Maximum power point tracking control method for photovoltaic system based on APSO-SA

Abstract

The present invention's APSO-SA-based maximum power point tracking control method for a photovoltaic system includes: constructing a photovoltaic system model and initializing parameters and particle swarms for a preset APSO-SA algorithm in MPPT control; collecting voltage and current from the photovoltaic system model in real time, inputting the initialized APSO-SA algorithm into an iterative optimization process to obtain an optimal duty cycle; and adjusting the operating state of a boost converter based on the optimal duty cycle, so that the photovoltaic system operates at the optimal duty cycle and achieves maximum power point tracking. This method enables the MPPT system to quickly relocate the maximum power point in the event of a sudden change in illumination, while maintaining high-precision tracking in steady state, significantly improving the energy capture efficiency of the photovoltaic system.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular to a maximum power point tracking control method for a photovoltaic system based on APSO-SA. Background Art

[0002] As a crucial component of clean and renewable energy, photovoltaic power generation and its efficient utilization have become a research hotspot in the energy field. PV systems convert solar energy into electricity through the photovoltaic effect of semiconductor materials. However, their output power exhibits significant nonlinear characteristics due to dynamic variations in environmental factors such as light intensity, temperature, and cloud cover. To ensure that PV systems consistently operate at their maximum power point (MPP), maximum power point tracking (MPPT) technology is required to maximize the power generated by the PV system.

[0003] Traditional MPPT methods, including perturb and observe (P&O) and incremental conductance (IC), offer good tracking performance in steady-state conditions. However, they are prone to local extrema and power oscillations under complex operating conditions, such as partial shading and rapid irradiance variations. To improve photovoltaic tracking speed, some researchers have proposed a robust improved perturb and observe maximum power point tracking (MPO-MPPT) algorithm based on the P&O algorithm. Experimental results demonstrate the effectiveness of the proposed MPO-MPPT algorithm compared to traditional techniques. An improved incremental conductance (IMP-IC) method has been proposed in the prior art. This method uses indirect control based on an adjustable-step-size PID controller. Results show that the IMP-IC algorithm achieves an average tracking time of 0.12 seconds and a tracking efficiency of 99.88%. While the improved traditional algorithm improves tracking speed and efficiency, it suffers from robustness issues in dynamic conditions. When partial shading causes multiple peaks in the power-voltage curve, the algorithm may converge to a local MPP.

[0004] In recent years, to overcome the limitations of traditional methods and improve photovoltaic system efficiency, MPPT control strategies based on intelligent optimization algorithms have attracted widespread attention due to their global search capabilities. Common optimization algorithms include particle swarm optimization (PSO), artificial neural network control, fractional fuzzy PID control, grey wolf optimizer (GWO), flower pollination algorithm (FPA), cuckoo search algorithm, and sparrow search algorithm. The PSO algorithm has been successfully applied to photovoltaic systems due to its low parameter count and fast convergence. However, its fixed inertia weight and lack of an effective escape mechanism can still lead to premature convergence and delayed dynamic response. A meta-heuristic MPPT technique based on the enhanced autonomous swarm PSO (EAGPSO) algorithm has been proposed for applications in rapidly varying solar radiation conditions. Results show that the EAGPSO technique achieves an MPPT efficiency of 98.6%. To reduce power losses in photovoltaic systems under partial shading, some researchers have proposed a Bayesian fusion technique. This hybrid optimization algorithm combines GWO and FPA to optimize the performance of solar panels in photovoltaic systems. A state-of-the-art method combines a modified FPA (MFPA) with a modified P&O (MP&O) method. Results show that MFPA-MP&O offers advantages over other methods. A key limitation of most MPPT techniques is their inability to track the MPP under conditions of slow or abrupt changes in ambient temperature and solar irradiance. Consequently, numerous research efforts are underway to improve tracking efficiency and the extracted MPP under both uniform and non-uniform atmospheric conditions.

[0005] In summary, traditional MPPT methods are prone to local extrema when faced with complex operating conditions such as partial shadows and rapid irradiance changes, leading to power oscillations and reduced tracking efficiency. Fixed inertia weights and the lack of an effective escape mechanism can lead to premature convergence and delayed dynamic response, making it impossible to maintain efficient tracking in complex environments. Summary of the Invention

[0006] This paper proposes a photovoltaic system maximum power point tracking control method based on APSO-SA to address the aforementioned issues in the existing technology. This method tracks the MPPT of a photovoltaic system by taking into account variations in ambient temperature and solar irradiance, and the SA mechanism helps the model escape local optima. The proposed method's primary contribution is an improvement over the traditional PSO algorithm based on an adaptive inertia weighting approach to reduce oscillations around the MPP, accelerate algorithm convergence, and reduce power losses.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] The maximum power point tracking control method for photovoltaic systems based on APSO-SA includes:

[0009] Construct a photovoltaic system model and initialize the parameters and particle swarm of the preset APSO-SA algorithm in MPPT control;

[0010] The voltage and current of the photovoltaic system model are collected in real time, and the initialized APSO-SA algorithm is input for optimization iteration to obtain the optimal duty cycle;

[0011] The working state of the boost converter is adjusted based on the optimal duty cycle, so that the photovoltaic system operates at the optimal duty cycle and achieves tracking of the maximum power point.

