Research On Operational Reliability Evaluation Approach Of Power Component Based On Imprecise Probability

With the construction of UHV,AC and DC mixes,development of new energy sources and the construction of the Global Energy Internet,to ensure the safe and stable operation for the power systems is facing great challenges.The safe operation of power component is an important basis for the safe operation of the power systems,as the key components,the overhead transmission line(OTL),large power transformer's health level is directly related to the operational security and stability level of power systems.In this regard,using advanced and reasonable approaches to evaluate the operational reliability level of electrical components,i.e.,to estimate the outage probability of the OTL,and to diagnose the transformer faults have important significances for security and steady operation of power systems.However,with extensive applications of new energies and new technologies many new components are lack of statistical data,and it is difficult to evaluate the operational reliability of power component with limited statistical samples.Under this condition,if given deterministic evaluation results,it is not only valuable,but also can mislead the dispatching strategy,and destroy the operational reliability of power systems.Imprecise probability is an effective approach to deal with the incomplete sample information,which can be used to deal with the disadvantage of the lack of samples in the component operational reliability evaluation,which takes the form of interval probability instead of deterministic single-valued probability.In this paper,imprecise probability theory is applied to the reliability evaluation of power component,and the fluctuation range of component reliability index is evaluated under the limited sample conditions,which can provide a more objective decision-making basis for power system operational risk control.In this paper,a novel approach was proposed to estimate the imprecise conditional outage probabilities of OTLs based on the credal network(CN),which could evaluate the interval ranges of conditional probabilities.Based on the historical outage statistics and the component operational conditions at the target period,a credal network,which used the imprecise Dirichlet model(IDM)to get the imprecise probabilistic dependencies,was set up to estimate the imprecise conditional outage probabilities.Therefore,the imprecise outage probabilities of OTLs can be obtained by using the reasoning algorithm of the credal network.The proposed approach can reflect the variation of the conditional outage probabilities with respected to the operational conditions of the OTLs,and breaks a new way for the reliability assessment of power components with limited outage observations.In this paper,a novel imprecise probability based approach for transformer fault diagnosis is proposed.The approach is composed of IDM and a credal classifier.Instead of providing each type of fault with its single-valued probability,the approach gives the interval probability for each type of transformer fault with limited historical fault sample records to explicitly indicate the uncertainty of the diagnostic result objectively.Then,according to whether there is an overlap between the interval probabilities of each type of fault,the proposed approach is selected to output either only one explicit fault or probable multiple faults occurred in the transformer.The approach proposed can make up for the defects of conventional methods which cannot deal with limited fault samples or detect the multiple incipient faults occurred in a transformer precisely,thereby,the accurate diagnostic rate can be effectively improved by the implement of the approach proposed.