A fuzzy logic framework for control of switched capacitors in distribution systems

This paper proposes a fuzzy expert system for the multilevel control of switched capacitors installed on a distribution system with a nonconforming load profile. The control objectives are minimization of power system losses without violating the voltage security of the power system. Expert systems enhanced by fuzzy sets are used to determine the control variables corresponding to the given load values. The rules are adapted using a neural learner to build the rule set and train the membership functions. A load flow determines the corresponding state of the power system. The knowledge base chooses the design from a set of suboptimal solutions obtained from the load flow. The method is based on the application of fuzzy sets to sensitivities in expert systems to refine the solution. Initial trial runs using the above approach on a 30-bus distribution system are very encouraging. Simplicity, processing speed and ability to model load uncertainities make this approach a viable option for online VAr control.