[0012] Optionally, initializing parameters and particle swarms for the preset APSO-SA algorithm in MPPT control includes:

[0013] Set the APSO population size, learning factor, initial inertia weight, as well as the initial temperature and annealing rate of SA. At the same time, configure the voltage / current sampling period and duty cycle adjustment range of MPPT. The initial position of the particle swarm is randomly distributed in the duty cycle search space, and the speed is initialized according to the preset range.

[0014] Optionally, inputting the initialized APSO-SA algorithm for optimization iteration includes:

[0015] Based on the real-time collected voltage and current, the output power of the current photovoltaic system model is calculated as the fitness value;

[0016] Based on the calculated fitness value, dynamically update the individual historical optimal solution of each particle and the group global optimal solution;

[0017] Based on the individual historical optimal solution and the group global optimal solution, the speed and position of the particle are updated;

[0018] The updated particle positions are processed by simulated annealing mechanism to obtain the optimized duty cycle;

[0019] Determine whether the APSO-SA algorithm has reached the preset termination condition. If not, re-collect the voltage and current of the photovoltaic system model, enter the next round of optimization cycle, recalculate the duty cycle, and finally output the optimal duty cycle until the preset termination condition is met.

[0020] Optionally, dynamically updating the individual historical optimal solution of each particle and the group global optimal solution includes:

[0021] Compare the current fitness value of each particle with the fitness value corresponding to its own historical optimal solution. If the current fitness value is better, the current solution will be used as the new individual historical optimal solution and the individual historical optimal position will be updated.

[0022] Among the individual historical optimal solutions of all particles, the solution with the highest fitness value is selected as the new global optimal solution, and the corresponding position is taken as the global optimal position.

[0023] Optionally, updating the speed and position of the particle based on the individual historical optimal solution and the swarm global optimal solution includes:

[0024] Based on the individual historical optimal solution and the group global optimal solution, the particle dispersion dual index and APSO adaptively adjust the inertia weight to update the particle speed and position.

[0025] Optionally, subjecting the updated particle positions to a simulated annealing mechanism includes:

[0026] After each update of the particle swarm global optimal solution, the APSO-SA algorithm accepts the suboptimal solution with a preset probability, where the preset probability is determined by the current temperature of SA and the objective function difference;

[0027] When the particle swarm converges to a local peak, the SA mechanism allows some particles to jump out of the current area in a probabilistic manner and turn to search other potential solution spaces.

[0028] Optionally, the preset termination condition is: if the maximum number of iterations is reached or the power change is less than a set threshold for multiple consecutive times, the search is stopped and the current optimal duty cycle is output.

[0029] Optionally, the photovoltaic system model includes: a photovoltaic cell equivalent circuit model and a boost converter model.

[0030] The beneficial effects of the present invention are:

[0031] The present invention proposes a photovoltaic system maximum power point tracking control method based on APSO-SA, which is used to achieve maximum power point tracking of the PV system. In order to enhance the global optimization ability of the algorithm, an APSO algorithm is proposed based on PSO, thereby improving the search accuracy and speed. At the same time, by introducing the SA mechanism, it is possible to perturb the APSO when it falls into a local optimum, helping it to escape the local optimum and enhance the global search ability. Simulation and experimental results show that under three experimental environments, the relative root mean square error and standard deviation of the algorithm proposed by the present invention are the smallest, and the success rate is the highest. The present invention can effectively track the maximum power point with a shorter convergence time and smaller oscillation, which has significant advantages over other methods.

[0032] The hybrid architecture of APSO-SA proposed in this invention dynamically balances global exploration and local development, enabling the MPPT system to quickly relocate the maximum power point when the illumination changes suddenly, while maintaining high-precision tracking in steady state, significantly improving the energy capture efficiency of the photovoltaic system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 This is a schematic diagram of the photovoltaic system structure according to an embodiment of the present invention;

[0035] Figure 2 Schematic diagram of an equivalent circuit of a photovoltaic cell according to an embodiment of the present invention;

[0036] Figure 3 Schematic diagram of IV and PV characteristics (T=25°C) at different irradiances at constant temperature according to an embodiment of the present invention;

[0037] Figure 4 The IV and PV characteristics of the embodiment of the present invention at different temperatures under equal irradiance (S = 1000W / m 2 ) Schematic diagram;

[0038] Figure 5 Schematic diagram of a DC-DC boost converter solution according to an embodiment of the present invention;

[0039] Figure 6-7 is a basic flow chart of the optimization algorithm of an embodiment of the present invention; wherein, Figure 6 is the PSO algorithm flow chart, Figure 7 It is a schematic diagram of the SA algorithm flow;

[0040] Figure 8 Schematic diagram of the overall framework of the APSO-SA algorithm for implementing MPPT in an embodiment of the present invention;

