Under-frequency load shedding (UFLS) is used in the power industry to rescue systems facing extreme disturbances to avoid system collapse. Traditionally, many computations are repeated to seek the proper power system settings such that the UFLS provides the desired good performance for selected scenarios. An adaptive UFLS method based on the genetic algorithm was developed to automate the finding of optimal parameters to minimize the repetitive trial-error calculations. Simulations demonstrate that the method has better performance than previous schemes and reduces the time and effort of the repetitive simulations.

Internal faults in three phase induction motors can result in serious performance degradation and eventual system failures if not properly detected and treated in time. Artificial intelligence techniques, the core of soft-computing, have numerous advantages over conventional fault diagnostic approaches; therefore, a soft-computing system was developed to detect and diagnose electric motor faults. The fault diagnostic system for three-phase induction motors samples the fault symptoms and then uses a fuzzy-expert forward inference model to identify the fault. This paper describes how to define the membership functions and fuzzy sets based on the fault symptoms and how to construct the hierarchical fuzzy inference nets with the propagation of probabilities concerning the uncertainty of faults. The designed hierarchical fuzzy inference nets efficiently detect and diagnose the fault type and exact location in a three phase induction motor. The validity and effectiveness of this approach is clearly shown from obtained testing results.