An Experience In Developing An Expert System ForScheduling Problem In Electric Power System

Abstract

The mathematical programming based scheduling algorithms have shortcomings in handling constraints that cannot be expressed in terms of the required mathematical form. Furthermore experienced system operators are often able to tune data so that a schedule can be improved. This paper presents the development of an expert system that has been used as a guide to a mathematical programming based generation scheduling algorithm. The expert system is developed using the knowledge of experience operators and information from the mathematical programming. The expert system takes care constraints that are difficult or impractical to be implemented in a mathematical programming based scheduling algorithm such as cycling of gas turbine and steam turbine units, etc. Expert system is also used to check the feasibility of the solution. The experiences in developing the expert system in the aspects of knowledge acquisition, knowledge base development, inference engine and user interface is discussed in the paper. The developed expert has been integrated into two mathematical programming based algorithms, the dynamic programming and Lagrangian relaxation. Results of the expert system integrated with Lagrangian relaxation algorithm is presented. It was found that the expert system performed its function within acceptable execution time. Transactions on Information and Communications Technologies vol 16, © 1996 WIT Press, www.witpress.com, ISSN 1743-3517