A coordinated power management system for multi-source distributed photovoltaic power generation system

Abstract

The application discloses a kind of for the coordinated power management system of multi-source distributed photovoltaic power generation system, belongs to power grid control field, especially related to the coordinated power management and power grid control technology of distribution network, the system includes: data processing module is used to determine the first and second state parameters of different material photovoltaic module;Power attenuation module is used to determine the first and second power attenuation based on equivalent power generation efficiency attenuation coefficient and state parameter respectively, and accurately predict the total output power of photovoltaic power generation system based on the first and second attenuation power;Control module is used to compare total output power with output power target value, generate power adjustment instruction, to control the output power of energy storage system, realize coordinated power management, therefore, by implementing the application, it can realize the smooth processing of photovoltaic power generation fluctuation and the balance control of power grid output power, to realize the power balance and stable operation of distribution network after multi-source distributed photovoltaic access power grid.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power grid control, and in particular to a coordinated power management system for a multi-source distributed photovoltaic power generation system. BACKGROUND

[0002] In the field of operation and control of alternating current distribution networks, the access of high proportion multi-source distributed photovoltaic power generation systems has brought great challenges to power flow management, voltage stability and dispatching plan of the distribution network. The coordinated power management system of the distribution network relies on accurate prediction of the output power of the distributed photovoltaic power generation system to formulate the charging and discharging strategy of the energy storage system, arrange the switching of reactive power compensation devices and optimize the distribution of network power flow.

[0003] In the prior art, the coordinated power management system of the distribution network in the photovoltaic power generation system, the energy storage system and other power grid systems is mainly based on the macroscopic prediction value of the overall output power of the photovoltaic power generation system to formulate the charging and discharging strategy of the energy storage, but the current power prediction ignores the differentiated performance of photovoltaic components of different materials in the multi-source distributed photovoltaic power generation system under specific environments, especially in dry environments, the problem of dust adsorption caused by static electricity accumulation on the surface of photovoltaic components, which makes the management system unable to timely adjust the judgment of the output power of the photovoltaic power generation system when facing environmental mutations, resulting in that the energy storage charging and discharging plan formulated based on the photovoltaic prediction value does not match the actual demand in the process of power system management, which cannot effectively smooth the power fluctuation injected into the distribution network, makes the line power flow deviate from the expectation, increases the risk of line overload or voltage out-of-limit, and the power grid is unstable. SUMMARY

[0004] The present application provides a coordinated power management system for a multi-source distributed photovoltaic power generation system, which can solve the problem of how to overcome the power prediction deviation caused by the state difference of the distributed photovoltaic power generation system material component, so as to improve the accuracy of the coordinated power management of photovoltaic and energy storage of the distribution network, and to maintain the power flow stability and voltage quality of the distribution network.

[0005] The present application provides a coordinated power management system for a multi-source distributed photovoltaic power generation system, comprising:

[0006] The data processing module is configured to obtain first state data of glass photovoltaic components and second state data of plastic photovoltaic components.

[0007] The power attenuation module is connected with the data processing module, and is used for determining a first equivalent power generation efficiency attenuation coefficient of the glass material photovoltaic module based on the first state data, and determining a first power attenuation amount according to the first equivalent power generation efficiency attenuation coefficient, determining a second equivalent power generation efficiency attenuation coefficient of the plastic material photovoltaic module based on the second state data, and determining a second power attenuation amount according to the second equivalent power generation efficiency attenuation coefficient, and determining an output power prediction value of the photovoltaic power generation system based on the first power attenuation amount, the second power attenuation amount, the first state data and the second state data;

[0008] The control module is connected with the power attenuation module, and is used for generating a charge-discharge control instruction of the energy storage control system according to the difference if the difference between the output power prediction value and the output power target value of the power distribution network is greater than a preset deviation threshold, and controlling the output power of the energy storage control system according to the charge-discharge control instruction, so as to realize the collaborative power management of the power distribution network on the photovoltaic power generation system and the energy storage control system.

[0009] The embodiment of the application realizes the differentiated fine perception of the dust accumulation characteristics of photovoltaic modules of different materials by acquiring the state parameters (such as electrostatic parameters and dust density) of the glass and plastic material photovoltaic modules, and provides a data basis for accurate evaluation. By determining the power attenuation amount of the photovoltaic modules of different materials respectively according to the state parameters thereof, the error of evaluating the photovoltaic array as a homogeneous whole is overcome, the power prediction accuracy of the photovoltaic power generation system in a specific environment (such as a dry environment) is improved, and the system-level prediction is realized by fusing the equivalent power generation efficiency attenuation coefficient, the power attenuation amount and the electrostatic parameter, so as to realize the forward-looking prediction from the microstate of the module to the macro output of the system. The difference between the output power prediction value of the photovoltaic power generation system and the output power target value corresponding to the power distribution network demand power is compared to control the energy storage system, the accuracy of the collaborative power management of the power distribution network on the photovoltaic and energy storage is improved, so as to maintain the power flow stability and voltage quality of the power distribution network.

[0010] Further, the charge-discharge control instruction of the energy storage control system is generated according to the difference, and the output power of the energy storage control system is controlled according to the charge-discharge control instruction, specifically:

[0011] The current state of charge and the maximum charge-discharge power of the energy storage control system are acquired, and if the current state of charge is greater than or equal to a preset safety threshold, and the maximum charge-discharge power is greater than a power adjustment amount corresponding to the difference, a charge-discharge control instruction corresponding to the power adjustment amount is generated;

[0012] If the current state of charge is less than the preset safety threshold, or the maximum charging and discharging power is less than or equal to the power adjustment amount, the maximum power adjustment amount of the energy storage control system is determined, the charging and discharging control instruction corresponding to the maximum power adjustment amount is generated, and the cooperative scheduling instruction is generated based on the second difference value between the power adjustment amount and the maximum power adjustment amount, and the micro-grid or distributed power source within the preset distance range is controlled according to the cooperative scheduling instruction.

[0013] In this way, by generating the charging and discharging control instruction corresponding to the power adjustment amount, accurate power instructions can be executed within the capacity range of the energy storage system, and optimal power support can be achieved. The cooperative scheduling instruction is generated, and the system-level redundant backup mechanism is started when the output power of the energy storage system is insufficient. By cooperating with the micro-grid or adjacent distributed power source, the resilience and reliability of the system in dealing with extreme situations are greatly improved, and the power supply safety of critical loads is ensured.

[0014] Further, the first state data of the glass material photovoltaic module and the second state data of the plastic material photovoltaic module are obtained, wherein the first state data includes a first electrostatic parameter and a first dust density, and the second state data includes a second electrostatic parameter and a second dust density, the first electrostatic parameter and the second electrostatic parameter are used to represent the change of electrostatic adsorption performance, specifically:

[0015] When the environmental humidity data is lower than the preset humidity threshold, a dry environment monitoring instruction is generated and responded to, and the electrostatic field intensity of the glass material photovoltaic module and the plastic material photovoltaic module is measured respectively to obtain a first electrostatic field intensity distribution and a second electrostatic field intensity distribution;

[0016] By analyzing the first electrostatic field intensity distribution, a static uniform area of the glass material photovoltaic module is determined, and a ratio of the electric field intensity of the static uniform area to a preset electric field intensity is calculated to obtain the first electrostatic parameter, wherein the static uniform area refers to an area in which the field intensity value changes continuously and the field intensity of any point deviates from the average field intensity of the area by not more than a first preset threshold;

[0017] By analyzing the second electrostatic field intensity distribution, a static non-uniform area of the plastic material photovoltaic module is determined, and based on the average electric field intensity of the static non-uniform area, the second electrostatic parameter is determined, wherein the static non-uniform area refers to an area in which the field intensity value is suddenly changed and the local point field intensity deviates from the average field intensity of the area by more than a second preset threshold;

[0018] The first dust accumulation thickness of the glass material photovoltaic module per unit area is detected by a laser dust detection sensor, and the first dust density is calculated in combination with the first electrostatic parameter;

[0019] The dust coverage is obtained by processing a photovoltaic surface image of the plastic photovoltaic module through an image recognition technology, and the second dust accumulation thickness and the second electrostatic parameter of the plastic photovoltaic module are measured through a laser triangulation method, and the second dust density is calculated.

