Long-Prediction Horizon FCS-MPC for Multiphase Electric Drives With a Selective Control Action Promotion
Juan Carrillo-Ríos, Ignacio González‐Prieto, Ángel González-Prieto, Mario J. Durán, Juan José Aciego
Abstract
Model predictive control (MPC) is an interesting regulation solution in the field of multiphase electric drives. Its performance has been improved from different perspectives, but most works implement a single-step horizon approach. The complexity of using long-prediction horizons along with the higher complexity of multiphase systems has hindered the real-time implementation of multistep strategies. In fact, the only attempt of designing a multistep solution for multiphase drives uses a smart selection of the voltage vector, reducing the computational burden. Unfortunately, this strategy completely disregards the voltage output selected as the better candidate in the second step (i.e., in <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\boldsymbol{k+3}$</tex-math></inline-formula> ), hence missing valuable information that can be inherited from one sampling period to the next one. This work suggests informing the MPC by creating a so-called hall of fame that stores the switching state determined as a suitable candidate in the previous sampling time. Such vector is promoted in the iterative process by adding a term in the cost function that is activated when the switching state under evaluation is within the hall of fame. Experimental results in a six-phase drive confirm the capability of the proposed scheme to simultaneously improve the current tracking and reduce the switching frequency.