A method and system for estimating distributed photovoltaic power
By establishing a spatial correlation model of the output power of distributed photovoltaic power plants, and utilizing historical data and characteristic parameters of neighboring photovoltaic power plants, the problem of photovoltaic power prediction in newly built power plants was solved, and accurate photovoltaic output estimation was achieved in the absence of historical data.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2023-01-10
- Publication Date
- 2026-07-31
AI Technical Summary
Existing technologies make it difficult to predict photovoltaic power in newly built or newly added distributed photovoltaic power stations, especially in the absence of historical data and meteorological information, making it impossible to accurately estimate photovoltaic output.
By establishing a spatial correlation model of the output power of distributed photovoltaic power plants, and using historical power data and spatial characteristic parameters of nearby monitored photovoltaic power plants, the correlation between the benchmark photovoltaic power plant and the neighboring power plants is analyzed, and a power estimation model for the target power plant is established.
It enables accurate estimation of photovoltaic output of distributed photovoltaic power stations in the absence of historical power data and meteorological information, thus improving the accuracy of photovoltaic power prediction.
Smart Images
Figure CN116054142B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of distributed photovoltaic power estimation technology, specifically relating to a distributed photovoltaic power estimation method and system. Background Technology
[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.
[0003] The large-scale and unrestrained use of traditional fossil fuels globally has led to a continuous depletion of non-renewable resources, a chronic energy shortage, and increasingly severe environmental pollution. The dwindling of non-renewable energy sources has become a pressing issue for humanity. Simultaneously, rapid economic development has increased overall human demand for energy, further exacerbating the energy supply-demand imbalance. Against this backdrop, developing green energy has become a primary approach and an inevitable trend for addressing the energy crisis and environmental problems. Solar energy, with its abundant reserves, environmental friendliness, wide distribution, and convenient access, is widely recognized as one of the most promising renewable energy sources for the future. As the primary method of utilizing solar energy, the photovoltaic industry has experienced rapid development.
[0004] In recent years, with the continuous advancement of photovoltaic technology, the penetration rate of distributed photovoltaic (PV) power systems has been increasing year by year. However, due to the volatility and uncertainty of PV power output, it can impact the stability of the power system. If accurate prediction of PV power output can be achieved, plans can be made in advance, grid dispatch can be coordinated, and the impact of PV power generation on the stable operation of the power system can be minimized. Therefore, PV power output prediction technology has become an important part of the field of PV power generation research.
[0005] Currently, photovoltaic (PV) power prediction methods are mainly divided into two categories: direct prediction and indirect prediction. Direct prediction methods are essentially based on historical data statistics. They establish power prediction models by analyzing a large amount of historical data from PV power plants, typically using historical power data or combining it with numerical weather prediction to directly predict PV output. Indirect prediction methods do not rely on historical PV output data. They first estimate irradiance data based on geographical and meteorological information, and then establish a model based on the physical characteristics of PV modules. The predicted irradiance value serves as the input value to the physical model, thereby achieving PV output prediction. Both of these methods often require solar radiation information, ambient temperature, meteorological information, or historical power data as data support for establishing the prediction model. For newly built sites or newly added distributed PV power plants, due to the lack of historical output data and the inability to obtain accurate numerical weather predictions, it is difficult to directly apply these methods to achieve power estimation. Summary of the Invention
[0006] To address the aforementioned problems, this invention proposes a method and system for estimating distributed photovoltaic (PV) power. This invention relies on historical power data from nearby monitored distributed PV power plants, without requiring power data or meteorological information from the target distributed PV power plant. When building the model, nearby monitored distributed PV power plants are used as target power plants for training, fully exploring the spatial correlation of distributed PV power output, resulting in accurate calculation results.
[0007] According to some embodiments, the present invention adopts the following technical solution:
[0008] A method for estimating distributed photovoltaic power includes the following steps:
[0009] Spatial characteristic parameters of distributed photovoltaic (PV) power plants are extracted. Historical power data of neighboring distributed PV power plants are used to analyze the correlation between the output power of the benchmark PV power plant and the output power and relative spatial characteristic parameters of neighboring PV power plants. A spatial correlation model of the output power of distributed PV power plants that takes into account spatial distribution characteristics is established.
