MPPT photovoltaic power optimization method and system based on photovoltaic module

Through deep learning models and multiple control strategies, accurate tracking of the maximum power point of the photovoltaic system is achieved, solving the problems of slow response speed and low tracking accuracy in complex environments, improving the energy efficiency of the photovoltaic system and reducing the Power loss.

CN118939073APending Publication Date: 2024-11-12HUANENG FUXIN WIND POWER GENERATION CO LTD

Patent Information

Application Number
CN202411434902.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-15
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing MPPT methods have slow response speed and low tracking accuracy in complex environments. Especially when light changes rapidly or components are partially blocked, it is difficult to accurately track the maximum power point, and the fixed step size adjustment leads to power fluctuations and efficiency losses.

Method used

Deep learning models (such as LSTM or Transformer) combined with multiple control strategies are adopted to collect and analyze the environmental data of the photovoltaic system in real time, predict future power changes, and dynamic step size adjustment and disturbance observation to optimize the installation layout of photovoltaic modules. To achieve accurate and efficient tracking of the maximum power point.

Benefits of technology

It significantly improves the tracking efficiency of the maximum power point, reduces power loss, improves the energy efficiency of the photovoltaic system in complex environments, and avoids the power oscillation problem caused by fixed step lengths.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939073A_ABST
    Figure CN118939073A_ABST
Patent Text Reader

Abstract

The invention discloses an MPPT photovoltaic power optimization method and system based on a photovoltaic module, and relates to the technical field of photovoltaic power optimization, and the method comprises the steps: collecting photovoltaic signals, and carrying out the analysis of the collected signals; based on a deep learning algorithm, predicting future illumination and temperature change trends through learning historical data and real-time environment data, and generating a corresponding power point prediction model; processing the data locally and implementing control by using edge computing equipment, and adjusting the output power of the photovoltaic module in real time according to the generated MPPT control strategy; and optimizing the installation layout of the photovoltaic module through a reinforcement learning algorithm. According to the method, the deep learning model is introduced to be combined with the dynamic step length adjustment and perturbation observation strategy, so that the photovoltaic system can accurately and efficiently track the maximum power point in a complex environment, and the stability and the response speed of power output are effectively improved. According to the method, power fluctuation and loss are reduced, and the overall energy efficiency of the photovoltaic power generation system is remarkably improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power optimization, and in particular to an MPPT photovoltaic power optimization method and system based on photovoltaic modules. Background Art

[0002] With the growing global demand for renewable energy, photovoltaic power generation systems have become an important part of modern power systems due to their clean, efficient and sustainable characteristics. However, the output power of photovoltaic power generation systems is affected by a variety of environmental factors, including changes in light intensity, temperature, component orientation, shadows, etc., which causes its maximum power point (MPP) to be in a state of dynamic fluctuation. In order to maximize the power output of photovoltaic systems under various complex conditions, maximum power point tracking (MPPT) technology has emerged. Existing MPPT methods, such as perturbation and observation (P&O) and incremental conductance methods, can track the maximum power point well in some cases, but their response speed and tracking accuracy are still insufficient in complex environments, especially when the light changes rapidly or the components are partially blocked, which can easily lead to a decrease in tracking efficiency or even failure. In addition, the step size adjustment of traditional methods is relatively fixed, resulting in large oscillations or efficiency losses during power tracking. To solve these problems, combining deep learning with intelligent control technology to further improve the MPPT efficiency of photovoltaic systems has become the focus of current research. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides an MPPT photovoltaic power optimization method based on photovoltaic modules, which mainly solves the problems of slow response speed, low tracking accuracy, and power fluctuation of traditional MPPT algorithms in existing photovoltaic systems under complex and changeable environmental conditions. When the existing MPPT method encounters rapidly changing lighting conditions or partial shading of components, its fixed step adjustment method may cause tracking to be too slow or generate large power losses, and it is impossible to achieve accurate and real-time tracking of the maximum power point. In addition, when processing multi-dimensional environmental data, traditional methods lack an effective prediction mechanism and only rely on currently collected data for adjustment, which makes it difficult to cope with power fluctuations in complex environments. The present invention introduces a deep learning model (such as LSTM or Transformer), learns the historical operation data and real-time environmental data of the photovoltaic system, predicts future power change trends, and combines multiple control strategies such as dynamic step adjustment and disturbance observation to achieve accurate and efficient tracking of the maximum power point, thereby effectively solving the problem of poor performance of traditional MPPT algorithms in complex environments.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions, based on the MPPT photovoltaic power optimization method of photovoltaic modules, including: The input and output signals of the photovoltaic components are collected in real time, and the signals are collected respectively by different sensors; the collected signals are transmitted to the data processing module, and the data processing module uses a neural network algorithm to analyze the collected data, evaluate the current state of the photovoltaic components, and determine the output power curve and maximum power point of the photovoltaic components; based on the deep learning model, by learning historical data and real-time environmental data, the future light and temperature change trends are predicted, and a corresponding power point prediction model is generated for dynamically adjusting the maximum power point tracking MPPT control strategy; using edge computing devices, the data is processed locally and control is implemented, and the output power of the photovoltaic components is adjusted in real time according to the generated MPPT control strategy to ensure that the power output is always close to the maximum power point; through the reinforcement learning algorithm, the installation layout of the photovoltaic components is optimized, including the installation density, orientation and tilt angle of the components, so as to achieve the optimization of the power output of the photovoltaic system.

