Research On Operational Condition Monitoring And Health Maintenance System For Wind Turbine

With the rapid development of China’s wind power industry,the installed capacity has increased year by year,and it has become China’s the third largest source of electricity.However,with the commissioning of large-scale wind farms,the proportion of aging units and quality assurance units has been rising.Also,with the high failure rate of the units and the backward operation and maintenance management level that have led to the unit operation capital investment increased with each passing year.So how to ensure the security of wind turbine operation,and improve the maintenance and management level has become an urgent problem for the wind power industry.Based on the research on the structure and working principle of grid-connected wind turbine,this paper deeply analyzes its operating characteristics,and studied the operational condition monitoring and health maintenance system for wind turbine.The main work is as follows:Firstly,based on the theory of system science,the occurrence and evolution of wind turbine faults are deeply studied.Under its guidance,the mechanism of wind turbine three systems and 42 typical fault modes were got a comprehensive analysis,which were from the view point of safety,economy of energy efficiency conversion capability and reliability of equipment operation performance.Meanwhile,the mapping relationship between fault modes and symptoms was clarified.Based on the FMEA and FTA methods,the unit information was comprehensively combed and expressed to form a knowledge system of wind turbine operational condition monitoring and health maintenance.Secondly,based on the analysis of the failure mechanism of wind turbine,a research on unit condition indicator analysis and system construction was carried out.Guided by the comprehensive evaluation theory,supported by the structural composition and functional characteristics of wind turbine,a indicators system was initially built,and the corresponding relationship between unit equipment and function was constructed.From the view of economy,safety and reliability,this paper comprehensively analyzed the operation state indicator characteristics of the unit,which established the "three in one"mechanism of unit operating state indicators.Then the corresponding relationship between unit indicators and equipment functions was clarified.A multi-block theoretical analysis method based on correspondence analysis was presented.By the analysis of correlation between unit operation indicators,which clarified the relationship between operation indicators and equipment,improves the content of index system,then the indicator system of operation status associated with "indicator-equipment-function" was build.Third,aiming at the problems of inaccurate state estimation and complex model structure in current intelligent fault diagnosis methods,a composite network diagnosis model for wind turbine that integrated multi-homogeneous and heterogeneous was proposed.(1)From the network structure,the multi-layer and multi-level topology structure forms the "whole to local" unit state analysis path,which decomposes the complex fault problem of the unit by layers,reduces the complexity of fault analysis and the network structure,and also improved the efficiency and feasibility of network analysis.(2)From the function of the model,the composite network model includes data preprocessing,feature extraction,fault early warning and diagnosis.which were related to state analysis,forming a systematic diagnosis process.(3)Form analysis reliability,the factor analysis-hierarchy nonlinear state estimate technology(FA-HNSET)is proposed in the early warning.Through the data characteristics on the degree of equipment operation state representation analysis,the early warning sample data was reasonably selected to improve the sample data characteristics and the accuracy of fault early warning.In the diagnosis section,the event tree analysis-fuzzy petri net(ETA-FPN)multi-layer network model was proposed,and a two-layer reasoning mechanism combining top-level fault event reasoning and local fault symptom reasoning was established.The network node information of the model was comprehensively analyzed by improving the information entropy method,which made the network structure was reasonably constructed to realize the description of the sequence and correlation of the unit fault events.(4)From the analysis effect,the network’s systematic diagnosis strategy could provide a complete evidence chain for the fault diagnosis results of unit,which described the fault comprehensively and provided key information support for the unit health maintenance.Fourth,a comprehensive study on the health maintenance of wind turbine was carried out.(1)A fuzzy evaluation method was proposed to evaluated the health of unit,which result is the necessity of making maintenance decision.(2)By the ARMA model,the reasonable deduction of unit operation state was carried out.Then the maintenance time was depended on deterioration trend prediction.(3)Guided by the theory of FMEA and FTA,the failure modes and the maintenance sequence of specific failure equipment were determined by statistical analysis.By using the method of logical decision diagram analysis and fault bottom event property division,the maintenance mode of specific fault event was determined,then the specific work of unit equipment health maintenance was completed,which was improved the pertinence of unit maintenance.Finally,the wind turbine operational condition monitoring and health maintenance technology framework was applied to engineering practice.A wind turbine operational condition monitoring and health maintenance system was developed based on the 1.5MW onshore wind turbine of Guohua Changzhou wind farm.This system,which adopted database MySQL and B/S modes,was promoted the transformation of results and engineering applications.