Research on Control Strategy of Three-phase Active Filter

pushi

Research on Control Strategy of Three-phase Active Filter [Copy link]

With the continuous development of power electronics technology, more and more power electronic devices are widely used in various fields. However, the inherent nonlinearity of power electronic devices makes its impact on the mains, such as harmonic pollution and input power factor problems, increasingly prominent. In the past, we used passive filtering networks to solve harmonic problems. However, the filtering characteristics of passive filtering networks are non-selective, and they also have some disadvantages such as large size, easy to cause resonance, and poor compensation characteristics. In recent years, many static VAR compensator topologies have been proposed to solve the power factor correction problem. However, some static VAR compensators themselves will generate low-order harmonics. Moreover, their dynamic response cannot meet the requirements of some special loads.

　　To this end, a series of active power filter (APF) schemes have been proposed to solve harmonic and reactive power compensation and improve power factor. For example, Akagi [16] proposed an active filter using a multi-voltage source converter and a delayed PWM scheme. In his subsequent paper [17], he described in detail the control circuit based on the instantaneous reactive power theory and discussed its transient response characteristics. Enslin [19] determined the compensation current by detecting and calculating the reactive power of the load and injected it into the power grid. Enslin and Hayafune [20] used a microprocessor to calculate the compensation current and generate a switching signal. Jou H.L. [24] determined the compensation current by calculating the fundamental active part of the load current and injected it into the power grid. All of the above active filter control methods require detecting the waveform of the load current and then performing corresponding processing to generate a reference for the compensation current. Wu J.C. and Jou H.L. [25] proposed a simplified single-phase active filter control method in 1996. This control method detects the waveform of the incoming power current and only generates a power current reference, not a compensation current reference. Therefore, the incoming power current is the directly controlled quantity.

　　The practical three-phase active filter control strategy described in this paper is an expansion and improvement of the "simplified control method" and is applied to the three-phase active filter. It is not only convenient and reliable to implement, but also the three-phase active filter after the dynamic current bandwidth control can not only realize the compensation of reactive power and the suppression of harmonic current, but also balance the three-phase current on the transmission line. Therefore, it can improve the power quality and the transmission efficiency of the transmission line.

pushi

The main circuit structure and working principle of the three-phase active filter Figure 1 The main circuit of the three-phase parallel active filter The main circuit of the active filter adopts a three-phase four-wire voltage-type converter, as shown in Figure 1. The DC capacitors C1 and C2 on the DC bus of the voltage-type converter are used as energy storage elements, and the capacitance of C1 and C2 is equal. Switches S1 and S4 form a half-bridge converter of phase A. When the bridge arm is turned on and off, the voltage on the inductor L1 makes ifa rise or fall linearly, thereby controlling the size, phase and waveform of ifa, so as to control the input current ia of phase A. This converter can work in the rectification state or in the inverter state. Similarly, the B-phase half-bridge converter composed of switches S3 and S6 and the C-phase half-bridge converter composed of switches S5 and S2 can control the input current of phase B and phase C respectively, and they can also work in the rectification and inverter states. Inductor L0 and capacitor C0 form a low-pass filter, which is used to eliminate the switching ripple generated by the converter. The basic idea of the control strategy of the practical three-phase active filter discussed in this paper is: set three phase current references for the three-phase input phase current, which are in phase with the phase voltage, contain only the fundamental wave, and have equal amplitudes to each other, and make the phase currents ia, ib, ic follow the references within a difference band by controlling the converters of each phase. Because the reference is a sine wave with equal amplitude and in phase with the phase voltage, when the switching band is filtered out by an additional passive filter, the phase currents iA, iB, iC are symmetrical currents with high power factor, low distortion. If the loss of the active filter is ignored, the converter neither absorbs nor releases energy. In steady state, the energy consumed by the load is equal to the energy provided by the power supply, that is, the input and output power remain balanced. The capacitor on the DC bus is an energy storage element, so the converter can exchange energy with the power supply and the load, and the reactive power required by the load is obtained by exchanging with the active filter. Reactive power is only exchanged between the active filter and the load. The asymmetric power required by the load is provided to the load by the power supply after the energy exchange between the converter phases. If the input and output power are out of balance, the unbalanced energy will be supplied or absorbed by the DC capacitor, causing a change in the capacitor voltage. If the active power supplied by the power supply is less than the load requirement, the DC capacitor voltage drops. After the control circuit detects the capacitor voltage drop, it will simultaneously increase the amplitude of the current reference of each phase to increase the active power supplied by the power supply. On the contrary, when the DC capacitor voltage rises, the amplitude of the current reference of each phase will be reduced at the same time, and finally the supply and demand of energy will be balanced. The average value of the capacitor voltage on the DC bus can provide information on whether the power is balanced. We use feedback control to determine the amplitude of the reference current to achieve a balance between input and output power.

