Rapid maximum power point tracking method based on optimization algorithm and variable step size strategy

Abstract

The invention relates to the technical field of photovoltaic power generation control, in particular to a rapid maximum power point tracking method based on an optimization algorithm and a variable step size strategy. According to the method, an improved intelligent optimization algorithm based on biological heuristic is adopted, and the performance is optimized through an adaptive weight factor and a dynamic search strategy; acquiring a current maximum power point by utilizing a self-adaptive step length adjustment strategy based on a power change rate and combining a momentum factor and a step length reset mechanism; identifying a global optimal region through second derivative information of a power-voltage curve and a sliding window technology; and a state evaluator based on fuzzy logic automatically selects an optimal tracking mode according to the illumination change rate and the power fluctuation amplitude. According to the method, the maximum power point of the photovoltaic system can be quickly and accurately tracked under complex and changeable environmental conditions, and the energy conversion efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation control technology, and in particular to a fast maximum power point tracking method based on optimization algorithms and variable step size strategies. Background Technology

[0002] With the rapid development of renewable energy, photovoltaic (PV) power generation is playing an increasingly important role in the global energy structure due to its clean and environmentally friendly characteristics. To fully utilize the power generation capacity of PV systems, maximum power point tracking (MPPT) technology has become an indispensable key technology in PV systems. The core of MPPT technology is to control the output voltage or current of the PV array so that the system always operates at its maximum power point, thereby improving the system's energy conversion efficiency.

[0003] In traditional MPPT methods, the perturbation and observation (P&O) method is widely used due to its simplicity. This method modifies the output power by perturbing the voltage of the photovoltaic cell during each operating cycle to achieve power optimization. However, the traditional P&O method uses a fixed step size, making it difficult to achieve a balance between tracking speed and steady-state accuracy. To address this issue, CN119248061A discloses a P&O method that combines particle swarm optimization with an adaptive step size adjustment coefficient. By introducing an adaptive step size adjustment coefficient, the step size of the P&O method can be dynamically adjusted to cope with changes in illumination and system dynamic response under different operating conditions.

[0004] As research progressed, researchers discovered that under complex lighting conditions such as partial shading, the PV curve of a photovoltaic array exhibits multi-peak characteristics. Traditional gradient-based MPPT methods are prone to getting trapped in local maximum power points and cannot accurately track the global maximum power point (GMPPT). To address this issue, MPPT methods based on intelligent optimization algorithms have been proposed and widely studied. CN114706445B discloses a photovoltaic maximum power point tracking method based on the DE-GWO algorithm. This method uses the Differential Evolution (DE) algorithm for global search and the Grey Wolf Optimization (GWO) algorithm for local search, enabling it to track the global maximum power point under different partial shading conditions.

[0005] To further improve the performance of intelligent optimization algorithms in MPPT, CN117572929A proposes a photovoltaic maximum power point tracking method based on an adaptive vulture search algorithm. Based on the BES algorithm, it introduces a Gaussian mixture adaptive walk strategy, a progressive dive adaptive switching strategy, and a vulture flock size adjustment mechanism, which can successfully track the global maximum power point under local shading conditions, and has faster tracking speed and higher accuracy.

[0006] Furthermore, CN119396246B discloses a photovoltaic power control method and system based on an improved sparrow search algorithm and a variable step-size perturbation observation method. This method first uses the improved sparrow search algorithm to quickly track to the vicinity of the maximum power point, and then uses the variable step-size perturbation observation method to perform a fine search for the maximum power point, thereby improving the tracking speed and tracking accuracy. CN113595132B proposes a photovoltaic online parameter identification method based on global maximum power point tracking and a hybrid optimization algorithm. This method combines the powerful global search capability of the quantum particle swarm optimization algorithm and the powerful local search capability of the Levenberg-Marquardt algorithm, further improving the speed, accuracy, stability, reliability, and convergence of photovoltaic model parameter identification.

[0007] However, existing MPPT technology still has the following problems.

[0008] 1. Traditional gradient-based algorithms, such as the perturbation-observation method, are prone to getting stuck in local extrema under complex lighting conditions (such as local shadows, rapid fluctuations, etc.), and cannot accurately track the global maximum power point.

