A Photovoltaic Power Generation MPPT Control Method Based on an Improved Particle Swarm Optimization Algorithm

Abstract

This invention relates to a photovoltaic power generation MPPT control method based on an improved particle swarm optimization algorithm, belonging to the field of photovoltaic power generation technology. It solves the problems of low tracking accuracy, slow tracking speed, and large power oscillations in existing photovoltaic power generation MPPT control methods when tracking the maximum power point of a photovoltaic array. This invention identifies illumination conditions and adopts different control strategies under different illumination conditions. It also incorporates a particle elimination mechanism to prevent the transmission of particles that do not provide new information about the photovoltaic array's power. By organically combining an improved particle swarm optimization algorithm and an adaptive step-size perturbation-observation method, it achieves accurate tracking of the maximum power point while ensuring global search capability. This invention is mainly used for tracking the maximum power point of a photovoltaic array.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic power generation technology, and specifically relates to a control method for maximum power point tracking. Background Technology

[0002] In recent years, the increasing demand for electricity has accelerated the consumption of fossil fuels. Due to the adverse environmental impacts and limited resources of fossil fuels, people have paid close attention to solar energy, a clean and renewable energy source. When photovoltaic (PV) power generation systems operate outdoors, the PV array is often in partial shadow due to the movement of clouds, dust, and building shadows. In this situation, the PV characteristic curve of the PV array exhibits multiple peaks. This complex PV characteristic curve causes traditional Maximum Power Point Tracking (MPPT) methods, such as the constant voltage method and the incremental conductance method, to become trapped in local peaks, leading to the failure of the MPPT method.

[0003] To overcome the shortcomings of traditional MPPT methods, an increasing number of intelligent algorithms are being applied to solve the MPPT problem, including particle swarm optimization (PSO), firefly swarm optimization (FSO), and cuckoo search swarm optimization (CSSO). PSO is widely used due to its simplicity and fast convergence. However, traditional PSO MPPT suffers from slow tracking speed, large power oscillations, and susceptibility to local peaks leading to low tracking accuracy. Some current research attempts to improve performance by changing the parameters of the PSO algorithm in real time; however, since the parameters are monotonically changing, excessive perturbation during the initial iterations may cause the algorithm to skip potential maximum power points. Furthermore, many studies fuse two types of MPPT algorithms to eliminate the limitations of individual algorithms; however, improper fusion can lead to excessive algorithm complexity, increasing cost and further complicating implementation. Therefore, the problems of slow tracking speed, low tracking accuracy, and large power oscillations in photovoltaic power generation MPPT control methods based on traditional PSO algorithms urgently need to be addressed. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of low tracking accuracy, slow tracking speed and large power oscillation in existing photovoltaic power generation MPPT control methods when tracking the maximum power point of a photovoltaic array. This invention proposes a photovoltaic power generation MPPT control method based on an improved particle swarm optimization algorithm, which improves tracking efficiency and accuracy while reducing power oscillation.

[0005] A photovoltaic power generation MPPT control method based on an improved particle swarm optimization algorithm, the method includes...

[0006] S1. The duty cycle of the switching signal output by the MPPT controller received by the DC / DC converter in the photovoltaic power generation system is used as the particle position. The particle size is set to N, and the positions of the first to Nth particles increase sequentially.

[0007] S2. Record the voltage and current output by the photovoltaic array when the DC / DC converter receives the position of each particle, and calculate the output power of the photovoltaic array. Use the output power of the photovoltaic array as the fitness of the particle. Sort the N fitness values ​​from largest to smallest, and select the particle position corresponding to the maximum value among the N fitness values ​​as the global best position.

[0008] S3. Based on the voltage and current corresponding to the positions of all particles, determine whether the illumination of the photovoltaic array is uniform. If the result is yes, proceed to step S8; if the result is no, proceed to step S4.

[0009] S4. Constructing a particle swarm:

[0010] From N fitness values ​​sorted from largest to smallest, select the M particles corresponding to the top M fitness values ​​to form a particle swarm.

[0011] S5. The t-th update initializes the particle swarm: the initial value of t is 1. The fitness and position of each particle in the particle swarm are respectively taken as the individual best fitness and individual best position of the particle. The maximum value of the M fitnesss in the particle swarm is taken as the global best fitness, and the position of the particle corresponding to the maximum fitness value is taken as the global best position. The initial values ​​of the velocities of the M particles in the particle swarm are all set to 0, M < N.

