Model based optimization of commutation for a linear hybrid stepper motor

In this paper, a model based method is presented to create an optimized commutation rule with the purpose to reduce force ripples and improve synchronism. The method comprises a model with force functions of each phase that are identified by experiments with several weights. The force functions include nonlinearities and allow a separation of the phases characteristics. Based on the identified force functions, a proposed optimization process is conducted. As a result, load dependent commutation rules are obtained in terms of a look-up table for each phase individually. To compare the results properly with the common sinusoidal commutation, an adaption function is introduced in order to transfer the control signal from current to force values. Implemented on real-time hardware, velocity controlled test runs are carried out enabling a validation of the optimized commutation rules. Besides proving force ripple reduction using an accelerometer as indicator for horizontal vibrations, restrictions of applicability are outlined. In this work, a three-phase linear hybrid stepper motor (LHSM) is considered for experimental research but the method is addressed to all sinusoidal type linear drives that do not show ideal behavior.