[0009] 2. Although existing intelligent optimization algorithms have global search capabilities, they are insufficient in terms of convergence speed and tracking accuracy, making it difficult to meet the requirements for dynamic response speed in practical engineering.

[0010] 3. In scenarios with rapidly changing lighting, algorithm parameters are often fixed and lack an adaptive adjustment mechanism, making it difficult to achieve rapid response while ensuring global search capabilities.

[0011] 4. Existing methods lack in-depth research on the mapping relationship between environmental parameters and key algorithm parameters, and cannot automatically optimize algorithm configuration according to environmental changes.

[0012] 5. Most algorithms lack a smooth transition strategy during mode switching, which can easily lead to fluctuations in system output power and affect system stability.

[0013] Therefore, there is an urgent need to develop a novel MPPT method that can quickly and accurately track the global maximum power point under complex lighting conditions. This method should have adaptive parameter adjustment capabilities, be able to automatically optimize the algorithm configuration according to environmental changes, and achieve a smooth transition during mode switching, thereby improving the energy conversion efficiency and stability of photovoltaic systems. Summary of the Invention

[0014] To address the technical challenges faced by traditional MPPT methods under complex lighting conditions, such as local extremum trapping, insufficient convergence speed and tracking accuracy of intelligent optimization algorithms, and insufficient algorithm response speed in scenarios with rapidly changing lighting, this invention provides a fast maximum power point tracking method based on optimization algorithms and a variable step size strategy. This method aims to effectively avoid trapping in local extrema, significantly improve tracking speed, achieve fast and accurate maximum power point tracking under complex lighting conditions, improve the overall power generation efficiency of photovoltaic systems, and ensure that the algorithm maintains optimal performance under different operating conditions.

[0015] The technical solution adopted by this invention to solve its technical problem is as follows: a fast maximum power point tracking method based on optimization algorithms and variable step size strategies is provided, including: employing a bio-inspired improved intelligent optimization algorithm, optimizing algorithm performance through adaptive weighting factors and dynamic search strategies, setting a population search space according to the photovoltaic array voltage range, and adjusting the search range using a nonlinear convergence factor; employing an adaptive step size adjustment strategy based on the power change rate, combined with momentum factors and a step size reset mechanism, to obtain the current maximum power point; identifying the global optimal region through the second derivative information of the power-voltage curve and sliding window technology, establishing a short-term prediction model based on historical data, and judging and triggering the global search mode; a state evaluator based on fuzzy logic automatically selects the optimal tracking mode according to the rate of change of illumination and the amplitude of power fluctuations, and adopting a transition interval buffer strategy to achieve smooth mode transition; establishing a mapping relationship model between environmental parameters and key algorithm parameters, and continuously optimizing parameter configuration through an online learning mechanism.

[0016] Optionally, the improved intelligent optimization algorithm based on bioinspiration includes any one of the sparrow search algorithm, hippo algorithm, gray wolf optimization algorithm, or snake optimization algorithm, wherein: in the initial iteration stage of the algorithm, an adaptive weight factor in the range of 0.9-1.2 is used to enhance the global search capability; as the iteration progresses, the adaptive weight factor is gradually reduced to improve the local search accuracy.

[0017] Optionally, the dynamic search strategy includes: dynamically adjusting the search neighborhood range according to the convergence trend of the current optimal solution; automatically expanding the search neighborhood range when the power change rate is detected to exceed a preset threshold, wherein: the adjustment range of the search neighborhood range is proportional to the magnitude of the power change rate.

[0018] Optionally, the adaptive step size adjustment strategy based on the power change rate includes: dynamically adjusting the perturbation step size according to the current power gradient value; automatically reducing the step size to improve tracking accuracy when approaching the maximum power point; and increasing the step size to accelerate the tracking speed when moving away from the maximum power point.

[0019] Optionally, the step size reset mechanism includes: triggering a step size reset when the output power is detected to drop more than a preset threshold within a single sampling period; and dynamically adjusting the size of the reset step size according to the power drop magnitude.

[0020] Optionally, identifying the globally optimal region includes: monitoring local extreme points of the power-voltage curve in real time using sliding window technology; identifying true peak points by combining second-order derivative information, and eliminating false peaks caused by noise.

