Dynamic contingency screening with simplified fuzzy inference in power systems

This paper proposes a fuzzy inference based method for dynamic security assessment in power systems. In recent years, the deregulated power market encourages IPP to serve power utilities so that less expensive generation cost is obtained in power system operation and planning. As a result, the generation scheduling becomes more complicated while the degree of uncertainty increases due to occurrence of potential faults. Thus, it is much more important to consider dynamic security assessment in power systems. In this paper, an efficient method is proposed to handle contingency screening for transient stability as one of the problems in dynamic security assessment. Contingency screening requires efficiency to process a bunch of contingencies. To meet the conditions, this paper presents simplified fuzzy inference to enhance the efficiency and accuracy of the solution. To optimize the fuzzy system, this paper determines the number, location and shape of the fuzzy membership functions with tabu search. The determination of them results in a complicated combinatorial problem. Tabu search has an adaptive memory called the tabu list that controls the sequence of the search process and obtains better solutions in terms of global optimization. After tuning up the fuzzy membership functions, the output of the consequence is tuned up by supervised learning. Also, the Akaike information criterion (AIC) is introduced to construct a reasonable model so that the model avoids the model overfitting. The proposed method is demonstrated in a sample system