[0006] As a preferred solution of the MPPT photovoltaic power optimization method based on photovoltaic modules described in the present invention, the input and output signals include current, voltage, light intensity, temperature, and wind speed signals.

[0007] As a preferred solution of the MPPT photovoltaic power optimization method based on photovoltaic modules described in the present invention, the real-time acquisition of the input and output signals includes multiple current sensors, voltage sensors, light intensity sensors, temperature sensors and wind speed sensors on the photovoltaic modules, each sensor collects corresponding data of the environment and the modules respectively, and the data is transmitted to the central processing unit via a wireless or wired network.

[0008] As a preferred solution of the MPPT photovoltaic power optimization method based on photovoltaic modules described in the present invention, the use of a neural network algorithm to analyze the collected data includes learning the historical operating data and real-time environmental data of the photovoltaic system, capturing the changing pattern of the power output of the photovoltaic modules under different environmental conditions, and predicting future power fluctuations through time series analysis; the prediction model is trained based on light intensity, temperature, and the voltage and current parameters currently output by the module.

[0009] As a preferred solution of the MPPT photovoltaic power optimization method based on photovoltaic modules described in the present invention, wherein: the use of a neural network algorithm to analyze the collected data also includes assuming that the input multi-sensor signal is , define the input signal matrix ,in is the number of sensors, is the number of sampling points in the time window, each sensor The signal is represented as .

[0010] CNN is used to extract features from input signals and define convolution kernels. As the layer convolution kernel, where is the width of the convolution kernel, is the height of the convolution kernel, and the convolution operation is expressed as: , in, For the The feature map of the layer, is the activation function, is the bias term, Represents the convolution operation. Through the multi-layer convolution and pooling operations of CNN, the features in the signal are extracted to generate a high-dimensional feature vector ,in is the feature dimension; Extract feature vectors Finally, the correlation between different sensor signals is analyzed using the Bayesian estimation method, assuming that the characteristic distribution of each sensor is a conditional probability distribution , the Bayesian formula is expressed as: , in, and They are from and The feature vector of each sensor is is the conditional probability, which means that In the case The probability of occurrence, for The probability of for probability.

[0011] In order to enhance the effect of Bayesian estimation, prior information is introduced , integrating the signal correlations of different sensors into a weighted model : , in, is the weight coefficient, which is used to adjust the balance between prior and observed data; Represents the prior distribution, which represents the prior assumption on the feature distribution of sensor i.