[0021] Optionally, the short-term prediction model adopts an autoregressive integral moving average model to predict future short-term power change trends using historical power data.

[0022] Optionally, the fuzzy logic state evaluator includes: defining a fuzzy set of illumination change rate and power fluctuation amplitude; establishing a fuzzy rule base to map input variables to the optimal tracking mode; using the centroid method for defuzzification; and outputting the tracking mode selection result.

[0023] Optionally, the transition interval buffering strategy includes: setting a transition interval during mode switching; and dynamically adjusting the length of the transition interval according to the current power change trend.

[0024] Optionally, the mapping relationship model between the environmental parameters and the key parameters of the algorithm adopts a radial basis function neural network. The inputs include: light intensity, component temperature, number of shaded areas, current operating voltage, and output power; the outputs include: population size, convergence factor, and mutation probability; the parameters of the mapping relationship model between the environmental parameters and the key parameters of the algorithm are updated online using the sliding window recursive least squares method.

[0025] The fast maximum power point tracking (MPPT) method based on optimization algorithms and variable step-size strategies provided by this invention has the following advantages: By organically combining an improved intelligent optimization algorithm with an adaptive perturbation-observation method, efficient MPPT is achieved under complex lighting conditions. Compared with the traditional perturbation-observation method, this invention effectively avoids the problem of getting trapped in local extrema and can accurately identify and track the global maximum power point under local shadow conditions. Compared with simply using intelligent optimization algorithms, this invention significantly improves the tracking speed through an adaptive step-size adjustment strategy and a fuzzy logic controller, maintaining stable tracking performance even under rapidly changing lighting conditions. By constructing a mapping model between environmental parameters and key algorithm parameters, dynamic adaptive adjustment of algorithm parameters is achieved, ensuring optimal performance under different operating conditions. Experimental results show that under standard test conditions, the tracking efficiency of this invention is 3.5% higher than the traditional P&O method and 2.1% higher than the basic PSO algorithm. Under complex lighting conditions, the tracking efficiency advantage of this invention is even more significant, exceeding 5.8% compared to the traditional method, while reducing the response time by approximately 40%, effectively improving the overall power generation efficiency of the photovoltaic system. Attached Figure Description

[0026] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application.

[0027] Figure 1 This is the overall flowchart of the present invention.

[0028] Figure 2 This is a schematic diagram of an electronic device. Detailed Implementation

[0029] 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.

[0030] Example 1

[0031] In one embodiment of the present invention, please refer to Figure 1 This paper presents a fast maximum power point tracking (MPPT) method based on optimization algorithms and a variable step size strategy. This method achieves efficient and stable MPPT through a multi-level optimization strategy. The method specifically includes the following steps.

[0032] S1. An improved intelligent optimization algorithm based on bioinspiration is adopted. The algorithm performance is optimized by adaptive weighting factors and dynamic search strategies. The population search space is set according to the voltage range of the photovoltaic array, and the search range is adjusted by nonlinear convergence factors.

[0033] In one embodiment of the present invention, the improved intelligent optimization algorithm adopts a bio-inspired intelligent optimization algorithm framework, specifically selecting any one of the Sparrow Search Algorithm, Hippo Algorithm, Gray Wolf Optimization Algorithm, or Snake Optimization Algorithm as the basic intelligent optimization algorithm. In algorithm implementation, the initial position distribution range of the population is first determined based on the current open-circuit voltage and maximum power point voltage range of the photovoltaic array. For example, for a photovoltaic module with a rated voltage of 30V, the search space is set to the 0-40V voltage range to ensure coverage of all possible power peak areas.

[0034] In one embodiment of the present invention, the core improvements of the bio-inspired improved intelligent optimization algorithm include three aspects, specifically: (1) introducing an adaptive weight factor, which is nonlinearly adjusted according to the number of iterations and the quality of the current optimal solution. A larger adaptive weight factor is used in the initial iteration stage of the algorithm, for example, an adaptive weight factor in the range of 0.9-1.2, to enhance the global exploration capability. As the iteration progresses, the weight factor is gradually reduced to the range of 0.4-0.6 to improve the local exploration accuracy; (2) using a nonlinear convergence factor, the mathematical expression of which is: This quadratic function decay curve can better balance the exploration and development stages; (3) Implement a dynamic neighborhood search strategy. When the power change rate is detected to be lower than the threshold, the search neighborhood radius is automatically reduced and the search is focused on the vicinity of the current optimal solution for fine search.

