Reliability Analysis Of Wind Power System Based On The Semi-Markov Process And Copula Function

With the continuous development of the wind power industry,the operational reliability of wind power system has become increasingly prominent.It has importantly theoretical guidance significance and practical application value for the optimization of the wind power system configuration and operation maintenance plan,to study the changes rules of performance state and reliability of wind power system during the actual operation process.Focusing on the randomness of the performance degradation,the fault recovery and the system requirements,as well as the correlation between each wind turbines in the wind power system during the actual operation,this dissertation reveals the change regulation of the operational reliabilities for the wind power system according to analysis the change rules of component performance states,based on the reliability analysis theory of multi-state systems,in accordance with the research route that is from component level to system level,from independent to dependent.The main contributions of this study include the three aspects:(1)A performance semi-Markov model for describing the randomness of the wind turbine performance degradation and the fault recovery degree is established.The changing rule of the wind turbine performance state is analyzed,too.The changing regulations of the performance state and instantaneous reliability index of wind power system during the actual operation process are revealed,using the universal generating function.The results show that both the changing rates of the performance state probability and the instantaneous availability of wind power system show a tendency to decline from large to small and terminally to a certain stable value;the system has higher instantaneous availability and higher operating safety,when the requirement of wind power system is in the intermediate state of system performance.At the same time,the relevant parameters in the model are derived from the fitting and estimation of field data directly.The state transition time distribution forms between the performance states are flexible and diverse,which can describe the performance state transition more accurately during the actual operation of the system,and also provides an effective approach to analyze the status changes and dynamic reliability of wind power systems more precisely.Through the combination of semi-Markov process and universal generating function,the complexity of wind power system reliability analysis is reduced effectively.(2)To solve the probabilities of the requirement state in actual operation process,a requirement semi-Markov model is established to describe the characteristics of random requirements of wind power systems.Combining with the performance distribution of the wind power system,the changing rules of the instantaneous availability of the system are analyzed under the condition of random changes in wind power system demand.The results show that the changing rate of the requirement state probability of wind power system shows a slowly decreasing,then gradually inceasing trend,and the decreasing rate is larger than the increasing one.It is different from the variation trend eventually reach a certain stable value when the random changes of the demand is not considered.That is mainly caused by that semi-Markov model has the nature of state transition times to be arbitrarily distributed among states,and the characteristics of random changes in system requirements during performance degradation process.Consequently,the result indicates that it is necessary to consider the randomness of the demand changes when analysis the reliability of wind power systems.(3)The correlations between wind turbines are described by Copula function,and the correlations between them in actual operation process of wind power system are revealed.And the results show that the correlations between wind turbines in the different states of wind power system are different each other under the multi-state background.On this basis,different Copula functions are selected to describe different related structures.Furthermore,the time distribution functions of the performance state transition for wind power systems are analyzed.Next,the performance semi-Markov model of wind power system under relevant conditions is established,and the changing rules of system performance state and instantaneous reliability indexes of wind power system are analyzed.The influence of reliability index of wind power system under different correlation degrees is studied,too.The results show that the instantaneous availability of wind power system increases due to the increasing of correlation degree between wind turbines.They are both between the instantaneous availabilities under the extreme conditions: “extremely independent” and “extremely dependent”.The approach that combined SMP and Copula function is applied to the reliability analysis of wind power system under relevant conditions,and the effectiveness and feasibility is verified.The research results of this study provide an effective way to analyse the reliability of multi-state system under the relevant conditions.