Photovoltaic global maximum power point tracking method based on fuzzy PI control

. At present, algorithms for finding the maximum power point of photovoltaic systems generally have some limitations, especially in cases of uneven light intensity, traditional algorithms often fall into local optima, with limited application space, and their tracking speed and stability are also not very good. To address this issue and improve tracking performance, the combination control of global scanning and fuzzy PI control is proposed. This combination control can avoid falling into local optima and achieve stable output of maximum power. This article first proves theoretically that the proposed method can achieve maximum power output; Then, an experimental platform was built for testing. Build control and sampling circuits and use a solar simulator to irradiate the photovoltaic array unevenly. In this environment, test the control effects of combination control, conductivity increment method, and disturbance observation method respectively. Through testing, it was found that compared to conductivity increment method and disturbance observation method, the proposed method in this article has faster tracking speed, higher output power, and better stability. The overall performance of combination control is better than other algorithms and has a relatively broad application space.


Introduction
With the cost of traditional energy extraction is becoming increasingly high, traditional energy can no longer meet development needs.It is necessary to develop renewable and pollution-free energy.Solar energy is ubiquitous, widely distributed, and pollution-free has always been people's first choice.From the famous discovery of the "photovoltaic effect" by French scientist Becquerel in 1839 to the first single crystal silicon solar cell made by American scientists Charbin and Pearson in 1954, and now to the ubiquitous solar panels, the application of photovoltaic power generation technology has become increasingly widespread.
In order to achieve higher efficiency in photovoltaic solar power generation, various algorithms have been proposed, among which many scholars have mentioned conductivity increment method, disturbance observation method, and fuzzy algorithm [1][2].The disturbance observation method and conductivity increment method can achieve good tracking results when the light intensity is relatively stable or changes slowly [3].However, when encountering extreme or cloudy weather, local shadows may appear, leading to many peaks, and these two algorithms may fall into local optima [4][5][6].In addition, the space for improving the tracking speed and tracking speed of these algorithms is limited, making it difficult to balance speed and accuracy.Fuzzy control can overcome the problem of multiple peaks, but it is susceptible to external factors that can cause output fluctuations.Addressing these drawbacks of the three control algorithms mentioned above, the combination control of global scanning and fuzzy PI control is proposed.This article first analyzes the principle of combination control, and then conducts comparative experiments.The PV array used in our practical engineering is composed of multiple photovoltaic panels in series and parallel.When encountering uniform radiation, its output P-V curve only has one peak.However, in practical application environments, due to factors such as dust, leaves, and buildings, the lighting received by the PV array is often uneven and dynamically changing.In order to find the global maximum power value among multiple peaks, this paper designs a global scanning circuit.

Analysis of global scanning principles
Figure 1 shows the global scanning control circuit diagram.The circuit structure includes three parts: PV array, Zeta chopper, and load.During the scanning process, the sampling circuit collects the voltage and current at both ends of the load of the PV array in realtime, as well as the output voltage and current of the load.The voltage and current data relate to the maximum power point are saved and used to calculate the duty cycle relate to the maximum power point.The specific steps can be divided into scanning preparation, global scanning, and duty cycle calculation.
The purpose of scanning preparation is to set output voltage of the PV array to zero.The specific operation step is to short circuit switch S, and capacitor 1 C begins to charge inductor 1 L .During this process, the voltage of the capacitor gradually approaches zero.When the voltage of capacitor 1 C approaches zero, the scanning preparation is completed, and the next step is global scanning.If the initial voltage at both ends of capacitor 1 C is small, skip the scanning preparation step and proceed directly to the next step.
Global scanning is to gradually increase the voltage of capacitor 1 C from zero to oc U .The specific steps are to disconnect switch S, so that capacitor 1 C is connected to the PV array.
If the power losses of 1 C , 2 C , 1 L , and 2 L in the circuit are not considered, according to the principle of energy conservation, The energy generated by PV array is equal to the energy consumed by the load, and the equation can be expressed as: The impedance of PV array can be expressed in the following form: By combining equations ( 1), (2), and (3), the duty cycle D can be obtained: ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Where ( ) Pi is the output power value, ( )  According to the above rules, establish a fuzzy control rule table, as shown in Table 1.

The combination of global scanning and fuzzy PI controller
The control system includes three parts: fuzzy control module, PI control module, and global scanning module.The mainly task of fuzzy control module is change the PI parameter size based on changes in external environment.The mainly task of global scanning module is find the global maximum power output value among multiple peaks.When the external environment is stable or slowly changing, only the fuzzy PI controller needs to be run.
When the system first starts running or the light intensity undergoes a sudden change, firstly, run the global scanning control module to locate a rough location of optimal solution; Then run the fuzzy PI controller for precise tracking.The advantage of this combination control system is that it can avoid falling into local optima, achieve faster tracking speed, and better stability, have strong adaptability to external natural environment.