[0041] Figure 9 Schematic diagram of the output power of photovoltaic modules under different environments according to an embodiment of the present invention; (a) is partial shading, (b) is uniform illumination, and (c) is a dynamic environment;

[0042] Figure 10 Schematic diagram comparing the statistical results of the four algorithms of the embodiments of the present invention under different environments; (a) is RE, (b) is RMSE, (c) is SD, and (d) is SR. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0045] This embodiment proposes a photovoltaic system maximum power point tracking control method based on APSO-SA, which mainly includes the following steps:

[0046] Construct a photovoltaic system model and initialize the parameters and particle swarm of the preset APSO-SA algorithm in MPPT control;

[0047] The voltage and current of the photovoltaic system model are collected in real time, and the initialized APSO-SA algorithm is input for optimization iteration to obtain the optimal duty cycle;

[0048] The working state of the boost converter is adjusted based on the optimal duty cycle, so that the photovoltaic system operates at the optimal duty cycle and achieves tracking of the maximum power point.

[0049] Furthermore, the photovoltaic system model includes: a photovoltaic cell equivalent circuit model and a boost converter model.

[0050] Specifically, in this embodiment, the photovoltaic cell equivalent circuit model includes a photovoltaic array, a boost converter, a maximum power point tracking (MPPT) control and a load circuit. Figure 1 It is worth noting that Figure 1 The MPPT controller in the PV module ensures maximum energy output from the PV system under different climatic conditions and optimizes power transmission from the PV panels to the load.

[0051] The photovoltaic cell equivalent circuit model in this embodiment is as follows Figure 2 As shown, a three-series-one-parallel model is used. ) and ( The model of PVG consisting of ) connected panels can be expressed by formula (1).

[0052]

[0053] In formula (1), is the output current of the solar cell, is the reverse saturation current of the actual battery, Photocurrent, is the output voltage of the solar cell, , are parallel and series resistors, respectively, is the electron charge (1.60222×10 -19 C), is the Boltzmann constant (1.381×10 -23 J / K), is the electron charge PN junction ideality factor, is the actual temperature of the solar cell (Kelvin). The saturation current of the solar cell It can be expressed by formula (2).

[0054]

[0055] In the above formula (2), is the reference cell reverse saturation current, is the reference solar cell temperature (Kelvin) and is the silicon band gap energy ( =1.12eV). where the variable source current depends on the solar cell temperature and irradiance It is expressed by equation (3).

[0056]

[0057] In the above formula (3), is the short-circuit current, is the temperature coefficient of the short-circuit current, and are irradiance and reference irradiance respectively, is the reference temperature of the solar cell. To apply the algorithm proposed in this embodiment, the photovoltaic panel cell model used is Ht-260W-60, and the manufacturing specifications of the photovoltaic panel used are shown in Table 1.

[0058] Table 1 Ht-260W-60 photovoltaic panel parameters

[0059]

[0060] In order to ensure that the photovoltaic cell can operate normally under varying temperatures or irradiances, the IU and PU characteristic curves under no-load conditions are as follows: Figure 3 、 Figure 4 shown.

[0061] The boost converter is very important for the implementation of the proposed algorithm in the MPPT control of photovoltaic systems. It generates a voltage greater than the photovoltaic system voltage. Output voltage , transferring energy from the photovoltaic system to the load. The boost converter tracks the MPP by changing the duty cycle D. Therefore, this embodiment uses a boost converter to model the dynamic characteristics of photovoltaics under different light and temperature conditions. Its equivalent method is as follows Figure 5 shown.

[0062] exist Figure 5 The boost converter includes a power switch Q driven by the pulse width modulation of the MPPT control algorithm, an inductor L, a capacitor C, a resistor inductor r, a fast recovery diode S and a load resistor Based on Kirchhoff’s voltage law and current law, we can get Figure 5 The mathematical model of the boost converter is defined as shown in equation (4).

[0063]

[0064]

[0065] In the above formula (4), when In the case of , the relationship between the input and output voltages of the boost converter is expressed as shown in Equation (5). In the design of the boost converter, in order to ensure that the minimum values ​​of the inductor L and capacitor C can be calculated for continuous conduction, Equations (6) and (7) are used to achieve this.

[0066]

[0067]

[0068]

[0069] In the above three formulas, is the switching frequency. According to the previous equation, the output ripple voltage can be obtained The definition of is shown in formula (8). You can also use the duty cycle D and the capacitance C, load resistance and frequency The relationship between them is obtained.

[0070] The APSO-SA hybrid algorithm proposed in this embodiment is based on the PSO algorithm. The PSO algorithm is an intelligent algorithm proposed by simulating the migration and flocking behavior of bird flocks during foraging. It can find the optimal solution through information exchange between particles. It has the advantages of simple algorithm and fast execution speed. The PSO algorithm is initialized as a group of particles at random positions and gives these particles an initialized speed. These particles update themselves by tracking two extreme values. One is the optimal solution found by the particle itself, called the individual extreme value; the other is the optimal solution found by the entire particle group, called the global extreme value. Particles update the first extreme value in the solution process. Iteration particle speed and location , as shown in formulas (9) and (10).