[0035] Taking the Sparrow Search algorithm as an example, it first initializes 20-50 voltage operating points (i.e., "sparrows") evenly distributed within a preset voltage range. Each operating point represents a possible power peak location, and the fitness value is calculated by measuring the output power at that voltage. In each iteration, the algorithm guides other individuals to move towards higher power regions based on the location information of the current optimal solution. The movement step size is dynamically controlled by an adaptive weight factor, and the individual positions are updated in conjunction with real-time power measurement data of the photovoltaic system. When the change in the optimal solution is less than 1% for three consecutive iterations, a neighborhood shrinking mechanism is triggered, narrowing the search range to within ±2V of the current optimal solution for fine-tuning. The entire process continues until the preset convergence condition is met (e.g., the number of iterations reaches 50 or the power change is less than 0.5W).

[0036] The specific implementation of the dynamic search strategy includes: dynamically adjusting the search neighborhood range based on the convergence trend of the current optimal solution; automatically expanding the search neighborhood range when the power change rate exceeds a preset threshold (e.g., 5% / second), with the adjustment magnitude proportional to the power change rate. For example, when the power change rate is 10% / second, the search neighborhood range is expanded to twice its original size; when the power change rate is 20% / second, the search neighborhood range is expanded to four times its original size.

[0037] S2. An adaptive step size adjustment strategy based on the power change rate is adopted, combined with momentum factor and step size reset mechanism, to obtain the current maximum power point.

[0038] In one embodiment of the present invention, this step designs an adaptive step size adjustment strategy based on the power change rate, using an exponentially decaying step size adjustment function whose decay coefficient is associated with the power gradient value, thereby achieving rapid approach and accurate tracking of the maximum power point.

[0039] In one embodiment of the present invention, the adaptive step size adjustment strategy based on the power change rate includes: dynamically adjusting the perturbation step size according to the current power gradient value; automatically reducing the step size to improve tracking accuracy when approaching the maximum power point; and increasing the step size to accelerate the tracking speed when moving away from the maximum power point.

[0040] In one embodiment of the invention, the perturbation step size is dynamically adjusted based on the current power gradient value. When approaching the maximum power point (e.g., power gradient value less than 0.5 W / V), the step size is automatically reduced to 0.01-0.05 V to improve tracking accuracy; when moving away from the maximum power point (e.g., power gradient value greater than 2 W / V), the step size is increased to 0.2-0.5 V to accelerate tracking speed. Simultaneously, a momentum factor is introduced to maintain the continuity of the search direction and avoid directional errors caused by instantaneous fluctuations.

[0041] In one embodiment of the present invention, the step size reset mechanism is specifically implemented by: triggering a step size reset when the output power is detected to decrease by more than a preset threshold (e.g., 3%) within a single sampling period; and dynamically adjusting the size of the reset step size according to the power decrease. For example, when the power decreases by 5%, the reset step size is 1.5 times the base step size; when the power decreases by 10%, the reset step size is 2 times the base step size. This mechanism can effectively cope with sudden changes in illumination and quickly relocate the maximum power point.