[0012] S6, the (t+1)th update includes:

[0013] Update the velocity and position of each particle in the particle swarm, the individual best fitness of each particle, the individual best position of each particle, and the global best position;

[0014] Particles whose positions changed before and after the update were considered valid particles, while particles whose positions did not change were considered invalid particles.

[0015] S7. Based on the maximum and minimum position values ​​among the effective particles, determine whether the update termination condition is met. If the result is yes, proceed to step S8. If the result is no, t = t + 1, and proceed to step S6.

[0016] S8. The global optimal position is used as the initial point of the adaptive step-size perturbation observation method and sent to the DC / DC converter through the MPPT controller. When the power change rate of the sampled photovoltaic array output is less than the preset upper limit of the power change rate, the search step size of the perturbation observation method is dynamically adjusted to achieve dynamic tracking of the maximum power point. During the dynamic tracking process, the search step size is dynamically adjusted from 0.005 to 0.0001, and the output power of the photovoltaic array when the search step size is 0.0001 is taken as the maximum power point. When the power change rate of the sampled photovoltaic array output is greater than or equal to the upper limit of the power change rate, the process returns to step S1.

[0017] Preferably, in step S6, the method for updating the velocity and position of each particle in the particle swarm is as follows:

[0018]

[0019] in,

[0020] This represents the position of the i-th particle after the (t+1)-th update, where i = 1, 2, ..., M. This represents the position of the i-th particle after the t-th update. This represents the velocity of the i-th particle after the (t+1)-th update. This represents the velocity of the i-th particle after the t-th update. c represents the weight factor of the i-th particle after the t-th update. 1,i c represents the cognitive learning factor of the i-th particle after the t-th update. 2,i c represents the social learning factor of the i-th particle after the t-th update. 2,i The value is 1.1. This represents the optimal position of the i-th particle after the t-th update. This is the globally optimal position after the t-th update.

[0021] As a preferred option, c 1,i <c 2,i .

[0022] Preferably, in step S6, the method for updating the individual optimal fitness of each particle in the particle swarm includes:

[0023] The positions of each effective particle after the t+1th update are sent to the DC / DC converter, the voltage and current output by the photovoltaic array are collected, and the fitness of each effective particle is calculated. The fitness of each effective particle after the t+1th update is compared with the best fitness of the corresponding particle after the tth update. If the difference is greater than 0, the fitness of the effective particle is taken as the individual best fitness of the corresponding effective particle after the t+1th update.

[0024] At the same time, the fitness and optimal fitness of each ineffective particle after the (t+1)th update remain unchanged from the fitness and optimal fitness of the corresponding particle after the tth update.

[0025] Preferably, in step S6, the method for updating the optimal individual position of each particle in the particle swarm includes:

[0026]

[0027] in, Let i be the position of the i-th particle after the (t+1)-th update, where i = 1, 2, ..., M. Let be the optimal position of the i-th particle after the (t+1)-th update. Let be the optimal position of the i-th particle after the t-th update. for The corresponding fitness for The corresponding fitness.

[0028] As a preferred embodiment, the implementation of updating the global optimal position in step S6 includes:

[0029]

[0030] in, This represents the globally optimal position after the (t+1)th update. This is the globally optimal position after the t-th update. Let i be the optimal position of the i-th particle after the t-th update, where i = 1, 2, ..., M. for The corresponding fitness for The corresponding fitness.

[0031] Preferably, in step S3, the method for determining whether the illumination of the photovoltaic array is uniform based on the voltage and current corresponding to all particle positions includes:

[0032] Using the voltage and current corresponding to the positions of the first and Nth particles as reference points, if the voltage and current corresponding to each particle position satisfy either of the two conditions, the photovoltaic array is determined to be under uniform illumination; otherwise, the photovoltaic array is determined to be in local shadow.

[0033] Condition 1: The voltage V(j) corresponding to the position of the j-th particle satisfies:

[0034] Condition 2: The current I(j) corresponding to the position of the j-th particle satisfies:

[0035] a is the upper limit of the percentage of voltage difference, b is the upper limit of the percentage of current difference, j = 2, ..., N-1.

[0036] As a preferred value, a = 0.2, b = 0.1.

