Research And Application Of Pump Turbine Fault Diagnosis Technology Based On Fault Tree Analysis

Pumped-storage power station is a way to indirectly store electrical energy in order to solve the contradiction between peaks and valleys in power grid.In addition,it is also responsible for dynamic functions such as frequency modulation,phase modulation and accident backup.Therefore,pumped-storage power station is an important part of the operation of power systems,and they are the pillars to ensure the safety and economic operation of power grids.Especially our country promote the green energy today,it is necessary to use this functions of the pumping power station to ensure the safe and stable of the power system.The pump turbine is the core device in the pumped storage unit.Its operating conditions are variable and frequently started.The mechanical,hydraulic and electrical factors are coupled with each other,and the causes of the fault are complicated.Therefore,the traditional planned maintenance and accident maintenance can not effectively prevent the failure of the pump turbine.And if the failure occurs,it is difficult to quickly and accurately locate.In order to construct a predictive maintenance system of pumped storage units and achieve the functions of diagnosis and pre-diagnosis,and change traditional planned maintenance to predictive maintenance,it is necessary to construct a fault diagnosis model for the pump turbine.It can diagnose fault accurately and analysis it effectively,and make maintenance decisions.It can take solve fault effectively before it expands.In order to solve the problem that the current fault diagnosis method of pump turbine is not comprehensive and accurate in engineering,this paper studies various diagnostic methods and model algorithms,and constructs a comprehensive fault diagnosis model for pump turbine.And the results of the project are used in the actual station by Java.The main contents and innovative achievements of this paper have the following three points:(1)In order to solve the problem that the current pump turbine fault diagnosis method is not closely related to engineering practice,by collecting the typical fault record of the pump turbine and analyzing the associated parameters behind the fault,a typical fault correlation system conforming to the engineering reality is established.Further,combined with the actual operation of the power station monitoring system and experience of experts,a fault tree analysis and diagnosis model was established.(2)In order to solve the problem,which is based on the view of expert and simple calculation sometimes,that the current pump turbine fault tree diagnosis method was not accurate,this paper used artificial neural network model.,Combined with the fault tree analysis,this paper used ANN to diagnosis pump turbines and calculate the bottom event probability Based on the result,fault tree quantitative analysis was used to calculate the critical importance which can give operator suggestions for maintenance,then the pump turbine fault diagnosis model is established.(3)Based on the existing database of a monitoring system of a pumped-storage power stations,the fault diagnosis model of the pump turbine is used to diagnose the historical data of the pump turbine,and the effectiveness of the comprehensive fault diagnosis model of the pump turbine is verified.Then a comprehensive fault diagnosis software system for pumped storage units based on the three-layer B/S architecture was developed,and the research results of the project were successfully applied in pumped-storage power stations.