Tracking configuration regulation and control method and system of inverter

Abstract

The invention discloses a tracking configuration regulation and control method and system of an inverter, and relates to the technical field of tracking configuration regulation and control of the inverter, and the method comprises the following steps: obtaining output power data in real time, constructing a power time sequence, carrying out differential processing, and generating a differential trend characteristic value; dynamically evaluating an illumination disturbance trend according to the difference trend characteristic value, and constructing a step length adjustment function and outputting an original step length in combination with the size of a current power error; smoothing the original step length, judging whether a plurality of extreme points exist in the current operation environment or not, and if so, identifying a global optimal power point direction; and generating a control adjustment parameter based on the global optimal power point direction, correcting the step length and the target voltage after smoothing processing, generating a fusion control quantity, and updating the inverter working point. According to the method, the differential trend characteristic value based on the power curvature is constructed, and the global optimal power point direction is accurately assisted to be recognized, so that the steady-state efficiency and the tracking robustness under multi-peak interference are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of inverter tracking configuration control technology, and more specifically, to an inverter tracking configuration control method and system. Background Technology

[0002] In applications such as distributed energy systems, microgrids, and smart grids, inverters serve as core equipment for power conversion and energy management, and their operational performance directly affects the overall efficiency and stability of the system. With the increasing proportion of renewable energy integration, inverters not only perform traditional DC-AC energy conversion functions but also need the ability to dynamically track external changes (such as sunlight intensity, wind speed, and load variations) and perform intelligent configuration and control. Especially in photovoltaic, wind power, and energy storage systems, inverters need to achieve real-time optimized management of system energy flow through multiple functions such as maximum power point tracking (MPPT), load tracking, and grid support.

[0003] In existing technologies, inverter tracking configuration control typically employs classic algorithms such as disturbance and observation (P&O) and incremental conductance (INC), as well as local control logic based on fixed parameter settings. These methods can adequately meet system operation requirements in static or slowly varying environments. However, in practical applications, renewable energy generation exhibits volatility, randomness, and intermittency, leading to more complex dynamic control challenges for inverters. For example, in multi-inverter parallel operation, multi-source heterogeneous systems, and bidirectional power flow conditions, traditional control methods are prone to problems such as insufficient tracking accuracy, lag in dynamic response, energy distribution imbalance, and system oscillations, severely impacting system operating efficiency and reliability.

[0004] Furthermore, as the power grid places higher demands on inverter grid-connected performance, such as the integration of functions like low voltage ride-through (LVRT), frequency regulation, and voltage reactive power support, relying solely on traditional configurations and local tracking is no longer sufficient to meet the system-level coordinated control requirements in complex environments. Therefore, there is an urgent need to research and develop a novel inverter tracking configuration and control method that addresses dynamic environmental changes while balancing local optimization and global coordination, in order to improve the intelligent adaptability and operational stability of inverters in multi-source, multi-load systems.

[0005] For example, the invention patent announcement CN117970781A discloses an intelligent control method and system for an inverter, comprising: acquiring the output voltage data sequence of the inverter and dividing it into several sequence segments; obtaining the degree of change of each sequence segment based on the difference between adjacent output voltage data within each sequence segment, thereby obtaining the contribution of each sequence segment; acquiring the reference degree of each output voltage data in the output voltage data sequence, thereby obtaining the weight of each output voltage data in the output voltage data sequence; obtaining a predicted value of the output voltage data based on the weight of each output voltage data in the output voltage data sequence; inputting the predicted value of the output voltage data into a PID controller; and outputting a control command for the inverter. This invention improves the accuracy of predicting the acquired voltage data using this model, thereby improving the accuracy of the inverter control system in regulating and controlling the output voltage.

