Genetic algorithm frequency-domain optimization of an anti-resonant electromechanical controller

Abstract The advancement of power electronics over the last two decades has opened the door for greater performance of electromechanical systems in space applications and manufacturing processes. Often, resonance problems limit controller effectiveness when high bandwidths are required, and can even result in component failure. These resonance problems often appear in the form of vibration problems in machine-tool applications and electromechanical actuators. Previous research has provided a method for reducing the effects of resonance in these electromechanical systems, commonly modeled as double-mass–spring–damper systems. Here, the electromechanical actuators specific to thrust vector control applications are addressed. The control solution involves a traditional PID controller with an additional control filter. Genetic algorithms are utilized to optimize the controller gains using a frequency-domain technique. This research provides an approach for multi-parameter controller design using the genetic algorithm. It incorporates traditional frequency-domain loop-shaping objectives directly in optimization.