[0020] In this way, by triggering the dry environment monitoring instruction when the environmental humidity is lower than the threshold value, trigger-type intelligent monitoring is realized, resource utilization is optimized, and data effectiveness is ensured. By analyzing the electrostatic field intensity distribution and distinguishing uniform and non-uniform areas to extract the electrostatic parameter, accurate quantification of the electrostatic spatial distribution characteristics is realized, and the parameter is more suitable for the physical properties of the material. For the glass module, the dust density is calculated by combining the laser dust detection and the electrostatic parameter, and for the plastic module, the image recognition and laser triangulation are combined to realize accurate and customized measurement of multi-sensor information fusion.

[0021] Further, the first equivalent power generation efficiency decay coefficient of the glass photovoltaic module is determined based on the first state data, specifically:

[0022] Based on the first dust density, the environmental particulate matter concentration is obtained by inversion through a settling model;

[0023] The dust adsorption amount growth rate is determined by correlation analysis of the environmental particulate matter concentration and the first electrostatic parameter;

[0024] Based on the dust adsorption amount growth rate and the first dust density, the first equivalent power generation efficiency decay coefficient is calculated through a light transmittance decay function.

[0025] In this way, by inversing the environmental particulate matter concentration based on the dust density, the environmental "reason" is deduced from the "result" on the module surface, providing input for predicting future dust accumulation trends. By correlating the environmental particulate matter concentration with the electrostatic parameter to determine the dust adsorption amount growth rate, quantitative prediction of future dust accumulation speed is realized. By combining the dust adsorption amount growth rate and the current dust density to calculate the first equivalent power generation efficiency decay coefficient, dynamic and nonlinear evaluation of the first equivalent power generation efficiency decay coefficient is realized, making the evaluation result more close to the actual physical process.

[0026] Further, the second equivalent power generation efficiency decay coefficient of the plastic photovoltaic module is determined based on the second state data, and the second power decay amount is determined according to the second equivalent power generation efficiency decay coefficient, specifically:

[0027] The surface image of the plastic photovoltaic module is binarized to obtain a dust area and a clean area, and the dust coverage is obtained according to the ratio of the area of the dust area to the total area;

[0028] query a preset mapping relationship table between dust coverage and average light transmission blockage rate based on the dust coverage, to determine the second equivalent power generation efficiency attenuation coefficient;

[0029] determine an actual output power based on the second equivalent power generation efficiency attenuation coefficient and a preset output power of the plastic material photovoltaic module under current irradiance;

[0030] calculate the actual output power and the preset output power to obtain the second power attenuation amount.

[0031] Thus, the dust coverage is obtained by binarizing the surface image, realizing fast and intuitive evaluation of the dust distribution area on the surface of the plastic module. The dust coverage is converted into the second equivalent power generation efficiency attenuation coefficient by querying the preset mapping relationship table, realizing fast and efficient conversion from spatial coverage to optical blocking effect. The actual output power is determined based on the second equivalent power generation efficiency attenuation coefficient, realizing direct calculation of power attenuation based on optical principles, with clear and accurate logic.

[0032] Further, the output power prediction value of the photovoltaic power generation system is determined based on the first power attenuation amount, the second power attenuation amount, the first state data, and the second state data, specifically:

[0033] determine a first capacity proportion of the glass material photovoltaic module in the photovoltaic power generation system and a second capacity proportion of the plastic material photovoltaic module, respectively;

[0034] weight the first electrostatic parameter according to the first capacity proportion to obtain a first electrostatic weighting result, and weight the second electrostatic parameter according to the second capacity proportion to obtain a second electrostatic weighting result, and combine the first electrostatic weighting result and the second electrostatic weighting result to obtain a comprehensive electrostatic parameter of the photovoltaic power generation system;

[0035] weight the first power attenuation amount according to the first capacity proportion to obtain a first power weighting result, and weight the second power attenuation amount according to the second capacity proportion to obtain a second power weighting result, and combine the first power weighting result and the second power weighting result to obtain a current power attenuation amount;

[0036] obtain environmental humidity prediction data, and input the current power attenuation amount, the comprehensive electrostatic parameter, the environmental humidity prediction data, and an initial output power prediction value of the photovoltaic power generation system into a time series prediction model to obtain the output power prediction value.

[0037] In this way, the weight factor is introduced through the first capacity ratio and the second capacity ratio, and it can be recognized that different component types have different contribution degrees in the system. The comprehensive electrostatic parameter and the current power attenuation are obtained by weighting the electrostatic parameter and the power attenuation according to the capacity ratio, so that the system state integration from the micro to the macro can be realized. Through the weighted average, multiple parameters representing different material components are combined into a single macro parameter that can represent the current state and future trend of the entire photovoltaic power generation system, providing accurate input for system-level power prediction. By obtaining the environmental humidity prediction data and inputting it into the time series prediction model, the response capability of the prediction model to the key environmental driving factor is enhanced, and the prediction result is more forward-looking.

[0038] Further, the time series prediction model is trained based on a historical data set, wherein the historical data set includes historical environmental humidity data, historical comprehensive electrostatic parameters, historical power attenuation, and corresponding historical power training data, specifically:

[0039] The historical data set is divided into a training set and a validation set according to the time sequence;

[0040] An initial time series prediction model is constructed based on an LSTM neural network, the root mean square error between the model prediction value and the historical power training data is minimized as a loss function, the training set is used to train the initial time series prediction model, and the model hyperparameters of the initial time series prediction model are adjusted through the validation set, to obtain the time series prediction model.

[0041] In this way, the time series prediction model is constructed based on the LSTM neural network, which can take advantage of the LSTM model's ability to handle long-term dependencies in time series, effectively learn the complex nonlinear relationship between power attenuation, environmental changes and power generation, and capture its dynamic evolution law, thereby significantly improving the prediction accuracy. Minimizing the root mean square error as a loss function can ensure that the training target is to directly optimize the deviation between the prediction value and the true value, so that the model converges in the most practical direction. By adjusting the model hyperparameters through the validation set, overfitting can be effectively avoided, ensuring that the model still maintains good generalization ability when facing unknown new data, and improving the robustness of the system.

[0042] Further, the first power attenuation is determined according to the first equivalent power generation efficiency attenuation coefficient, specifically:

[0043] According to the first equivalent power generation efficiency attenuation coefficient, a first initial power attenuation is determined through a first mapping relationship, and at the same time, a mass density difference value is calculated according to the first dust density, and the mass density difference value, the first electrostatic parameter and the power attenuation linear coefficient are calculated to obtain a first theoretical power attenuation.

