Research On Dynamic Model Reduction Of Regional Power Grid Using Balanced Realization

With the increase of the power system size and complexity, the dynamic equations of the model dimension increases rapidly, calculation of exponential volume growth, making the power system analysis, simulation and controller design problems are becoming increasingly acute, and sometimes even real-time calculation can’t be achieved. Therefore, the study simplified model, reducing the dimensions of the model has important academic significance and practical value.This paper introduces the research actuality of the domestic and international model reduction. By analyzing and comparing the variety of model reduction techniques, we draw the conclusion that balanced realization theory is very suitable for the regional power grid model reduction.To solve the problem of model reduction in regional power grid of linear dynamic models,we studies the Low Rank Choleski Factor-Alternating Direction Implicit Method, and studies an efficient linear system reduction method that quickly to solve approximations to the controllability and observability gramians of large power system models.To solve the problem of model reduction in regional power grid of nonlinear dynamic models we presents an efficient reduction method that computers approximation of high-order nonlinear power control system using the balanced empirical controllability and observability gramians. The method, which involves the formation of controllability and observability gramians by Karhunen-Loeve decomposition technology, is an extension of the balanced truncation method that has been applied extensively in the reduction of linear system.At last, some test regional power grid systems were analyzed for model reduction by the proposed method. The results of simulation show that the proposed method is very efficient.