[0037] Preferably, in step S7, the condition for updating the termination condition is:

[0038] Where, d max d represents the maximum position among the effective particles. min It represents the minimum position among the effective particles.

[0039] Preferably, in step S8, the method for dynamically adjusting the search step size of the perturbation observation method is as follows:

[0040]

[0041] λ k Let λ be the search step size corresponding to the time interval from the k-th sampling to the (k+1)-th sampling. k-1 Let ΔP be the search step size corresponding to the time interval from the (k-1)th sample to the kth sample, Δλ be the search step size adjustment magnitude, and ΔP be the search step size adjustment magnitude. k Let be the rate of change of power output from the photovoltaic array during the k-th sampling.

[0042] The beneficial effects of this invention are:

[0043] 1. It can identify the type of illumination (illumination conditions) and adopt different control strategies under different illumination conditions, which effectively improves the efficiency of maximum power point tracking.

[0044] 2. A selection mechanism is adopted to obtain high-quality initial particles, effectively reducing the number of particle iterations. The iterative formula of the traditional particle swarm optimization algorithm is improved and a particle elimination mechanism is incorporated to avoid sending particles that cannot provide new information about the photovoltaic array power, thereby accelerating the convergence speed of the algorithm and reducing power oscillations.

[0045] 3. In the process of maximum power point tracking, the improved particle swarm optimization algorithm and the adaptive step-size perturbation observation method are organically combined, which ensures the global search capability of the algorithm while achieving accurate power tracking. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of a photovoltaic power generation system; where I PV V is the output current of the photovoltaic array. PV This is the voltage output by the photovoltaic array.

[0047] Figure 2This is a flowchart of the photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm described in this invention;

[0048] Figure 3 This is a PV curve of a photovoltaic array under different illumination conditions; that is, a curve showing the relationship between the power and voltage of the photovoltaic array.

[0049] Figure 4 This is a power tracking diagram of the present invention under illumination condition 1.

[0050] Figure 5 This is a power tracking diagram of the present invention under illumination condition 2.

[0051] Figure 6 This is a power tracking diagram of the present invention under illumination condition 3.

[0052] Figure 7 This is a comparison chart of the power tracking method of this invention and the PSO algorithm under sudden changes in illumination. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0055] This invention is used to evaluate photovoltaic power generation systems based on the proposed algorithm, such as... Figure 1 As shown, it includes a DC / DC converter consisting of an inductor L, capacitors C1 and C2, a switching transistor Q and a diode D, a photovoltaic array connected to the input terminal of the DC / DC converter, and a resistive load R connected to the output terminal.

[0056] Maximum power point tracking (MPPT) technology in photovoltaic (PV) power generation systems adjusts the external equivalent resistance of the PV array using an MPPT controller and a DC / DC circuit under varying light and temperature conditions, thereby changing the output power of the PV array. For the PV power generation MPPT control method based on an improved particle swarm optimization algorithm described in this embodiment, please refer to [link to relevant documentation]. Figure 2 The method includes:

[0057] S1. The duty cycle of the switching signal output by the MPPT controller received by the DC / DC converter in the photovoltaic power generation system is used as the particle position. The particle size is set to N, and the positions of the first to Nth particles increase sequentially.

[0058] S2. Record the voltage and current output by the photovoltaic array when the DC / DC converter receives the position of each particle, and calculate the output power of the photovoltaic array. Use the output power of the photovoltaic array as the fitness of the particle. Sort the N fitness values ​​from largest to smallest, and select the particle position corresponding to the maximum value among the N fitness values ​​as the global best position.

[0059] S3. Based on the voltage and current corresponding to the positions of all particles, determine whether the illumination of the photovoltaic array is uniform. If the result is yes, proceed to step S8; if the result is no, proceed to step S4.

[0060] In step S3, the method for determining whether the illumination of the photovoltaic array is uniform includes:

[0061] Using the voltage and current corresponding to the positions of the first and Nth particles as reference points, if the voltage and current corresponding to each particle position satisfy either of the two conditions, the photovoltaic array is determined to be under uniform illumination; otherwise, the photovoltaic array is determined to be in local shadow.

