A PWM method using vector division for neutral point voltage control in 3-Level inverter

Abstract

In this paper, a new method to control the neutral point voltage of a three-level inverter is proposed. The advantage of a 3-level inverter is lower Total Harmonic Distortion(THD) of output and lower voltage stress of each power semiconductor than a conventional two-level inverter. However, since the line current flows to the neutral point at the specific switching state, there is a problem that the neutral point voltage fluctuates due to charge and discharge of the DC-link capacitors. the neutral point voltage fluctuation can increase the distortion of the output and the voltage stress of the power semiconductor. Therefore, it is also very important to stabilize the neutral point voltage because the increase of the neutral point current causes heat to the DC-link capacitors and affects the lifetime. Although there are various methods for stabilizing the neutral point voltage, the method of the zero-sequence voltage injection is the most widely used. his is because it is simple to implement and the ability to stabilize the neutral point voltage at a relatively high power factor is excellent.[3] However, the modulation index and the power factor limits the range of a zero-sequence voltage that can be injected, accordingly, the neutral point voltage control is also limited. It is possible to show the area where the neutral point voltage control is limited on the space vector diagram according to the modulation index and the power factor, and a change in the Pulse Width Modulation(PWM) method is required to stabilize the neutral point voltage in this area. In this paper, a new PWM method was proposed for dividing a reference vector into two vectors, although It is the same to control the neutral point voltage using the zero-sequence voltage injection method. In order to output two reference vectors, the sampling frequency was reduced to half, and one of the two vectors was set so as to minimize the neutral point voltage fluctuation. Therefore, it is possible to minimize the fluctuation of the neutral point voltage.