[0044] a third difference value between the first theoretical power attenuation amount and the first initial power attenuation amount is calculated, and if the third difference value is greater than a second preset threshold value, the power attenuation linear coefficient is adjusted according to the first initial power attenuation amount to obtain a target power attenuation linear coefficient;

[0045] the first power attenuation amount is calculated according to the target power attenuation linear coefficient.

[0046] In this way, the first initial power attenuation amount is determined according to the first equivalent power generation efficiency attenuation coefficient, and the first theoretical power attenuation amount is obtained according to the mass density difference value, so that a double verification mechanism can be established. The first equivalent power generation efficiency attenuation coefficient measurement serves as a direct and reliable "baseline truth", and the mass density model serves as a "theoretical prediction" of adjustable parameters. The third difference value is calculated to determine whether to adjust the power attenuation linear coefficient, so that online self-calibration of the model can be realized. When the theoretical prediction deviates significantly from the baseline truth, the system automatically adjusts the model parameters, so that the theoretical model continuously approximates the real physical process. This enables the system to adapt to the environmental characteristics of different locations and different seasons, has the ability of continuous learning and optimization, and maintains high accuracy for a long time.

[0047] Further, the power attenuation linear coefficient is obtained by linear regression analysis on a historical data set, wherein the historical data set includes multiple groups of dust density sequence data, environmental humidity sequence data, and corresponding actual power attenuation sequence data, and specifically comprises:

[0048] The environmental humidity sequence data is divided into multiple environmental humidity intervals, and the dust density sequence data and the actual power attenuation sequence data are divided according to each of the environmental humidity intervals, and multiple linear regression analysis is performed on the divided data, with dust density and electrostatic parameters as independent variables and actual power attenuation value as dependent variable, to obtain a preset power attenuation linear coefficient corresponding to each of the environmental humidity intervals.

[0049] In this way, by dividing the environmental humidity sequence data into multiple intervals, it can be recognized that humidity is a key factor affecting the "dust-power attenuation" relationship, and a "divide and conquer" strategy is adopted. Multiple linear regression analysis is performed on the data in each humidity interval to obtain the corresponding linear coefficient, which establishes a customized evaluation model for different humidity environments. For example, the coefficient in a dry environment (low humidity) can amplify the effect of electrostatic adsorption, and the coefficient in a humid environment (high humidity) can better reflect the effect of gravity sedimentation. The formation of the preset power attenuation linear coefficient lookup table can enable the system to quickly query and call the most suitable parameters in real time during operation, so that the power attenuation evaluation can maintain high accuracy under different weather conditions, greatly enhancing the environmental adaptability of the system.

[0050] Further, the collaborative power management system for the multi-source distributed photovoltaic power generation system further comprises:

[0051] a cleaning module configured to calculate a cleaning benefit according to the first power attenuation amount and the second power attenuation amount, and generate a cleaning scheduling instruction when a fourth difference between the cleaning benefit and a cleaning cost is greater than a third preset threshold value and a time during which the output power prediction value continuously falls below an output power target value is predicted to be greater than a preset time threshold value, wherein the cleaning scheduling instruction is configured to start a cleaning device to clean a surface of the photovoltaic component and determine a cleaning sequence according to the first power attenuation amount and the second power attenuation amount.

[0052] In this way, the cleaning benefit is calculated according to the power attenuation amount, the cleaning scheduling instruction is generated when the fourth difference between the cleaning benefit and the cleaning cost is greater than the threshold value, and the cleaning decision can be changed from time driving or experience driving to data-driven economic decision. Only when the power generation income brought by cleaning is obviously higher than the cost, the system starts cleaning, and the investment return maximization of the operation and maintenance activity is realized. The cleaning sequence is determined according to the first and second power attenuation amounts, and the optimal allocation of cleaning resources can be realized. The component type with more serious power attenuation is preferentially cleaned, the system power generation capacity can be restored fastest with the least cleaning cost, and the operation and maintenance efficiency is improved. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0054] Figure 1 is a structural schematic diagram of a collaborative power management system for a multi-source distributed photovoltaic power generation system provided by an embodiment of the present application;

[0055] Figure 2 is a structural schematic diagram of another collaborative power management system for a multi-source distributed photovoltaic power generation system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of this application; the use of the terms "including," "comprising," or "having" and variations thereof herein is intended to be broad and encompass the terms "consisting of" and "consisting essentially of" and variations thereof. Unless otherwise required by context, singular terms shall include pluralities and vice versa. Unless otherwise required by context, the use herein of the singular is also to be construed as a use of the plural and vice versa.

[0058] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise explicitly specified.

[0059] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily independent or alternative embodiments to each other. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0060] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.

[0061] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two), and similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0062] In the description of the embodiments of the present application, unless otherwise explicitly specified and limited, the technical terms "mounting", "connecting", "connecting", "fixing" and the like should be broadly understood, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanical connection, or it can be electrical connection; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0063] Reference Figure 1To address the issue of power prediction deviations caused by differences in the material and component conditions of distributed photovoltaic (PV) power generation systems in existing technologies, thereby improving the accuracy of power management of PV and energy storage in the distribution network to maintain power flow stability and voltage quality, an embodiment of the present invention provides a collaborative power management system 100 for multi-source distributed PV power generation systems, comprising:

[0064] The data processing module 102 is used to acquire the first electrostatic parameter and the first dust density of the glass photovoltaic module, and the second electrostatic parameter and the second dust density of the plastic photovoltaic module, wherein the first electrostatic parameter and the second electrostatic parameter are used to characterize the change in electrostatic adsorption performance.

[0065] As an example of an embodiment of the present invention, the acquisition of first state data of a glass photovoltaic module and second state data of a plastic photovoltaic module, wherein the first state data includes a first electrostatic parameter and a first dust density, and the second state data includes a second electrostatic parameter and a second dust density, both the first and second electrostatic parameters being used to characterize changes in electrostatic adsorption performance, specifically: when the ambient humidity data is lower than a preset humidity threshold, a dry environment monitoring command is generated and responded to, and the electrostatic field strength of the glass photovoltaic module and the plastic photovoltaic module are measured respectively to obtain a first electrostatic field strength distribution and a second electrostatic field strength distribution; by analyzing the first electrostatic field strength distribution, the electrostatic uniform region of the glass photovoltaic module is determined, and the ratio of the electric field strength of the electrostatic uniform region to the preset electric field strength is calculated to obtain the first electrostatic parameter, wherein the electrostatic uniform region refers to the electric field strength... The region is defined as follows: The electrostatic field strength distribution is analyzed to determine the electrostatic non-uniform region of the plastic photovoltaic module. Based on the average electric field strength of the electrostatic non-uniform region, the second electrostatic parameter is determined. The electrostatic non-uniform region refers to a region where there are sudden changes in field strength and where the deviation between the field strength at a local point and the average field strength of the region exceeds the second preset threshold. A laser dust detection sensor is used to detect the first dust accumulation thickness per unit area of ​​the glass photovoltaic module. Combined with the first electrostatic parameter, the first dust density is calculated. Image recognition technology is used to process the photovoltaic surface image of the plastic photovoltaic module to obtain the dust coverage rate. A laser triangulation method is used to measure the second dust accumulation thickness and the second electrostatic parameter of the plastic photovoltaic module, and the second dust density is calculated.