[0012] In order to filter out noise, the system performs the feature vector Perform Kalman filtering, the state equation of the Kalman filter and the observation equation They are: , , in, is the state transfer matrix, is the observation matrix, is the process noise, To observe the noise, the update equation of the Kalman filter is: Prediction Steps: , Update steps: , in, is the Kalman gain, is the process noise covariance matrix, is the observation noise covariance matrix; is the Kalman filter The covariance matrix of the eigenvectors at each moment; is the updated feature vector, combining the correction of the prediction and observation data; is the Kalman filter The predicted value of the feature vector at the moment; is a matrix with diagonal elements of 1 and off-diagonal elements of 0; the random noise in the signal is filtered out by the current filter to obtain a smooth eigenvector , providing input data for subsequent power prediction and MPPT control.

[0013] As a preferred solution of the MPPT photovoltaic power optimization method based on photovoltaic modules described in the present invention, wherein: the generation of the corresponding power point prediction model includes that the MPPT control strategy is based on the prediction results of the deep learning model, and the system can adjust the working voltage and current of the photovoltaic modules in real time so that the photovoltaic modules are always close to or reach the maximum power point under the current environmental conditions; the control strategy includes step adjustment, disturbance observation and dynamic step adjustment; the deep learning model generates the maximum power point tracking adjustment control instructions according to the light intensity, temperature and output characteristics of the photovoltaic modules.

[0014] The reinforcement learning algorithm includes optimizing and adjusting the installation layout of photovoltaic modules through simulation experiments and field tests; the optimization and adjustment includes dynamic adjustment of the module spacing, orientation and tilt angle, continuously testing and updating the layout strategy through the reinforcement learning model, selecting the arrangement with the best power output, and automatically adjusting the module installation plan according to different geographical environments and lighting conditions.

[0015] As a preferred solution of the MPPT photovoltaic power optimization method based on photovoltaic modules described in the present invention, the reinforcement learning algorithm also includes that the optimization process is based on the real-time power output data and environmental parameters of the photovoltaic modules, including light intensity, geographical location, and shadow distribution factors.

[0016] As a preferred solution of the MPPT photovoltaic power optimization system based on photovoltaic components described in the present invention, the system includes a data acquisition module, a data processing module, a central processing unit, and an edge computing module; the data acquisition module is used to collect input and output signals of photovoltaic components in real time, collect the signals respectively through different sensors, and collect component characteristic information; the data processing module includes analyzing the collected data using a neural network algorithm, and implementing the algorithm; the central processing unit is used for synchronously calibrating data from different sensors, and ensuring the timeliness and accuracy of the data through time series analysis; the edge computing module is used to carry an edge computing device including an embedded processor and a local memory, the processor is used to calculate and execute the MPPT control algorithm in real time, and the local memory is used to cache input data from sensors and deep learning models.

[0017] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of an MPPT photovoltaic power optimization method based on photovoltaic modules when executing the computer program.

[0018] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a MPPT photovoltaic power optimization method based on photovoltaic modules.

[0019] Beneficial effects of the present invention: The present invention proposes a MPPT photovoltaic power optimization method based on photovoltaic modules, which has the following beneficial effects. First, by introducing a deep learning model, the system can learn and predict the power output change law of the photovoltaic system under different lighting, temperature, and shadow conditions in real time, and realize more accurate power point prediction. Compared with the traditional MPPT algorithm that only relies on current data for adjustment, deep learning can capture long-term dependencies in complex environments, identify power fluctuation trends in advance, and significantly improve the tracking efficiency of the maximum power point. Secondly, the present invention combines step adjustment, perturbation observation method and dynamic step adjustment strategy. The system can flexibly respond to rapid changes in the environment, especially in the case of rapid changes in light and partial shading. It can adjust the operating voltage and current faster, so that the photovoltaic system is always close to the maximum power output and reduce power loss. In addition, the dynamic step adjustment mechanism avoids the power oscillation problem caused by the fixed step size by automatically adjusting the step size according to the power change amplitude, which not only improves the tracking response speed, but also reduces the energy loss caused by improper adjustment. Finally, the entire MPPT optimization strategy makes full use of multi-dimensional real-time sensor data and historical data, and can maintain efficient operation in complex and changeable photovoltaic environments, greatly improving the overall energy efficiency of the photovoltaic system and having broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0021] Figure 1 A schematic flow chart of a method for optimizing photovoltaic power using MPPT based on photovoltaic modules provided in accordance with an embodiment of the present invention.

