Photovoltaic array dynamic reconstruction adaptive control method and system

Abstract

The invention discloses a photovoltaic array dynamic reconstruction adaptive control method and system, and the method comprises the steps: collecting the related data of a photovoltaic array, carrying out the preliminary analysis of the related data at an edge end, and recognizing the state of a photovoltaic panel; mPPT control is carried out on each photovoltaic panel through a perturbation and observation method; and the photovoltaic array end carries out shadow detection through clustering analysis, evaluates the state of the photovoltaic array and carries out dynamic reconstruction. According to the method, the advantages of edge calculation and centralized control are combined, so that rapid and accurate shadow detection and efficient dynamic reconstruction control are realized. The method belongs to the technical field of photovoltaic power generation system control, and has the advantages of improving the overall performance of the system, enhancing the environmental adaptability, improving the expandability and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular to a method and system for dynamically reconfiguring and adaptively controlling a photovoltaic array. Background Art

[0002] As a clean and renewable form of energy, photovoltaic power generation plays an increasingly important role in the transformation of the global energy structure. With the continuous advancement of photovoltaic technology, the efficiency and reliability of photovoltaic arrays have been significantly improved. However, in practical applications, photovoltaic arrays still face many challenges, the most prominent of which is the performance degradation under partial shading and uneven lighting conditions. Traditional fixed topology photovoltaic arrays often perform poorly in these conditions, resulting in a significant reduction in the overall output power of the system. In order to address this problem, researchers have proposed dynamic reconstruction technology to optimize system performance by adjusting the connection method of photovoltaic arrays in real time. Early dynamic reconstruction methods were mainly based on preset rules or simple heuristic algorithms. Although they can improve system performance in some cases, they still have shortcomings such as slow response speed and poor adaptability.

[0003] With the development of artificial intelligence and big data technology, the dynamic reconstruction method of photovoltaic arrays based on machine learning has gradually become a research hotspot. These methods analyze a large amount of historical data and establish complex mathematical models to predict the optimal reconstruction strategy. However, the existing intelligent reconstruction methods still have some limitations: first, most methods require a lot of computing resources and complex central control systems, which are difficult to be widely used in actual engineering; second, these methods often ignore the real-time status of each component in the photovoltaic array and cannot respond quickly to local environmental changes; third, the existing methods are not scalable enough when dealing with large-scale photovoltaic arrays, and it is difficult to meet the needs of photovoltaic power stations of different sizes. In addition, the existing technology still has room for improvement in the accuracy of shadow detection and the optimization of reconstruction strategies, especially under complex and changing environmental conditions, it is difficult to achieve true adaptive control. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the existing photovoltaic array dynamic reconstruction technology mainly has the problems of slow response speed, poor adaptability, insufficient scalability and poor performance in complex environments.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for adaptive control of dynamic reconstruction of a photovoltaic array, comprising: collecting relevant data of the photovoltaic array, preliminarily analyzing the relevant data at the edge end, and identifying the state of the photovoltaic panel; performing MPPT control on each photovoltaic panel through the perturbation observation method; performing shadow detection through cluster analysis at the photovoltaic array end, evaluating the state of the photovoltaic array and performing dynamic reconstruction.

[0007] As a preferred solution of the photovoltaic array dynamic reconstruction adaptive control method described in the present invention, the relevant data includes illumination data, temperature data, electrical data, and system topology and reconstruction state data.

[0008] As a preferred solution of the photovoltaic array dynamic reconstruction adaptive control method of the present invention, the edge end corresponds to a single photovoltaic panel, is responsible for data collection and calculation of performance ratio to judge the photovoltaic panel state, which is expressed as:

[0009]

[0010] Among them, PR represents performance ratio; P represents output power; G represents light intensity; P 0 Indicates the nominal power of the photovoltaic panel; calculates the difference between the current PR and the average PR of the same period in the past 7 days, C Z , and the difference rate C between the current panel PR and the average PR of the adjacent panels X , when C Z <-20% and C X > -20%, it is considered as shadow occlusion; when C Z <-20% and C X When the value is less than -20%, it is judged as a photovoltaic panel failure, the photovoltaic panel is immediately isolated, and a fault alarm is sent to the photovoltaic array end; when C Z >-20% and C X When the power consumption is less than -20%, it is judged that the photovoltaic panel has long-term low power, and no isolation measures are taken, and a fault alarm is sent to the photovoltaic array end.

