Skip to Main Content
In this paper, a generic approach to predictive control of inverter-fed loads, called ldquodirect controlrdquo, is presented. The controlled variable can be either current or flux, depending on the type of the load (e.g., passive or active). Nevertheless, in both cases the same basic principle is applied. The prediction algorithm dictates the inverterpsilas voltage space vector to be applied in order to achieve the smallest possible error vector at the end of the sampling interval. Two variants of modulation principle are proposed, depending on the priorities to be followed: low ripple or low switching frequency. The validity of the approach has been demonstrated on laboratory models with three different types of loads: active power filter, synchronous and induction motor. The presented results show good performance, regardless on the operating point.