[0062] Condition 1: The voltage V(j) corresponding to the position of the j-th particle satisfies:

[0063] Condition 2: The current I(j) corresponding to the position of the j-th particle satisfies:

[0064] 'a' represents the upper limit of the voltage difference percentage, 'b' represents the upper limit of the current difference percentage, and 'j' = 2, ..., N-1. As a preferred example, when N is 8, the MPPT controller outputs 8 particle positions. The positions of the first particle P1 and the eighth particle P8 sent by the controller are used as reference points for illumination recognition. The positions of particles P2 to P7 are compared with these reference points, and the following criteria are given:

[0065]

[0066] The value of 'a' can be set to 0.2, and the value of 'b' can be set to 0.1. These values ​​of 'a' and 'b' are based on the maximum power point voltage V of the photovoltaic module. MPP and maximum power point current I MPP The values ​​are approximately 80% and 90% of the open-circuit voltage and short-circuit current of the I-V curve, respectively, based on experimental observations. Substituting the voltage and current values ​​of P2 to P7 into the above criteria, if all six particles satisfy at least one of the two criteria, it proves that the photovoltaic array is under uniform illumination; otherwise, it is in local shadow. If uniform illumination is determined, the adaptive step-size perturbation observation method is directly entered from the globally optimal position, avoiding subsequent iterations of the particle swarm optimization algorithm and significantly reducing power oscillations and the algorithm's convergence time.

[0067] S4. Constructing a particle swarm:

[0068] From N fitness values ​​sorted from largest to smallest, select the M particles corresponding to the top M fitness values ​​to form a particle swarm.

[0069] S5. The t-th update initializes the particle swarm: the initial value of t is 1. The fitness and position of each particle in the particle swarm are respectively taken as the individual best fitness and individual best position of the particle. The maximum value of the M fitnesss in the particle swarm is taken as the global best fitness, and the position of the particle corresponding to the maximum fitness value is taken as the global best position. The initial values ​​of the velocities of the M particles in the particle swarm are all set to 0, M < N.

[0070] S6, the (t+1)th update includes:

[0071] Update the velocity and position of each particle in the particle swarm, the individual best fitness of each particle, the individual best position of each particle, and the global best position;

[0072] This invention proposes a particle elimination mechanism, which requires a judgment of position information after particle iteration. Because particles whose positions remain unchanged cannot provide new information about the power of the photovoltaic array, the transmission of positions of particles whose positions have not changed is paused until the position of the particle changes; therefore, particles whose positions have changed before and after the update are considered valid particles, and particles whose positions have not changed are considered invalid particles.

[0073] The method for updating the velocity and position of each particle in the particle swarm is as follows:

[0074]

[0075] This represents the position of the i-th particle after the (t+1)-th update, where i = 1, 2, ..., M. This represents the position of the i-th particle after the t-th update. This represents the velocity of the i-th particle after the (t+1)-th update. This represents the velocity of the i-th particle after the t-th update. c represents the weight factor of the i-th particle after the t-th update. 1,i c represents the cognitive learning factor of the i-th particle after the t-th update. 2,i This represents the social learning factor of the i-th particle after the t-th update. This represents the optimal position of the i-th particle after the t-th update. This is the globally optimal position after the t-th update. 2,i The value can be 1.1.

[0076] This invention improves the weighting factor ω of the traditional particle swarm optimization algorithm. ω has the ability to keep particles moving in their original direction. This effect is advantageous during the initial search, preventing particles from getting trapped in local peaks. However, in the later stages of the search, particles oscillate around the maximum power point, increasing the algorithm's convergence time. Therefore, for faster convergence, this invention adopts an adaptive ω value. Some existing studies set ω as a constant, which reduces the algorithm's complexity but fails to address the particle oscillation problem in the later stages of optimization. Other studies, while using an adaptive ω value, have a linear convergence and are directly related to the iteration number t, which may cause ω to decrease to a small value too early or too late, thus failing to meet requirements. To address these issues, this invention proposes an improved weighting factor scheme that adaptively changes the ω value based on the distance between the particle's position and the global optimum. The specific calculation formula is as follows:

[0077]

[0078] As can be seen from the above formula, each particle has a different weight factor ω value than other particles, and ω can appropriately converge to a smaller value when approaching the maximum power point, so that the particles move quickly in the early stage of the iteration and avoid the particles oscillating around the maximum power point in the later stage of the iteration.