[0011] As a preferred solution of the photovoltaic array dynamic reconstruction adaptive control method described in the present invention, the MPPT control includes: the edge end performs MPPT control through the perturbation observation method, adjusts the working point of the photovoltaic panel in real time, calculates the power P(k) according to the collected voltage V(k) and current I(k), performs a small amplitude perturbation on the voltage, collects the voltage V(k+1) and current I(k+1) after the disturbance to calculate the power P(k+1) after the disturbance, compares the power and the power after the disturbance, and determines the next disturbance direction, which is expressed as:

[0012] ΔP=P(k+1)-P(k)

[0013] ΔV=V(k+1)-V(k)

[0014] Wherein, AP represents the power disturbance change; ΔV represents the voltage disturbance change; when AP / ΔV>0, V(k+2)=V(k+1)+ΔV; when AP / ΔV<0, V(k+2)=V(k+1)-ΔV; the step size is adjusted by the adaptive mechanism, the step size increases when it is far away from the maximum power point, and decreases when it is close to the maximum power point, which is expressed as:

[0015] ΔV(k+1)=ΔV(k)×(1+α×|ΔP / ΔV|)

[0016] Among them, α represents the step adjustment parameter; the voltage, current, power, temperature, MPPT efficiency and accumulated energy output data of the photovoltaic panel are uploaded to the photovoltaic array end regularly.

[0017] As a preferred solution of the photovoltaic array dynamic reconstruction adaptive control method described in the present invention, the shadow detection includes: comprehensively analyzing the output conditions of all panels of the photovoltaic array, when there are shadow obstructions or panels with long-term low power, performing shadow detection and identification of the photovoltaic array through a clustering algorithm, and calculating the voltage and current deviation rate and temperature deviation of each panel based on the data uploaded by each edge end in the photovoltaic array, which is expressed as:

[0018]

[0019]

[0020] T d =TT e

[0021] Among them, V d represents the voltage deviation rate; V represents the photovoltaic panel voltage; V 0 Indicates the nominal voltage of the photovoltaic panel; I d represents the current deviation rate; I represents the photovoltaic panel current; I 0 Indicates the nominal current of the photovoltaic panel; T d represents temperature deviation; T represents photovoltaic panel temperature; T e Indicates the ambient temperature;

[0022] Create a 4D feature vector for each photovoltaic panel

[0023] Set the number of clusters K = 3, representing no shadow, partial shadow, and full shadow, initialize the center point, and assign each panel to the nearest center point:

[0024] cluster(i)=argmin j ||X i -cj || 2

[0025] Where cluster(i) represents the cluster number to which panel i is assigned; X i represents the feature vector of panel i; c j Represents the center point of cluster j; recalculate the center point of each cluster:

[0026]

[0027] Among them, cluster j Represents cluster j; the iteration is terminated when the center point moves less than the set threshold or reaches the maximum number of iterations, and the clusters are sorted according to the PR value of the cluster center, and marked as no shadow, partial shadow, and full shadow from high to low according to the PR value; a photovoltaic panel position matrix is ​​created to record the row and column positions of each photovoltaic panel. For each panel in the cluster, the adjacent panels are checked and the proportion of adjacent panels belonging to the same cluster is calculated. The spatial continuity score of the panels within each cluster is calculated, which is expressed as:

[0028]

[0029] Among them, S k represents the spatial continuity score; n j represents the number of adjacent panels in the same cluster j; n represents the total number of adjacent panels; when S k When >0.7, the confidence of cluster j shadow judgment increases by z 1 ; Save the clustering results of the last three time points. When the clustering results are the same for three consecutive times, the confidence of the shadow judgment increases by z 2 ; For the basic confidence of each cluster j, the overall average of the PR of all panels in the PV array is calculated, the absolute difference between the average PR of cluster center j and the overall average is calculated, and the absolute difference is normalized as the basic confidence z 0 , total confidence Z = z 0 +z 1 +z 2 When Z < 0.8, the clustering confidence is judged to be too low and clustering is performed again. When Z ≥ 0.8, a photovoltaic array shadow status report is generated, and the shadow status of each panel and the number and proportion of panels in each shadow status are output. The clustering results are visualized in combination with the created photovoltaic panel position matrix.

[0030] As a preferred solution of the photovoltaic array dynamic reconstruction adaptive control method described in the present invention, the dynamic reconstruction includes: when a faulty photovoltaic panel or shadow is detected, dynamic reconstruction is performed, and the photovoltaic panels affected by the shadow are divided into partial shadow groups and full shadow groups according to the partial shadows and full shadows marked by the shadow detection, and the photovoltaic panels are connected in parallel in each shadow group, and the photovoltaic panels judged as faulty at the edge are removed from the series group and isolated separately; for photovoltaic panels without shadows, all panels are arranged in descending order by PR value, starting from the high PR value, and the panels are allocated to the series group one by one until the voltage upper limit is reached or the photovoltaic panels are used up. If there are still photovoltaic panels remaining when the voltage upper limit is reached, a new series group is started.