[0071]

[0072]

[0073] in 、 They are Iteration particle speed and position; 、 are the individual optimal value and the global optimal value respectively; is the inertia weight coefficient; 、 They are Random numbers within; 、 is the learning factor. In the MPPT of photovoltaic power generation, the standard PSO algorithm flow chart is as follows Figure 6 shown.

[0074] The inertia weight in the PSO algorithm regulates the particle's search capability. A large inertia weight is beneficial for global search, while a small inertia weight is beneficial for local search. In the APSO algorithm, the inertia weight automatically changes with the particle's objective function value. When the particle's objective function value approaches the local optimal value, the inertia weight increases; when the particle's objective function value is relatively dispersed, the inertia weight decreases. Therefore, the nonlinear dynamic inertia weight of the APSO algorithm can effectively improve both global and local search capabilities. The inertia weight is shown in Equation (11).

[0075]

[0076] in and are the maximum and minimum values ​​of inertia weight, Represented as the current objective function value of the particle, and are the average target value and the minimum target value of all particles at present. At the same time, the learning factor of APSO algorithm is 、 Using asynchronous changes, the algorithm is in the early stage When , it is beneficial to search for the optimal value of the particle itself, and the algorithm is When , it is beneficial to search for the global optimal value and accelerate the convergence speed. The asynchronous learning factor of APSO is shown in formula (12).

[0077]

[0078] in and They are The maximum and minimum values ​​of and They are The maximum and minimum values ​​of is the current iteration number, The APSO algorithm uses an asynchronously changing learning factor to improve global optimization. As the particle's position and velocity are iteratively updated, the inertia weight and learning factor also change simultaneously, allowing the algorithm to escape local optima and find the MPP more quickly.

[0079] Although the APSO algorithm can improve the global search capability, it may sometimes fall into areas where non-global optimal values ​​are located, or may exhibit phenomena such as "premature maturity". Therefore, this embodiment introduces the SA mechanism on this basis, which allows the algorithm to select updated particles and enable it to jump out of the local optimum, further enhancing the global search capability. In simple terms, the simulated annealing algorithm process is to accept the new solution if it is better than the original solution; if the new solution is not better than the original solution, it is not directly discarded, but accepted with a certain probability. In a sense, the effect of particle diversification or variation is achieved, achieving the purpose of breaking free from the local optimal solution during the algorithm process. The process of the simulated annealing algorithm can be simply described by Figure 7 express.

[0080] In the SA algorithm, the most important concept is the probability of accepting a worse solution , called the transition probability. The acceptance of the new solution at this time is often implemented using the standard Metropolis criterion. The specific value calculation formula is shown in formula (13).

[0081]

[0082] in is the function when searching for a new solution, is the function of the original solution, It is the temperature of the current search. It can be seen that the transition probability in the SA algorithm is the new particle generated by each PSO update. , to determine the original particle and The determined transition probability Is it satisfied , for If the random number between , then accept the new solution ; If not satisfied, the current solution remains unchanged.

[0083] Compared to the classic MPPT algorithm, the conventional PSO algorithm can track the MPP under varying irradiance and temperature conditions and exhibits superior dynamic tracking speed. However, low-frequency oscillations exist in the steady state around the MPP. To improve the performance of the conventional PSO algorithm, this embodiment proposes an improved APSO-SA hybrid algorithm. This algorithm integrates the SA algorithm on the basis of APSO, probabilistically jumping out of the local optimal solution and ultimately approaching the global optimal solution.

[0084] Furthermore, the initialization of parameters and particle swarm of the preset APSO-SA algorithm in MPPT control includes:

[0085] Set the APSO population size, learning factor, initial inertia weight, as well as the initial temperature and annealing rate of SA. At the same time, configure the voltage / current sampling period and duty cycle adjustment range of MPPT. The initial position of the particle swarm is randomly distributed in the duty cycle search space, and the speed is initialized according to the preset range.

[0086] Furthermore, the optimized iteration of the initialized APSO-SA algorithm includes:

[0087] Based on the real-time collected voltage and current, the output power of the current photovoltaic system model is calculated as the fitness value;

[0088] Based on the calculated fitness value, dynamically update the individual historical optimal solution of each particle and the group global optimal solution;

[0089] Based on the individual historical optimal solution and the group global optimal solution, the speed and position of the particle are updated;

[0090] The updated particle positions are processed by simulated annealing mechanism to obtain the optimized duty cycle;

[0091] Determine whether the APSO-SA algorithm has reached the preset termination condition. If not, re-collect the voltage and current of the photovoltaic system model, enter the next round of optimization cycle, recalculate the duty cycle, and finally output the optimal duty cycle until the preset termination condition is met.