[0006] For example, the invention patent announcement CN111474978B discloses a photovoltaic MPPT control method with intelligent step size adjustment and its system storage medium. It uses a particle swarm optimization (PSO) algorithm to improve a neural network (RBF / BP) and combines it with a perturbation-observation method. The method switches from multiple power peaks to the peak with the global maximum MPP. On the peak with the global maximum MPP, the algorithm continuously narrows the convergence step size to approach the peak value, thus continuously approaching the MPP and stabilizing it within the allowable error range, reducing oscillations at the maximum power point. Currently, the common perturbation-observation method is generally used for convergence towards the maximum power point. To improve convergence efficiency and save time, we use a BP neural network to switch the step size of the perturbation-observation method. Therefore, the convergence step size needs to be changed constantly. Large step sizes are used in the early stages of ramp-up convergence, but later, at a certain point, the BP neural network is switched to a smaller step size to continue convergence, preventing excessive oscillations at the maximum MPP.

[0007] The aforementioned disclosed technical solutions have at least the following technical problems: Traditional variable step size methods (such as dynamically adjusting the step size based on the absolute value of the power error) require a pre-set threshold and can only adjust the step size based on historical errors, failing to perceive environmental change trends in real time (such as the rate and direction of sudden changes in illumination). When illumination decreases rapidly, the error threshold may not trigger a large step size adjustment in time, leading to tracking delay. Simultaneously, the step size is usually adjusted with a fixed gradient, causing frequent switching between "large step size oscillations" and "small step size stagnation" during periods of frequent illumination fluctuations (such as cloudy weather), resulting in additional power fluctuations. When local shading occurs in the photovoltaic array, leading to multiple peak values, the variable step size strategy may fall into a local optimum due to misjudging the direction of the main peak value, and small step sizes are unlikely to escape local traps, while large step sizes may exacerbate oscillations.

[0008] To address the above problems, this invention proposes a solution. Summary of the Invention

[0009] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a tracking configuration control method and system for inverters. By constructing a dual-channel control mechanism based on curvature-sensing differential trend characteristic values, disturbance level mapping, continuously differentiable step size adjustment function, and fusion of global power topology map, the method addresses the problem that the inverter operating point is prone to getting trapped in local optima under multi-peak disturbances or local shadowing conditions and cannot accurately and quickly jump to the global optimal power point.

[0010] To achieve the above objectives, the present invention provides the following technical solution: A tracking configuration control method for an inverter includes the following steps: acquiring real-time output power data of a photovoltaic array, constructing a power time series, performing differential processing on the power time series to generate differential trend feature values; dynamically evaluating the illumination disturbance trend based on the differential trend feature values, constructing a step size adjustment function and outputting the original step size by combining the error between the current power measurement value and the target power value; performing smoothing filtering on the original step size, and determining whether there are multiple extreme points in the current operating environment of the photovoltaic system. If multiple extreme points exist, the direction of the globally optimal power point is identified; generating control adjustment parameters based on the globally optimal power point direction, using the adjustment parameters to correct the smoothed step size and target voltage, generating a fused control quantity, and updating the inverter operating point based on the fused control quantity.

[0011] In a preferred embodiment, the differential processing of the power time series to generate differential trend feature values ​​specifically involves: continuously collecting voltage and current values ​​at the input of the photovoltaic inverter within at least three sampling periods; calculating real-time power values ​​based on the voltage and current values ​​to construct a power time series; combining the voltage values ​​with the corresponding spatiotemporal power time series to form a trajectory curve; performing local quadratic curve fitting on the current operating interval of the trajectory curve; calculating the curvature value at the current operating point based on the fitted quadratic curve, and using the curvature value as the differential trend feature value; identifying the turning trend in power disturbances based on the numerical change of the differential trend feature value; and triggering a switching of the tracking strategy when the absolute value of the curvature corresponding to the turning trend exceeds a preset threshold.