[0031] As a preferred solution of the photovoltaic array dynamic reconstruction adaptive control method of the present invention, the dynamic reconstruction also includes defining the domain operation of the unshaded photovoltaic panel as exchanging the positions of two unshaded panels and moving one unshaded panel to another series group, taking the total output power as the main goal and the inter-string performance balance as the secondary goal to construct a fitness function, which is expressed as:

[0032]

[0033] Where D is the fitness score; w is the weight function; M is the total number of unshaded panel series groups; P j represents the power of the jth series group; Represents the power mean of all series groups; initialized by simulated annealing idea, setting the initial temperature T init and the termination temperature T final , cooling rate a, iteration N times at each temperature; at temperature T>T final When , a new solution is generated through domain operation and the fitness of the new solution D is calculated new , when D new >D, accept the new solution, when D new ≤D with probability (D new -D) / T accepts the new solution. If the fitness score of the new solution is greater than the highest fitness score in history, the new solution is updated as the optimal solution. The process of generating new solutions and calculating fitness is repeated N times, and the temperature is lowered to T new =T×a; when the termination temperature is reached, the optimal solution is output, and based on the optimal solution, a minimized reconstruction operation sequence from the current state to the target state is generated.

[0034] In the second aspect, the present invention also provides a photovoltaic array dynamic reconstruction adaptive control system, including a data acquisition module, where the edge end collects relevant data of each photovoltaic panel and analyzes it to preliminarily determine whether the photovoltaic panel has a fault or shadow; a disturbance control module, which performs MPPT control on each photovoltaic panel through a disturbance observation method, adjusts the working point of the photovoltaic panel in real time, and uploads the data of the photovoltaic panel to the photovoltaic array end; and a dynamic reconstruction module, where the photovoltaic array end identifies shadows based on the data uploaded by the edge end, and dynamically reconstructs based on the detected shadows to improve the performance of the photovoltaic system.

[0035] In a third aspect, the present invention further provides a computing device, comprising: a memory and a processor;

[0036] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the photovoltaic array dynamic reconstruction adaptive control method are implemented.

[0037] In a fourth aspect, the present invention further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the photovoltaic array dynamic reconstruction adaptive control method.

[0038] Beneficial effects of the invention: The method of the invention realizes fast and accurate shadow detection and efficient dynamic reconstruction control by combining the advantages of edge computing and centralized control. The method belongs to the field of photovoltaic power generation system control technology and has the advantages of improving the overall performance of the system, enhancing environmental adaptability and improving scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0040] Figure 1 An overall flow chart of a photovoltaic array dynamic reconstruction adaptive control method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0042] Example 1, reference Figure 1 , as an embodiment of the present invention, provides a photovoltaic array dynamic reconstruction adaptive control method, comprising:

[0043] S1: Collect relevant data of the photovoltaic array, and the edge end preliminarily analyzes the relevant data to identify the status of the photovoltaic panel.

[0044] A photovoltaic array is a system consisting of multiple solar panels (also called photovoltaic panels or photovoltaic modules) connected in a specific way. These solar panels are usually installed in a certain arrangement to maximize the absorption and conversion efficiency of solar energy.

[0045] Dynamic Reconfiguration of Photovoltaic Arrays During the operation of a photovoltaic system, the electrical connections between photovoltaic panels are dynamically changed according to real-time conditions, so as to optimize the performance of the entire system when part of the array is blocked or the efficiency of some components decreases.

[0046] Furthermore, the relevant data include illumination data, temperature data, electrical data, and system topology and reconstruction status data. The temperature data is divided into the temperature of each photovoltaic panel and the ambient temperature of the photovoltaic field. The temperature data and electrical data of each photovoltaic panel are collected through sensors installed on each photovoltaic panel.

[0047] Light data mainly includes light intensity and light incident angle. Light sensors (such as silicon photocells or thermopile pyranometers) are installed near photovoltaic panels or integrated into the panels to measure light intensity and incident angle in real time.

[0048] The system topology and reconstruction status data are not directly measured data, but status information maintained within the system. After each reconstruction operation, the system will update the current connection topology and working status information.

[0049] Furthermore, the edge end corresponds to a single photovoltaic panel, which is responsible for data collection and calculating the performance ratio to determine the status of the photovoltaic panel, expressed as:

[0050]

[0051] Among them, PR represents performance ratio; P represents output power; G represents light intensity; P 0 Indicates the nominal power of the photovoltaic panel.