[0092] Furthermore, dynamically updating the individual historical optimal solution of each particle and the group global optimal solution includes:

[0093] Compare the current fitness value of each particle with the fitness value corresponding to its own historical optimal solution. If the current fitness value is better, the current solution will be used as the new individual historical optimal solution and the individual historical optimal position will be updated.

[0094] Among the individual historical optimal solutions of all particles, the solution with the highest fitness value is selected as the new global optimal solution, and the corresponding position is taken as the global optimal position.

[0095] Furthermore, based on the individual historical optimal solution and the group global optimal solution, updating the speed and position of the particle includes:

[0096] Based on the individual historical optimal solution and the group global optimal solution, the particle dispersion dual index and APSO adaptively adjust the inertia weight to update the particle speed and position.

[0097] Furthermore, the updated particle positions are processed by a simulated annealing mechanism including:

[0098] After each update of the particle swarm global optimal solution, the APSO-SA algorithm accepts the suboptimal solution with a preset probability, where the preset probability is determined by the current temperature of SA and the objective function difference;

[0099] When the particle swarm converges to a local peak, the SA mechanism allows some particles to jump out of the current area in a probabilistic manner and turn to search other potential solution spaces.

[0100] Specifically, in this embodiment, the process of implementing the APSO-SA algorithm in MPPT control is as follows: First, the improved APSO-SA algorithm completes the initialization of parameters and particle swarms in MPPT control. The system sets the APSO population size, learning factor, initial inertia weight, and SA's initial temperature, annealing rate and other parameters, and configures the MPPT's voltage / current sampling period, duty cycle adjustment range, etc. Then the initial position of the particle swarm is randomly distributed in the duty cycle search space, and the speed is initialized according to the preset range. Then in each iteration, the algorithm collects the voltage and current of the photovoltaic array in real time, calculates the current output power as the fitness value, and dynamically updates the individual historical optimal solution of each particle. and the global optimal solution of the group . By introducing the dual indicators of number of iterations and particle dispersion, APSO adaptively adjusts the inertia weight: maintains a higher weight in the early stage of the search to enhance the global exploration capability, and reduces the weight in the later stage to accelerate local convergence, thereby reducing dependence on manual parameter adjustment. When updating the speed and position of particles based on the individual historical optimal solution and the global optimal solution of the group, the dual indicators of particle dispersion and APSO adaptively adjust the inertia weight are introduced. The introduction of the dual indicators of particle dispersion further enhances the global search capability of the algorithm, helps the particle swarm better explore the solution space, and avoids falling into local optimality. The adaptive adjustment mechanism enables the APSO algorithm to effectively balance global exploration and local development in different search stages, thereby improving the performance of the algorithm.

[0101] Secondly, in order to avoid the problem that traditional PSO is prone to falling into local optimality, the APSO-SA algorithm embeds the simulated annealing (SA) selection mechanism. Afterwards, the algorithm accepts a suboptimal solution with a certain probability, determined by the current temperature of the SA and the objective function difference. When the particle swarm converges to a local peak, the SA mechanism allows some particles to probabilistically jump out of the current region and search other potential solution spaces. As the temperature gradually decreases according to the annealing schedule, the algorithm gradually focuses on the refined search for a more optimal solution. This hybrid strategy retains the adaptive and rapid convergence characteristics of APSO while enhancing the global optimization capability through the randomness of the SA, significantly improving the robustness of MPPT under complex lighting conditions (such as partial shading).

[0102] The algorithm's closed-loop control achieves maximum power point tracking by continuously adjusting the PWM duty cycle. After each iteration, the system detects the termination condition: if the maximum number of iterations is reached or the power change is less than a set threshold for multiple consecutive times, the search stops and the current optimal duty cycle is output. The driver circuit adjusts the switching state of power devices (such as the DC-DC boost converter). Otherwise, the algorithm resamples the PV system data and enters the next optimization cycle. APSO-SA's hybrid architecture dynamically balances global exploration and local exploitation, enabling the MPPT system to quickly relocate the maximum power point in the event of sudden changes in illumination while maintaining high-precision tracking in steady state, significantly improving the energy capture efficiency of the PV system.

[0103] Figure 8This paper demonstrates the implementation of the APSO-SA hybrid algorithm in MPPT control. The APSO algorithm's adaptive adjustment mechanism reduces the algorithm's reliance on manual parameter settings to a certain extent. During the search process, parameters such as the particle swarm's velocity and inertia weight can be automatically adjusted based on the search situation, reducing the difficulty of parameter adjustment. The SA algorithm's probabilistic jump mechanism can, to a certain extent, prevent the algorithm from falling into local optimality and causing search stagnation, allowing the algorithm to converge more quickly towards the global optimal solution. Furthermore, the APSO-SA algorithm incorporates an adaptive adjustment mechanism to continuously optimize the search strategy during the search process, further improving the convergence speed.

[0104] To verify the feasibility of the proposed algorithm, this example compares the proposed algorithm with other algorithms and performs an MPPT performance evaluation. The statistical indicators used for this evaluation include relative error (RE), root mean square error (RMSE), standard deviation (SD), and success rate (SR). The formulas used for the evaluation are shown in Equations (14), (15), (16), and (17), respectively.