[0066] In this embodiment, when the environmental humidity sensor monitors that the environmental humidity data is below a preset threshold, the system determines that the current photovoltaic module is in a dry environment and automatically generates a dry environment monitoring instruction. In response to this instruction, the system starts to measure the electrostatic field intensity of the photovoltaic module. For a glass material photovoltaic module, a non-contact electrostatic meter is used to scan and measure along the predetermined grid path on the surface of the module, a set of discrete field intensity values are obtained, and a continuous first electrostatic field intensity distribution is generated by a spatial interpolation algorithm. When analyzing this distribution, the sliding window method is used to calculate the average field intensity and standard deviation of the local area, and those areas where the field intensity values change continuously and the deviation of the field intensity of any measurement point in the window from the average field intensity of the window is not more than a first preset threshold (for example, ±5%) are identified as electrostatic uniform areas. Then, the average electric field intensity of the electrostatic uniform area is calculated and compared with an ideal electric field intensity value preset based on material characteristics, and the ratio of the two is determined as the first electrostatic parameter, which quantifies the deviation of the current electrostatic adsorption performance from the ideal state.

[0067] For a plastic material photovoltaic module, a second electrostatic field intensity distribution is also obtained by scanning with a non-contact electrostatic meter. When analyzing this distribution, the focus is on identifying areas where the field intensity values change abruptly, i.e., areas where the deviation of the local point field intensity from the average field intensity of the surrounding larger area exceeds a second preset threshold (for example, ±15%), such areas are defined as electrostatic non-uniform areas. The average electric field intensity of the electrostatic non-uniform area is calculated and directly used as the basis for calculating the second electrostatic parameter, which reflects the electrostatic adsorption characteristics of the plastic surface due to material unevenness or local charge accumulation.

[0068] At the same time of obtaining the electrostatic parameters, dust density measurement is performed. For a glass material photovoltaic module, a laser dust detection sensor is used, which emits a laser beam and detects the light intensity scattered by the dust on the surface of the module, combines with the known optical model, and calculates the first dust accumulation thickness per unit area. Then, the first dust accumulation thickness obtained is multiplied by the first electrostatic parameter calculated previously, to comprehensively reflect the dust adhesion enhanced by the electrostatic adsorption effect, and finally the first dust density is obtained.

[0069] For plastic material photovoltaic modules, the determination of dust density combines two techniques. First, by collecting photovoltaic surface images through high-definition cameras, image recognition technology is used to conduct grayscale analysis and contour recognition on the images, calculate the percentage of dust coverage area to total area, and obtain the dust coverage rate. At the same time, by using laser triangulation method, the second dust accumulation thickness is accurately measured by emitting laser to the surface of the module and detecting the position offset of the reflected light spot on the sensor. Finally, the second dust accumulation thickness, the calculated dust coverage rate, and the second electrostatic parameter are fused (for example, weighted average or combination based on empirical formula) to obtain a more comprehensive and accurate second dust density, which takes into account the physical coverage thickness, the proportion of coverage area, and the influence of electrostatic adsorption on the dust adhesion state.

[0070] The power attenuation module 104 is connected with the data processing module 102, and is used for determining a first equivalent power generation efficiency attenuation coefficient of the glass material photovoltaic module based on the first state data, and determining a first power attenuation amount according to the first equivalent power generation efficiency attenuation coefficient, determining a second equivalent power generation efficiency attenuation coefficient of the plastic material photovoltaic module based on the second state data, and determining a second power attenuation amount according to the second equivalent power generation efficiency attenuation coefficient, and determining an output power prediction value of the photovoltaic power generation system based on the first power attenuation amount, the second power attenuation amount, the first state data, and the second state data.

[0071] In this embodiment, the glass material surface is smooth and the electrostatic distribution is uniform, and the dust adsorption forms a uniform coverage layer that affects the overall incident light flux, so the first equivalent power generation efficiency attenuation coefficient uses the light transmittance index to represent the attenuation of light transmission efficiency; while the surface characteristics of the plastic material easily lead to uneven electrostatic distribution and local dust accumulation, and its influence on power generation mainly manifests as the shielding of the effective light receiving area, so the area loss is reflected by the dust coverage rate and the average light transmittance block rate. Specifically, for the glass material photovoltaic module, first, a mapping relationship between the first dust density and the environmental particulate matter concentration is established, and the current environmental particulate matter concentration level is derived through the mapping relationship; then the first electrostatic parameter is associated with the environmental particulate matter concentration for correlation analysis, and the enhancement effect of electrostatic adsorption on dust accumulation is quantified, and then the comprehensive first equivalent power generation efficiency attenuation coefficient of the module surface is calculated through the optical attenuation model; finally, based on the linear relationship between the first equivalent power generation efficiency attenuation coefficient and the power output in the photovoltaic power generation characteristics, the first power attenuation amount is determined through the first equivalent power generation efficiency attenuation coefficient value. For the plastic material photovoltaic module, the dust coverage rate is identified and calculated by analyzing the coverage feature information contained in the second dust density; then according to the area loss model, the dust coverage rate is directly mapped to the reduction degree of power generation efficiency, so as to determine the second power attenuation amount.

[0072] The meteorological prediction data of irradiance intensity, ambient temperature and the like in a specific period in the future are acquired, and based on this, an initial output power prediction value of the photovoltaic power generation system in an ideal clean state of the component surface is calculated. Subsequently, based on the capacity proportion of glass and plastic photovoltaic components in the system, the first electrostatic parameter, the second electrostatic parameter, the first power attenuation amount and the second power attenuation amount converted from the first equivalent power generation efficiency attenuation coefficient are fused respectively to obtain a comprehensive electrostatic parameter and a current overall power attenuation amount representing the overall dust adsorption characteristics of the system. Finally, the initial output power prediction value, the overall power attenuation amount and the comprehensive electrostatic parameter are input into a time sequence prediction model in combination with the prediction data of the future environmental humidity change trend. The model quantifies the influence of the comprehensive electrostatic parameter and the humidity change on the future dust accumulation speed or dissipation potential, dynamically adjusts the prediction value of the power attenuation amount, and then deducts the attenuation amount from the initial output power prediction value to calculate the final output power prediction value of the photovoltaic power generation system.

[0073] The control module 106 is connected with the power attenuation module 104, and is used for generating a charge-discharge control instruction of an energy storage control system according to the difference between the output power prediction value and the output power target value of the power distribution network, and controlling the output power of the energy storage control system according to the charge-discharge control instruction, so as to realize the cooperative power management of the photovoltaic power generation system and the energy storage control system by the power distribution network.

[0074] As an example of an embodiment of the present application, the charge-discharge control instruction of the energy storage control system is generated according to the difference, and the output power of the energy storage control system is controlled according to the charge-discharge control instruction, specifically: the current state of charge and the maximum charge-discharge power of the energy storage control system are acquired, if the current state of charge is greater than or equal to a preset safety threshold, and the maximum charge-discharge power is greater than the power adjustment amount corresponding to the difference, a charge-discharge control instruction corresponding to the power adjustment amount is generated; if the current state of charge is less than the preset safety threshold, or the maximum charge-discharge power is less than or equal to the power adjustment amount, the maximum power adjustment amount of the energy storage control system is determined, and a charge-discharge control instruction corresponding to the maximum power adjustment amount is generated, and a cooperative scheduling instruction is generated based on the second difference between the power adjustment amount and the maximum power adjustment amount, and the micro-grid or distributed power source within a preset distance range is controlled according to the cooperative scheduling instruction.