[0052] Calculate the difference between the current PR and the average PR of the same period in the past 7 days C Z , and the difference rate C between the current panel PR and the average PR of the adjacent panels X , expressed as:

[0053]

[0054]

[0055] Among them, PR Z It represents the average PR of the same period in the past 7 days; PRX represents the average PR of adjacent panels.

[0056] When C Z <-20% and C X When >-20%, it is judged as shadow occlusion.

[0057] When C Z <-20% and C X When the value is less than -20%, it is judged as a photovoltaic panel failure, the photovoltaic panel is immediately isolated, and a fault alarm is sent to the photovoltaic array end.

[0058] When C Z >-20% and C X When the power consumption is less than -20%, it is judged that the photovoltaic panel has long-term low power, and no isolation measures are taken, and a fault alarm is sent to the photovoltaic array end.

[0059] It should be noted that when C Z When it is less than -20%, it means that the performance of the current photovoltaic panel is much lower than the average performance at the same time in history. There may be two situations: shadow blocking or photovoltaic panel failure. If the performance of adjacent panels is also reduced, that is, C X >-20%, it is very likely that shadows are blocking this group of photovoltaic arrays. If the performance of the adjacent panels remains basically unchanged and only this photovoltaic panel has declined, it is very likely that the photovoltaic panel is faulty and needs to be isolated and an early warning issued, and staff should be dispatched to inspect and repair it.

[0060] When C Z >-20% and C X When it is less than -20%, it means that the performance of this photovoltaic panel is declining over a period of time. In this case, it generally does not affect the performance of adjacent photovoltaic panels, so there is no need to take isolation measures. The impact can be reduced through dynamic reconstruction of the photovoltaic array end, but manual inspection is required to determine whether there are causes such as aging or failure.

[0061] S2: Perform MPPT control on each photovoltaic panel through the perturbation and observation method.

[0062] The edge end uses the perturbation and observation method to perform MPPT control and adjust the working point of the photovoltaic panel in real time. The basic principle of the perturbation and observation method (P&O) MPPT is to determine the maximum power point by making a small disturbance to the working voltage of the photovoltaic panel and then observing the change in output power.

[0063] The edge end uses the perturbation observation method to perform MPPT control, adjust the working point of the photovoltaic panel in real time, calculate the power P(k) based on the collected voltage V(k) and current I(k), make a small disturbance to the voltage, collect the disturbed voltage V(k+1) and current I(k+1) to calculate the disturbed power P(k+1), compare the power and the power after the disturbance, and decide the next disturbance direction, which is expressed as:

[0064] ΔP=P(k+1)-P(k)

[0065] ΔV=V(k+1)-V(k)

[0066] Wherein, AP represents the power disturbance change; ΔV represents the voltage disturbance change.

[0067] When AP / ΔV>0, V(k+2)=V(k+1)+ΔV; when AP / ΔV<0, V(k+2)=V(k+1)-ΔV.

[0068] The step size is adjusted through an adaptive mechanism. The step size increases when it is far away from the maximum power point and decreases when it is close to the maximum power point. It is expressed as:

[0069] ΔV(k+1)=ΔV(k)×(1+α×|ΔP / ΔV|)

[0070] Among them, α represents the step adjustment parameter; the voltage, current, power, temperature, MPPT efficiency and accumulated energy output data of the photovoltaic panel are uploaded to the photovoltaic array end regularly.

[0071] The sampling frequency is 100Hz-1kHz, and the voltage regulation is achieved by adjusting the duty cycle of the DC-DC converter. The fault detection is assisted by MPPT data. When the power P(k) is lower than the preset threshold P min After a period of time, a fault warning is triggered. min Set to the theoretical minimum output power of the panel under specific light and temperature conditions. If the actual output power P(k) is lower than this minimum value for a long time, it is likely that the panel has some kind of fault or abnormality. The judgment for a period of time can exclude the impact of short-term environmental fluctuations (such as cloud cover). The specific duration can be set according to the situation of the photovoltaic site.

[0072] When |ΔP / ΔV|<ε for a period of time, it means that the panel is aging or needs to be cleaned. |ΔP / ΔV| represents the rate of change of power to voltage, that is, the slope of the PV curve. Under normal circumstances, the PV curve should have an obvious slope, and ε is the minimum acceptable slope.

[0073] When the panel ages, the efficiency of the photovoltaic panel will gradually decrease over time. Aging will cause the PV curve to become flatter and the maximum power point will no longer be so obvious. The accumulation of dust and dirt will reduce the photoelectric conversion efficiency of the panel, which will also cause the PV curve to become flatter. The change in the slope of the PV curve can detect the efficiency decline problem early and distinguish between the need for replacement (aging) and the need for maintenance (cleaning), avoiding the output power from dropping to P min the following.

