Table of contents

1. Introduction 2

2.Team introduction 2

3. Question analysis 2

4. Theoretical analysis and calculation 3

4.1 Demonstration and selection of DC-DC conversion circuit topology 3

4.2 Demonstration and selection of buck-boost circuit topology 3

4.3 Demonstration and selection of MPPT scheme 3

4.4 Select the appropriate switching frequency 4

4.5 Selection of power section components 4

4.6 Methods to improve power efficiency 4

4.7 Calculation of inductor and capacitance of four-switch buck-boost DC-DC converter 4

4.8 Current and voltage measurement and control circuit parameters 5

5. Circuit analysis of each schematic diagram 5

5.1 Overall system structure 5

5.2 Circuit Design 6

5.2.1 Circuit presidential frame design 6

Preface

    This work designs a four-switch buck-boost DC-DC conversion device and a half-bridge bidirectional DC-DC converter. The system controls the output of the DC-DC converter in a closed loop based on the PI control algorithm; the main control chip STM32F411CEU6 controls the gate driver EG3112 and 023N10N5 to form a topological bridge, which realizes automatic switching of Boost, Buck and buck-boost transition modes in different modes. Function: Use INA181 and INA180 to sample the current and voltage, adjust and feedback the voltage and current of each port, and finally realize the regulation of voltage and current at each port. The following are relevant arguments and choices.

2.Team introduction

    The team consists of three members, all of whom are from Wuyi University. They have many award-winning experiences and are all electronic students. Two of the members are brothers from the class of 2019, and one is from the class of 2020. Some of the team members are familiar with the use of STM32, and some are good at PID adjustment. The members of the team are relatively experienced in component selection, because we have all worked on some small projects for several years. So this time the e-sports team joined forces and finally produced the work in three days and four nights. Although the process was quite bumpy, it also gained a lot for us.

3. Question analysis

    The meaning of the question is relatively clear. When the power output by the solar cell cannot make the voltage at the load end equal to 30V, the battery needs to be boosted to provide power compensation to stabilize the output at 30V. On the contrary, when the input power is too large, the excess is used. power to charge the battery. When the input conditions change, the controller will use the real-time values ​​of current and voltage sampling to stabilize the output voltage at 30V forever. The power supply will continuously switch between boost and buck modes, so bidirectional DC-DC is definitely needed. Topology, after determining the basic topology, what remains is the problem of parameter adjustment of the CNC power supply.

To maximize the power tracking of the part, in fact, as long as the guaranteed voltage is equal to half of the input voltage at all times, the duty cycle of the tubes in the topology can be continuously adjusted on the CNC. The specific adjustment method will be mentioned later. . As for efficiency, in fact, as long as synchronous rectification and better-performing tubes are used, it can basically meet the needs of the question. As for how to select, it will also be mentioned below. But if you want higher efficiency, you may need to introduce soft switching mode.

4. Theoretical analysis and calculation

4.1 Demonstration and selection of DC-DC conversion circuit topology

    Three-port converter topology can be divided into two categories: non-isolated type and fully isolated type.

The non-isolated converter topology can be simplified by multiple DC conversion units sharing a DC bus or from multiple two-port topologies. It is easy to expand the ports, has a simple topology, high power density, and no electrical isolation between each port. There are certain safety risks.

The fully isolated converter topology achieves electrical isolation between input and output through an isolation transformer, and the system has a high safety factor. This structure facilitates circuit analysis and achieves approximate decoupling by adding hardware decoupling units ^[1] . However, this topology has shortcomings such as a large number of switching devices, low integration, complex control, and difficulty in improving power density.

Solution summary: Since the voltage given in the question is not high, in order to facilitate control and obtain higher power density, while ensuring sufficient parameter margin of all components, this three-port converter solution uses a non-isolated type.