[0075] In the embodiment, when the difference between the output power prediction value and the target value exceeds the preset threshold, the power compensation decision mechanism is started. The mechanism first evaluates the real-time running state and the upper limit of the capacity of the energy storage control system, and generates a preliminary charging and discharging control instruction on this basis. The core principle is to ensure that the instruction is within the safe and feasible operating range of the energy storage system. If the evaluation finds that the energy storage system alone cannot completely compensate for the power difference, the system coordination mechanism is further started, the remaining difference between the maximum compensation power provided by the energy storage system and the total demand power is calculated, and the coordination scheduling instruction for other distributed power sources or microgrids is generated accordingly. Finally, through the execution of this series of instructions, the coordinated control of the output power of the energy storage system and other controllable units is realized, thereby ensuring the stability and reliability of the total output power of the photovoltaic power generation system, and completing the coordinated power management.

[0076] As an example of the embodiment of the present application, as shown in Figure 2 Another coordinated power management system 200 for a multi-source distributed photovoltaic power generation system is provided, which also includes a cleaning module 108 for calculating a cleaning benefit according to the first power attenuation and the second power attenuation. When a fourth difference between the cleaning benefit and a cleaning cost is greater than a third preset threshold, and the time during which the output power prediction value is continuously lower than the output power target value is expected to be greater than a preset time threshold, a cleaning scheduling instruction is generated. The cleaning scheduling instruction is used to start the cleaning device to clean the surface of the photovoltaic component, and to determine the cleaning sequence according to the size of the first power attenuation and the second power attenuation.

[0077] In the present embodiment, the module receives the first power attenuation amount and the second power attenuation amount from the power attenuation module. The calculation of the cleaning benefit is based on the power loss represented by these attenuation amounts, which is converted into economic value. Specifically, the module multiplies the power attenuation amount (in kilowatts) by the local real-time electricity price (yuan / kilowatt-hour) and estimates the total loss of electric energy caused by the continuous attenuation within the next planned cleaning period (for example, 24 hours), thereby calculating the economic benefit that can be recovered if cleaning is performed. At the same time, the module obtains cleaning cost parameters from the system database, including water resource consumption, labor cost, and equipment wear and tear, etc. The difference between the cleaning benefit and the cleaning cost is the net benefit. When the net benefit is greater than a third preset threshold (determined according to a preset return on investment rate) and the output of the power prediction module indicates that the state in which the power prediction value is lower than the target value is expected to last for more than a preset time threshold (for example, continuously for 3 hours), the cleaning module determines to trigger the cleaning operation. When generating the cleaning scheduling instruction, the module determines the cleaning sequence according to the size relationship between the first power attenuation amount and the second power attenuation amount. The logic is to clean the component area with more serious power attenuation first to restore the maximum power generation capacity as quickly as possible. For example, if the first power attenuation amount of the glass material component is significantly higher than the second power attenuation amount of the plastic material component, the cleaning sequence will instruct the cleaning device to clean the glass material component area first. The scheduling instruction is sent to the cleaning device (such as an automatic cleaning robot or a spraying system) to start the cleaning operation on the surface of the photovoltaic component. Through this intelligent decision-making based on economic efficiency and power recovery urgency, the optimal allocation of cleaning resources is achieved.

[0078] As an example of an embodiment of the present application, the first equivalent power generation efficiency attenuation coefficient of the glass material photovoltaic component is determined based on the first state data, specifically: the environmental particulate matter concentration is obtained by inversion through a settling model based on the first dust density; the dust adsorption amount growth rate is determined by correlation analysis of the environmental particulate matter concentration and the first electrostatic parameter; and the first equivalent power generation efficiency attenuation coefficient is calculated through a light transmittance attenuation function based on the dust adsorption amount growth rate and the first dust density.

[0079] In this embodiment, first, the ambient particulate matter concentration is obtained by inversion based on the first dust density through a settling model. This process is achieved by establishing a physical model based on the theory of settling, which takes the accumulated mass of dust collected per unit area on the surface of the photovoltaic module (i.e., the first dust density) as the core input parameter, while integrating real-time meteorological monitoring data, including wind speed, wind direction, and air turbulence intensity. The model is inverted by solving the balance equation between the dust settling flux and the atmospheric particulate matter concentration, where the settling velocity is determined according to the particle size distribution and aerodynamic characteristics of typical dust particles. In specific implementation, an iterative optimization algorithm is used to continuously adjust the assumed value of the ambient particulate matter concentration in the model until the error between the theoretical settling amount calculated by the model and the measured first dust density reaches a minimum, at which time the corresponding concentration value is the inverted ambient particulate matter concentration.

[0080] Next, correlation analysis is performed between the ambient particulate matter concentration and the first electrostatic parameter to determine the dust adsorption growth rate. This step is completed by a pre-calibrated adsorption kinetics model. The model takes the ambient particulate matter concentration and the first electrostatic parameter (representing the enhancement factor of electrostatic adsorption capacity) as independent variables. Its functional relationship is generally that the dust adsorption growth rate is positively correlated with the ambient particulate matter concentration and is proportional to a certain power of the electrostatic parameter (determined by experimental data fitting). By substituting the real-time acquired particulate matter concentration value and the first electrostatic parameter into this model, the dust mass increase rate per unit area per unit time under the current conditions, i.e., the dust adsorption growth rate, can be calculated.

[0081] Finally, based on the calculated dust adsorption growth rate and the measured first dust density, the first equivalent power generation efficiency decay coefficient of the glass material photovoltaic module is calculated through the transmittance decay coefficient. This transmittance decay coefficient is constructed based on the radiation transfer theory of light passing through a turbid medium (dust layer), and its core is to regard the dust layer as a homogeneous medium with absorption and scattering effects on light. The input parameters of the function include the total mass of the current dust layer (which can be represented by the first dust density), the dust adsorption growth rate (which reflects the dynamic trend of dust accumulation and its potential impact on the shape of the particle accumulation), and the optical property parameters of the dust particles determined by laboratory measurement (such as the extinction coefficient). This function directly outputs the real-time first equivalent power generation efficiency decay coefficient value of the module by calculating the attenuation ratio of incident light intensity after passing through the dust layer. Through this series of continuous calculations based on physical models and experimental calibration, an accurate and reproducible conversion from dust density and electrostatic parameters to the first equivalent power generation efficiency decay coefficient is achieved.

[0082] As an example of an embodiment of the present application, the first power attenuation amount is determined according to the first equivalent power generation efficiency attenuation coefficient, specifically: according to the first equivalent power generation efficiency attenuation coefficient, a first initial power attenuation amount is determined through a first mapping relationship, at the same time, a mass density difference value is calculated according to the first dust density, and the mass density difference value, the first electrostatic parameter and a power attenuation linear coefficient are calculated to obtain a first theoretical power attenuation amount; a third difference value between the first theoretical power attenuation amount and the first initial power attenuation amount is calculated, if the third difference value is greater than a second preset threshold value, the power attenuation linear coefficient is adjusted according to the first initial power attenuation amount to obtain a target power attenuation linear coefficient; the first power attenuation amount is calculated according to the target power attenuation linear coefficient.