[0042] In one embodiment of the present invention, the steps of implementing the variable step size perturbation observation mechanism include: (1) calculating the power change rate: real-time acquisition of the output power of the photovoltaic array. and operating voltage The formula for calculating the power change between adjacent sampling periods is: , ,in, This represents the change in output power between adjacent sampling periods. The output power of the previous sampling period. This represents the change in operating voltage between adjacent sampling periods. The voltage is the operating voltage of the previous sampling period; the power gradient is calculated using the center difference method to improve accuracy, and the formula is: ,in, For power gradient, The working voltage for the next sampling period. (2) Design a step size adjustment strategy based on the hyperbolic tangent function to generate an adaptive step size, the formula is: ,in, This is the sensitivity adjustment factor, with a value range of 0.5-2.0. and These are the maximum and minimum step size limits, respectively. This is an adaptive step size function. Its characteristic is that when the power gradient is large, the step size approaches... The gradient automatically decays to a small value. (3) An improved gradient sign detection method is used to generate direction decisions, and the formula is: ,in, The dead zone threshold is set to 0.02-0.05V to avoid oscillations near extreme points. (4) A momentum factor is introduced to accelerate convergence, and the formula is: Where β is the momentum coefficient, ranging from 0.6 to 0.8, which can smooth the process of step size change. For the final adaptive step size, This is the adaptive step size used in the previous convergence; when a continuous decrease in power is detected, a step size reset mechanism is triggered, as shown in the formula: , where η is the backoff coefficient, with a value range of 0.3-0.5.

[0043] Given that the maximum power point voltage of a photovoltaic module under standard test conditions is 18V, and the system parameter settings include: It is 1.2V. It is 0.05V. It is 1.0. Taking 0.03V as an example, the specific implementation process includes: (1) the working voltage at the initial working point. The initial output power was measured at 15V. It is 85w. (2) Adjust to 16.2V, initial adaptive step size It was measured to be 1.2V. The value is 92W. (3) Calculate the power gradient: , maintain positive adjustment. (4) According to Calculate the new step size: (5) Applying the momentum factor (β = 0.7): (6) Operating voltage at the next operating point The output power at this operating point was measured. The value is 96W. (7) Continue iterating until the power gradient is reached. It eventually stabilized at 17.8V (actual MPPT point).

[0044] When encountering sudden changes in light intensity (such as a 30% drop in power due to cloud cover), the specific implementation process includes: (1) Current operating voltage It is 17.8V. The power was changed from 96W to 67W. (2) If the power drop exceeded 20%, the step size reset mechanism was immediately triggered. (3) A backoff step size of η was adopted, which was 0.4. The value is 1.2, so we get It is 0.48V. (4) Reverse adjustment to (5) Recalculate the power gradient and proceed with the normal tracking process.

[0045] S3. Identify the global optimal region by using the second derivative information of the power-voltage curve and the sliding window technique, and establish a short-term prediction model by combining historical data to determine and trigger the global search mode.

[0046] In one embodiment of the present invention, the globally optimal region is identified using the second derivative information of the power-voltage curve and a sliding window technique. Specifically, this includes: real-time monitoring of local extreme points of the power-voltage curve using a sliding window technique, with the window size typically set to 5-10 sampling points; and identifying true peak points by combining second derivative information with noise-induced false peaks. When the second derivative changes from positive to negative and its absolute value is greater than a preset threshold, it is determined to be a true peak point.

[0047] In one embodiment of the present invention, the short-term prediction model employs an autoregressive integral moving average model to predict future short-term power change trends using historical power data. The model parameters are set as follows: autoregressive term order p=3, difference order d=1, and moving average term order q=2. This model can predict power change trends for the next 3-5 sampling points based on historical data from the past 15-20 sampling points. When the prediction results indicate that multiple power peaks may occur or the current peak is not globally optimal, a global search mode is triggered.

[0048] In one embodiment of the present invention, the multimodal peak identification and decision-making implementation steps include: (1) constructing the second derivative matrix of the power-voltage (PV) curve, and calculating the curvature characteristics using the five-point central difference method, the formula being: ,in, This is a curvature characteristic. When When the value is greater than γ, it is marked as a potential peak point, where γ is the curvature threshold, ranging from 0.5 to 1.5 W / V. 2 (2) Sliding window peak detection: Design a dynamic width sliding window, with the voltage span within the window being... With current step size The correlation formula is: ,in, and These are the maximum and minimum operating voltage limits, respectively; an improved Z-score algorithm is used within the window to identify local peaks, as shown in the formula: ,in, This is a local peak value. The average power within the window. Standard deviation This represents the first i-th power value within the window. When... θ is the local peak threshold, ranging from 2.5 to 3.5, and satisfies... When < -γ, it is confirmed as a valid peak value. (3) Establish the peak quality evaluation function, the formula is: ,in, This is the peak quality assessment value. , , These are all weighting coefficients, with values ​​of 0.6, 0.3, and 0.1 respectively. This represents the maximum power within the window. This represents the maximum value of the peak neighborhood stability index. The peak neighborhood stability index is calculated using the following formula: , This is the j-th power value within the window. Select... The largest peak value is used as the tracking target. (4) The power trend is predicted using the time series ARIMA(2,1,1) model, and the formula is: ,in, This is a power trend prediction value. and All are weighting coefficients. This is the predicted power trend value at time point t-1. This is the predicted power trend value at time point t-2. This represents the impact of the error term at time point t-1 on the current difference value. For the random error term at time point t-1, This represents the random error term at time point t. A global search mode is triggered when three consecutive power trend predictions deviate from the current value by more than 15%.

