Sequential short-term reliability evaluation considering repair time distribution

Abstract

The performance of a power system in an operational time frame can be evaluated by the short-term reliability assessment. The component reliability model usually uses exponential distribution for the repair time because of its simplicity. However, this assumption may lead to the inaccuracy in evaluation results, especially large errors in short-term evaluation. Therefore, the non-exponential distribution of repair time should be modeled. Unlike the exponential distribution model, the non-exponential distribution model has the property of “memory”, i.e., historical repair time has great influence on the component's availability in a very short time interval. This influence cannot be identified by traditional non-sequential evaluation methods. Thus, this paper proposes a sequential short-term reliability evaluation method to obtain more accurate indices by considering all the factors described above. A comparison study between the proposed method and non-sequential method is present and the method's advantages are illustrated using IEEE Reliability Test System.