AN AGENT TECHNOLOGY BASED METHOD FOR POWER SYSTEM TOPOLOGY VERIFICATION WITH USE OF RADIAL BASIS FUNCTION NETWORKS

摘要

The paper deals with the power-system topology verification, which is important from the viewpoint of dispatchers and all applications, which use the power-system topology model. Taking into account that the topology verification in state estimation can fail when there are multiple topology errors, in the paper the topology verification before state estimation is considered. For the aim of topology verification, the knowledge on a power system, which is in relationships among measured quantities (active and reactive power flows, voltage magnitudes) is taken into account. Using this knowledge, special so-called unbalance indices are defined. Values of these indices create sets characteristic for different cases of occurrence of topology errors. The verification problem is especially complex when there are multiple topology errors and there are errors burdening measurement data. Then, radial basis function networks are used to solve the considered problem. The topology verification based on the unbalance indices can be considered as many linked processes. In this situation an agent approach to the verification is proposed. Using the described method, the verification is not only very efficient process but also it becomes much more flexible, e.g. in terms of scalability, distributed processing, parallel computation.