Saber Simulation Analysis of Digitally Controlled Full-Bridge Soft-Switching Power Supply

Digitalization is the development trend of switching power supply. It can realize fast and flexible control design, improve the transient response performance of the circuit, and make it faster, more accurate and more reliable. Therefore, this paper models and simulates a high-power phase-shifted full-bridge ZVS power supply system (12 V / 5 000 A) using digital control based on Saber simulation software, and analyzes the simulation results.

1 Main circuit modeling

The phase-shifted controlled full-bridge ZVS2PWM converter circuit is simple to implement, reliable in operation, and fully utilizes the parasitic parameters of the device. It does not require the addition of auxiliary circuits and is more suitable for high-power, low-voltage, and high-current applications. Its main circuit structure is shown in Figure 1.

Figure 1 Main circuit of phase-shift controlled full-bridge ZVS2PWM power supply system

Saber software provides a power device modeling tool Model Architect. Figure 2 shows the IGBT equivalent circuit model provided by the tool. The modeling can be completed by adjusting the parameter values ​​in Figure 2 according to the parameters of the actual device. The IGBT model used in this system is CM400HA-24E, and its rated parameters are 1 200 V /400 A. Capacitors c1~c4 are external resonant capacitors, where c1 = c3, c2 = c4.

The high-frequency transformer uses a combination of two unit transformers in series and parallel, which can achieve automatic current sharing between the parallel output rectifier diodes, modularize the transformer design, simplify the transformer manufacturing process, and reduce losses. The primary side uses a series inductor lr as the equivalent leakage inductance of the transformer, and uses a current-controlled voltage source (CCVS) module to replace the Hall current sensor with current sampling function.

[page]

Figure 2 IGBT equivalent structure diagram

The secondary output adopts a current-doubler rectifier circuit structure, in which the inductor current and transformer secondary current are small, and the rectifier tube conduction loss and transformer copper loss are small; this structure has dual-inductor interleaved filtering, which can reduce current ripple and improve dynamic response performance under the premise of a small inductance value.

2 Modeling of digital controller

2.1 Peak current control method

The transient values ​​of internal variables such as the conduction current of the power switch device of the switching power supply are relatively independent. Only by directly controlling the transient peak value of the current can the power switch device be effectively and quickly protected, while overcoming the magnetic bias problem of the full-bridge converter and improving its dynamic response speed and reliability. Therefore, this system adopts the peak current control mode. The system structure diagram of the peak current control mode switching power supply is shown in Figure 3, and the system control mathematical model is shown in Figure 4.

Figure 3 Switching power supply system structure diagram

Figure 4 System control mathematical model

2.2 PI Regulator Modeling

PI regulation is the most mature and widely used regulation method in control systems. The expression of discrete PI controller is:

For systems using peak current mode control, when the duty cycle is greater than 0.5, instability will occur. Slope compensation can improve system performance and increase system stability. According to other information, in control engineering practice, the rise rate of the slope compensation voltage is generally designed to be 70%~80% of the converted value of the output inductor current detection signal drop rate.

In formula (1): k is the sampling number; U (k) is the offset of the PI regulator output at the Kth sampling; Kp is the proportional coefficient of the PI regulator; T is the sampling period; Ti is the PI regulator integration time; E (k) is the deviation value of the kth sampling. From formula (1), the discrete PI incremental formula can be deduced as:

In formula (2): U (k-1) is the offset of the PI regulator output at the k-1th sampling; E (k-1) is the deviation value of the k-1th sampling; Ki is the integral parameter of the PI regulator.

The PI regulator model is shown in Figure 5, and its implementation process is:

[page]The AD voltage sampling link is implemented by an analog-to-digital conversion interface "a2z". The sampling value is Z0 (k), and the voltage reference Zref is provided by the given signal module "zdata". The difference between the two is the error term E (k); the amplifier module "zamp" is used to amplify the deviation value E (k) by the integral coefficient Ki times to obtain the integral correction value ΔI (k); the deviation value E (k) is subtracted from the k-1th deviation value E (k-1) maintained by the delay module "zdelay" through the subtraction module "zsub", and the above difference is amplified by the proportional parameter Kp times by the amplifier module to obtain the proportional correction value ΔP (k); finally, the addition module "zadd" adds the integral correction value ΔI (k), the proportional correction value ΔP (k), and the k-1th result U (k-1) maintained by the delay module to obtain the Kth sampling result U (k).

