EFWA as a Method of Optimizing Model Parameters

Abstract

The Fireworks Algorithm (EFWA) is studied as a method to optimize the noise covariance parameters in an induction motor system model to control the motor speed without a speed sensor. The authors considered a system that employs variable frequency drives (VFDs) and executes an extended Kalman filter (EKF) algorithm to estimate the motor speed based on other measured values. Multiple optimizations were run, and the authors found that the EFWA optimization provided, on average, better solutions than the Genetic Algorithm (GA) for a comparable number of parameter set trials. However, EFWA parameters need to be selected carefully; otherwise, EFWA's early performance advantage over GA can be lost.