[0022] Figure 2 A schematic diagram of an IV curve of a photovoltaic module and an annotation of an MPP definition of a photovoltaic power optimization method based on an MPPT photovoltaic module provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the "one embodiment" or "embodiment" referred to herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. The "in one embodiment" that appears in different places in this specification does not all refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments. The present invention is described in detail in conjunction with schematic diagrams. When describing the embodiments of the present invention in detail, for the sake of convenience, the cross-sectional view showing the device structure will not be partially enlarged according to the general proportion, and the schematic diagram is only an example, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional spatial dimensions of length, width and depth should be included.

[0026] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0027] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0028] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and provides a MPPT photovoltaic power optimization method based on photovoltaic modules, comprising: S1: Real-time collection of input and output signals of photovoltaic modules, and collection of the signals by different sensors.

[0029] S2: The collected signals are transmitted to the data processing module. The data processing module uses a neural network algorithm to analyze the collected data, evaluate the current state of the photovoltaic module, and determine the output power curve and maximum power point of the photovoltaic module.

[0030] S3: Based on the deep learning model, by learning historical data and real-time environmental data, the future light and temperature change trends are predicted, and the corresponding power point prediction model is generated to dynamically adjust the maximum power point tracking MPPT control strategy.

[0031] S4: Utilize edge computing devices to locally process the data and implement control, and adjust the output power of the photovoltaic module in real time according to the generated MPPT control strategy to ensure that the power output is always close to the maximum power point.

[0032] S5: Through reinforcement learning algorithms, the installation layout of photovoltaic modules is optimized, including the installation density, orientation and tilt angle of the modules, to achieve the optimal power output of the photovoltaic system.

[0033] The input and output signals include current, voltage, light intensity, temperature, and wind speed signals.

[0034] The real-time collection of the input and output signals includes multiple current sensors, voltage sensors, light intensity sensors, temperature sensors and wind speed sensors on the photovoltaic components. Each sensor collects corresponding data of the environment and components respectively, and the data is transmitted to the central processing unit via a wireless or wired network.

[0035] The use of a neural network algorithm to analyze the collected data includes learning the historical operating data and real-time environmental data of the photovoltaic system, capturing the changing pattern of the power output of the photovoltaic components under different environmental conditions, and predicting future power fluctuations through time series analysis; The prediction model is trained based on light intensity, temperature, and the voltage and current parameters currently output by the component.

[0036] Assume that the input multi-sensor signal is , define the input signal matrix ,in is the number of sensors, is the number of sampling points in the time window, each sensor The signal is represented as .

[0037] CNN is used to extract features from input signals and define convolution kernels. As the layer convolution kernel, where is the width of the convolution kernel, is the height of the convolution kernel, and the convolution operation can be expressed as: , in, For the The feature map of the layer, is the activation function, is the bias term, Represents the convolution operation. Through the multi-layer convolution and pooling operations of CNN, the features in the signal are extracted to generate a high-dimensional feature vector ,in is the feature dimension.

[0038] Extract feature vectors Finally, the correlation between different sensor signals is analyzed using the Bayesian estimation method, assuming that the characteristic distribution of each sensor is a conditional probability distribution , the Bayesian formula is expressed as: , in, and They are from and The feature vector of each sensor is is the conditional probability, which means that In the case The probability of occurrence, for The probability of for probability.

[0039] In order to enhance the effect of Bayesian estimation, prior information is introduced , integrating the signal correlations of different sensors into a weighted model : , in, is the weight coefficient, which is used to adjust the balance between prior and observed data; Represents the prior distribution, which represents the prior assumption on the feature distribution of sensor i.