[0049] A photovoltaic array exhibits bimodal characteristics under partial shading, and the parameter settings include: γ = 0.8 W / V. 2 θ is 3.0. Taking 2V as an example, the specific implementation process includes: (1) Current working point It was measured to be 16V. 120W. (2) Within the sliding window [14V, 16V], the following were detected: a. Peak value 1: operating voltage 15.2V, output power 125W, curvature characteristics -1.2; b, Peak value 2: Operating voltage It is 17.5V, and the output power is... 135W, curvature characteristics =-0.9. (3) Calculate the peak mass: , (4) Choose As the tracking target, the ARIMA model is used to predict the future power change rate at 3 points as [0.8%, 0.6%, 0.7%]. (5) The actual measured power change rate is [-12%, -15%, -18%], triggering a global search. (6) A new global peak is found during the rescan. It is 18.2V, and the output power is... 145W. (7) Update the working point and enter stable tracking mode.

[0050] S4. The fuzzy logic-based state evaluator automatically selects the optimal tracking mode based on the rate of change of illumination and the amplitude of power fluctuations, and adopts a transition interval buffer strategy to achieve smooth mode transition.

[0051] In one embodiment of the present invention, this step establishes a state estimator based on fuzzy logic to automatically select the optimal tracking mode based on the rate of change of illumination and the amplitude of power fluctuation. The fuzzy logic state estimator includes: defining fuzzy sets for the rate of change of illumination and the amplitude of power fluctuation, with the rate of change of illumination divided into three levels: "low," "medium," and "high," and the amplitude of power fluctuation divided into three levels: "small," "medium," and "large"; establishing a fuzzy rule base to map the input variables to the optimal tracking mode; using the centroid method for defuzzification; and outputting the tracking mode selection result.

[0052] In one embodiment of the present invention, the steps for switching the hybrid control strategy include: (1) defining fuzzy input variables: designing two input variables, namely the rate of change of illumination and the power fluctuation amplitude, wherein the rate of change of illumination is calculated using the difference in irradiance between adjacent sampling periods, and the formula is: ,in, The rate of change of illumination, The irradiance at time point t. The irradiance at time point t-1 The sampling period is [value], and the power fluctuation amplitude is the normalized value of the power standard deviation within the sliding window, as shown in the formula: ,in, For power fluctuation amplitude, The standard deviation of power within the sliding window. (2) The triangular membership function is used for fuzzification. The domain of illumination change rate is divided into {"low", "medium", "high"}, and the illumination change rate at the boundary points is [0, 50, 100, 200] W / m 2 / s, the domain of power fluctuation amplitude is divided into {"small", "medium", "large"}, and the power fluctuation amplitude at the boundary point is [0, 0.05, 0.1, 0.2]. (3) Construct a fuzzy rule base and design control rules, for example: when the rate of change of illumination is "low" and the power fluctuation amplitude is "small", select the P&O mode; when the rate of change of illumination is "high" or the power fluctuation amplitude is "large", select the global search mode; otherwise, select the hybrid search mode. (4) Use the centroid method to calculate the control output and generate the defuzzified decision, the formula is: ,in, For the clarity of the conclusions of each rule, For rule activation strength, Select the result for the output tracking mode. (5) Mode smooth transition mechanism: Design a transition interval buffer strategy, the formula is: ,in, To dynamically adjust the load factor, The load factor under the old model. The load factor is the old model, where α is the mixing coefficient, calculated gradually using the sigmoid function. The formula is: , where k is the control of the transition speed, and its value ranges from 5 to 10.