Demonstration and selection of buck-boost circuit topology

    For DC-DC peripheral topology circuits, the options include Sepic converter and Cuk converter, but the above two types of converters have their own shortcomings: Sepic converter uses too many peripheral components ^[2] ; Cuk converter The converter has many peripheral components, the output voltage is reversed, and the coupling capacitor on the Cuk converter has poor working conditions and is easily damaged.

In order to ensure that the output and input voltages are in the same direction, and at the same time realize the bidirectional flow of energy in a topological structure, two topological structures are selected in this solution: a four-switch buck-boost DC-DC converter and a half-bridge bidirectional DC -DC converter. The voltage variation range of simulated photovoltaic cells is large, so a topology structure that can switch between Boost and Buck at any time is required. The input voltage of photovoltaic cells is higher when no load is applied, depending on the function and the voltage on each switch tube. According to the situation and the convenience of control, choose the four-switch buck-boost DC-DC converter. For the charge and discharge management part of the lithium battery, a half-bridge DC-DC converter is used because the function is relatively simple.

Demonstration and selection of MPPT plan

    Option 1: Conductance increment method. The conductance increment method outputs a control signal by comparing the conductance increment and instantaneous conductance of the solar panel. When the change in output conductance is equal to the negative value of the output conductance, the solar panel operates at the maximum power point. This method can achieve high tracking with a low false positive rate, but it has the disadvantages of high hardware requirements and complex algorithms ^[3] .

    Option 2: Interference observation method. By comparing the output power of the solar panel this time with the last time, it is determined to increase or decrease the operating voltage of the solar panel to implement MPPT. The hardware construction of this solution is relatively simple and the algorithm is easy to implement, but it is prone to oscillation.

Solution selection: The question has given that the internal resistance of the simulated photovoltaic cell is 10Ω, so you only need to ensure that the equivalent resistance of the circuit connected to the output end of the simulated photovoltaic cell is 10Ω to achieve maximum power tracking.

Choose the right switching frequency

    The higher the switching frequency, the resolution of PWM will decrease, and the switching loss caused by MOS will be greater, and switching at high frequency will increase the copper loss and iron loss of the magnetic core. Therefore, in a lower frequency operating circuit, switching losses can be reduced, but too low a switching frequency will cause the output voltage ripple to increase when loading. If you need to reduce the output voltage ripple, you need to use a larger inductance. Large inductance and larger capacitance in parallel, so after considering the above factors and actual testing, this system selected a switching frequency of 40KHZ, that is, the microcontroller outputs a complementary symmetrical square wave of 40KHZ.

Selection of power section components

    In this system, IPP023N10N5 is selected as the power device. The drain-source withstand voltage of this MOS tube is 100V. Rdson measured about 1.5mΩ. It has extremely low drain-source on-resistance, low gate body resistance, and the power required to drive the MOS. Small, the gate charging time constant is small, and the rise and fall times, turn-on and turn-off delay times of the MOS tube are also within the acceptable range of the PWM signal cycle. The driver chip is EG3112 from Yijing Microelectronics. The IO output source and sink current is fully sufficient to drive the NMOS selected for this system at the frequency set by this system. At the same time, the turn-on delay, turn-off delay and rise and fall time do not affect the performance of the PWM output.

Ways to improve power efficiency

    Compared with asynchronous rectification, full-bridge rectification uses low-on-resistance MOS tubes to replace the rectifier diodes, which can effectively avoid the problem of heat generation and waste of efficiency caused by freewheeling diodes during conduction. Replace the diode with a MOS tube of several milliohms to reduce the on-state voltage drop and significantly reduce the conduction loss of the rectifier circuit, thereby achieving high efficiency. ,

Four-switch buck-boost DC-DC converter inductor and capacitance calculation

    The input voltage, output voltage, and output current of this circuit are constant, the switching frequency is, the output current ripple and voltage ripple are, and are the corresponding PWM mode duty cycles ^[4] .

Inductance value calculation

    Therefore, after calculation, the inductance value is 220μH, and the material is iron-silicon-aluminum.