[0040] In order to filter out noise, the system performs the feature vector Perform Kalman filtering, the state equation of the Kalman filter and the observation equation They are: , , in, is the state transfer matrix, is the observation matrix, is the process noise, To observe the noise, the update equation of the Kalman filter is: Prediction Steps: , Update steps: , in, is the Kalman gain, is the process noise covariance matrix, is the observation noise covariance matrix; is the Kalman filter The covariance matrix of the eigenvectors at each moment; is the updated feature vector, combining the correction of the prediction and observation data; is the Kalman filter The predicted value of the feature vector at the moment; is a matrix with diagonal elements of 1 and off-diagonal elements of 0; the random noise in the signal is filtered out by the current filter to obtain a smooth eigenvector , providing input data for subsequent power prediction and MPPT control.

[0041] High-quality data by integrating CNN feature extraction, Bayesian estimation correlation analysis and Kalman filtering denoising , power prediction can be performed. Assume that the power output prediction model is , which is represented by a simple linear regression model or LSTM neural network as: , in, is the prediction function, is the prediction error.

[0042] According to the predicted power value, the MPPT algorithm adjusts the working point of the photovoltaic system in real time to make it close to the maximum power point. The adjustment rule of MPPT can be expressed as: , in, is the optimal operating voltage of the photovoltaic module at time t, is the optimal operating voltage of the photovoltaic module at time t, It is the voltage change during the MPPT adjustment process.

[0043] The generation of the corresponding power point prediction model includes that the MPPT control strategy is based on the prediction results of the deep learning model, and the system can adjust the working voltage and current of the photovoltaic components in real time so that the photovoltaic components are always close to or reach the maximum power point under the current environmental conditions; the control strategy includes step regulation, disturbance observation and dynamic step size adjustment.

[0044] The deep learning model generates a maximum power point tracking adjustment control instruction based on light intensity, temperature and output characteristics of the photovoltaic module.

[0045] Predicting future moments using deep learning models Power output . Assume that at time , the actual power output of the system is , calculate the current power error: , in, Maximum power point predicted for deep learning models The corresponding power, Indicates the deviation between the current power and the maximum power point. Through this error feedback, the system is guided to make adjustments.

[0046] Step regulation method adjusts the voltage or current To approach the maximum power point . Assume that the voltage at the current moment is , the voltage at the next moment is adjusted according to the step adjustment method: , in, is the step adjustment amount. It can be adjusted according to the power error To dynamically adjust the step size: , in, is the proportional coefficient of the step adjustment. In this way, the system gradually adjusts the voltage to approach the maximum power point.

[0047] Traditional The algorithm perturbs the voltage And observe the power change to determine the adjustment direction. If the power increases, continue to perturb in that direction, otherwise change the perturbation direction. The improved algorithm introduces the prediction information of deep learning. Adjust step direction: , in, is the perturbation step size, It is a sign function that determines the voltage adjustment direction based on the difference between the predicted power and the current power. Unlike the traditional P&O algorithm, this method uses the predicted power point instead of the actual measured power as a reference.

[0048] The dynamic step size adjustment method dynamically adjusts the step size based on the power error and the voltage / current change trend. Its goal is to adjust the step size when the system is close to the formula: , in, is the adjustment coefficient, step length Dynamic adjustment is made according to the size of the power error. When the error is large, the step size increases; when the error is small, the step size decreases, ensuring the fineness of the adjustment when approaching the maximum power point.

[0049] Combined with the deep learning model, we can use the power gradient information to adjust the system parameters more accurately. Let the gradient of power relative to voltage predicted by the deep learning model be , the system can use the gradient information for dynamic adjustment: , in, is the learning rate, which is used to control the magnitude of the adjustment. Through this gradient-based method, the system can be adjusted directly in the direction of power increase, improving the efficiency of regulation.

[0050] According to the real-time power error Choose between different tuning strategies: When the power error is large, step regulation and dynamic step adjustment methods are used to quickly adjust the voltage to approach the maximum power point.

[0051] When the power error is small, a gradient-based tracking method is used for fine tuning.

[0052] When the power change trend is unclear, the perturbation observation method is used in combination with the prediction results of deep learning to determine the adjustment direction.

[0053] Therefore, the final voltage adjustment formula can be expressed as: , in, The value of is determined by the current power error, gradient, and perturbation observation results: , in, and is the error threshold, which is used to distinguish different adjustment strategies.