pushi

Control block diagram Figure 2 shows the control block diagram of the active filter. In the control of the active filter, the quantities to be detected are: phase current ik, phase voltage uk, and voltages Uc1 and Uc2 on the upper and lower DC capacitors. The three-phase incoming line voltage is detected to generate a unit sine wave Sk in phase with the phase voltage. Uref is the expected value of the set capacitor voltage. The level R obtained by proportional integration of the error e between it and (Uc2 + Uc1) is used to adjust the amplitude of the reference to balance the energy between the power supply and the load. Feedback to the current balance setting circuit generates a fine-tuning level Tk to fine-tune the reference amplitude so that the three-phase current reference amplitudes are completely equal. This makes the three-phase power on the incoming line symmetrical when the three-phase voltage is symmetrical. The difference between Uc2 and Uc1 is proportionally integrated to obtain an error ε, which is added to the current reference so that the phase current has an adjustable DC component to ensure that the average values of the voltages on C1 and C2 are equal during operation. (This will be discussed in detail later).

pushi

The reference current floats up and down by δ, forming a bandwidth of 2δ. We use the hysteresis control pulse modulation method to control the opening and closing of the upper and lower arms of the half-bridge converter, so that the phase current ik rises and falls within the difference band, thereby realizing the three major functions of the active filter. Figure 2 Control block diagram Assuming that the system voltage is constant and the load is constant, the load current i1k (k=a, b, c) is stable. i1k can be an asymmetric, nonlinear current, containing reactive power and harmonics. The phase current ik is the sum of the load current i1k and the compensation current ifk. Therefore, if the compensation current ifk can be controlled, the phase current ik can be controlled. The typical working mode of the hysteresis control pulse modulator is shown in Figure 3. Switching devices S1 and S4 form a half-bridge converter of phase A. By adjusting the reference amplitude of the phase current, the average value of Uc1 and Uc2 is constant. In order for the active filter to work properly, Uc1 and Uc2 must be greater than the peak value of the input voltage. Open S4 and close S1. If ifa>0, ifa flows through S4 to discharge C2, Uc2 decreases, ifa increases, and ia increases; if ifa<0, ifa flows through the lower bridge arm reverse diode to forward charge C2, Uc2 increases, ifa increases, and ia increases. Therefore, ifa and ia increase when the lower bridge arm is turned on (dia/dt>0). When ia increases to slightly greater than +δ, turn off S4 and open S1. If ifa>0, ifa flows through the upper bridge arm reverse diode to forward charge C1, Uc1 increases, ifa decreases, and ia decreases; if ifa<0, ifa flows through S1 to discharge C1, UC1 decreases, ifa decreases, and ia decreases. Therefore, when the upper bridge arm is turned on, ifa and ia decrease (dia/dt<0). When ia drops to slightly less than -δ, S4 is turned on again, S1 is turned off, and so on. Through the opening and closing of the upper and lower bridge arms of this half-bridge converter, the phase current ia can follow the phase current reference change within a difference band 2δ, as shown in Figure 3. The working principle of the hysteresis control pulse modulator is explained above using phase A as an example. The main circuit of the three-phase active filter is a three-phase half-bridge type. The hysteresis control pulse modulation of each phase is relatively independent. Because the reference current (k=a, b, c) is a three-phase symmetrical sine fundamental wave, and the main current ik (k=a, b, c) tracks the reference current under dynamic current bandwidth control, after passive filtering, the switching frequency is easily filtered out, leaving only the required three-phase symmetrical fundamental wave, so the current on the transmission line at the input end is three-phase balanced, no zero-sequence current flows on the neutral line of the power supply, and the zero-sequence current required by the nonlinear load is all provided by the active filter. In the process of the upper and lower bridge arms of the converter being turned on in sequence, whether ifk flows into or out of the converter, it will cause fluctuations in UC1 and UC2. Table 1 summarizes the voltage changes on capacitors C1 and C2 caused by various conditions of current ifk. Table 1 Capacitor voltage UC1 and UC2 changes ifk>0 and difk/dt<0 UC1 increases UC1 increases ifk<0 and difk/dt<0 UC1 decreases UC1 decreases ifk<0 and difk/dt>0 UC2 increases ifk>0 and difk/dt>0 UC2 decreases Figure 3 Typical operating mode of hysteresis control pulse modulator