[0083] In the embodiment, according to the calculated first equivalent power generation efficiency attenuation coefficient, a pre-generated transmittance-power attenuation mapping relationship table is queried, the mapping relationship is established based on the photoelectric conversion principle of the photovoltaic module, and reflects a linear positive correlation between the first equivalent power generation efficiency attenuation coefficient and the output power. A preliminary power attenuation estimate, i.e. the first initial power attenuation, is directly obtained through the mapping relationship. At the same time, a mass density difference value is calculated according to the first dust density. The calculation process is to subtract a standard cleaning state reference dust density value of the module surface calibrated through experiments from the currently measured first dust density value, and the difference value is the mass density difference value, which directly reflects the deviation degree of the current dust adhesion relative to the cleaning state. Then, the mass density difference value, the first electrostatic parameter and a pre-set power attenuation linear coefficient are multiplied to obtain a first theoretical power attenuation. The power attenuation linear coefficient is a proportional constant obtained by regression analysis of a large amount of historical data, and its physical meaning is the power attenuation caused by unit mass density difference value under unit electrostatic adsorption capacity. Next, a third difference value between the first theoretical power attenuation and the first initial power attenuation is calculated. If the absolute value of the third difference value is greater than a second pre-set threshold, it indicates that there is a significant deviation between the theoretical calculation based on the physical model and the initial estimate based on the direct mapping of photoelectricity. At this time, the system starts a coefficient adjustment process: taking the first initial power attenuation as a reference benchmark closer to the actual working condition, according to the size and direction of the third difference value, the power attenuation linear coefficient is adjusted in a specific proportion in the opposite direction, so that the theoretical value calculated by the power attenuation linear coefficient can converge to the initial estimate value. After the adjustment, a new and more accurate target power attenuation linear coefficient is obtained. Finally, the target power attenuation linear coefficient is used to re-execute the previous calculation process, i.e. multiplying it by the mass density difference value and the first electrostatic parameter to calculate the final confirmed first power attenuation. This method of mutual verification between initial estimate and theoretical calculation and dynamic correction of key coefficients effectively improves the accuracy and reliability of the power attenuation evaluation result.

[0084] As an example of an embodiment of the application, the power attenuation linear coefficient is obtained by linear regression analysis of a historical data set, wherein the historical data set includes multiple groups of dust density sequence data, environmental humidity sequence data, and corresponding actual power attenuation sequence data. Specifically, the environmental humidity sequence data is divided into multiple environmental humidity intervals, and the dust density sequence data and the actual power attenuation sequence data are divided according to each environmental humidity interval. Multivariate linear regression analysis is performed on the divided data, with dust density and electrostatic parameter as independent variables and actual power attenuation value as dependent variable, to obtain a pre-set power attenuation linear coefficient corresponding to each environmental humidity interval.

[0085] In this embodiment, first, the system extracts a historical dataset from the long-term operation database, which contains multiple sets of sequence data aligned by timestamps, including the dust density sequence periodically collected by the dust sensor, the environmental humidity sequence collected by the environmental sensor, and the actual power attenuation sequence recorded by the power monitoring unit within the corresponding time period.

[0086] Subsequently, the numerical range of the entire environmental humidity sequence data is divided into several consecutive intervals, such as a dry interval with humidity below 30%, a moderate humidity interval between 30% and 60%, and a high humidity interval above 60%. The basis for division is that humidity has a phased impact on the physical properties of static electricity generation and dissipation.

[0087] Then, according to this humidity interval division, the original dust density sequence data and actual power attenuation sequence data are also divided correspondingly, forming multiple data subsets associated with specific humidity intervals. For example, all dust density data and actual power attenuation data collected under the condition of humidity below 30% are classified into the dry interval subset.

[0088] Next, multiple linear regression analysis is performed on each data subset corresponding to a humidity interval. In the regression model, the dust density value and the static electricity parameter value at the same time are taken as two independent variables, and the corresponding actual power attenuation value is taken as the dependent variable. The goal of regression analysis is to fit a linear equation of the form actual power attenuation = k1 * dust density + k2 * static electricity parameter + b.

[0089] Through regression algorithms such as least squares, the corresponding regression coefficients k1 and k2 for each humidity interval subset can be calculated. Among them, k1 is the basic influence coefficient of dust density on power attenuation within a specific humidity interval, while k2 reflects the additional influence of static electricity parameters. The combination of these two coefficients, or a comprehensive coefficient integrated according to the system model, is defined as the preset power attenuation linear coefficient corresponding to the humidity interval.

[0090] Finally, the system stores a series of preset power attenuation linear coefficients determined according to different humidity intervals in the coefficient database. In actual application, the system automatically selects the preset power attenuation linear coefficient corresponding to the humidity interval according to the real-time monitored environmental humidity value for calculation, thereby ensuring the environmental adaptability and accuracy of the power attenuation evaluation model.

[0091] As an example of an embodiment of the present application, the second equivalent power generation efficiency decay coefficient of the plastic photovoltaic module is determined based on the second state data, and a second power decay amount is determined according to the second equivalent power generation efficiency decay coefficient. Specifically, the surface image of the plastic photovoltaic module is binarized to obtain a dust area and a clean area, and the dust coverage rate is obtained according to the ratio of the area of the dust area to the total area. Based on the dust coverage rate, a preset mapping relationship table between the dust coverage rate and the average light transmission blocking rate is queried to determine the current second equivalent power generation efficiency decay coefficient of the plastic photovoltaic module. Based on the current second equivalent power generation efficiency decay coefficient and the preset output power of the plastic photovoltaic module under the current irradiance, the actual output power is determined. The actual output power and the preset output power are calculated to obtain the second power decay amount.

[0092] In this embodiment, first, a high-resolution industrial camera installed above the photovoltaic array is used to collect the surface image of the plastic photovoltaic module, and an adaptive threshold algorithm is used to binarize the image. This algorithm dynamically determines the segmentation threshold value according to the gray scale distribution characteristics of the local area of the image, and determines the pixel points with a gray scale value higher than the threshold value as the dust coverage area, and the pixel points with a gray scale value lower than the threshold value as the clean area. By counting the ratio of the total number of pixel points in the dust coverage area to the total number of pixel points on the module surface, the dust coverage rate is accurately calculated.

[0093] Subsequently, based on the calculated dust coverage rate, a mapping relationship table established in advance through experiments is queried. This mapping table is constructed by measuring the first equivalent power generation efficiency decay coefficient of the plastic photovoltaic module under different dust coverage rates in the laboratory, and calculating the light blocking degree relative to the clean state. The system finds the second equivalent power generation efficiency decay coefficient value corresponding to the current dust coverage rate in the mapping table through linear interpolation.

[0094] Then, combined with the current real-time solar irradiance monitoring data, the preset output power value of the plastic photovoltaic module under the standard clean state is obtained. The current second equivalent power generation efficiency decay coefficient is used as a power decay factor to calculate the actual output power of the module through a photoelectric conversion model. This model takes into account the attenuation effect of light transmission blocking on incident light intensity, as well as the current-voltage characteristic changes of the photovoltaic cell itself.

[0095] Finally, the calculated actual output power is compared with the preset output power under the clean state, and the difference is the second power decay amount. This calculation process fully considers the uneven distribution of dust on the surface of the plastic photovoltaic module, and realizes the accurate quantification of the power loss caused by local shading through the combination of image recognition technology and optical characteristic mapping.

