Investigation of the Accuracy Improvement of Advanced Dominant Eigenvalue Estimation Method for Electric Power Systems

The authors have been developing an estimation method of a dominant eigenvalue in large scale power systems using the measured data of the power system. A dominant eigenvalue means the eigenvalue due to long cycle oscillation mode between large scale power systems and has an important information on stability. This paper describes the basic principle of eigenvalue estimation method, the analysis of factors giving influence to accuracy and the accuracy improvement method. The result from the basic principle of eigenvalue estimation method shows that the estimated eigenvalue fluctuates in certain time domains and the accuracy is not satisfactory. The studies were made and it was clarified that frequencies lower than the frequency of the dominant eigenvalue are major sources for fluctuation and noise. In order to improve accuracy, a new method was proposed by applying maximum entropy method (MEM) to obtain power spectrum of maseaued data and by introducing new argorithm for descriminating signal from noise using MEM results. Results of applying the proposed method to dominant eigenvalue estimation show the accuracy was improved and the validity of the proposed method was verified.