Capacitance value calculation

Current and voltage measurement and control circuit parameters

    四开关升降压输出端、电池组电压、模拟光伏电池电流电压采样均采用电阻分压实现，考虑到电阻的耗散，分压电阻取36K和2K，通过分压实现电压闭环控制，电流采样采用简单的专用电流检测放大器进行采样。

5.各原理图电路分析

5.1系统整体结构

    系统由硬件电路与软件程序组成。通过软件程序输出数字PWM波，实现Buck/Boost电路的升降压。其中在模拟光伏电池，电池组以及负载一侧均设置有电流电压电路，通过对采样点的数据进行分析，应用PI调节数字PWM波的占空比，实现恒流恒压的功能，系统框图组成如图1所示。

图1 系统组成

电路设计

5.2.1电路总统框架设计

    系统由两个DC-DC变换器，电流电压采样电路与STM32小板组成。MOS驱动芯片为EG3112，MOS管为IPP023N10N5。电路组成框图见图2。

图2 系统电路组成

半桥式DC-DC变换器

    在电路中，使用自举电桥驱动器为上下管提供可靠的驱动信号。使用双向DC-DC半桥拓补结构，当输入端的电压大于电池电压时，通过调整占空比使工作模式为同步Buck，对进行降压给电池充电。当模拟光伏电源的供电功率不足时，电池相当于电源，工作模式为同步Boost，为负载提供升压输出，见图3。

 

图3 半桥DC-DC

四开关升降压DC-DC变换器

    采用该结构可以使输入电压在低于或者高于输出电压时都能正常工作，当然，在题目的条件中，只用到了升压部分。为了让该电路可以在其他地方或者条件使用，提高泛用性，使用了升降压拓扑。空载情况下输出电压=输入电压*半桥1的占空比/半桥2的占空比。通过软件控制两侧占空比以实现最大功率跟踪。见图4。

图4 四开关升降压DC-DC

（3）辅助电源

辅助电源由XL7005和7805构成，通过该电路从电源输入端取电，降压至12v为电桥驱动器供电，再通过7805将12V转为5V，给单片机小板供电，3.3V稳压芯片集成在STM32小板上，见图5。

图5 辅助电源

（4）运放及单片机最小系统

运放采用INA181A1和INA180A1，分别负责双向电池电流采样和单向输入端电流采样，因为其有固定的放大倍数，外围元件极少，因此采用。STM32小板及其电路图如下图所示。

图6 运放及单片机最小系统

6.电路PCB设计分析

    在此次电源题目中，由于是一个相对较低压的系统，因此在设计中并没有将功率地和数字地隔开，但是倘若在电磁环境较复杂的电路中，就需要注意两地的隔离。由于电赛时间比较仓促，所以PCB未投入使用，我们的实物时用洞洞板搭出来的。PCB设计图如图7。

图7 PCB设计图

     需要注意的是，在本设计中对于信号部分进行了铺铜处理，这样可以保证导线在传输过程中收到的外界干扰比较少，但是在不能确保地是比较“干净”时，普铺铜可能会起到反作用；另外，由于本电路中的开关频率较低，因此布线没有太讲究，但是一定要注意采样电阻的输出引线是从电阻焊盘最内侧引出，同时以最短路线接到放大器，以减少干扰。

7.实物展示

    实物为洞洞板作品，如图8所示。制作的PCB由于比完赛后有考试，现在刚送打样，希望各位看官不要嫌弃。

图8 作品实物展示

8.程序设计

核心程序如下：

ii=(int)(ADC_Value[3]-3)*0.00402;

ib=(int)(ADC_Value[4]-2047)*0.00408;

ui=ADC_Value[0]*0.01555+ii*0.03;

uo=ADC_Value[1]*0.01574-(ib)*0.03;

ub=ADC_Value[2]*0.01543;

us=ii*10.25+ui+offset+0.2;

p+=ui*ii;

if(ui>=60){fault|=(1<<1);}//过压过流保护和故障错误码，极大减少了炸机的可能性

if(uo>=50){fault|=(1<<2);}

if(ub>=50){fault|=(1<<3);}

if(ii>=10){fault|=(1<<4);}

if(ib>=4.5){fault|=(1<<5);}

 

if (uo>32.0){

buost=CV_Loop(uo,35.0);

