different states of the monitored mine,the mathematical model of tension in normal and typical fault states is established,and the corresponding data are extracted to study the algorithm.For the model of normal operation,considering the influence of the coupling vibration on the dynamic tension,the transverse-longitudinal coupling kinetic model of hoisting rope with an arbitrarily varying length and double excitation in mine hoisting system was established by Hamilton principle.The finite difference method was used to discretize the partial differential equation to ordinary differential equation.And the Matlab simulation was applied to analysis results.On this basis,the dynamic tension in normal operation of hoisting system is deduced and verified by experiments.For the model of fault,the fault tree of hoist is analyzed.And the three typical failures of card tank,over-winding and rope slipping are simulated.The tension characteristics of faults at different locations and speeds are also compared.A training data set is built by combining normal and faults data.It can provide a basis for the research of intelligent diagnosis algorithm.After preprocessing the tension data collected on the field,it is used as the input of prediction data of the machine learning algorithm.The running state of hoisting system is forecasted by RBF-BP neural network through calling the training model.This model is formed by training the data set.And forecast states include card tank,over-winding and rope slipping.This paper compares the prediction results with the LSSVM algorithm after selecting the RBF function as the kernel function.On this basis,the two fixed parameters of penalty algorithm and RBF kernel parameters in LSSVM are further optimized by artificial bee colony algorithm.The prediction accuracy and the mean square error(MSE)performance of the optimized algorithm(ABC-LSSVM algorithm)have been further improved.And it can be used for the design of the fault diagnosis system of tension as an intelligent diagnosis algorithm.In order to popularize the intelligent diagnosis algorithm to the industrial field,the fault diagnosis system of tension based on ABC-LSSVM diagnosis algorithm is achieved by LabVIEW.It can be used for monitoring by the signal of tension in real time.And the intelligent diagnosis algorithm based on ABC-LSSVM is used to diagnose the faults which are difficult to monitor in the past,such as card tank,over-winding and rope slipping.The system will alarm and the fault database is updated when the fault occurs.Meanwhile,the Web technology is published to the external network to achieve the remote access of the fault diagnosis system in the external network.The running of double thread program is executed to achieve the display,storage and reading of tension data in real time,disk management and other functions while intelligent forecasting.The system also provides a data interface and can be connected with the existing monitoring system to establish a complete fault diagnosis system.