Figure 5 Peak current type control principle diagram

The current loop control adopts P regulation, and its implementation process is as follows: after the Hall current sensor samples, the sampled value is converted into a discrete signal by the analog-to-digital conversion interface, and after a certain multiple amplification, slope compensation is performed. The slope compensation link is implemented by the "z_pulse" module to generate a triangle wave with a certain frequency and slope according to the above compensation law.

The slope-compensated current signal is compared with the result of voltage PI regulation to obtain the final error adjustment value. Finally, the comparison module "zcmp" forms a saturation link to prevent the output phase shift value from exceeding the achievable phase shift range.

2.3 Phase-Shifted Full-Bridge PWM Waveform Modulation

Saber and Simulink can realize collaborative simulation, which can give full play to the advantages of Simulink in software algorithms and generate phase-shifted PWM signals through custom S functions. Saber is used as the host to call Simulink, and the two exchange data at a fixed time step.

Figure 6 shows the principle diagram of phase-shift PWM pulse implementation. The main principle is: when the corresponding front drive waveform jumps to high, the phase shift value U (k) obtained by the digital PI controller subtracts a constant k at a fixed time much smaller than the cycle. When the difference is zero, a pair of pulse waves with the same width as the corresponding front bridge arm drive are generated. The t shown in the figure is the phase shift time.

Figure 6 Phase shift principle.

Figure 7 shows the Saber model for implementing the phase shift process. The "z_pulse" module generates a PWM signal with a fixed frequency and a duty cycle of 50%, which is consistent with the driving timing of the system's super-forearm. The "switchpwm1" module in the figure is equivalent to a multi-way switch. Its working process is: when the super-forearm pulse changes from low to high, the input end is connected, the offset of the feedback is sampled, and then the pulse module immediately changes from high to low to connect the delay module "zdelay" with discrete holding function. Finally, the fixed constant k (generated by the "z_dc" module) is subtracted through the subtraction module "zsub". After the holding time t set by the delay module, the subtraction result is subtracted by the constant k, and the subtraction result is transmitted to the phase shift module "shiftpwm1".

[page]

Fig. 7 Phase-shifted PWM modulation model.

Both "switchpwm1" and "shiftpwm1" modules work together through Saber and Simulink. They implement specific functions by calling S2function. Simply modify sys = mdlOutputs(t, x, u) in the S-function sample file.

3 Simulation Results

The system input DC voltage is 580 V, the operating frequency is 20 kHz, the switch tube parallel capacitors c1~c4 are 47 nF, the leakage inductance lr is set to 10 μH, the proportional parameter Kp = 1, the integral parameter Ki = 0.15, the output filter inductance lo1 = lo2 = 0.5 μH, the filter capacitor co = 82 mF, and the transformer turns ratio n = 10. The load is set to 2.4 mΩ, the output voltage vo = 12 V, and the output current io = 5 000 A.

Figure 8 Switch tube drive waveform

Figure 8 shows the driving waveform of the switch tube. Q1 and Q3 are leading arm switch tubes, which conduct 180° complementary (with a certain dead time), and Q2 and Q4 are lagging arm switch tubes, which have a certain phase shift time for Q1 and Q3 respectively.

FIG9 shows the transformer primary voltage and current waveforms. Analysis shows that the primary voltage and current waveforms of the simulation system are consistent with the working principle of the phase-shift controlled full-bridge ZVS2PWM converter.

Figure 9 Transformer primary voltage and current waveforms

FIG10 shows the conduction and shutdown conditions of the leading arm switch tube q1 and the lagging arm switch tube q2 when the output is 12 V/5 000 A.

Figure 10 The on and off conditions of the switches q1 and q2.

It can be seen from FIG10 that, regardless of the switch tube q1 or q2, before being turned on, the voltage uds across D and S has dropped to zero, indicating that the switch tube has achieved zero voltage turn-on; after the switch tube is turned off, uds begins to rise linearly, indicating that the switch tube has achieved zero voltage turn-off.

Figure 11 Output voltage and current waveforms.

Figure 11 shows the output voltage and current waveforms of the simulation system. It can be seen from the results that the output voltage reaches a steady-state value of 12V at about 112 ms, and the output current reaches a steady-state value of 5000 A. The overshoot of the voltage waveform is less than 0.124 V, and the overshoot of the current waveform is less than 100 A, meeting the performance index of voltage fluctuation of 2%.

4 Conclusion

Through simulation research, we can clearly understand the working process and working characteristics of high-power switching power supply systems, provide an important reference for the development of digital power supplies, and effectively save development costs and shorten the R&D cycle.