[0012] In a preferred embodiment, the step of dynamically assessing the illumination disturbance trend based on the differential trend feature value, and constructing a step size adjustment function and outputting the original step size by combining the error between the current power measurement value and the target power value, specifically involves: continuously recording curvature values ​​within several sampling periods to construct a differential trend feature value sequence; calculating the moving average rate of change based on the differential trend feature value sequence; mapping the differential trend feature values ​​to a preset set of standardized disturbance levels according to the numerical relationship between the current curvature value and the moving average rate of change; and constructing a continuously differentiable step size adjustment function based on the standardized disturbance level and the current power error, and outputting the original step size.

[0013] In a preferred embodiment, the step of smoothing and filtering the original step size and determining whether there are multiple extreme points in the current operating environment of the photovoltaic system specifically involves: acquiring a continuously collected original step size sequence within a preset time window; smoothing the current original step size using an exponential moving average function to output a smoothed step size; constructing a step size response sequence based on the smoothed step size and calculating the difference sequence of the sequence; calculating a disturbance offset rate index based on the difference sequence, and determining that abnormal step size oscillation behavior exists when the index exceeds a preset offset rate threshold; synchronously calculating the power direction change sequence within the corresponding window and counting the number of positive and negative power direction jumps; and determining that there are multiple extreme points in the current operating environment of the photovoltaic system when both the abnormal step size oscillation behavior determination and the number of power direction jumps exceed a frequency threshold are met.

[0014] In a preferred embodiment, the identification of the global optimal power point direction specifically involves: if it is determined that there are multiple extreme points in the current operating environment of the photovoltaic system, then a path separation mechanism is activated, and a power topology map is constructed using historical power-voltage data to assist in identifying the global optimal power point direction; if it is determined that there are no multiple extreme points in the current operating environment, then the main path tracking strategy is maintained, and the operating point is updated based on the current smoothed step size and power change direction.

[0015] In a preferred embodiment, the auxiliary identification of the global optimal power point direction specifically involves: acquiring historical power values ​​and corresponding voltage sampling data within the target identification period; constructing directed edges based on the power change trends between adjacent nodes, using voltage sampling points as nodes, to form a power topology graph; calculating the power gain gradient and direction consistency score for each path; selecting the optimal path based on the score, and using the direction of this path as the global optimal power point indication direction.

[0016] In a preferred embodiment, the step of generating control adjustment parameters based on the direction of the globally optimal power point specifically involves: extracting the unit direction vector of the path from the current operating point to the globally optimal power point according to the power topology map, and obtaining the current power change trend vector; calculating the deviation angle between the unit direction vector and the power change trend vector, normalizing the angle into a direction probability component, and calculating the direction information entropy term; extracting the average power gain rate and trend consistency score of the optimal path from the power topology map, and fusing them into a path confidence score; standardizing the power change rate within the sliding window, constructing a trend fluctuation probability distribution, and calculating the trend information entropy term; and weightedly fusing the direction information entropy term, the path confidence score, and the trend information entropy term, and generating continuously differentiable control adjustment parameters through a Sigmoid function mapping.

[0017] In a preferred embodiment, the step of correcting the smoothed step size and target voltage using adjustment parameters to generate a fused control quantity, and updating the inverter operating point based on the fused control quantity, specifically involves: constructing a dual-weighted channel fusion model, wherein: the main channel input is the current smoothing step size, and the auxiliary channel input is the target voltage change generated based on the direction of the globally optimal power point; dynamically allocating the output weights of the main channel and the auxiliary channel according to the dynamic amplitude of the control adjustment parameters; performing regularized fusion processing on the weighted dual-channel outputs to generate a fused control function; and correcting the current operating parameters based on the fused control function to update the inverter operating point.

[0018] The technical effects and advantages of the following invention: A tracking configuration control method and system for inverters. 1. This invention constructs differential trend feature values ​​based on power curvature and combines them with the moving average rate of change to classify disturbance trends, effectively identifying abrupt changes and trend inflection points in illumination disturbances. Simultaneously, a power error-driven step-size adjustment mechanism is introduced to achieve continuously differentiable step-size control. By comprehensively judging the step-size response and the power direction jump frequency, it can accurately identify whether the system is in a multi-peak operating environment. When in a multi-peak operating environment, a power topology map is constructed based on historical power-voltage data to accurately assist in identifying the direction of the globally optimal power point, thereby significantly improving steady-state efficiency and tracking robustness under multi-peak interference.

