Application of Log-Normal Distribution and Data Mining Method in Component Repair Time Calculation of Power System Operation Risk Assessment

摘要

Power system operation risk assessment can comprehensively take into account the occurrence possibility and severity of disturbances, and it is an effective complement to traditional deterministic security analysis. Calculating transient state probability of components in future short time duration is one of the key issues in power system operation risk assessment. In this paper, a novel probability distribution named'log-normal distribution'was presented for component fault repair time. The shape of probability density of log-normal distribution can describe the distribution characteristics of component fault repair times well. When the history data is limited, it can be used to calculate the instantaneous state probability, the method is simple and easy to use. But, if the system has a large amount of historical data, it will be suitable to establish the discrete model of repair time with data mining methods. Based on the information entropy weight allocation method, a real-time discrete model was established in this paper. The model can relatively accurate predict component repair time under various operation conditions. The results of example show that, if historical data is limited, a log-normal distribution can better meet characteristics of probability distribution of the actual sample, the transient state probability will be got accurately. When there is a large amount of historical data, discrete model established by data mining method is accurate and effective, can be very good to meet the field.