Experimental verification
Conduct experimental verification on the combination control of global scanning and fuzzy PI control proposed in this article.The sunlight is provided by the LGH-MTI high-precision artificial solar simulator.This simulator can provide a lighting environment that is very close to natural light and is not affected by external factors such as temperature, climate, and time.The simulator has a radiation area of 1-100 square meters, a radiation intensity of 400-1800/ 2 , and a light instability of ±10%.The photovoltaic array board consists of two polycrystalline solar PV panels with a peak power of 40, a peak voltage of 36V, a peak current of 2.1A, a short circuit current of 3.54A, and an open circuit voltage of 44V.The fuzzy PI control circuit is designed according to Figure 4. Use a solar simulator to irradiate the PV array unevenly, resulting in multiple peaks in power output.The P-V characteristic curve is shown in Figure 5.For this curve, traditional tracking algorithms may fall into local optima, leading to tracking failure.The combined control system proposed in this article can find the global optimal solution through global scanning.The results show that the disturbance observation method takes 70ms to get the local optimal solution.The output power is only 10W, and the fluctuation error after stabilization is 0.5W.The conductivity increment method takes about 100ms to get the local optimal solution.And the output power is only 10W, and the fluctuation error after stabilization is 0.5W.The combination control of global scanning and fuzzy PI control only takes 40ms to complete the global scanning, and then directly enters fuzzy PI controller to achieve relatively stable output.The output power is 23W, and the fluctuation error after stabilization is 0.15W.Compare the corresponding speed and stable fluctuation error of the three control algorithms.The combination control of global scanning and fuzzy PI control can find the global maximum power output value among multiple peaks, overcome the problem of falling into local optima.Compared with the other two algorithms, the tracking speed and stability have been improved.

Conclusion
To address the shortcomings of traditional algorithms that cannot adapt to local shadows and external environmental interference.This article combines the global scanning method with the fuzzy PI control algorithm to propose a control method that can quickly track and resist interference.Through the above experimental comparison, it can be found that compared to traditional control algorithms, combined control can obtain the optimal solution, and the tracking speed is faster, with less fluctuation after stability.
Compared to traditional control algorithms, the combined control has more comprehensive functions and larger application space.The combination control includes fuzzy control module, PI control module, and global scanning module.The three modules work together to enhance strengths and avoid weaknesses.Among them, the global scanning module has strong global search ability, indicating that the system has global control ability.The PI control module can make the system have good stability, greatly reduce system volatility.The fuzzy module improves the maneuverability of the PI control module, making it always in the best control state.It is precisely these advantages that make the system can adapt to more complex environments.
Although combination control has many advantages, there are still some improvements in this control system, such as the need for rich control experience in the fuzzy PI module, which is not convenient for expansion and application.If this problem can be solved, combined control will have greater application space.In addition, this article does not integrate solar position tracking into the combined control system.If integrated, the power output efficiency of the control system will be further improved, and further research will be conducted in the future.

Capacitor 1 C
is charged by the PV array, and the voltage of capacitor 1 C gradually increases from zero to oc U .During the charging process, the mpp U and mpp I relate to the maximum output power point are saved and recorded through the sampling and holding circuit.When the voltage of capacitor 1 C approaches oc U , the global scanning is completed, and the next step is to calculate the duty cycle.As shown in Figure1, the duty cycle related to mpp U is represented as D. At the maximum power point, the equation relationship between

Fig. 2 .
Fig. 2. Combination control system.The principles diagram of the combination control system of global scanning and fuzzy PI control is shown in Figure 2. The combined control system consists of three modules, namely the global scanning module, fuzzy control module, and PI control module.Among them, the input of the global scanning module is the rate of change in light intensity.When the rate of change in light intensity exceeds the set value, the global scanning module is activated.The input of the fuzzy control module is ( ) Ei ,


Pi− is the output power value of the previous moment; ( ) Ii is the output current value, and ( ) 1 Ii − is the output current value of the previous moment.The output of the fuzzy controller is a base value respectively to get a new set of PI controller parameters.The equation can be expressed as: In this way, the parameter values of the PI controller can be flexibly adjusted.Improved control capability of PI controller.The global scanning module and PI control module adjust the output power by changing the duty cycle D. E3S Web of Conferences 466, 01012 (2023) https://doi.org/10.1051/e3sconf/202346601012ICAEER & CEEST 2023 The fuzzy sets of input variables ( ) Ei and ( ) dE i are NB, NS, ZO, PS, PB, while the fuzzy sets of output variables are NB, NM, NS, ZO, PS, PM, PB.The fuzzy domain of ( )

Fig. 3 .---
Fig. 3. P-V characteristic curve.As shown in Figure 3, based on the values of variables ( ) Ei and


from left to right.

Fig. 5 .Fig. 6 .
Fig. 5. P-V characteristic curve of photovoltaic array under local shading.Experimental tests were conducted under these environmental conditions, and the test results are shown in the following figure 6-8:

Fig. 8 .
Fig. 8.The combination control of global scanning and fuzzy PI control.

Table 1 .
Fuzzy control rule table Ei NB NS ZO PS PB NB PB/NB PM/NM PM/NM PS/NS ZO/ZO NS PM/NB PM/NS PS/NS ZO/ZO NS/PS ZO PM/NM PS/NS ZO/ZO NS/PS NM/P M PS PS/NM ZO/ZO NS/PS NS/PS NM/PB PB ZO/ZO NS/PS NM/PM NM/PM NM/PB Note: Each group of data in the table is represented by