[0074] S3: The photovoltaic array performs shadow detection through cluster analysis, evaluates the status of the photovoltaic array and performs dynamic reconstruction.

[0075] Furthermore, considering the output of all panels in the photovoltaic array, when there are shadows or panels with long-term low power, the shadow detection and identification of the photovoltaic array is performed through the clustering algorithm. According to the data uploaded by each edge end in the photovoltaic array, the voltage and current deviation rate and temperature deviation of each panel are calculated, which is expressed as:

[0076]

[0077]

[0078] T d =TT e

[0079] Among them, V d represents the voltage deviation rate; V represents the photovoltaic panel voltage; V 0 Indicates the nominal voltage of the photovoltaic panel; I d represents the current deviation rate; I represents the photovoltaic panel current; I 0 Indicates the nominal current of the photovoltaic panel; T d represents temperature deviation; T represents photovoltaic panel temperature; T e Indicates the ambient temperature.

[0080] Create a 4D feature vector for each photovoltaic panel

[0081] Set the number of clusters K = 3, representing no shadow, partial shadow, and full shadow, initialize the center point, and assign each panel to the nearest center point:

[0082] cluster(i)=argmin j ||X i -c j || 2

[0083] Where cluster(i) represents the cluster number to which panel i is assigned; X i represents the feature vector of panel i; cj Represents the center point of cluster j, and recalculates the center point of each cluster:

[0084]

[0085] Among them, cluster j represents cluster j.

[0086] The iteration is terminated when the center point moves less than the set threshold or reaches the maximum number of iterations. The clusters are sorted according to the PR values ​​of the cluster centers and marked as no shadow, partial shadow, and full shadow from high to low according to the PR value.

[0087] Create a photovoltaic panel position matrix, record the row and column positions of each photovoltaic panel, for each panel in the cluster, check the adjacent panels and calculate the proportion of adjacent panels belonging to the same cluster, and calculate the spatial continuity score of the panels within each cluster, expressed as:

[0088]

[0089] Among them, S k represents the spatial continuity score; n j represents the number of adjacent panels in the same cluster j; n represents the total number of adjacent panels; when S k When >0.7, the confidence of cluster j shadow judgment increases by z 1 .

[0090] Save the clustering results of the last three time points. When the clustering results are the same for three consecutive times, the confidence of the shadow judgment increases by z. 2 .

[0091] For the basic confidence of each cluster j, the overall average of the PR of all panels in the PV array is calculated, and the absolute difference between the average PR of cluster center j and the overall average is calculated, and the absolute difference is normalized as the basic confidence z 0 , total confidence Z = z 0 +z 1 +z 2 When Z < 0.8, the clustering confidence is judged to be too low and clustering is performed again. When Z ≥ 0.8, a photovoltaic array shadow status report is generated, and the shadow status of each panel and the number and proportion of panels in each shadow status are output. The clustering results are visualized in combination with the created photovoltaic panel position matrix.

[0092] It should be noted that z 1 and z 2 It can be simply set to a smaller constant, such as 0.1 or 0.2, to satisfy S k When the total confidence is calculated, z is available only when the total confidence is calculated. 1 The added value of2 Only when the clustering results are the same for three consecutive times will they be included in the total confidence.

[0093] Furthermore, when a faulty PV panel failure or shadow is detected, dynamic reconstruction is performed to divide the PV panels affected by the shadow into partial shadow groups and full shadow groups based on the partial shadows and full shadows marked by the shadow detection, connect the PV panels in parallel within each shadow group, and remove the PV panels judged as faulty at the edge ends from the series connection group and isolate them separately.

[0094] For unshaded PV panels, arrange all panels in descending order of PR value, starting from high PR value, and assign panels to series groups one by one until the voltage upper limit is reached or the PV panels are used up. If there are still PV panels left when the voltage upper limit is reached, a new series group is started.

[0095] It should be noted that the series connection of photovoltaic panels means that the panels are connected end to end, the voltage is added, and the current remains unchanged. This can increase the system voltage, reduce the current, and thus reduce line losses. However, if the performance of one panel decreases, it will affect the performance of the entire series chain. Therefore, when there is no shadow or fault, the photovoltaic panels are connected in series to form multiple series groups to reduce line losses and maximize system performance.

[0096] The parallel connection of photovoltaic panels is to connect the positive pole of the panel to the positive pole, and the negative pole to the negative pole, add the current, and keep the voltage unchanged. Problems with a single panel will not significantly affect the overall output. Larger currents need to be handled, which may increase line losses.

