Multivector model predictive control for multiphase induction machines with dead-time knowledge

As requirements in the field of energy efficiency and transport have become more demanding, multiphase technology has begun to emerge in commercial applications in order to meet these demands. In terms of multiphase electric drive regulation, model predictive control has gained traction thanks to its fast dynamic response, flexibility and easy control design. Furthermore, the development of implicit modulators such as multivector solutions has provided these schemes with competitive current quality results. However, the use of higher sampling frequencies in the pursuit of an improved performance presents a blind spot for these control solutions. More concretely, the dead time of voltage source converters (VSCs) has an increasing impact on multiphase electric drives as the sampling frequency becomes higher. The present work first analyses how this phenomenon affects the uniquely available subspaces which characterize multiphase systems. Following these results, dead-time information is added in a multivector predictive scheme and the cost function is reformulated to mitigate its impact. Different operating metrics have been employed to illustrate the effectiveness of the proposed algorithm to increase current quality in a six-phase drive when dead-time effect appears.