[0096] As an example of an embodiment of the present application, the output power prediction value of the photovoltaic power generation system is determined based on the first power attenuation amount, the second power attenuation amount, the first state data and the second state data, specifically: the first capacity proportion of the glass material photovoltaic module in the photovoltaic power generation system and the second capacity proportion of the plastic material photovoltaic module are determined respectively; the first static electricity parameter is weighted according to the first capacity proportion to obtain a first static electricity weighted result, and the second static electricity parameter is weighted according to the second capacity proportion to obtain a second static electricity weighted result, and the first static electricity weighted result and the second static electricity weighted result are combined to obtain a comprehensive static electricity parameter of the photovoltaic power generation system; the first power attenuation amount is weighted according to the first capacity proportion to obtain a first power weighted result, and the second power attenuation amount is weighted according to the second capacity proportion to obtain a second power weighted result, and the first power weighted result and the second power weighted result are combined to obtain a current power attenuation amount; environmental humidity prediction data is obtained, and the current power attenuation amount, the comprehensive static electricity parameter, the environmental humidity prediction data, and the initial output power prediction value of the photovoltaic power generation system are input into a time series prediction model to obtain the output power prediction value.

[0097] In the present embodiment, first, the installation capacities of the glass material photovoltaic module and the plastic material photovoltaic module are obtained from the configuration database of the photovoltaic power generation system, and their proportions in the total capacity of the system are calculated, respectively denoted as the first capacity proportion and the second capacity proportion. Then, the parameters of photovoltaic modules of different materials are processed by using a weighted fusion method: the first static electricity parameter is multiplied by the first capacity proportion to obtain a first static electricity weighted result, the second static electricity parameter is multiplied by the second capacity proportion to obtain a second static electricity weighted result, and the two weighted results are added to obtain a comprehensive static electricity parameter reflecting the static electricity adsorption characteristics of the system as a whole; at the same time, the first power attenuation amount is multiplied by the first capacity proportion to obtain a first power weighted result, the second power attenuation amount is multiplied by the second capacity proportion to obtain a second power weighted result, and the two weighted results are added to obtain the current overall power attenuation amount of the system.

[0098] Then, the environmental humidity prediction data sequence of a future specific period is obtained from the weather forecasting system, and the initial output power prediction value under the ideal clean state and the predicted irradiance and temperature conditions is obtained from the photovoltaic power generation performance model. Finally, the current power attenuation, the comprehensive electrostatic parameter, the environmental humidity prediction data sequence, and the initial output power prediction value are input into the pre-trained time series prediction model as input features. The model captures the time sequence dependency relationship between the feature parameters and the power change through the long short-term memory network architecture, comprehensively analyzes the influence of the historical power attenuation trend, the electrostatic adsorption potential, and the humidity change on the dust accumulation, and finally outputs the power prediction value sequence of the photovoltaic power generation system in the future specified period. This process realizes the comprehensive consideration of the characteristics of components of different materials, environmental factors, and system configuration, and ensures the accuracy and practicability of the power prediction.

[0099] As an example of an embodiment of the present application, the time series prediction model is trained based on a historical data set, wherein the historical data set includes historical environmental humidity data, historical comprehensive electrostatic parameters, historical power attenuation, and corresponding historical power training data. Specifically, the historical data set is divided into a training set and a validation set according to the time sequence; an initial time series prediction model is constructed based on an LSTM neural network, the root mean square error between the model prediction value and the historical power training data is minimized as a loss function, the training set is used to train the initial time series prediction model, and the model hyperparameters of the initial time series prediction model are adjusted through the validation set to obtain the time series prediction model.

[0100] In this embodiment, first, historical operation data is collected and organized to construct a historical data set, which includes historical environmental humidity data aligned by time stamp, historical comprehensive electrostatic parameters calculated by the weighting method, historical power attenuation, and corresponding historical power training data measured actually. These data cover different seasons, weather conditions, and component pollution states to ensure the generalization ability of the model.

[0101] Subsequently, the entire historical data set is divided in chronological order, usually reserving the last 20% of the data as the validation set and the first 80% of the data as the training set to ensure that the continuity of the time sequence is not destroyed. An initial time series prediction model is constructed based on a long short-term memory neural network, and the input layer of the model is designed to receive a feature window containing multiple time steps. Each time step feature includes historical environmental humidity, historical comprehensive electrostatic parameters, historical power attenuation, and the historical power value of the previous time.

[0102] The training of the model takes the root mean square error between the predicted value and the historical power training data as the loss function, and uses the back propagation algorithm and adaptive matrix estimator optimizer to update the parameters. During the training process, the performance of the model is continuously monitored through the validation set. When the validation set loss does not decrease for consecutive multiple training periods, the early stopping mechanism is triggered to prevent overfitting. At the same time, the grid search method is used to optimize and adjust the key hyperparameters, including the number of neurons in the LSTM hidden layer, the learning rate, the dropout rate, and the length of the input time window. Through this systematic training and tuning process, a time series prediction model is finally obtained, which can accurately capture the dynamic characteristics of the system and has good prediction accuracy. The model can effectively integrate the influence of environmental humidity trends, the overall static state of the system, and existing power attenuation on future power generation, achieving high-precision power prediction.

[0103] It should be noted that the system embodiments described above are only schematic and some or all of the modules thereof can be selected to achieve the purpose of the embodiment. In addition, the connection relationship between the modules in the system embodiments provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0104] The above describes the preferred embodiments of the present application. It should be noted that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which are also considered within the scope of protection of the present application.

Claims

1. A coordinated power management system for a multi-source distributed photovoltaic power generation system, characterized in that, The method comprises the following steps: a data processing module is used to acquire first state data of a glass photovoltaic module and second state data of a plastic photovoltaic module; a power attenuation module is connected to the data processing module and is used to determine a first equivalent power generation efficiency attenuation coefficient of the glass photovoltaic module based on the first state data, determine a first power attenuation amount according to the first equivalent power generation efficiency attenuation coefficient, determine a second equivalent power generation efficiency attenuation coefficient of the plastic photovoltaic module based on the second state data, determine a second power attenuation amount according to the second equivalent power generation efficiency attenuation coefficient, and determine an output power prediction value of a photovoltaic power generation system based on the first power attenuation amount, the second power attenuation amount, the first state data, and the second state data; a control module is connected to the power attenuation module and is used to generate a charge and discharge control instruction of an energy storage control system according to a difference between the output power prediction value and an output power target value of a power distribution network if the difference is greater than a preset deviation threshold, and control an output power of the energy storage control system according to the charge and discharge control instruction to realize collaborative power management of the photovoltaic power generation system and the energy storage control system by the power distribution network; the first state data of the glass photovoltaic module and the second state data of the plastic photovoltaic module are acquired, wherein the first state data comprises a first electrostatic parameter and a first dust density, and the second state data comprises a second electrostatic parameter and a second dust density, the first electrostatic parameter and the second electrostatic parameter are both used to represent changes in electrostatic adsorption performance, and specifically: when environmental humidity data is lower than a preset humidity threshold, a dry environment monitoring instruction is generated and a measurement is performed on electrostatic field intensity of the glass photovoltaic module and the plastic photovoltaic module respectively to obtain a first electrostatic field intensity distribution parameter and a second electrostatic field intensity distribution; by analyzing the first electrostatic field intensity distribution, a static uniform area of the glass photovoltaic module is determined, a ratio of an electric field intensity of the static uniform area to a preset electric field intensity is calculated to obtain the first electrostatic parameter, wherein the static uniform area refers to an area in which field intensity values change continuously and a deviation of field intensity at any point from an average field intensity of the area does not exceed a first preset threshold; by analyzing the second electrostatic field intensity distribution, a static non-uniform area of the plastic photovoltaic module is determined, and the second electrostatic parameter is determined based on an average electric field intensity of the static non-uniform area, wherein the static non-uniform area refers to an area in which field intensity values change suddenly and a deviation of local point field intensity from the average field intensity of the area exceeds a second preset threshold; a first dust accumulation thickness of the glass photovoltaic module per unit area is detected by a laser dust detection sensor, and the first dust density is calculated in combination with the first electrostatic parameter. The dust coverage is obtained by processing a photovoltaic surface image of the plastic photovoltaic module through an image recognition technology, and the second dust accumulation thickness and the second electrostatic parameter of the plastic photovoltaic module are obtained by laser triangulation, and the second dust density is calculated; The first equivalent power generation efficiency attenuation coefficient of the glass photovoltaic module is determined based on the first state data, specifically: Based on the first dust density, the environmental particulate matter concentration is obtained by inversion through a settling model; The dust adsorption amount growth rate is determined by correlation analysis of the environmental particulate matter concentration and the first electrostatic parameter; Based on the dust adsorption amount growth rate and the first dust density, the first equivalent power generation efficiency attenuation coefficient is calculated through a light transmittance attenuation function.