[0105]

[0106]

[0107]

[0108]

[0109] in, Indicates the number of data after SIMULINK model runs, is the theoretical MPP value from the photovoltaic power curve, represents the MPP value actually measured from the PV system using the algorithm, and SR refers to the number of times the MPP is met ( ) and the total number of iterations ( ) percentage.

[0110] In order to verify the effectiveness of the APSO-SA algorithm on MPPT, this embodiment uses MATLAB / SIMULINK software to simulate under three environmental conditions, with a sampling period of 1 . The simulation mainly consists of four parts, namely photovoltaic array, boost conversion circuit, MPPT algorithm and load. Among them, the photovoltaic array is composed of three Ht-260W-60 solar panels installed in series. The most important module in the boost conversion circuit is the IGBT device, which can be connected in parallel with the series RC snubber circuit. When the IGBT is turned on, the inductor is charged. When the IGBT is turned off, the inductor discharges and transfers energy to the load, thereby increasing the voltage. The load mainly includes capacitors and resistors. The PSO, APSO, APSP-GWO and APSO-SA used in this embodiment are all implemented in MATLAB / Code in the MPPT algorithm. Table 2 shows the parameters of the APSO-SA hybrid algorithm in MPPT implementation.

[0111] Table 2 APSO-SA hybrid algorithm parameters

[0112]

[0113] In this example, the three experimental environments were all maintained at a constant temperature of 25°C. The first experimental environment was partially shaded for MPPT, while the second and third experimental environments were uniformly illuminated and dynamically changing, respectively. The performance of the proposed APSO-SA optimization algorithm on MPPT was investigated using the varying light intensities in these three experimental environments.

[0114] In order to study the MPPT effect based on the proposed APSO-SA optimization algorithm, the MPPT experiment was first conducted under partial shading conditions. In this environment, the illumination intensity of the three series-connected photovoltaic panels was set to 1000W / m 2 , 800W / m 2 and 600W / m 2 In addition, this work compares the photovoltaic output power of the four proposed optimization algorithms: PSO, APSO, APSP-GWO and APSO-SA. The output power of the photovoltaic module under partial shading is as follows: Figure 9 (a-1). Table 3 shows Figure 9 (a-1) Quantitative comparison of the performance of the four MPPT algorithms, the tracking time, MPP value, and system efficiency of the MPPT algorithms used in this work are given in Table 3. Compared with the other three algorithms, the proposed APSO-SA algorithm can achieve better MPPT control performance, which shows the effectiveness and feasibility of the proposed technique under partial shading conditions.

[0115] Table 3 Comparison of MPPT performance of different algorithms under partial shading.

[0116]

[0117] In order to evaluate the performance of the proposed MPPT method, this embodiment conducts a statistical analysis to compare the results obtained by the four algorithms under partial shading. The four performance indicators RE, RMSE, SD, and SR used for evaluation are shown in Table 4.

[0118] Table 4 Statistical performance indicators of the MPPT algorithm under partial shading conditions.

[0119]

[0120] Observe Table 3 and Figure 9 From the experimental results of (a-1), it can be found that the proposed APSO-SA algorithm can reach MPP in about 0.03s. In addition, it is worth noting that Figure 9 (a-2), Figure 9 Figures (a-3) and (a-4) show that the algorithm can quickly track the MPP under partial shading, outperforming the other three algorithms. The performance comparison in Table 4 shows that under partial shading, the APSO-SA algorithm improves the RMSE accuracy by 2.001W, 0.512W, and 1.183W, respectively, compared to the PSO, APSO, and APSO-GWO algorithms. The RE accuracy improves by 0.75%, 0.44%, and 0.27%, respectively, with the APSO-SA algorithm performing best on all other two metrics. Two reasons contribute to this: First, during the search process, the APSO-SA algorithm dynamically adjusts the particle search behavior based on the current search status and results. Second, the SA algorithm's probabilistic acceptance mechanism allows the algorithm to accept a poor solution with a certain probability, thereby avoiding falling into a local optimum and better adapting to the complex PV curve characteristics of photovoltaic systems under partial shading.

[0121] MPPT experiments were conducted on the four algorithms under uniform illumination. The illumination intensity of the three photovoltaic panels in series was set to 1000W / m 2 , other conditions remain unchanged. The output power of the photovoltaic module under uniform illumination is as follows Figure 9 (b-1) It is worth noting that as the illumination intensity of the three photovoltaic panels increases, the output power increases significantly. Figure 9 (b-3) and (b-4) show that, compared with the other three algorithms, the APSO-SA algorithm has the smallest power fluctuation amplitude after tracking the MPP. Table 5 shows Figure 9 (b-1) Quantitative comparison of the performance of the four MPPT algorithms. The table shows the tracking time, MPP value and system efficiency of the MPPT algorithms used in this work.