[0097] The purpose of connecting shaded panels in parallel is to avoid the "short panel effect" and maximize the remaining output. In a series configuration, the weakest panel will limit the current of the entire series chain. If one panel is severely shaded, it may significantly reduce the output of the entire series group. By connecting in parallel, each shaded panel can independently output its maximum possible current without being limited by other panels.

[0098] Furthermore, the domain operation of unshaded photovoltaic panels is defined as exchanging the positions of two unshaded panels and moving one unshaded panel to another series group. The total output power is taken as the main goal and the performance balance between strings is taken as the secondary goal to construct the fitness function, which is expressed as:

[0099]

[0100] Where D is the fitness score; w is the weight function; M is the total number of unshaded panel series groups; P j represents the power of the jth series group; Represents the average power of all series groups.

[0101] Initialize with simulated annealing idea and set the initial temperature T init and the termination temperature T final , cooling rate a, iteration N times at each temperature.

[0102] At temperature T>T final When , a new solution is generated through domain operation and the fitness of the new solution D is calculated new , when D new >D, accept the new solution, when D new ≤D with probability (D new -D) / T accepts the new solution. If the fitness score of the new solution is greater than the highest fitness score in history, the new solution is updated as the optimal solution. The process of generating new solutions and calculating fitness is repeated N times, and the temperature is lowered to T new =T×a.

[0103] When the termination temperature is reached, the optimal solution is output, and based on the optimal solution, a minimized reconstruction operation sequence from the current state to the target state is generated.

[0104] The present embodiment also provides a photovoltaic array dynamic reconstruction adaptive control system, including a data acquisition module, where the edge end collects relevant data of each photovoltaic panel and analyzes it to preliminarily determine whether the photovoltaic panel has a fault or shadow; a disturbance control module, which performs MPPT control on each photovoltaic panel through a disturbance observation method, adjusts the working point of the photovoltaic panel in real time, and uploads the data of the photovoltaic panel to the photovoltaic array end; and a dynamic reconstruction module, where the photovoltaic array end identifies shadows according to the data uploaded by the edge end, and dynamically reconstructs according to the detected shadows to improve the performance of the photovoltaic system.

[0105] This embodiment also provides a computing device, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to implement the method for realizing dynamic reconstruction adaptive control of photovoltaic arrays as proposed in the above embodiment.

[0106] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for adaptive control of dynamic reconstruction of a photovoltaic array proposed in the above embodiment is implemented.

[0107] The storage medium proposed in this embodiment and the photovoltaic array dynamic reconstruction adaptive control method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0108] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccessMemory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0109] Embodiment 2, below is an embodiment of the present invention, which provides a method for dynamic reconstruction adaptive control of a photovoltaic array. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0110] In order to verify the effectiveness and superiority of the proposed method for dynamic reconstruction of photovoltaic arrays, a 30-day experiment was conducted in a photovoltaic power station. A simulation experiment was conducted using 20 photovoltaic panels in the photovoltaic power station, and the nominal power of each panel was 350W.

[0111] The experiment is divided into three stages: the first stage (1-10 days) uses the traditional MPPT control method; the second stage (11-20 days) uses the edge MPPT control method of the present invention, but does not perform shadow detection and dynamic reconstruction; the third stage (21-30 days) uses the complete dynamic reconstruction adaptive control method of the present invention.

[0112] In the first stage, the traditional perturbation and observation method was used for MPPT control with a fixed step size of 0.5 V. In the second stage, edge computing devices were installed on each photovoltaic panel to implement the edge MPPT control method described in this invention, including adaptive step size adjustment and performance ratio calculation. In the third stage, a complete dynamic reconstruction adaptive control system was deployed, including shadow detection and dynamic reconstruction functions.

[0113] In order to simulate the actual working environment, some abnormal conditions were artificially created at different time periods: on the 15th day, a thin film was covered on one panel to simulate mild pollution; on the 25th day, two panels were covered with a sunshade cloth to simulate partial shadow; on the 28th day, two panels were artificially caused to fail. Data was collected every 10 minutes, and the average value of the data for the day was calculated at the end of each day. Some experimental data were rounded and displayed, as shown in Table 1.

[0114] Table 1 Experimental data table

[0115]

[0116] It can be seen from the data that under similar light intensity and temperature conditions, the total power generation is significantly improved after using the dynamic reconstruction adaptive control method of the present invention. For example, under similar light intensity on the 5th day (traditional MPPT) and the 30th day (dynamic reconstruction), the total power generation increased from 1180kWh to 1248kWh, an increase of about 5.8%, proving that the method of the present invention can more effectively utilize photovoltaic resources and improve the overall power generation efficiency of the system.