[0053] Based on the parameters of a certain photovoltaic system: It is 300W. 50W, sampling period Taking 0.1s as an example, the specific implementation process includes: (1) the currently measured 850W / m 2 , 830W / m 2 ,but 200W / m 2 / s. (2) Power data within the sliding window [280, 275, 285, 270] W, power standard deviation within the sliding window is 6.45 W, power fluctuation amplitude is 0.0258. (3) Fuzzification result: For "high", (4) Rule activation: Based on the fuzzification result, match the corresponding control rule and select the corresponding search mode. (5) Calculate the control output using the centroid method, and obtain the result. The value is 0.8. (6) The mixing coefficient α is calculated to be 0.91. (7) Finally, a mixed strategy of 91% improved intelligent optimization algorithm and 9% P&O is adopted. (8) When subsequent detection Reduced to 30W / m 2 When the value is / s, α is automatically adjusted to 0.12, switching to a P&O-based mode.

[0054] In one embodiment of the present invention, the dynamic adjustment process includes: when the cloud layer moves rapidly, causing... >150W / m 2 When the value is / s and lasts for 3 consecutive cycles, it is forced to enter the global search mode; when switching modes, the optimal solution of the previous mode is retained as the initial population center point of the new mode; the minimum residence time is set to 0.5s to prevent high-frequency oscillation switching.

[0055] S5. Establish a mapping model between environmental parameters and key algorithm parameters, and continuously optimize parameter configuration through an online learning mechanism.

[0056] In one embodiment of the present invention, this step establishes a mapping model between environmental parameters and key algorithm parameters, using a radial basis function-based neural network. The inputs include: light intensity, component temperature, number of shaded areas, current operating voltage, and output power; the outputs include: population size, convergence factor, and mutation probability. The neural network employs a three-layer structure with 15 hidden nodes, and uses a Gaussian radial basis function as the activation function.

[0057] This model continuously optimizes parameter configuration through an online learning mechanism, employing a sliding window recursive least squares method to update the mapping relationship between environmental parameters and key algorithm parameters online. The sliding window size is set to 100 sampling points, and the initial learning rate is 0.05, gradually decreasing to 0.01 over time. This adaptive parameter adjustment mechanism enables the algorithm to maintain optimal performance under different environmental conditions.

[0058] Through the synergistic effect of the above five steps, this method can quickly and accurately track the maximum power point of a photovoltaic system under various complex lighting conditions, significantly improving energy conversion efficiency. Experimental results show that compared with the traditional perturbation-observation method, this method improves steady-state tracking accuracy by 15%, increases dynamic response speed by 30%, and effectively avoids getting trapped in local optima under partially shaded conditions.

[0059] Example 2

[0060] Based on Embodiment 1 described above, this embodiment also provides an electronic device, please refer to the appendix. Figure 2, Figure 2 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0061] like Figure 2 As shown, an electronic device may include a processing unit (such as a central processing unit, graphics processing unit, etc.) that can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) or a program loaded from a storage device into random access memory (RAM). The RAM also stores various programs and data required for the operation of the electronic device. The processing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0062] Typically, the following devices can be connected to an I / O interface: input devices such as touchscreens, touchpads, keyboards, mice, and cameras; output devices such as liquid crystal displays (LCDs) and speakers; storage devices such as magnetic tapes and hard drives; and communication devices. Communication devices allow electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 2 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively. Figure 2 Each box shown can represent a device or multiple devices as needed.

[0063] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processing device, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0064] Example 3

[0065] Based on Embodiment 1 above, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above method.

[0066] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), or any suitable combination thereof.

[0067] In some embodiments, the client and server may communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and may interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0068] The aforementioned computer-readable medium may be included in the aforementioned device or may exist independently without being assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire training data and transform the training data to obtain initial data; determine an initial rule base based on the initial data and optimize the parameters of the initial rule base to obtain a target rule base; calculate activation weights for the rules in the target rule base according to a preset activation weight calculation formula; and determine abnormal information based on test data and the activation weights.