[0019] 2. This invention designs a control parameter with fusion characteristics of directional information entropy, path confidence, and trend information entropy. This factor is mapped to a continuously smooth control weight through a sigmoid function, driving a dual-weight channel fusion control model. The main channel focuses on the stable update of the local step size, while the auxiliary channel guides the operating point to jump towards the globally optimal power direction. Through dynamic weight allocation and regularized fusion strategies, it effectively balances "local tracking stability" and "global optimization jump performance," significantly improving the inverter's ability to identify and escape local power traps, and enhancing energy utilization efficiency. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a tracking configuration control method for an inverter according to the present invention. Figure 2 This is a schematic diagram of the structure of a tracking configuration control system for an inverter according to the present invention. Detailed Implementation

[0021] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0022] Example 1, Figure 1 The present invention provides a tracking configuration control method for an inverter, comprising the following steps: S1, acquire the output power data of the photovoltaic array in real time, construct a power time series, perform differential processing on the power time series, and generate differential trend feature values, specifically: The voltage and current values ​​at the input of the photovoltaic inverter are continuously collected for at least three sampling periods. The real-time power value is calculated based on the voltage and current values ​​to construct a power time series. The voltage value is combined with the corresponding power time series to form a trajectory curve, and a local quadratic curve fitting is performed on the current working interval of the trajectory curve. The curvature value at the current working point is calculated based on the fitted quadratic curve, and the curvature value is used as the differential trend feature value. The turning trend in power disturbance is identified based on the numerical change of the differential trend feature value. When the absolute value of curvature corresponding to the turning trend exceeds a preset threshold, the tracking strategy is switched.

[0023] The tracking strategies include low-response tracking strategies and high-response tracking strategies.

[0024] Low-response tracking strategy: power is stable without any transition (absolute value of curvature ≤ preset threshold, such as when there is stable lighting on a sunny day and no sudden temperature change).

[0025] High-response tracking strategy: When power shows a turning point (the absolute value of curvature exceeds the preset threshold, such as when cloud cover or sudden changes in illumination cause a sudden rise or fall in power).

[0026] The method of identifying the power disturbance transition trend based on the numerical change of the differential trend feature value, and triggering a tracking strategy switch when the absolute value of curvature exceeds a preset threshold, has the following specific purpose: When photovoltaic power is disturbed by factors such as sunlight (e.g., cloud cover, light spot movement), temperature (e.g., sudden rises or falls), and shadows, it can shift from a "stable change" to a "drastic change" (e.g., a sudden drop / rise in power, or a sudden change in the rate of change). The differential trend characteristic value (i.e., the curvature value after fitting a quadratic curve) is essentially the "degree of curvature of the power-voltage trajectory curve": the smaller the absolute value of curvature, the smoother the power change (e.g., under stable sunlight on a sunny day, power changes approximately linearly with voltage); the larger the absolute value of curvature, the more critical the power is at a "turning point" (e.g., when cloud cover suddenly blocks power, the power changes from stable to a rapid decline, and the trajectory curve bends sharply). By monitoring changes in curvature values ​​to identify turning trends, strategy switching can be triggered in advance when power is "about to change drastically," avoiding the "tracking lag" caused by the slow response of traditional fixed strategies (e.g., the power has already decreased but the strategy continues to track at the original rate, missing the maximum power point).

