A knowledge-based system for power system weekly scheduling

Abstract A new knowledge-based system for weekly power system operation scheduling is introduced. There are a number of improperly structured constraints which are difficult to solve using an algorithmic approach in the weekly power system operations scheduling domain. The knowledge-based system is proposed to handle normal system behaviour as well as emergency cases such as lack of water inflow, generator outage and network problems. The proposed weekly scheduler includes the following functions: load prediction, inflow prediction, possible run-of-river hydro production, storage hydro production, unit commitment, generator maintenance and power interchange between interconnected power systems. The need for the proposed ‘intelligent’ tool has been recognized by the Croatian power utility due to expected benefits such as: minimization of generation costs and increase of reliability level. The real power system is utilized to test the proposed knowledge-based system.