Speed control of PMSM drives by generalized predictive algorithms

Abstract

The paper deals with generalized predictive algorithms applied to the industrial drives employing Permanent Magnet Synchronous Motors (PMSM). The Generalized Predictive Control (GPC) belongs to the multistep model-based control design. The presented GPC algorithms are arranged in the form suitable for direct use in a real PMSM drive application. Here, the usual form of GPC action calculating formula is specifically decomposed in a summation of products of tabulated GPC gains with real values of PMSM topical outputs, state, future required reference and possibly with values of previous control actions. The formulated GPC algorithms are experimentally compared with a standard cascade vector PI control. A speed control task is considered for the comparison. The experiments are documented by the time histories in oscillogram screenshots and appropriate figures of applied GPC gains.