Application of knowledge engineering for decomposition and parallel processing in power systems

点击翻译按钮获取中文摘要

As a result of rapid growth and interconnection of electric power systems, various analysis related to them have become costly and time consuming processes as well as the fault tolerance capacity of the system has been reduced resulting a higher risk of total blackout. One way to over come this problem is employment of knowledge engineering based autonomous decentralized type control scheme which requires efficient decomposition of the system and parallel computation within each subsystem.Keeping the autonomous decentralized control as the main objective, research was carried out in various directions. They can be divided into following three topics mainly. (1) Non-uniform decomposition of power systems for parallel processing: This work developed a method which is different from the existing methods which are mainly based on uniform decomposition. But this method is proven to be more efficient than the existing methods. (2) Building an expert system using object oriented knowledge representation to restore power systems: This work developed a method to restore power systems by changing over transmission lines when they are over loaded. Object oriented knowledge representation makes it possible to represent power systems in a very natural way, to process data in parallel and to increase software productivity. (3) Identification of coherent machines and grouping them to construct equivalents: An efficient method which is based on modal analysis, to identify coherent machine groups in power systems was developed. This enables to construct a simplified model for the external part of a particular sub system. Hence reducing the computational time and memory requirement for various analysis related to each sub system.