[0054] The reinforcement learning algorithm includes optimizing and adjusting the installation layout of photovoltaic modules through simulation experiments and field tests; the optimization and adjustment includes dynamic adjustment of the module spacing, orientation and tilt angle, continuously testing and updating the layout strategy through the reinforcement learning model, selecting the arrangement with the best power output, and automatically adjusting the module installation plan according to different geographical environments and lighting conditions.

[0055] Through multiple iterations and feedback, the system gradually optimizes the installation density and arrangement of components to ensure maximum power output in different environments while taking into account the shading problems between photovoltaic components and land utilization.

[0056] The reinforcement learning algorithm also includes that the optimization process is based on the real-time power output data of the photovoltaic components and environmental parameters, including light intensity, geographical location, and shadow distribution factors.

[0057] The optimized installation layout is determined through multiple iterative calculations of the reinforcement learning algorithm. The system will adjust the layout strategy in real time according to the geographical environment, component characteristics and meteorological data, and verify its effectiveness through simulation and field testing to ensure the optimal power output efficiency of photovoltaic components in complex geographical environments.

[0058] Example 2, reference Figure 2 , which is an embodiment of the present invention, provides an MPPT photovoltaic power optimization method based on photovoltaic modules. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0059] The neural network algorithm in the present invention is a long short-term memory network (LSTM) or Transformer structure, which predicts future power fluctuations by learning the historical data and real-time environmental data of the photovoltaic system. In order to accurately track the maximum power point, the model will process the input multi-dimensional time series data such as light intensity, temperature, output voltage and current, and capture the dynamic changes of photovoltaic system power under complex environmental conditions through time series analysis methods.

[0060] Assume that the input multidimensional time series data is ,in Indicates the number of sensors (e.g. light intensity, temperature, voltage, current), Represents the number of time steps. The data for each time step can be represented as a vector: , in, Indicates that from Sensors at time The measured value at time.

[0061] The basic state equation of LSTM is as follows: Forget Gate: , in, is the output of the forget gate, is the weight matrix of the forget gate, is the bias term, is the hidden state of the previous time step, is the input of the current time step, is the activation function (usually the Sigmoid function).

[0062] Input Gate: , in, is the input gate, is the candidate memory cell state, and is the weight matrix, and is the bias term.

[0063] Memory unit update: , in, is the updated memory cell state, is the memory cell state at the previous time step.

[0064] Output Gate: , in, is the output gate, is the hidden state of the current time step. The final output of LSTM Represents the power output prediction for the next time step: , in, is the weight matrix of the output layer, is the bias term.

[0065] Transformer is a neural network structure based on the attention mechanism, which is suitable for processing time series problems with long-term dependencies. In order to capture the dynamic characteristics of power fluctuations under complex environmental conditions, we can apply the Transformer structure to power prediction.

[0066] The core of Transformer lies in the self-attention mechanism, which learns important dependencies by comparing the similarity of each element in the input sequence with other elements. The self-attention mechanism can be expressed as: , in, are query matrix, key matrix and value matrix respectively, is the dimension of the key vector.

[0067] In our application scenario, the input signal After linear transformation, we get : , Transformer's multi-head self-attention mechanism computes multiple attention heads in parallel to capture different dependencies: , Finally, the feed-forward network in the Transformer is applied to the output at each time step: , Transformer output prediction results Represents the power prediction for the next time step: , Whether it is LSTM or Transformer, the predicted power value It is based on the input multi-dimensional time series data The loss function uses the mean square error (MSE) to optimize the parameters of the model: , in, is the actual power output, is the power value predicted by the model.

[0068] Future power prediction based on LSTM or Transformer , the system can dynamically adjust the working point (voltage, current) of the photovoltaic module to track the maximum power point. The MPPT adjustment strategy is: , in, To adjust the step size, Indicates the gradient of power relative to voltage and is used to guide voltage adjustment.

