Base On Reliability Study Of Improved Bayesian Network GO Method In HVDC Transmission System

With the rapid development of China's national economy and power industry,the power system is facing an unprecedented development situation.In recent years,with the development of electric power resources in the west,to send electrical distance more and more far,power transmission capacity is more and more,HVDC(high-voltage direct current)in the development of the power grid construction plays a more and more important position,combined with AC and DC power transmission project construction cycle is long,high cost,running in the special environment thereby increasing the difficulty of maintenance,therefore on HVDC system reliability analysis is very important.The main method of DC transmission system reliability analysis of frequency-duration method,fault tree method and GO method etc..Due to the HVDC system with a large number of components,flexible operation,the method in the analysis there are deficiencies:frequency continuous time method for complex system reliability analysis will suffer from the curse of dimension problem;fault tree modeling and analysis using the minimal cut set approximation errors;go method of total signal enables the quantitative calculation process is cumbersome.Therefore,we need to seek a balance point between the accuracy and the computation of the algorithm.Based on the above analysis,this paper introduces the Bayesian network GO method,which combines the GO method with the powerful reasoning ability of Bayesian network.In view of the shortcomings of the simple and non repairable system,the improvement is put forward.For the existence of complex correlation of HVDC system on parallel correlation structure equivalent to the element nodes and according to the corresponding algorithm in Bayesian network construction,to solve the equivalent unit reliability characteristic quantity,update model of the whole system,so as to simplify the system model for precise reasoning.The traditional go method with an example of HVDC system of the same example of reliability analysis results were compared to verify the correctness of the algorithm and fast,also Bayesian network two-way reasoning ability,and can also identify the weak link in the system.Taking into account the repairable system reliability is in constant change,in order to more accurate analysis of the availability of system with the trajectory of the time,then introduces the dynamic Bayesian go method for analyzing system reliability index at any point in time.