[0079] This invention relates to the cognitive learning factor c of particles. 1,i Social learning factor c 2,i The value of c has been improved to some extent. 1,i Social learning factor c 2,i These factors respectively affect the speed at which particles move towards their individual optimal position and the speed at which they move towards the global optimal position. The PV curve under local shading may exhibit multiple peaks and multiple minima. When a particle is closer to the maximum power point and near a minima, because the particle is farther from the local peak, the cognitive component of the particle's motion will dominate, causing the particle to move in a direction away from the maximum power point, thus prolonging the optimization time of the particle swarm optimization algorithm. After analysis, c... 1,i The value should be less than c 2,i The value should be adaptively changed as the algorithm runs.

[0080] Unlike some existing studies that linearly optimize cognitive and social learning factors, the present invention... 2,i Adopt a constant value of 1.1, while c 1,i By c 2,i The position of the globally optimal particle, together with the position of the particle itself, determines the effectiveness of the algorithm at different stages. 1,i The values ​​are set as follows:

[0081]

[0082] Among them, c 1,i Let represent the cognitive learning factor of the i-th particle after the t-th update.

[0083] A larger value for c indicates that there are many local minima between the two particles, weakening the particle's global optimization ability. In this case, a smaller c is needed. 1,i This reduces the impact of local factors, thereby enhancing the ability to find global optimization. When the value of c is small, the distance between particles is small, and at this time c 1,i It is approximately equal to 0.5c². To enhance... For c 1,i The weight of the influence will The value is increased by a factor of 5. Finally, the random numbers r1 and r2 are discarded, which reduces the amount of hardware computation while ensuring that particle iteration is predictable.

[0084] The methods for updating the individual optimal fitness of each particle in the particle swarm include:

[0085] The positions of each effective particle after the (t+1)th update are sent to the DC / DC converter. The voltage and current output by the photovoltaic array are collected, and the fitness of each effective particle is calculated. The fitness of each effective particle after the (t+1)th update is compared with the best fitness of the corresponding particle after the (t)th update. If the difference is greater than 0, the fitness of the effective particle is taken as the individual best fitness of the corresponding effective particle after the (t+1)th update. At the same time, the fitness and individual best fitness of each non-effective particle after the (t+1)th update are kept unchanged from the fitness and individual best fitness of the corresponding particle after the (t)th update.

[0086] The methods for updating the optimal individual positions of each particle in the particle swarm include:

[0087]

[0088] in, Let i be the position of the i-th particle after the (t+1)-th update. Let be the optimal position of the i-th particle after the (t+1)-th update. Let be the optimal position of the i-th particle after the t-th update. for The corresponding fitness for The corresponding fitness.

[0089] The implementation methods for updating the global optimal position include:

[0090]

[0091] in, This represents the globally optimal position after the (t+1)th update. This is the globally optimal position after the t-th update. Let be the optimal position of the i-th particle after the t-th update. for The corresponding fitness for The corresponding fitness.

[0092] S7. Based on the maximum and minimum position values ​​among the effective particles, determine whether the update termination condition is met. If the result is yes, proceed to step S8. If the result is no, t = t + 1, and proceed to step S6.

[0093] The conditions for updating the termination conditions are:

[0094] Where, d max d represents the maximum position among the effective particles. min This represents the minimum position among the effective particles. If the update termination condition is met, it means that the positions of all particles are sufficiently close and have converged towards the maximum power point.

[0095] S8. The global optimal position is used as the initial point of the adaptive step-size perturbation observation method and sent to the DC / DC converter through the MPPT controller. When the power change rate of the sampled photovoltaic array output is less than the preset upper limit of the power change rate, the search step size of the perturbation observation method is dynamically adjusted to achieve dynamic tracking of the maximum power point. During the dynamic tracking process, the search step size is dynamically adjusted from 0.005 to 0.0001, and the output power of the photovoltaic array when the search step size is 0.0001 is taken as the maximum power point. When the power change rate of the sampled photovoltaic array output is greater than or equal to the upper limit of the power change rate, the process returns to step S1.

[0096] Furthermore, the method for dynamically adjusting the search step size of the perturbation-observation method is as follows:

[0097]

[0098] λ k Let λ be the search step size corresponding to the time interval from the k-th sampling to the (k+1)-th sampling. k-1 Let ΔP be the search step size corresponding to the time interval from the (k-1)th sample to the kth sample, Δλ be the search step size adjustment magnitude, and ΔP be the search step size adjustment magnitude. k Let Δλ be the rate of change of power output from the photovoltaic array at the k-th sampling. The value of Δλ is 0.0002, and λ... k The value range is [0.0001, 0.005].

