The Research And Application Of Vibration Online And Fault Diagnosis For Hydroelectric Generating Sets

Hydropower plays an important role in clean energy.The research on the monitoring and controlling of hydropower equipment is an important foundation of the hydropower development.Hydroelectric generating unit is one of the key equipment of hydropower stations.The stabilization of Hydroelectric generating unit physically guarantees the safety,quality and economies of the power supply.In recent years,the development trend of small and medium-sized hydropower stations is "unattended"(few people on duty),which requires to improve the level of intelligence to achieve the unit maintenance.However,the present monitoring and fault diagnosis system and the corresponding network configuration for Hydroelectric generating units are complex,costly and inconveniently maintained,which cannot meet the needs of small and medium-sized hydropower station that is cost-effective,simple operation.Therefore,it is of great significance to research the on-line monitoring and fault diagnosis system of hydroelectric generating units in small and medium-sized hydropower stations.Since vibration is an important factor that influences the security and reliability of hydroelectric generating units,the core technology of vibration fault diagnosis of hydroelectric generating units is the signal feature extraction technology and fault diagnosis methods.The specific contents of this paper are as follows:Based on the generalized S-transform,via analyzing the energy spectrum of the generalized S-transform,an energy feature extraction method is proposed.This method uses the generalized S-transform to calculate the energy spectrum of the vibration signal to extract the energy value of the signal with specific frequency and complete the feature extraction of the vibration signal of the unit.By comparing with the standard energy distribution of the original signal,the result shows that the method can extract energy characteristics,which has high precision.Combining generalized S-transform with singular value decomposition(SVD),a SVD extraction method based on generalized S-transform for impulse feature is proposed.In this method,the spectrum matrix of the vibration signal is calculated using the generalized S transform.The singular entropy is calculated using the singular value decomposition.The threshold,being less than or equal to which the singular value was set zero,is determined by the singular entropy.Finally,the spectrum matrix is inversed to obtain the time domain impulse characteristics of the signal.Compared with themethod of impact feature extraction based on S-transform spectrum analysis,the result shows that the method used in this paper has better performance.Combining quantum particle swarm optimization(QPSO)and support vector machine theory(SVM),a support vector machine algorithm based on quantum particle swarm optimization is proposed.The quantum particle swarm optimization algorithm is used to optimize the parameters of the support vector machine which are the penalty factor C and the kernel function parameter g.Then the support vector machine adopts the optimized C and g to complete the classification model training.The simulation results show that QPSO-SVM algorithm doesn't only provide favorable classification results,also has better diagnosis precision than methods such as improved particle swarm-support vector machine algorithm(PSO-SVM)and genetic-support vector machine(GA-SVM).For the actual engineering demand of on-line monitoring and fault diagnosis device for hydro-generating units of small and medium-sized hydropower stations,a tentative solution is designed on the hardware composition,sensor selection and the main structure of the software system.