}else{

buost=-CV_Loop(ui,us/2);

bost=BAT_Loop(uo,30.0);

}

 

 

if (cnt==5000){cnt=0;p/=5000;//扰动观察MPPT，以较低频率运行微调

if(ib>0.2){HAL_GPIO_WritePin(GPIOC,GPIO_PIN_13,GPIO_PIN_SET);}

if(ib<-0.2){HAL_GPIO_WritePin(GPIOC,GPIO_PIN_13,GPIO_PIN_RESET);}

if (p_last>p){

if (dir_last==0){dir=1;}

if (dir_last==1){dir=0;}

}

if (p_last<p){

if (dir_last==0){dir=0;}

if (dir_last==1){dir=1;}

}

if (dir==0){

offset+=0.01;

p_last=p;

dir_last=dir;

}

if (dir==1){

offset-=0.01;

p_last=p;

dir_last=dir;

}

}cnt++;

if(fault==0x00){

BAT_Boost(bost);

Buck_Boost(buost);

}else{

Stop_pwm();

fault_cnt++;

if (fault_cnt>=50000){fault_cnt=0;fault=0x00;Start_pwm();}//故障恢复等待时间

}

void Buck_Boost(int duty)

{

int CR1,CR2,max;

max=0.98*TIM1->ARR;

if(duty>=0){CR1=TIM1->ARR-duty;CR2=max;}//正为降压，负为升压，0为不升不降

if(duty<0){CR2=TIM1->ARR+duty;CR1=max;}

if(CR1>max){CR1=0.98*TIM1->ARR;}//限制最高占空比

if(CR2>max){CR2=0.98*TIM1->ARR;}

if(CR2<0.2*TIM1->ARR){CR2=0.2*TIM1->ARR;}

TIM1->CCR1=CR1; TIM1->CCR2=CR2;//限制升压过高

}

void BAT_Boost(int duty)

{

int CR3;

CR3=duty;

if(CR3>0.98*TIM1->ARR){CR3=0.98*TIM1->ARR;}

if(CR3<0.3*TIM1->ARR){CR3=0.3*TIM1->ARR;}

TIM1->CCR3=CR3;

}

float CV_Loop(float now,float target)

{//最最简单的增量pi电压环控制

float kp=1,ki=3;

static float error,output,last_error;

error=now-target;

output+=kp*(error-last_error)+ki*error;

 

if(output<-2400){output=-2400;}

if(output>2400){output=2400;}

last_error=error;

return output;

}

float BAT_Loop(float now, float target)

{

float kp=1,ki=4;

static float error,output,last_error;

error=now-target;

output+=kp*(error-last_error)+ki*error;

 

if(output<0){output=0;}

if(output>2400){output=2400;}

last_error=error;

return output;

}

    Due to the rush of time, some variable types are not used well. At the same time, I personally feel that the final voltage adjustment accuracy of the work is not enough because there is no external high-precision ADC.

9. Summary

    During the e-sports, every teammate put in a lot of time and energy. We all slept in the laboratory, and then resumed work after waking up. I am very grateful to my brothers for their efforts during this time. At the same time, we are also very grateful to Lichuang Platform for giving us this opportunity to share our participation experience. At the same time, due to my limited knowledge and limited knowledge, please correct me if I have any omissions or mistakes when writing the report. Let us all learn from each other and make progress together! Thank you everyone~

 

Video link: https://www.bilibili.com/video/BV1AM4y1c7Ar?spm_id_from=333.851.dynamic.content.click