Research On Condition Assessment And Maintenance Decision Of Wind Turbine Based On Operation And Maintenance Data

In recent years,the wind energy utilization industry in the world has made great strides,but the complicated wind turbine system and poor working conditions lead to frequent failures.Therefore,it is one of the key issues that need to be solved in the healthy development of wind power industry that how to comprehensively study the status assessment of wind turbines and promote the development,operation and maintenance of intelligent operation and maintenance of wind farms based on the operation and maintenance data of in-service generating units.Aiming at the above problems,this paper puts forward the methods of state evaluation and maintenance decision-making of wind turbines.As a prerequisite for condition assessment which is to construct an evaluation system that reflects the status of generating units,the importance of the components of the generating units needs to be evaluated to select the components that have an important influence on the operating status of the generating units.Therefore,a gray-based fuzzy importance evaluation method based on interval numbers is proposed.Based on the evaluation system formed by the key components or systems,this paper presents a model of wind turbine condition assessment based on the fusion theory,and determines the condition of the turbine by the assessment results.For maintenance decision-making,this paper proposes a multi-component opportunity maintenance decision based on preventive maintenance.Firstly,for the importance evaluation model,aiming at the uncertainty of membership and gray value in gray fuzzy evaluation,and without comprehensively considering the relationship between grayness and fuzziness,this paper proposes a grey fuzzy importance evaluation method of interval number.The method expands the point values gray and membership into the interval form,which increases the objectivity of the evaluation process.By introducing the OWA operator,the influence between the gray part and the modal part is quantitatively analyzed.The case analysis shows that this method can effectively evaluate the importance and has the advantages of high dispersion of evaluation results and not easily affected by gray value than traditional method.Secondly,for the condition assessment model,in view of the fact that the state monitoring results of a single system or component are difficult to fully express the condition of the generating unit,a condition assessment model of the wind turbine generating unit is proposed in this paper.The model builds the evaluation index system based on the evaluation results of importance degree.Based on the normal cloud model,taking into account that the threshold of temperature index which accounts for the majority of SCADA data is greatly influenced by the ambient temperature and power,this paper proposes a modified temperature.The concept of DSmT is introduced to eliminate the influence of power and ambient temperature,and a more objective threshold is obtained.Considering the subjective and objective fusion of strongly conflicting evidence types,this paper presents the DSmT evidence fusion theory.For the cloud model parameters,a method to determine the digital features based on the 3En principle is given,and a solution to the problem of overlay and relative closeness of the cloud model membership curve is given.The case study shows that the model can alarm the early defects of the unit and get the trend of deterioration of unit status before failure,which is of guiding significance to operation and maintenance.Finally,for the decision-making model,this paper selects the representative repairable parts or systems in the state assessment,and proposes a multi-part opportunity maintenance decision-making based on the periodic preventive maintenance.The decision model considers the economic dependencies between components and optimizes the opportunity maintenance and preventive maintenance time with the objective of minimizing the total maintenance cost of multi-component unit time.The case study shows that the model can reduce the cost of multi-component maintenance and optimize the maintenance cycle.