The matrix converter (MC) with an LC filter at source side suffers from system instability, thereby requiring damping control. To improve system stability, this paper proposes a novel active damping control strategy. Firstly, the method of generating space vector modulation signals is improved, in order to make the amplitude and phase angle of input currents directly controllable without affecting the priority of output voltage control. Then the proposed strategy, which is realized by injecting damping signals into input reference currents, is presented. In this way, the proposed strategy can suppress the oscillations in source currents directly. Besides, it is effective with source voltages or capacitor voltages sampled for modulation, regardless of the operation mode of MC. Furthermore, it is applicable to most of existing modulation algorithms. Finally, experimental results of four-quadrant operation on a 2.4kW prototype illustrate that, under the condition of the same parameters, the proposed active damping control strategy performs better than passive damping control in filtering and the same in damping at source side, without sacrificing the driving performance at output side. Index Terms-Matrix Converter (MC), LC Filter, ActiveDamping Control, Input currents control.

No abstract

When matrix converter (MC) works under unbalanced input voltages, sinusoidal input currents could be obtained if the input power factor angle in conventional control strategy is modified based on the positive and negative sequence components of input voltages. Some existing literature realized this goal by using sequence component separation, which was usually complicated and/or storage-consuming. Instead, a simple realization method is proposed by this letter to obtain sinusoidal input currents, which constructs the expected input power factor angle by filtering the square of input voltage vector with a notch filter. The proposed method is easy to implement since no sequence extraction algorithm is required. Experimental results demonstrate that, compared with the existing realization methods, the proposed method reduces the computation time and storage space with identical performance of improving input currents quality.

No abstract

No abstract