2. The coordinated power management system for multi-source distributed photovoltaic power generation system of claim 1, wherein, The charge and discharge control instruction of the energy storage control system is generated according to the difference, and the output power of the energy storage control system is controlled according to the charge and discharge control instruction, specifically: The current state of charge and the maximum charge and discharge power of the energy storage control system are obtained, if the current state of charge is greater than or equal to a preset safety threshold, and the maximum charge and discharge power is greater than the power adjustment amount corresponding to the difference, a charge and discharge control instruction corresponding to the power adjustment amount is generated; If the current state of charge is less than the preset safety threshold, or the maximum charge and discharge power is less than or equal to the power adjustment amount, the maximum power adjustment amount of the energy storage control system is determined, and a charge and discharge control instruction corresponding to the maximum power adjustment amount is generated, and a cooperative scheduling instruction is generated based on a second difference between the power adjustment amount and the maximum power adjustment amount, and the microgrid or distributed power source within a preset distance range is controlled according to the cooperative scheduling instruction.

3. The coordinated power management system for multi-source distributed photovoltaic power generation systems of claim 1, wherein, The second equivalent power generation efficiency attenuation coefficient of the plastic photovoltaic module is determined based on the second state data, and a second power attenuation amount is determined according to the second equivalent power generation efficiency attenuation coefficient, specifically: The surface image of the plastic photovoltaic module is binarized to obtain a dust area and a clean area, and the dust coverage is obtained according to the ratio of the area of the dust area to the total area; Based on the dust coverage, a preset mapping relationship table between dust coverage and average light transmittance blocking rate is queried to determine the second equivalent power generation efficiency attenuation coefficient; Based on the second equivalent power generation efficiency attenuation coefficient and the preset output power of the plastic photovoltaic module under the current irradiance, the actual output power is determined; The actual output power and the preset output power are calculated to obtain the second power attenuation amount.

4. The coordinated power management system for multi-source distributed photovoltaic power generation system of claim 1 or 3, wherein, The output power prediction value of the photovoltaic power generation system is determined based on the first power attenuation amount, the second power attenuation amount, the first state data, and the second state data, specifically: The first capacity proportion of the glass photovoltaic module in the photovoltaic power generation system and the second capacity proportion of the plastic photovoltaic module are determined respectively; The first static electricity parameter is weighted according to the first capacity proportion to obtain a first static electricity weighted result, and the second static electricity parameter is weighted according to the second capacity proportion to obtain a second static electricity weighted result, and the comprehensive static electricity parameter of the photovoltaic power generation system is obtained by combining the first static electricity weighted result and the second static electricity weighted result; The first power attenuation amount is weighted according to the first capacity proportion to obtain a first power weighted result, and the second power attenuation amount is weighted according to the second capacity proportion to obtain a second power weighted result, and the current power attenuation amount is obtained by combining the first power weighted result and the second power weighted result; The environment humidity prediction data is obtained, and the output power prediction value is obtained by inputting the current power attenuation amount, the comprehensive static electricity parameter, the environment humidity prediction data and an initial output power prediction value of the photovoltaic power generation system into a time sequence prediction model.

5. The coordinated power management system for a multi-source distributed photovoltaic power generation system of claim 4, wherein, The time sequence prediction model is obtained by training based on a historical data set, wherein the historical data set includes historical environment humidity data, historical comprehensive static electricity parameters, historical power attenuation amounts and corresponding historical power training data, and specifically includes: The historical data set is divided into a training set and a verification set according to a time sequence; An initial time sequence prediction model is constructed based on an LSTM neural network, a root mean square error between a model prediction value and the historical power training data is taken as a loss function, the training set is used to train the initial time sequence prediction model, and model hyperparameters of the initial time sequence prediction model are adjusted through the verification set, so as to obtain the time sequence prediction model.

6. The coordinated power management system for multi-source distributed photovoltaic power generation systems of claim 1 or 3, wherein, The first power attenuation amount is determined according to the first equivalent power generation efficiency attenuation coefficient, and specifically includes: The first initial power attenuation amount is determined through a first mapping relationship according to the first equivalent power generation efficiency attenuation coefficient, and a mass density difference value is calculated according to the first dust density, and the mass density difference value, the first static electricity parameter and a power attenuation linear coefficient are calculated to obtain a first theoretical power attenuation amount; A third difference value between the first theoretical power attenuation amount and the first initial power attenuation amount is calculated, if the third difference value is greater than a second preset threshold value, the power attenuation linear coefficient is adjusted according to the first initial power attenuation amount to obtain a target power attenuation linear coefficient; The first power attenuation amount is calculated according to the target power attenuation linear coefficient.

7. The coordinated power management system for a multi-source distributed photovoltaic power generation system of claim 6, wherein, The power attenuation linear coefficient is obtained by linear regression analysis on a historical data set, wherein the historical data set includes a plurality of groups of dust density sequence data, environment humidity sequence data and corresponding actual power attenuation sequence data, and specifically includes: The environmental humidity sequence data is divided into a plurality of environmental humidity intervals, and the dust density sequence data and the actual power attenuation sequence data are divided according to each of the environmental humidity intervals, and a multiple linear regression analysis is performed on the divided data, taking the dust density and the electrostatic parameter as independent variables and taking the actual power attenuation value as a dependent variable, to obtain a preset power attenuation linear coefficient corresponding to each of the environmental humidity intervals.

8. The coordinated power management system for multi-source distributed photovoltaic power generation systems of claim 1, wherein, Also includes: A cleaning module is configured to calculate a cleaning benefit according to the first power attenuation amount and the second power attenuation amount, and generate a cleaning scheduling instruction when a fourth difference between the cleaning benefit and a cleaning cost is greater than a third preset threshold value and a time during which the output power prediction value is continuously lower than an output power target value is expected to be greater than a preset time threshold value, wherein the cleaning scheduling instruction is configured to start a cleaning device to clean the surface of the photovoltaic module and determine a cleaning sequence according to the size of the first power attenuation amount and the second power attenuation amount.

Images