[0099] When the output power of the photovoltaic array varies significantly, the algorithm converges quickly to the vicinity of the maximum power point with a large step size. When the output power of the photovoltaic array varies slightly, the algorithm gradually reduces the step size to a level that can ignore steady-state oscillations, while continuously tracking the maximum power point that changes with small variations in illumination or load.

[0100] In practical applications, the upper limit of the preset power change rate is set manually. When the power change rate of the sampled photovoltaic array output is greater than or equal to the upper limit, it is determined that a sudden change in the illumination situation has occurred. If the restart condition is met, the process returns to step S1; otherwise, dynamic tracking of the maximum power point continues. Preferably, the upper limit of the preset power change rate α = 5%.

[0101] Verification experiment:

[0102] use Figure 1 A simulation analysis was conducted on a photovoltaic power generation system. The photovoltaic array consists of four photovoltaic modules connected in series. The maximum power P of the photovoltaic modules is... MPP =200W, maximum power point voltage V MPP =18.2V, maximum power point current I MPP =11A, open circuit voltage V oc =22.5V, short-circuit current I sc =11.6A, ambient temperature is 25℃. Using... Figure 3 The PV curves shown were tested under the following lighting condition 1: the light intensity applied to the four photovoltaic modules was 700 W / m². 2 700W / m 2 700W / m 2 700W / m 2 Lighting condition 2: The light intensity applied to the four photovoltaic modules is 1000 W / m². 2 800W / m 2 600W / m 2 400W / m 2 Lighting condition 3: The light intensity applied to the four photovoltaic modules is 900 W / m². 2 300W / m 2 200W / m 2 100W / m 2 .

[0103] Figures 4-6The diagram shows the power tracking results of the method of this invention under three illumination conditions. Under illumination condition 1, the tracking time of this invention is 0.26 s, and the tracking accuracy is 99.9%. Under illumination condition 2, the tracking time of this invention is 0.48 s, and the tracking accuracy is 99.9%. Under illumination condition 3, the tracking time of this invention is 0.48 s, and the tracking accuracy is 99.9%. It can be seen that under different illumination conditions, the method of this invention can accurately track the maximum power point and has the advantages of fast convergence speed and small power oscillation. Figure 7 This image shows a power tracking comparison between the method of this invention and the existing Particle Swarm Optimization (PSO) algorithm. To verify the working capability of the method of this invention in complex environments, a more extreme lighting scenario was simulated, involving transitions from lighting condition 1 to lighting condition 2 and from lighting condition 2 to lighting condition 3, with each transition occurring after 3 seconds. It can be seen that the method of this invention still exhibits good dynamic performance and tracking effect under abrupt changes in lighting conditions, demonstrating a significant advantage over the traditional PSO algorithm.

[0104] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A photovoltaic power generation MPPT control method based on an improved particle swarm optimization algorithm, characterized in that, The method comprises: S1, taking the duty cycle of the switching signal output by the MPPT controller and received by the DC / DC converter in the photovoltaic power generation system as a particle position, setting the number of particles as N, and sequentially increasing the first to Nth particle positions; S2, recording the voltage and current output by the photovoltaic array when the DC / DC converter receives each particle position, and calculating the power output by the photovoltaic array, taking the power output by the photovoltaic array as the fitness of the particle; sorting the N fitnesses from large to small, and selecting the particle position corresponding to the maximum value in the N fitnesses as the global optimal position; S3, judging whether the illumination of the photovoltaic array is uniform illumination according to the voltage and current corresponding to all particle positions, and executing step S8 if the result is yes, and executing step S4 if the result is no; S4, constructing a particle swarm: selecting the first fitnesses from N fitnesses sorted in descending order particles corresponding to the first fitnesses respectively S5, No. The next update initializes the particle swarm: The initial value is 1. The fitness and position of each particle in the particle swarm are taken as the individual optimal fitness and position of that particle, respectively. The maximum fitness among the swarm swarm is taken as the global best fitness, and the position of the particle corresponding to the maximum fitness is taken as the global best position. Furthermore, the particle swarm swarm saturates... The initial velocity of each particle is set to 0. ; S6、the subsequent update includes: updating the speed and position of each particle in the particle swarm, the individual optimal fitness of each particle, the individual optimal position of each particle, and the global optimal position; taking the particles whose positions change before and after updating as effective particles, and taking the particles whose positions do not change as non-effective particles; S7、According to the maximum and minimum values of the position in the effective particle, it is judged whether the update termination condition is satisfied. If yes, step S8 is executed; if no, step S6 is executed. , step S6 is executed. S8, taking the global optimal position as the initial point of the adaptive step perturb and observe method, sending it to the DC / DC converter through the MPPT controller, when the power change rate of the photovoltaic array output obtained by sampling is less than the preset upper limit of the power change rate, realizing dynamic tracking of the maximum power point by dynamically adjusting the search step of the perturb and observe method, and during the dynamic tracking process, the search step is down-regulated from 0.005 to 0.0001 in a dynamic adjustment mode, and the photovoltaic array output power when the search step is 0.0001 is taken as the maximum power point; when the power change rate of the photovoltaic array output obtained by sampling is greater than or equal to the upper limit of the power change rate, returning to step S1.

