Feedforward neural fuzzy control of electrical power systems containing highly varying loads

Control areas that contain a significant amount of industrial load are often subject to highly varying demand profiles on their systems. Arc furnaces, rolling mills and other large motors can create large demands on the system which result in an unsatisfactory area control error (ACE). Studies have shown that very-short term load prediction can be incorporated into control schemes which are then able to compensate for the highly varying demand. Working with a sponsoring utility, the authors have developed a method of controlling these systems. Using a neural network prediction of the area load, ACE and its integral, ∫ ACE, a new fuzzy logic controller adjusts the set point of the area generation to attempt to match the upcoming changes on the system. Performance of the neural-fuzzy controller in a two-area tie-line model with actual load data from a collaborating utility is demonstrated and compared with the present AGC system through simulations. Results show that the proposed neural-fuzzy controller matches the demands of highly varying loads and significantly improves area control error on the system.