[0027] Meanwhile, photovoltaic power is subject to "instantaneous noise disturbances" (such as millisecond-level voltage fluctuations and sampling errors). These disturbances are not true "trend reversals" (the absolute value of curvature is small and the duration is short). By setting a "preset threshold for the absolute value of curvature" (e.g., calibrated based on the rated power of the photovoltaic array and historical disturbance data, such as setting the threshold to 0.02-0.05), switching can be triggered only when there is a "true trend reversal" (curvature exceeds the threshold), avoiding frequent strategy switching due to "instantaneous noise"—frequent switching increases the inverter's computational energy consumption and may also cause output power fluctuations, affecting grid connection stability.

[0028] The step of performing local quadratic curve fitting on the current working interval of the trajectory curve specifically involves: The voltage-power data of the current sampling point and n sampling points before and after it are acquired to form a local data window. The window length 2n+1 is adaptively adjusted according to a preset rule to balance fitting accuracy and real-time performance. The voltage and power data within the local data window are fitted using a quadratic polynomial based on the least squares method to obtain the fitting function. Weighting coefficients are introduced during the fitting process, giving higher weights to data points closer to the current sampling point in order to reduce the interference of distant points on the fitting results; Calculate the curvature value at the current sampling voltage using the first and second derivative formulas of the fitted function.

[0029] The preset rules are as follows: Calculate the current power change rate The absolute value; (If the rate of change is greater than the preset first threshold), it indicates that the power curve is steep, so choose a smaller n (faster response). (If the rate of change is less than the preset second threshold), it means the curve is flat, so choose a larger n (strong noise resistance).

[0030] The fitting function is specifically:

[0031]

[0032]

[0033]

[0034]

[0035]

[0036] The specific formula for calculating the curvature value is as follows:

[0037]

[0038]

[0039] in, , , These are the fitting coefficients. For power, The curvature value, The first derivative of the fitted function. The second derivative of the fitted function. To construct a matrix, This is the weight matrix. It is a power sequence. For the number of data points, The voltage at the sampling point, Let be the weighting coefficient at position i. These are the preset adjustment parameters.

[0040] S2, dynamically evaluate the illumination disturbance trend based on the differential trend characteristic value, combine the error between the current power measurement value and the target power value, construct the step size adjustment function and output the original step size; The process involves dynamically evaluating the illumination disturbance trend based on the differential trend characteristic value, combining the error between the current power measurement value and the target power value, constructing a step size adjustment function, and outputting the original step size. Specifically: Record the curvature values ​​within several sampling periods, construct a differential trend feature value sequence, and calculate the moving average rate of change based on the differential trend feature value sequence; Trend classification is performed based on curvature and moving average rate of change, and differential trend feature values ​​are mapped to standardized perturbation levels. Based on the standardized disturbance level and the current power error, a continuously differentiable step size adjustment function is constructed, and the original step size is output.

[0041] The trend classification based on curvature and moving average rate of change maps the differential trend feature values ​​to standardized perturbation levels, specifically as follows: A curvature threshold is set based on historical data. If the curvature is greater than the curvature threshold and the moving average rate of change is greater than 0, the disturbance level is high disturbance. If the curvature is greater than the curvature threshold and the moving average rate of change is less than 0, the disturbance level is moderate. If the curvature is less than the curvature threshold, the disturbance level is low disturbance.

[0042] The moving average rate of change is specifically:

[0043] The step size adjustment function is as follows:

[0044]

[0045] in, The moving average rate of change of curvature. The number of sampling periods. To integrate the disturbance factors, , and are weights, respectively. This is the current power error. For the target power, It is a very small positive number, used to prevent the denominator from being 0. The curvature of the power curve at the current sampling point. The original step size, For the minimum step size, This represents the maximum step size.