[0117] The efficiency of the traditional MPPT method is 95.2%, while the efficiency is increased to 97.8% after using the edge MPPT. Finally, after using the complete dynamic reconstruction method, the MPPT efficiency is stabilized at more than 98.5%. This significant efficiency improvement is due to the adaptive step adjustment mechanism and edge real-time control strategy proposed by the present invention, which enables the system to track the maximum power point faster and more accurately.

[0118] In summary, the experimental results of this embodiment fully verify the effectiveness and superiority of the photovoltaic array dynamic reconstruction adaptive control method. This method not only improves the power generation efficiency of the system, but also greatly improves the accuracy of fault detection and shadow detection, and at the same time enhances the adaptability and stability of the system through the dynamic reconstruction mechanism. These advantages make this method have great potential in practical applications and can provide strong support for the performance optimization and intelligent management of photovoltaic power generation systems.

[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A photovoltaic array dynamic reconstruction adaptive control method, characterized in that: include: Collect relevant data of the photovoltaic array, and the edge end preliminarily analyzes the relevant data to identify the status of the photovoltaic panel; MPPT control of each photovoltaic panel is performed through the perturbation and observation method; The photovoltaic array side performs shadow detection through cluster analysis, evaluates the status of the photovoltaic array and performs dynamic reconstruction.

2. The photovoltaic array dynamic reconstruction adaptive control method according to claim 1, characterized in that: The relevant data includes illumination data, temperature data, electrical data, and system topology and reconstruction status data.

3. The photovoltaic array dynamic reconstruction adaptive control method according to claim 2, characterized in that: The edge end corresponds to a single photovoltaic panel, which is responsible for data collection and calculating the performance ratio to determine the status of the photovoltaic panel, expressed as: Among them, PR represents performance ratio; P represents output power; G represents light intensity; P0 represents the nominal power of the photovoltaic panel; Calculate the difference between the current PR and the average PR of the same period in the past 7 days C Z , and the difference rate C between the current panel PR and the average PR of the adjacent panels X , when C Z <-20% and C X When >-20%, it is considered as shadow occlusion; When C Z <-20% and C X When the value is less than -20%, it is judged as a photovoltaic panel failure, the photovoltaic panel is immediately isolated, and a fault alarm is sent to the photovoltaic array end; When C Z >-20% and C X When <-20%, it is judged that the photovoltaic panel has long-term low power, no isolation measures are taken, and a fault alarm is sent to the photovoltaic array end.

4. The photovoltaic array dynamic reconstruction adaptive control method according to claim 3, characterized in that: The MPPT control includes performing MPPT control at the edge through the perturbation observation method, adjusting the working point of the photovoltaic panel in real time, calculating the power P(k) according to the collected voltage V(k) and current I(k), performing a small amplitude perturbation on the voltage, collecting the voltage V(k+1) and current I(k+1) after the disturbance to calculate the power P(k+1) after the disturbance, comparing the power with the power after the disturbance, and determining the next disturbance direction, which is expressed as: ΔP=P(k+1)-P(k) ΔV=V(k+1)-V(k) Among them, ΔP represents the power disturbance change; ΔV represents the voltage disturbance change; When ΔP / ΔV>0, V(k+2)=V(k+1)+ΔV; When ΔP / ΔV<0, V(k+2)=V(k+1)-ΔV; The step size is adjusted through an adaptive mechanism. The step size increases when it is far away from the maximum power point and decreases when it is close to the maximum power point. It is expressed as: ΔV(k+1)=ΔV(k)×(1+α×|ΔP / ΔV|) Among them, α represents the step adjustment parameter; the voltage, current, power, temperature, MPPT efficiency and accumulated energy output data of the photovoltaic panel are uploaded to the photovoltaic array end regularly.