[0122] Table 5 MPPT comparison of different algorithms under uniform illumination

[0123]

[0124] In order to evaluate the performance of the proposed MPPT method, this embodiment conducts a statistical analysis to compare the results obtained by the four algorithms under uniform illumination. The four performance indicators RE, RMSE, SD, and SR used for evaluation are shown in Table 6.

[0125] Table 6 Statistical performance indicators of the MPPT algorithm under uniform illumination conditions

[0126]

[0127] Observe Table 5 and Figure 9 (b-1), Figure 9 The experimental results in (b-2) show that the proposed APSO-SA algorithm achieves the fastest MPP under uniform illumination. Compared to the other three algorithms, this algorithm produces satisfactory results, achieving an efficiency of 99.38%. Table 4 shows that under uniform illumination, the APSO-SA algorithm achieves RMSE improvements of 8.509 W, 7.97 W, and 8.111 W compared to the PSO, APSO, and APSO-GWO algorithms, respectively. The APSO-SA algorithm performs best in all three other metrics. Notably, while the RMSEs of the other three algorithms are similar under uniform illumination, the APSO-SA algorithm achieves a significant improvement of approximately 8%. This improvement is primarily due to the increased power of all three PV panels and system efficiency under uniform illumination. Furthermore, the APSO component rapidly explores the general range of the solution space, while the SA component conducts a refined search within a localized area and escapes local optima. This enables the APSO-SA algorithm to more comprehensively and efficiently search the PV curve of the PV system under uniform illumination, thereby finding the MPP.

[0128] In order to ensure the reliability of the algorithm, this embodiment also conducts MPPT experiments on the four algorithms in a dynamically changing environment. First, at t=0s, the illumination intensity of the three series-connected photovoltaic panels in this environment is set to 1000W / m 2 After 0.2s, reduce 200W / m at a fixed temperature 2 , then drop to 200W / m at 0.3s 2 , and finally rise to 300W / m at 0.4s and 0.5s respectively 2 and 100W / m 2 , and the subsequent time remains unchanged. The output power of the photovoltaic module in the dynamic environment is as follows Figure 9 (c-1) It is worth noting that even in a dynamically changing environment, the proposed APSO-SA algorithm can track the MPP at the fastest speed and has better performance. Figure 9 (c-3), 9(c-4) It can be seen that compared with the other three algorithms, the APSO-SA algorithm can always fluctuate around the MPP and maintain the maximum output power. Table 7 shows Figure 9 (c-1) Quantitative comparison of the performance of the four MPPT algorithms. The table gives the tracking time, MPP value and system efficiency of the MPPT algorithms used in this work.

[0129] Table 7 MPPT comparison of different algorithms in dynamic environment

[0130]

[0131] In order to evaluate the performance of the proposed MPPT method, this embodiment conducts a statistical analysis to compare the results obtained by the four algorithms in a dynamic environment. The four performance indicators RE, RMSE, SD, and SR used for evaluation are shown in Table 8.

[0132] Table 8 Statistical performance indicators of MPPT algorithm under dynamic environmental conditions

[0133]

[0134] From Table 7 and Figure 9 (c-1), Figure 9 The experimental results in (c-2) show that the proposed APSO-SA algorithm can track the MPP at the fastest speed, even in a dynamically changing environment. The results also demonstrate that the algorithm can effectively track the maximum power in the shortest possible time, with a tracking time of 0.029 seconds. The performance comparison in Table 8 reveals an important insight: in a dynamic environment, the APSO-SA algorithm achieves the lowest RE and RMSE compared to the PSO, APSO, and APSO-GWO algorithms. It is worth noting that through adaptive parameter adjustment and the probabilistic acceptance strategy of the SA mechanism, the APSO-SA algorithm can improve convergence accuracy while maintaining convergence speed. In the initial search phase, the solution space is rapidly explored; when approaching the optimal solution, the SA mechanism conducts a refined search to avoid falling into local optima, thereby finding the MPP of the photovoltaic system more quickly and accurately.

[0135] In order to more directly compare all the algorithms used in this embodiment, APSO-SA and PSO algorithms are tested with the same number of initial particle swarms, which is set to 50. In addition, the number of iterations of each algorithm is set to 100 times, and the number of runs of each algorithm is set to 90 times. Figure 10 (a) is RE, Figure 10 (b) is RMSE, Figure 10 (c) is SD, Figure 10(d) is SR. From Table 4, Table 6, Table 8 and Figure 10 It is clear that the APSO-SA algorithm, based on the PSO, achieves the lowest RE and RMSE compared to other algorithms, regardless of uniform or non-uniform conditions. It is also noteworthy that the SD value of the APSO-SA algorithm demonstrates its stability over 90 runs compared to the other algorithms. Furthermore, the MPPT algorithm using the proposed APSO-SA achieves an average success rate (SR) improvement of 5.82%, 3.29%, and 3.6% over the PSO, APSO, and APSO-GWO algorithms, respectively. These results are attributed to the APSO-SA algorithm's adaptive mechanism and the probabilistic jump characteristics of simulated annealing, which provide it with a certain degree of resistance to noise and interference. When the search process is disturbed by external factors, the algorithm is able to maintain good optimization performance by adjusting its search strategy.