[0069] Furthermore, by real-time tracking of changes in the output voltage and current of the photovoltaic array due to changes in the working environment, relevant control means or technologies can be used to ensure that the photovoltaic array operates at the maximum output power point (MPP) of its I or PV curve. This behavior of automatically adjusting and obtaining maximum power output is called Maximum Power Point Tracking (MPPT) control technology in the industry. Figure 2 This is a typical photovoltaic module volt-ampere characteristic and power output characteristic curve. It can be seen that the power curve has a peak point, which is the maximum power point.

[0070] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0071] Embodiment 3, the third embodiment of the present invention, is different from the first two embodiments in that: If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0072] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0073] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0075] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0076] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

[0077] Embodiment 4 is the fourth embodiment of the present invention. This embodiment provides an MPPT photovoltaic power optimization system based on photovoltaic modules, including: the system includes a data acquisition module, a data processing module, a central processing unit, and an edge computing module; The data acquisition module is used to collect input and output signals of photovoltaic components in real time, collect the signals through different sensors, and collect component characteristic information; The data processing module includes analyzing the collected data using a neural network algorithm and implementing the algorithm. At the same time, the data processing module also receives system control instructions, which contain voltage and current adjustment values. Signal transmission is achieved through a digital signal processor DSP or a microcontroller. The edge computing device locally parses and executes these control instructions to ensure the system's real-time responsiveness and power optimization effects.

[0078] The central processing unit is used to synchronize and calibrate the data of different sensors and ensure the timeliness and accuracy of the data through time series analysis; The edge computing module is used to carry edge computing devices including embedded processors and local memory. The processor is used to calculate and execute MPPT control algorithms in real time, and the local memory is used to cache input data from sensors and deep learning models. The edge computing device exchanges data with the central control system through a low-latency communication network, but in the case of network delays or disconnections, the edge device can operate independently and maintain real-time control of the power output of the photovoltaic module.

Claims

1. The MPPT photovoltaic power optimization method based on photovoltaic modules is characterized by: include, The input and output signals of the photovoltaic modules are collected in real time, and the collected signals are transmitted to the data processing module. The data processing module uses the neural network algorithm to analyze the collected data and evaluate the current status of the photovoltaic modules; Based on the deep learning model, by learning historical data and real-time environmental data, the future light and temperature change trends are predicted, and the corresponding power point prediction model is generated; Using edge computing devices to locally process the data and implement control, and adjust the output power of the photovoltaic module in real time according to the generated MPPT control strategy; Through reinforcement learning algorithms, the installation layout of photovoltaic modules is optimized to achieve the optimal power output of the photovoltaic system; The use of a neural network algorithm to analyze the collected data includes learning the historical operating data and real-time environmental data of the photovoltaic system, capturing the changing pattern of the power output of the photovoltaic components under different environmental conditions, and predicting future power fluctuations through time series analysis; The prediction model is trained based on light intensity, temperature, and the voltage and current parameters currently output by the component; The use of a neural network algorithm to analyze the collected data also includes assuming that the input multi-sensor signal is , define the input signal matrix ,in is the number of sensors, is the number of sampling points in the time window, each sensor The signal is represented as ; CNN is used to extract features from input signals and define convolution kernels. As the layer convolution kernel, where is the width of the convolution kernel, is the height of the convolution kernel, and the convolution operation is expressed as: , in, For the The feature map of the layer, is the activation function, is the bias term, Represents convolution operation; Through CNN's multi-layer convolution and pooling operations, the features in the signal are extracted to generate a high-dimensional feature vector ,in is the feature dimension; Extract feature vectors Finally, the correlation between different sensor signals is analyzed using the Bayesian estimation method, assuming that the characteristic distribution of each sensor is a conditional probability distribution , the Bayesian formula is expressed as: , in, and They are from and The feature vector of each sensor is is the conditional probability, which means that In the case The probability of occurrence, for The probability of for probability; In order to enhance the effect of Bayesian estimation, prior information is introduced , integrating the signal correlations of different sensors into a weighted model : , in, is the weight coefficient, which is used to adjust the balance between prior and observed data; represents the prior distribution, which represents the prior assumption on the feature distribution of sensor i; In order to filter out noise, the system performs the feature vector Perform Kalman filtering, the state equation of the Kalman filter and the observation equation They are: , , in, is the state transfer matrix, is the observation matrix, is the process noise, To observe the noise, the update equation of the Kalman filter is: Prediction Steps: , Update steps: , in, is the Kalman gain, is the process noise covariance matrix, is the observation noise covariance matrix; is the Kalman filter The covariance matrix of the moment eigenvectors; is the updated feature vector, combining the correction of the prediction and observation data; is the Kalman filter The predicted value of the feature vector at the moment; is a matrix with diagonal elements of 1 and off-diagonal elements of 0; the random noise in the signal is filtered out by the current filter to obtain a smooth eigenvector , providing input data for subsequent power prediction and MPPT control.