5. The photovoltaic array dynamic reconstruction adaptive control method according to claim 4, characterized in that: The shadow detection includes: comprehensively analyzing the output of all panels in the photovoltaic array; when there is a shadow or a panel with long-term low power, the shadow detection and identification of the photovoltaic array is performed through a clustering algorithm; and the voltage, current deviation rate and temperature deviation of each panel are calculated based on the data uploaded by each edge end in the photovoltaic array, which is expressed as: T d =T-T e Among them, V d Indicates the voltage deviation rate; V indicates the photovoltaic panel voltage; V0 indicates the photovoltaic panel nominal voltage; I d represents the current deviation rate; I represents the photovoltaic panel current; I0 ​​represents the photovoltaic panel nominal current; T d represents temperature deviation; T represents photovoltaic panel temperature; T e Indicates the ambient temperature; Create a 4D feature vector for each photovoltaic panel Set the number of clusters K = 3, representing no shadow, partial shadow, and full shadow, initialize the center point, and assign each panel to the nearest center point: cluster(i)=argmin j ||X i -c j || 2 Where cluster(i) represents the cluster number to which panel i is assigned; X i represents the feature vector of panel i; c j represents the center point of cluster j; Recompute the centroids of each cluster: Among them, cluster j represents cluster j; When the center point moves less than the set threshold or reaches the maximum number of iterations, the iteration is terminated, and the clusters are sorted according to the PR value of the cluster center, and marked as no shadow, partial shadow, and full shadow from high to low according to the PR value; Create a photovoltaic panel position matrix, record the row and column positions of each photovoltaic panel, for each panel in the cluster, check the adjacent panels and calculate the proportion of adjacent panels belonging to the same cluster, and calculate the spatial continuity score of the panels within each cluster, expressed as: Among them, S k represents the spatial continuity score; n j represents the number of adjacent panels in the same cluster j; n represents the total number of adjacent panels; when S k When >0.7, the confidence of the shadow judgment of cluster j increases by z1; Save the clustering results of the last three time points. When the clustering results are the same for three consecutive times, the confidence of the shadow judgment increases by z2. For the basic confidence of each cluster j, the overall average PR of all panels in the PV array is calculated, and the absolute difference between the average PR of cluster center j and the overall average is calculated. The absolute difference is normalized as the basic confidence z0, and the total confidence Z = z0 + z1 + z2. When Z < 0.8, it is judged that the clustering confidence is too low and clustering is performed again. When Z ≥ 0.8, a PV array shadow status report is generated, and the shadow status of each panel and the number and proportion of panels in each shadow status are output. The clustering results are visualized in combination with the created PV panel position matrix.

6. The photovoltaic array dynamic reconstruction adaptive control method according to claim 5, characterized in that: The dynamic reconstruction includes, when a faulty photovoltaic panel failure or shadow is detected, performing dynamic reconstruction, dividing the photovoltaic panels affected by the shadow into a partial shadow group and a full shadow group according to the partial shadow and full shadow marked by the shadow detection, connecting the photovoltaic panels in parallel in each shadow group, and removing the photovoltaic panels judged as faulty at the edge end from the series group and isolating them separately; For unshaded PV panels, arrange all panels in descending order of PR value, starting from high PR value, and assign panels to series groups one by one until the voltage upper limit is reached or the PV panels are used up. If there are still PV panels left when the voltage upper limit is reached, a new series group is started.

7. The photovoltaic array dynamic reconstruction adaptive control method according to claim 6, characterized in that: The dynamic reconfiguration further includes defining the domain operation of the unshaded photovoltaic panel as exchanging the positions of two unshaded panels and moving one unshaded panel to another series group, taking the total output power as the primary goal and the inter-string performance balance as the secondary goal to construct a fitness function, which is expressed as: Where D is the fitness score; w is the weight function; M is the total number of unshaded panel series groups; P j represents the power of the jth series group; represents the power mean of all series groups; Initialize with simulated annealing idea and set the initial temperature T init and the termination temperature T final , cooling rate a, iteration N times at each temperature; At temperature T>T final When , a new solution is generated through domain operation and the fitness of the new solution D is calculated new , when D new >D, accept the new solution, when D new ≤D with probability (D new -D) / T accepts the new solution. If the fitness score of the new solution is greater than the highest fitness score in history, the new solution is updated as the optimal solution. The process of generating new solutions and calculating fitness is repeated N times, and the temperature is lowered to T new =T×a; When the termination temperature is reached, the optimal solution is output, and based on the optimal solution, a minimized reconstruction operation sequence from the current state to the target state is generated.

8. A photovoltaic array dynamic reconstruction adaptive control system using the method according to any one of claims 1 to 7, characterized in that: include, The data acquisition module collects and analyzes the relevant data of each photovoltaic panel at the edge to preliminarily determine whether there is a fault or shadow on the photovoltaic panel; The disturbance control module performs MPPT control on each photovoltaic panel through the disturbance observation method, adjusts the working point of the photovoltaic panel in real time, and uploads the data of the photovoltaic panel to the photovoltaic array end; Dynamic reconstruction module: The photovoltaic array end identifies shadows based on the data uploaded by the edge end, and dynamically reconstructs according to the detected shadows to improve the performance of the photovoltaic system.

9. A computing device comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images