[0069] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0070] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0071] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a data acquisition unit, a rule determination unit, a weight calculation unit, and an anomaly determination unit. The names of these units do not necessarily limit the specific unit; for example, a data acquisition unit may also be described as a "unit for acquiring training data."

[0072] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), etc.

[0073] Obviously, those skilled in the art will understand that the various steps of the present invention described above can be performed in a manner different from that described above, and the simulation methods and experimental equipment include, but are not limited to, the above description. The steps of the present invention described above can be performed in a different order in certain circumstances, and the steps shown or described above can be performed separately. Therefore, the present invention is not limited to any particular combination of hardware and software.

[0074] The above description, in conjunction with specific embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such deductions or substitutions should be considered within the scope of protection of the present invention.

Claims

1. A fast maximum power point tracking method based on optimization algorithms and variable step size strategies, applied to photovoltaic power generation systems, characterized in that... include: An improved intelligent optimization algorithm based on bioinspiration is adopted. The algorithm performance is optimized by adaptive weighting factors and dynamic search strategies. The population search space is set according to the voltage range of the photovoltaic array, and the search range is adjusted by nonlinear convergence factors. An adaptive step size adjustment strategy based on the power change rate is adopted, combined with momentum factor and step size reset mechanism, to obtain the current maximum power point; The global optimal region is identified by using the second derivative information of the power-voltage curve and the sliding window technique. A short-term prediction model is established by combining historical data to determine and trigger the global search mode. The fuzzy logic-based state evaluator automatically selects the optimal tracking mode based on the rate of change of illumination and the amplitude of power fluctuations, and adopts a transition interval buffering strategy to achieve smooth mode transition. Establish a mapping model between environmental parameters and key algorithm parameters, and continuously optimize parameter configuration through an online learning mechanism.

2. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The improved intelligent optimization algorithm based on bioinspiration includes any one of the following: sparrow search algorithm, hippo algorithm, gray wolf optimization algorithm, or snake optimization algorithm, wherein: In the initial iteration phase of the algorithm, an adaptive weight factor ranging from 0.9 to 1.2 is used to enhance the global search capability; As the iterations proceed, the adaptive weight factor is gradually reduced to improve the accuracy of the local search.

3. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The dynamic search strategy includes: The search neighborhood range is dynamically adjusted based on the convergence trend of the current optimal solution; When the power change rate is detected to exceed a preset threshold, the search neighborhood range is automatically expanded, wherein the adjustment range of the search neighborhood range is proportional to the magnitude of the power change rate.

4. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The adaptive step size adjustment strategy based on power change rate includes: The perturbation step size is dynamically adjusted based on the current power gradient value; Automatically reduce the step size to improve tracking accuracy when approaching the maximum power point; Increase the step size as you move away from the point of maximum power to speed up the tracking.

5. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The step size reset mechanism includes: When the output power is detected to drop more than a preset threshold within a single sampling period, the step size is reset. The reset step size is dynamically adjusted based on the power decrease rate.

6. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The identification of the globally optimal region includes: Real-time monitoring of local extreme points on the power-voltage curve using sliding window technology; By combining second derivative information, the true peak points are identified, and false peaks caused by noise are eliminated.

7. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The short-term prediction model adopts an autoregressive integral moving average model, which predicts the future short-term power change trend through historical power data.

8. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The fuzzy logic state estimator includes: Define a fuzzy set of illumination change rate and power fluctuation amplitude; Establish a fuzzy rule base to map input variables to the optimal tracking mode; The centroid method is used for deblurring, and the tracking mode selection result is output.

9. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The transition interval buffering strategy includes: Set a transition interval during mode switching; The length of the transition range is dynamically adjusted based on the current power change trend.

10. The fast maximum power point tracking method based on optimization algorithm and variable step size strategy as described in claim 1, characterized in that, The mapping relationship model between the environmental parameters and the key parameters of the algorithm adopts a radial basis function neural network. The inputs include: light intensity, component temperature, number of shaded areas, current operating voltage, and output power; the outputs include: population size, convergence factor, and mutation probability. The parameters of the mapping relationship model between the environmental parameters and the key parameters of the algorithm are updated online using the sliding window recursive least squares method.

Images