Research on Matrix Three-Phase/Single-Phase Power Supply

At present, matrix inverters have attracted much attention because of the use of matrix converters (MC) with adjustable input power factor, continuous output frequency, bidirectional power flow and no DC bus. Although three-phase electrical equipment is widely used in the production field, single-phase power is still required in some industries (such as induction heating and induction melting), and the electricity used in these industries causes serious pollution to the power grid. If matrix converters (MC) are used in these industries, it will have a profound impact on the new generation of "green" power. Here, considering that the low-frequency band and the high-frequency band have different resource occupancy rates for the system due to different control strategies, different control strategies are adopted. The CPU adopts DSP and CPLD joint control to realize a three-phase-single-phase power supply with safe commutation and corresponding protection functions. This power supply is well applied in corresponding occasions and gives full play to the excellent characteristics of the matrix power supply.

l Main circuit structure and commutation strategy

1.1 Main circuit structure

The system circuit adopts one of the simpler topological structures (with neutral line) of the three-phase to single-phase conversion circuit as shown in Figure 1. The state in which both S1+ and S1- are turned on is called the S1 state. In order to filter out as many high-frequency harmonic components as possible in the high-frequency harmonics generated by the switching action in the input current, reduce the high-frequency pollution on the grid side, and improve the input power factor, a filter is introduced. The damping resistor Rd is conducive to the attenuation of the high-frequency current after the turning frequency point, and the incorporation of the capacitor is conducive to reducing the coupling between the switching devices. The circuit uses reverse parallel IGBTs to form a bidirectional switch, and the target voltage is achieved by controlling the time of each switching state.

1.2 Flow switching strategy

The basic characteristics of the main circuit and its application in the induction heating industry determine that the matrix converter must follow two principles during operation: any two phases of the three-phase input of the matrix converter cannot be short-circuited to avoid overcurrent caused by short-circuiting the voltage source. The output of the matrix converter cannot be disconnected to avoid overvoltage caused by sudden disconnection of the inductive load. It can be seen that a reliable commutation strategy must be selected during the commutation process. In order to solve this problem, it is more appropriate to adopt the traditional four-step commutation strategy based on current detection. This method must be added with current detection elements (current transformers, Hall sensors, etc.). In order to ensure the reliable opening and closing of the IGBT, the control voltage is set to: opening voltage +15 V (recorded as 1), and closing voltage -5 V (recorded as O). For the convenience of explanation, the specified current is recorded as I(+) when shown in Figure 1, and vice versa I(-). The four-step commutation switch conversion process is shown in Figure 2. The four processes of commutation from S1 to S2 are now explained. It is assumed that the output current direction is I(+) at this time. The first step is to turn off S1- before turning on S2-, otherwise U1 and U2 will form a loop through S2+ and S1-; the second step is to turn on S2-, if U2>U1, the load current will immediately transfer from S1- to S2-, otherwise the load current will continue to pass through S1+; the third step is to turn off S1+ before turning on S2-, at which time the load current has been transferred to S2+; the fourth step is to turn on S2-.

When the current is reversed, the same method is used, but the opening sequence is different. It can be seen that the four-step commutation method not only prohibits the combination that may cause a short circuit in the power supply, but also ensures that there is at least one path at any time, thereby improving the safety of the circulation. It is worth noting that in order to avoid commutation errors during the commutation process, the information on the current direction needs to be latched.

2 Control strategy

Due to the structure of the system, the space vector modulation method and the dual voltage control method cannot be directly applied to the three-phase-single-phase matrix converter. In order to make the system run more reliably and reasonably, it is now necessary to solve the distribution and control of the on and off of the bidirectional switch to achieve the output requirements. The input fitting method is used in this system, which sets the output voltage as the target, determines the appropriate selection principle, and selects the corresponding input voltage in each sampling cycle based on the principle to fit the target voltage. As for the two control strategies currently used, the maximum phase and the minimum phase of the input three-phase voltage are used to fit the set output voltage. The output voltage is relatively stable, but the control strategy has a large CPU resource overhead in the high-frequency band. The difference between the input voltage and the output voltage is used as the selection basis. Its algorithm is simple and the resource occupancy rate is low in the high-frequency band, but the voltage output fluctuates greatly in the low-frequency band.

In order to achieve better performance of the system, a control strategy combining the two is adopted, the first control strategy is used in the low frequency band, and the second strategy is used in the high frequency band.

Assuming that the input of the converter is a three-phase ideal power supply voltage, then:

For the first strategy, in each sampling period, only the maximum phase Umax and the minimum phase Umin of the input voltage are used to synthesize the target output voltage U0.

Correspondingly, the maximum phase switching functions Smax and Smin are defined. In one sampling period, the on-times T1 and T2 of the two switches are:

Where: U0 is the output voltage reference value; Ts is the sampling cycle time length.

Under the corresponding control algorithm, the fitting schematic diagram is shown in Figure 3. It is essentially similar to a DC chopper circuit, but here it is AC chopping. The voltage obtained by the fitting method is relatively stable. The second control strategy is relatively simple and will not be described in detail here. The conversion between high-frequency and low-frequency control strategies is achieved through software, and the frequency f0 of the output U0 can be set through the human-computer interaction device (if the setting is below 50 Hz as low frequency and above as high frequency), and its subroutine structure block diagram is shown in Figure 4.

3 Digital control system components

There are many signals to be detected and high precision is required, and the corresponding control signals generated require good real-time performance. This determines that the CPU requirements are particularly high. In order to meet this requirement, the system adopts a CPLD+DSP digital control system as the CPU (see Figure 5). In order to give full play to their respective advantages, the analog input channel of the DSP (TMS32LF2407) is used to receive the input/output signals from the signal detection and modulation signal module to calculate and execute the control strategy (input fitting method) in real time, and then send the results of the operation to the CPLD. The CPLD performs logical operations according to the corresponding signals to realize the logic commutation function.

During the operation of the CPU, the CPLD and DSP receive input/output voltage and current signals at the same time, but the functions they implement are different: the signal received by the DSP is for the calculation of the control strategy, while the signal received by the CPLD is to ensure the accuracy of the control signal sent at each moment. When the CPLD detects a fault, it will perform corresponding processing and display the fault location.

4 Experimental analysis of matrix converter (MC) system

In the design of this system, the CPU module uses the SY-XDS510USB 2.0 DSP simulator to control the bidirectional switch tube, thereby realizing some experiments of the MC system. The following are the waveforms of the voltage and current experiments at different frequencies, as shown in Figure 6.

In the low frequency band, the maximum phase and the minimum phase fit the set output voltage control strategy, which is similar to the DC chopping method, so the output waveform is a chopping waveform. Since the capacitor is connected to the load, the voltage across the load is relatively stable. For the high frequency band, the voltage approximation principle is adopted, so there are certain fluctuations in the output voltage and current, but it saves CPU resources and improves the reliability of the system.

5 Conclusion

The system performs three-phase/single-phase power conversion for industries such as induction heating and induction melting, and adopts a frequency-segment control strategy to achieve a reasonable coordination of stability and resources, achieving very good results. Although the control method and cost are relatively high, it is still far superior to the existing conversion methods in terms of power factor and impact on the power grid. With the advancement of integrated modules and control methods, matrix converters will surely be applied in a wider range of fields.