Research On Wind Turbine Condition Evaluation And Maintenance Strategy Based On Multi-Source Operational And Maintenance Data

With the implementation of the "carbon peak,carbon neutral" strategy,new energy generation represented by wind power will usher in rapid development.By the end of 2022,China’s installed wind power has exceeded 365 million kilowatts,ranking first in the world,which has strongly promoted the clean and green transformation of China’s power system.However,because wind turbines are mostly in harsh operating environments,maintenance and replacement costs are expensive.At the same time,a large amount of multi-source operation and maintenance data from wind farms is in a dormant state,and its value in wind turbine condition evaluation and maintenance has not been fully explored.Therefore,how to accurately evaluate the health status of wind turbines based on the multi-source O&M data of wind farms and then develop maintenance strategies for wind turbines to meet the refined O&M needs of wind farms and reduce O&M costs has become a hot issue for research.In this regard,this paper empowers the condition evaluation and maintenance strategy development of wind turbines based on the multi-source operation and maintenance data collected by wind farm SCADA system.Fuzzy comprehensive evaluation is used to study the evaluation of wind turbine health status,and opportunity maintenance and preventive maintenance are introduced on the basis of reliability theory to explore wind turbine multi-component maintenance strategies.The research work in this paper is as follows:(1)The basic structure and working principle of wind turbine are summarized,and the functions of engine room,wind wheel system,variable pitch system,electrical system,transmission chain system and other important components are analyzed.On this basis,taking the wind farm SCADA system as the object,the basic framework of collecting wind farm multi-source operation and maintenance data is introduced in detail,and the multi-source operation and maintenance data of wind turbines monitored by the SCADA system are classified and summarized according to the Angle quantity,pressure force,temperature quantity,velocity quantity,vibration quantity,electric gas quantity and discrete quantity.It provides data support for subsequent state evaluation of wind turbine.(2)In view of the lack of comprehensiveness of the single-assignment method and the limitations of the current mainstream decision-making method to judge the state of wind turbines.In this paper,we propose a method to evaluate the condition of wind turbines based on the comprehensive weighting of indicators,considering the asymmetric closeness.Firstly,the wind turbine condition evaluation indexes are constructed considering the performance and output status of wind turbines,and the comprehensive weights of each index are determined based on the principle of minimum discriminative information.Secondly,the affiliation function of the index layer is constructed by set pair analysis,and the affiliation degree of the target layer is forwarded by the weighted average operator.The asymmetric proximity method is used to calculate the proximity between the state level and the affiliation degree and determine the wind turbine state according to the principle of proximity selection.Finally,the monitoring data of a1.5 MW wind turbine in a wind farm are analyzed as an example.The results show that the integrated weighting method and the asymmetric closeness method can effectively improve the accuracy of wind turbine condition assessment.(3)Aiming at the deficiencies of the current stochastic maintenance plan caused by ignoring the differences of various components and optimizing based on the unified reliability threshold,a reliability based preventive stochastic maintenance method for wind turbine multi-components was proposed.Firstly,Weibull distribution is used to describe the reliability distribution of fan gear box,blade,main bearing and generator.Then,the opportunity replacement threshold and the opportunity incomplete maintenance threshold of each component are regarded as independent optimization variables.The goal is to minimize the sum of equipment rental cost,personnel cost,maintenance adjustment cost and shutdown loss cost during the whole life cycle of the fan.A multi-component preventive opportunity model for wind turbine based on reliability is constructed considering reliability constraint and maintenance scope constraint.Finally,genetic algorithm is used to solve the problem.An example analysis shows that the maintenance cost and shutdown loss of the unit can be significantly reduced by optimizing the maintenance threshold of each component separately.