2. The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that, In step S6, the implementation manner of updating the speed and position of each particle in the particle swarm comprises: ； wherein , ; denotes the position of the i-th particle after the t-th update, denotes the position of the i-th particle after the t-th update, denotes the velocity of the i-th particle after the t-th update, denotes the velocity of the i-th particle after the t-th update, denotes the weight factor of the i-th particle after the t-th update, denotes the cognitive learning factor of the i-th particle after the t-th update, denotes the social learning factor of the i-th particle after the t-th update, is set to 1.1, denotes the individual best position of the i-th particle after the t-th update, is the global best position after the t-th update.​​​​​ 3.The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm of claim 2, characterized in that, 。 4. The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that, In step S6, the implementation manner of updating the individual optimal fitness of each particle in the particle swarm comprises: make The updated positions of each effective particle are sent to the DC / DC converter, the voltage and current output by the photovoltaic array are collected, and the fitness of each effective particle is calculated. The fitness of each effective particle after the second update is the same as that of the third update. After the update, the best fitness of the corresponding particles is compared. If the difference is greater than 0, the fitness of the effective particle is taken as the best fitness. The individual optimal fitness of the corresponding effective particles after the next update; Meanwhile, the fitness of each non-active particle and the individual best fitness after the first update remain unchanged. second update.

5. The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that, In step S6, the implementation manner of updating the individual optimal position of each particle in the particle swarm comprises: ； in, For the first The position of the i-th particle after the next update. , For the first The optimal position of the i-th particle after the next update. Let i be the optimal position of the i-th particle after the t-th update. for The corresponding fitness for The corresponding fitness level.

6. The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that, In step S6, the implementation manner of updating the global optimal position comprises: ； in, For the first The optimal global position after the next update. For the first The optimal global position after the next update. Let i be the optimal position of the i-th particle after the t-th update. , for The corresponding fitness for The corresponding fitness level. 7.The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm of claim 1, wherein, In step S3, the implementation manner of judging whether the illumination of the photovoltaic array is uniform illumination according to the voltage and current corresponding to all particle positions comprises: taking the voltage and current corresponding to the first and Nth particle positions as reference points, and the voltage and current corresponding to each particle position, when any one of the two conditions is met, it is determined that the photovoltaic array is under uniform illumination, otherwise, it is determined that the photovoltaic array is in local shadow; wherein, Condition 1, the first Voltage corresponding to the position of the particle satisfies: ; Condition 2, the first Current corresponding to the position of the particle satisfies: ; is an upper limit for the voltage difference percentage, is an upper limit for the current difference percentage, .

8. The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm according to claim 7, characterized in that, a=0.2, b=0.

1. 9.The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm of claim 1, wherein, In step S7, the condition for updating the termination condition is: ; wherein is the maximum value of the position in the effective particle, is the minimum value of the position in the effective particle.

10. The photovoltaic power generation MPPT control method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that, In step S8, the implementation manner of dynamically adjusting the search step of the perturb and observe method comprises: ； a search step size corresponding to the time period from the kth sampling to the k+1th sampling, a search step size corresponding to the time period from the k-1th sampling to the kth sampling, a search step size adjustment magnitude, a rate of change of the power output of the photovoltaic array at the kth sampling.

Images