[0046] S3, perform smoothing filtering on the original step size, and determine whether there are multiple extreme points in the current operating environment of the photovoltaic system. If there are multiple extreme points, identify the direction of the global optimal power point. The process of smoothing and filtering the original step size and determining whether there are multiple extreme points in the current operating environment of the photovoltaic system is as follows: Obtain continuous raw step size values ​​within a preset period, smooth the current raw step size using an exponential moving average function, and output the smoothed step size. ,in The smoothing step size for the k-th period. As a smoothing factor, This represents the voltage adjustment step size within the k-th sampling period; Construct a step-size response sequence based on a smooth step size and calculate the difference sequence. ; A perturbation offset rate index is constructed based on the difference sequence. If the perturbation offset rate is greater than the preset offset rate threshold, then there is abnormal oscillation behavior at the current step size. Simultaneously calculate the power direction change sequence within the corresponding period and determine its positive and negative jump frequencies; When the step size disturbance offset rate exceeds the limit and the power direction changes frequently, the current operating state is determined to be in a multi-extreme environment.

[0047] The offset rate metric is specifically:

[0048] in, The offset rate metric It is a difference sequence. , This represents the number of smoothing steps.

[0049] The identification of the direction of the globally optimal power point is specifically as follows: If the environment is determined to be a multi-polarity environment, the path separation mechanism is activated, and a power topology map is constructed using historical power-voltage data to help identify the direction of the globally optimal power point. If the environment is not determined to be a multi-extreme environment, the main path tracking strategy is maintained, and the operating point is updated based on the current smoothed step size and power change direction without the need to construct a power topology graph.

[0050] If the environment is determined to be a multi-polarity environment, a path separation mechanism is activated. A power topology map is constructed using historical power-voltage data to assist in identifying the direction of the globally optimal power point. Specifically: Acquire historical power and corresponding voltage sampling data within the target identification period; Using the sampled voltage points as nodes in the topology graph, directed edges are constructed based on the power change trends between adjacent nodes to form a power topology graph; Calculate the power gain gradient and directional consistency score for each path, select the optimal path based on the weighted comprehensive score, and determine the direction of the path as the indicator direction of the global optimal power point.

[0051] The power gain gradient is specifically as follows:

[0052] The directional consistency score is specifically as follows:

[0053] in, For the power gain gradient, Let j be the power value corresponding to node j. Let i be the power value corresponding to node i. This represents the number of edges in the path. Scoring is given for directional consistency.

[0054] S4 generates control adjustment parameters based on the global optimal power point direction, uses the adjustment parameters to correct the smoothed step size and target voltage, generates a fused control quantity, and updates the inverter operating point based on the fused control quantity.

[0055] The control adjustment parameters generated based on the globally optimal power point direction are as follows: Based on the path between the current operating point and the global optimal power point in the power topology graph, extract the unit direction vector and obtain the power change trend vector of the current operating point. Calculate the angle of deviation between the current direction vector and the trend vector, and normalize it into a direction probability component to calculate the direction information entropy term; The average power gain rate and trend consistency score of the optimal path are extracted from the power topology map and used as the path confidence score to measure the reliability of the direction as an "adjustment benchmark". The power change rate within the sliding window is standardized and its distribution is constructed. The trend fluctuation probability is defined, and the trend entropy term is calculated. By weighting and fusing the directional information entropy term, path confidence, and trend information entropy term, and then mapping them using the Sigmoid function, a continuously differentiable control adjustment parameter is obtained.

[0056] The process involves adjusting parameters to correct the smoothed step size and target voltage, generating a fused control quantity, and updating the inverter operating point based on this fused control quantity. Specifically: A dual-weighted channel fusion model is constructed based on control and regulation parameters. The current smoothing step size is input into the main channel, while the target voltage change derived from the direction of the global optimal power point is input into the auxiliary channel. Based on the dynamic amplitude changes of the adjustment parameters, the output results of the main channel and the auxiliary channel are proportionally weighted. When the adjustment parameters are small, the main channel is enhanced to maintain local stability, and when the adjustment parameters are large, the auxiliary channel is enhanced to accelerate the global jump. The fusion results of the two channels are weighted using regularization to construct a fusion control function, thus obtaining the fusion control variables; Based on the fusion of control variables, the current step size and target voltage are corrected respectively, and the operating point of the inverter is updated, so that it can escape the local power trap while ensuring dynamic stability.