2. The MPPT photovoltaic power optimization method based on photovoltaic modules according to claim 1, characterized in that: The input and output signals include current, voltage, light intensity, temperature, and wind speed signals.

3. The MPPT photovoltaic power optimization method based on photovoltaic modules according to claim 2, characterized in that: The real-time collection of the input and output signals includes multiple current sensors, voltage sensors, light intensity sensors, temperature sensors and wind speed sensors on the photovoltaic components. Each sensor collects corresponding data of the environment and components respectively, and the data is transmitted to the central processing unit via a wireless or wired network.

4. The MPPT photovoltaic power optimization method based on photovoltaic modules according to claim 3, characterized in that: The generating of the corresponding power point prediction model includes that the MPPT control strategy is based on the prediction result of the deep learning model, and the system can adjust the working voltage and current of the photovoltaic module in real time so that the photovoltaic module always approaches or reaches the maximum power point under the current environmental conditions; the control strategy includes step regulation, disturbance observation and dynamic step adjustment; The deep learning model generates a tracking adjustment control instruction for the maximum power point according to the light intensity, temperature and output characteristics of the photovoltaic module; The reinforcement learning algorithm includes optimizing and adjusting the installation layout of photovoltaic modules through simulation experiments and field tests; the optimization and adjustment includes dynamic adjustment of the module spacing, orientation and tilt angle, continuously testing and updating the layout strategy through the reinforcement learning model, selecting the arrangement with the best power output, and automatically adjusting the module installation plan according to different geographical environments and lighting conditions.

5. The MPPT photovoltaic power optimization method based on photovoltaic modules according to claim 4, characterized in that: The reinforcement learning algorithm also includes that the optimization process is based on the real-time power output data of the photovoltaic components and environmental parameters, including light intensity, geographical location, and shadow distribution factors.

6. A system using the MPPT photovoltaic power optimization method based on photovoltaic modules as described in any one of claims 1 to 5, characterized in that: The system includes a data acquisition module, a data processing module, a central processing unit, and an edge computing module; The data acquisition module is used to collect input and output signals of photovoltaic components in real time, collect the signals through different sensors, and collect component characteristic information; The data processing module includes analyzing the collected data using a neural network algorithm and implementing the algorithm; The central processing unit is used to synchronize and calibrate the data of different sensors and ensure the timeliness and accuracy of the data through time series analysis; The edge computing module is used to carry edge computing devices including an embedded processor and a local memory. The processor is used to calculate and execute the MPPT control algorithm in real time, and the local memory is used to cache input data from sensors and deep learning models.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Citation Information

Patent Citations

  • Underground complex flood evolution prediction method

    CN114626591A

  • Air pre-heater soot blowing strategy adaptive to deep peak regulation mode of thermal power generating unit

    CN115562012A

  • Energy storage battery capacity estimation method based on wavelet packet transformation-deep convolution gated neural network

    CN117233627A

  • Photovoltaic power generation short-term prediction method and system based on neural network

    CN117556958A

  • Intelligent control method and system of photovoltaic inverter

    CN118214078A

Cited By

  • Advanced power control method suitable for marine photovoltaic power generation system

    CN119134535A

  • Energy regulation and control method of photovoltaic system

    CN119209625A

  • Photovoltaic energy storage high-efficiency energy conversion system and method

    CN119382580A

  • Self-adaptive photovoltaic tracking control method and system

    CN119882838A

  • Adaptive photovoltaic tracking control method and system

    CN119882838B