An Intelligent Hierarchical Voltage Control based on the Dynamic Partition of Control Device Groups

Abstract

Optimal coordinated voltage control (CVC) for large scale power systems is a hard combinatorial optimization problem with fast response requirements. Intelligent techniques have been widely applied and proved with good performance for hard combinatorial optimization problems. However, with an increase in system scales, the knowledge-based techniques may lack adaptivity meeting enormous dynamics and changes, and the global search techniques cannot meet the quick response of CVC. A hierarchical CVC based on control device groups and an on-line supervised search is proposed. With multi-objective optimization, the control knowledge is obtained from the set of Pareto solutions by which the control devices are grouped into a hierarchical structure of basic, supplementary, and inactive control devices. A set of control knowledge are thus prepared by an off-line search, dynamically adjusted by an on-line search, and saved in a long-term memory. Based on the control knowledge an on-line hierarchical optimal search is applied to provide a quick real-time optimal control. Not only does the control knowledge provide active control devices considered for serious situations, but also works to supervise a global search including the inactive control devices for a non-serious situation. With this scheme, the optimal CVC is fast and adaptive to new situations.