[0057] The directional information entropy term is specifically:

[0058]

[0059] The control adjustment parameters are specifically as follows:

[0060] in, For probability, For the deviation angle, For directional information entropy, To control and adjust parameters, For path confidence, For the trend entropy term, , , These are the preset empirical adjustment coefficients.

[0061] Example 2, Figure 2 The present invention provides a tracking configuration control system for an inverter, comprising the following modules: Differential trend feature value extraction module: used to acquire the output power data of the photovoltaic array in real time, construct the power time series, perform differential processing on the power time series, and generate differential trend feature values; Step size adjustment function construction module: used to dynamically evaluate the illumination disturbance trend based on the differential trend characteristic value, combine the error between the current power measurement value and the target power value, construct the step size adjustment function and output the original step size; Multi-extreme value discrimination and direction recognition module: used to perform smoothing filtering on the original step size and determine whether there are multiple extreme values ​​in the current operating environment of the photovoltaic system. If multiple extreme values ​​exist, the direction of the global optimal power point is identified. Control Fusion Correction Module: Used to generate control adjustment parameters based on the global optimal power point direction, use the adjustment parameters to correct the smoothed step size and target voltage, generate fused control quantity, and update the inverter operating point according to the fused control quantity.

[0062] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0063] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0064] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0065] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0066] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0067] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A tracking configuration control method for an inverter, characterized in that, Includes the following steps: Real-time acquisition of output power data of photovoltaic array, construction of power time series, differential processing of power time series, and generation of differential trend feature values; The illumination disturbance trend is dynamically evaluated based on the differential trend characteristic value. The step size adjustment function is constructed and the original step size is output by combining the error between the current power measurement value and the target power value. The original step size is smoothed and filtered, and it is determined whether there are multiple extreme points in the current operating environment of the photovoltaic system. If multiple extreme points exist, the direction of the global optimal power point is identified. Control parameters are generated based on the global optimal power point direction. The smoothed step size and target voltage are corrected using the adjustment parameters to generate a fused control quantity. The inverter operating point is then updated based on the fused control quantity.

2. The inverter tracking configuration control method according to claim 1, characterized in that, The step of performing differential processing on the power time series to generate differential trend feature values ​​specifically involves: The voltage and current values ​​at the input of the photovoltaic inverter are continuously collected for at least three sampling periods. The real-time power value is calculated based on the voltage and current values ​​to construct a power time series. The voltage value is combined with the corresponding power time series to form a trajectory curve, and a local quadratic curve fitting is performed on the current working interval of the trajectory curve. The curvature value at the current working point is calculated based on the fitted quadratic curve, and the curvature value is used as the differential trend feature value. The turning trend in power disturbance is identified based on the numerical change of the differential trend feature value. When the absolute value of curvature corresponding to the turning trend exceeds a preset threshold, the tracking strategy is switched.

3. The inverter tracking configuration control method according to claim 2, characterized in that, The process involves dynamically evaluating the illumination disturbance trend based on the differential trend characteristic value, combining the error between the current power measurement value and the target power value, constructing a step size adjustment function, and outputting the original step size. Specifically: By continuously recording the curvature values ​​within several sampling periods, a sequence of differential trend feature values ​​is constructed. Calculate the moving average rate of change based on the differential trend eigenvalue sequence; Based on the numerical relationship between the current curvature value and the moving average rate of change, the differential trend characteristic value is mapped to a preset set of standardized disturbance levels; Based on the standardized disturbance level, a continuously differentiable step size adjustment function is constructed in conjunction with the current power error, and the original step size is output.