[0010] A power prediction model is established based on historical power data of nearby distributed photovoltaic power stations to obtain predicted power data of nearby distributed photovoltaic power stations;
[0011] Based on the spatial correlation model of the output power of the distributed photovoltaic power station and the power prediction data of neighboring photovoltaic power stations, the power estimation results of the target distributed photovoltaic power station are obtained.
[0012] As an alternative implementation, the spatial correlation model of the output power of the distributed photovoltaic power station is such that there is a numerical correlation between the output trajectories of distributed photovoltaic power stations with different spatial characteristics under the same time dimension.
[0013] As an alternative implementation method, the specific process of establishing a correlation model between the output power of a distributed photovoltaic power station and its spatial distribution includes:
[0014] One of the neighboring distributed photovoltaic (PV) power plants is selected as the benchmark PV power plant. The spatial characteristic difference sequence between the benchmark PV power plant and other neighboring distributed PV power plants is obtained. The data is normalized. The normalized spatial characteristic difference sequence and the historical power sequence of each PV power plant are used as a set of training data for the spatial correlation model of distributed PV output power. Each of the m+1 neighboring distributed PV power plants is used as the benchmark PV power plant, and m+1 sets of training data are obtained. The model is trained in sequence, and the correlation between the output power of the benchmark PV power plant and the output power of the neighboring PV power plants and the relative spatial characteristic parameters is analyzed to obtain the spatial correlation model of distributed PV power output power.
[0015] As a further constraint, the normalized installation characteristic difference sequence and the power data of other neighboring distributed photovoltaic power plants are used as model inputs, and the normalized baseline photovoltaic power data are used as model outputs.
[0016] As an alternative implementation method, the specific process of obtaining the power estimation result of the target distributed photovoltaic power station includes applying the established spatial correlation model of the output power of the distributed photovoltaic power station to the target photovoltaic power station, treating the target photovoltaic power station as the benchmark photovoltaic power station, eliminating the distributed photovoltaic power station that is farthest from the target photovoltaic power station, using the corresponding data of the remaining m neighboring distributed photovoltaic power stations as model input, and estimating the power value of the target power station.
[0017] As a further limitation, the corresponding data includes normalized power prediction values of neighboring power stations and a sequence of installation location differences between neighboring power stations and the target power station.
[0018] As a further constraint, the estimated power value of the target power plant is normalized data, and the actual estimated power is calculated after combining it with the rated installed capacity of the target photovoltaic power plant.
[0019] A distributed photovoltaic power estimation system, comprising:
[0020] The module for constructing a spatial correlation model of the output power of distributed photovoltaic power plants is configured to extract spatial characteristic parameters of distributed photovoltaic power plants, use historical power data of neighboring distributed photovoltaic power plants to analyze the correlation between the output power of the benchmark photovoltaic power plant and the output power of neighboring photovoltaic power plants, as well as the relative spatial characteristic parameters, and establish a spatial correlation model of the output power of distributed photovoltaic power plants that takes into account spatial distribution characteristics.
[0021] The neighboring distributed photovoltaic power prediction module is configured to establish a power prediction model for the neighboring distributed photovoltaic power station based on historical power data, and obtain the predicted power data of the neighboring distributed photovoltaic power station.
[0022] The estimation module is configured to obtain the power estimation result of the target distributed photovoltaic power station based on the spatial correlation model of the output power of the distributed photovoltaic power station and the power prediction data of neighboring photovoltaic power stations.
[0023] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the steps in the above method.
[0024] An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the steps in the method described above.
[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0026] This invention establishes a spatial correlation analysis model for the output power of distributed photovoltaic (PV) power plants using historical power data and spatial characteristic parameters. This model obtains the spatial correlation relationship of the output power of distributed PV power plants and applies this correlation relationship and power prediction data of neighboring distributed PV power plants to establish a power estimation model for the target distributed PV power plant. This model only uses historical power data and spatial characteristic parameters of neighboring distributed PV power plants to establish a power estimation model for the target distributed PV power plant, without relying on historical power data or meteorological information of the target distributed PV power plant, resulting in more accurate calculation results.
[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description
[0028] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.