[0136] Rapidly tracking the MPP of photovoltaic panels is of great significance for photovoltaic systems. To examine the impact of varying environmental conditions, a general MATLAB-Simulink model was constructed to simulate a solar photovoltaic generator. Different environmental conditions were considered in the model. Traditional optimization algorithms face difficulties in tracking the MPP under changing environmental conditions due to insufficient convergence accuracy and local extrema. This embodiment proposes a fused MPPT control method that utilizes both the APSO (adaptive adjustment mechanism) and the APSO-SA (self-adaptive adjustment mechanism) optimization algorithm.

[0137] In simulations, the proposed algorithm was tested using three types of solar photovoltaic modules connected in series under different environmental conditions. Comprehensive simulation experiments demonstrated that the proposed APSO-SA algorithm significantly improved system efficiency compared to the PSO, APSO, and APSO-GWO control algorithms. Simulation results show that the proposed APSO-SA algorithm performs better in effectively tracking the MPP, with shorter tracking times and smaller oscillation amplitudes. Ultimately, the average tracking efficiency of the proposed photovoltaic system increased from 97.24%, 97.97%, and 98.24% for the PSO, APSO, and APSO-GWO algorithms, respectively, to 99.25% for the APSO-SA algorithm.

[0138] Compared with existing research results, this embodiment proposes a dual-modal inertia weight adjustment function (11) for dynamically adjusting the inertia weight in the particle swarm optimization algorithm to achieve a better balance between global and local search capabilities; and constructs a correlation model between the SA perturbation probability and the particle aggregation degree to enhance the effectiveness of local extreme value escape.

[0139] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. The maximum power point tracking control method of photovoltaic system based on APSO-SA is characterized by: include: Build a photovoltaic system model and initialize the parameters and particle swarm of the preset adaptive particle swarm optimization-simulated annealing APSO-SA algorithm in MPPT control; The voltage and current of the photovoltaic system model are collected in real time, and the initialized APSO-SA algorithm is input for optimization iteration to obtain the optimal duty cycle; adjusting the operating state of the boost converter based on the optimal duty cycle, so that the photovoltaic system operates at the optimal duty cycle and achieves maximum power point tracking; The optimization iteration of the APSO-SA algorithm after input initialization includes: Based on the real-time collected voltage and current, the output power of the current photovoltaic system model is calculated as the fitness value; Based on the calculated fitness value, dynamically update the individual historical optimal solution of each particle and the group global optimal solution; Based on the individual historical optimal solution and the group global optimal solution, the speed and position of the particle are updated; The updated particle positions are processed by simulated annealing mechanism to obtain the optimized duty cycle; Determine whether the APSO-SA algorithm has reached the preset termination condition. If not, re-collect the voltage and current of the photovoltaic system model, enter the next round of optimization cycle, recalculate the duty cycle, and finally output the optimal duty cycle until the preset termination condition is met. Dynamically updating the individual historical optimal solution of each particle and the group global optimal solution includes: Compare the current fitness value of each particle with the fitness value corresponding to its own historical optimal solution. If the current fitness value is better, the current solution will be used as the new individual historical optimal solution and the individual historical optimal position will be updated. Among the individual historical optimal solutions of all particles, the solution with the highest fitness value is selected as the new global optimal solution, and the corresponding position is taken as the global optimal position; Based on the individual historical optimal solution and the group global optimal solution, updating the particle speed and position includes: Based on the individual historical optimal solution and the group global optimal solution, the particle dispersion dual index and APSO adaptively adjust the inertia weight to update the particle speed and position; The updated particle positions are processed by a simulated annealing mechanism including: After each update of the particle swarm global optimal solution, the APSO-SA algorithm accepts the suboptimal solution with a preset probability, where the preset probability is determined by the current temperature of SA and the objective function difference; When the particle swarm converges to a local peak, the SA mechanism allows some particles to jump out of the current area in a probabilistic manner and turn to search other potential solution spaces.

2. The photovoltaic system maximum power point tracking control method based on APSO-SA according to claim 1, characterized in that: The initialization of parameters and particle swarm for the preset APSO-SA algorithm in MPPT control includes: Set the APSO population size, learning factor, initial inertia weight, as well as the initial temperature and annealing rate of the simulated annealing algorithm SA. At the same time, configure the voltage / current sampling period and duty cycle adjustment range of the MPPT. The initial position of the particle swarm is randomly distributed in the duty cycle search space, and the speed is initialized according to the preset range.

3. The photovoltaic system maximum power point tracking control method based on APSO-SA according to claim 1, characterized in that: The preset termination condition is: if the maximum number of iterations is reached or the power change is less than the set threshold for multiple consecutive times, the search is stopped and the current optimal duty cycle is output.

4. The photovoltaic system maximum power point tracking control method based on APSO-SA according to claim 1, characterized in that: The photovoltaic system model includes: a photovoltaic cell equivalent circuit model and a boost converter model.

Images