4. The inverter tracking configuration control method according to claim 3, characterized in that, The process of smoothing and filtering the original step size and determining whether there are multiple extreme points in the current operating environment of the photovoltaic system is as follows: Obtain the raw step size sequence continuously collected within a preset time window; The current original step size is smoothed using an exponential moving average function, and the smoothed step size is output. Construct a step-size response sequence based on a smooth step size, and calculate the difference sequence of the sequence; The perturbation offset rate index is calculated based on the difference sequence. When the index exceeds the preset offset rate threshold, it is determined that there is abnormal step size oscillation behavior. Simultaneously calculate the power direction change sequence within the corresponding window and count the number of positive and negative power direction jumps; When the abnormal oscillation behavior of the step size is met and the number of power direction jumps exceeds the frequency threshold, it is determined that there are multiple extreme points in the current operating environment of the photovoltaic system.

5. The inverter tracking configuration control method according to claim 4, characterized in that, The identification of the direction of the globally optimal power point is specifically as follows: If it is determined that there are multiple extreme points in the current operating environment of the photovoltaic system, the path separation mechanism is activated to construct a power topology map through historical power-voltage data to help identify the direction of the globally optimal power point. If it is determined that there are no multiple extreme points in the current operating environment, the main path tracking strategy is maintained, and the operating point is updated based on the current smoothed step size and power change direction.

6. The inverter tracking configuration control method according to claim 5, characterized in that, The auxiliary identification of the global optimal power point direction is specifically as follows: Acquire historical power values ​​and corresponding voltage sampling data within the target identification period; Using voltage sampling points as nodes, directed edges are constructed based on the power change trends between adjacent nodes to form a power topology graph; Calculate the power gain gradient and directional consistency score for each path; The optimal path is selected based on the weighted comprehensive score, and the direction of this path is used as the indicator direction of the global optimal power point.

7. The inverter tracking configuration control method according to claim 6, characterized in that, The control adjustment parameters generated based on the globally optimal power point direction are as follows: Based on the power topology map, extract the unit direction vector of the path from the current operating point to the global optimal power point, and obtain the current power change trend vector; Calculate the angle of deviation between the unit direction vector and the power change trend vector, normalize the angle into direction probability components, and calculate the direction information entropy term; The average power gain rate and trend consistency score of the optimal path are extracted from the power topology map and fused into a path confidence score. The power change rate within the sliding window is standardized to construct a trend fluctuation probability distribution and calculate the trend information entropy term. The directional information entropy term, path confidence, and trend information entropy term are weighted and fused, and then mapped by the Sigmoid function to generate continuously differentiable control parameters.

8. The inverter tracking configuration control method according to claim 7, characterized in that, The process involves adjusting parameters to correct the smoothed step size and target voltage, generating a fused control quantity, and updating the inverter operating point based on this fused control quantity. Specifically: A dual-weighted channel fusion model is constructed, wherein the main channel input is the current smoothing step size, and the auxiliary channel input is the target voltage change generated based on the direction of the globally optimal power point. The output weights of the main channel and the auxiliary channel are dynamically allocated based on the dynamic amplitude of the control adjustment parameters. The weighted dual-channel outputs are then fused using regularization to generate a fusion control function. The inverter operating point is updated by correcting the current operating parameters based on the fusion control function.

9. A system using the tracking configuration control method for an inverter as described in any one of claims 1-8, characterized in that, Includes the following modules: Differential trend feature value extraction module: used to acquire the output power data of the photovoltaic array in real time, construct the power time series, perform differential processing on the power time series, and generate differential trend feature values; Step size adjustment function construction module: used to dynamically evaluate the illumination disturbance trend based on the differential trend characteristic value, combine the error between the current power measurement value and the target power value, construct the step size adjustment function and output the original step size; Multi-extreme value discrimination and direction recognition module: used to perform smoothing filtering on the original step size and determine whether there are multiple extreme values ​​in the current operating environment of the photovoltaic system. If multiple extreme values ​​exist, the direction of the global optimal power point is identified. Control Fusion Correction Module: Used to generate control adjustment parameters based on the global optimal power point direction, use the adjustment parameters to correct the smoothed step size and target voltage, generate fused control quantity, and update the inverter operating point according to the fused control quantity.

Images