A distributed inverse iteration method for eigenvalue analysis of interconnected power systems

Since China power grids have a hierarchical architecture in operation and management, centralized computation patterns are difficult to meet the demands of small-signal-stability analysis of the bulk interconnected power systems. A distributed eigenvalue algorithm derived from the inverse iteration method is proposed. It can not only obtain eigenvalues and eigenvectors from power system state matrix but also provide participation factors of all generators. In the computing process, the algorithm only requires exchanging data of boundary nodes and a small amount of other information of different regions. Therefore, it is very suitable to be deployed in a WAN (wide area network) based distributed environment. The algorithm has been tested on an IEEE39 system.