An improved robust predictive current regulation algorithm

Current regulation techniques for pulsewidth-modulated (PWM) voltage source inverters (VSIs) can be classified as either linear or nonlinear. Linear techniques consist principally of either a proportional-integral (PI) or a predictive current control strategy, while nonlinear schemes are usually based on a hysteresis strategy. Of the two linear strategies, predictive current control offers the advantages of precise current tracking with minimal distortion and can also be fully implemented on a digital platform. However, the conventional implementation of the predictive current regulation (PCR) algorithm is sensitive to noise and errors in the load inductance estimate, particularly when the back EMF is also estimated. This paper presents an improved PCR algorithm that retains all the benefits associated with PCR while achieving significantly increased robustness to load parameter mismatch and reduced zero current clamped oscillation effects. It is also relatively insensitive to noise in the sampled current measurements. The algorithm is equally applicable to variable fundamental frequency applications such as variable speed drives and to fixed fundamental frequency applications such as PWM rectifier systems or active filters. Simulation and experimental results are presented to confirm the improved robustness of the new algorithm.