[0029] Figure 1 This is a schematic diagram of the prediction model based on spatial correlation analysis of the present invention;
[0030] Figure 2 This is a schematic diagram of the spatial correlation analysis of the output power of distributed photovoltaic power stations according to the present invention, wherein (a): PV0 is used as the reference power station; (b): PV1 is used as the reference power station; (c): PV3 is used as the reference power station;
[0031] Figure 3 This is a schematic diagram of the spatial correlation analysis model of PV0 as a reference station in this invention;
[0032] Figure 4 This is a schematic diagram illustrating the power estimation of the target photovoltaic power station of the present invention;
[0033] Figure 5 This is a flowchart illustrating the present invention. Detailed Implementation
[0034] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0035] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, 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 invention pertains.
[0036] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.
[0037] Example 1
[0038] Because distributed photovoltaic (PV) power plants exhibit spatial output correlations, power estimation can be achieved by studying the correlations between PV plants with different spatial characteristics. Extensive research has been conducted on power prediction models that account for spatiotemporal correlations. Currently, spatial correlation analysis often involves calculating the correlation coefficients of each power plant or classifying their correlations through correlation clustering, and then using the classified data to build a power prediction model. However, correlation analysis requires the target power plant to have a small or large amount of historical power data, making it difficult to apply to newly built or added distributed PV plants. This invention establishes a spatial correlation analysis model using historical power data and spatial characteristic parameters to obtain the spatial correlation relationships of distributed PV plant outputs. This correlation relationship, along with power prediction data from neighboring distributed PV plants, is then used to build a power estimation model for the target distributed PV plant, such as... Figure 1 As shown. This model only uses historical power data and spatial characteristic parameters of nearby distributed photovoltaic (PV) power plants to establish a power estimation model for the target distributed PV power plant, and does not rely on historical power data or meteorological information of the target distributed PV power plant.
[0039] Distributed photovoltaic (PV) power stations with similar or identical geographical environments, weather conditions, and installation methods should exhibit a certain correlation. The PV output sequences of geographically proximate distributed PV power stations also show a high degree of spatial correlation. In this invention, spatial correlation refers to the numerical correlation between the output trajectories of distributed PV power stations with different spatial characteristics within the same time dimension.
[0040] Since the target photovoltaic (PV) power plant lacks historical power data and cannot be used for model training, historical power data from neighboring distributed PV power plants are used for spatial correlation analysis. Assume there are m+1 neighboring distributed PV power plants (PV0-PVm) with historical power data and spatial characteristic parameters within the spatial range, and a target distributed PV power plant (PVx) at an arbitrary installation location without historical power data. When studying spatial correlation, one of these power plants is randomly selected as the benchmark power plant. Then, the power output of the remaining m neighboring distributed PV power plants is spatially correlated with that of the benchmark power plant, such as... Figure 2 As shown.
[0041] The spatial correlation analysis model of distributed photovoltaic power output takes PV0 as the benchmark power station as an example. The historical power data of the benchmark power station is represented by P. 0 Spatial features are represented by a sequence. It means that, among them These represent the spatial characteristic parameters of the benchmark power station installation, including but not limited to the longitude, latitude, altitude, and installation angle of the benchmark station. Then, the power data of any nearby distributed photovoltaic power station is P. i The spatial feature sequence is
[0042] Since the differences between various spatial feature parameters may have large numerical differences, in order to make different data have the same measurement scale, the input data needs to be normalized before the model is trained so that the input data is between [0,1], so as to eliminate the prediction error caused by the different units of the original data.
[0043] Since the differences in installation characteristic parameters are known, the maximum, minimum, mean, and standard deviation are easily obtained. Therefore, the normalization method can be Z-Score standardization or other normalization methods. For example, using Z-Score standardization, the normalization formula is:
[0044]
[0045] Where x is the original sample value, μ is the data mean, and σ is the standard deviation. In this patent, the spatial characteristic parameters are normalized to include the target site's spatial characteristic parameters; that is, the target site parameters must be taken into account when calculating μ and σ. Therefore, the difference sequence of spatial characteristics between the benchmark site and neighboring distributed photovoltaic sites is denoted as:
[0046]
[0047] Since there is no historical data for the target photovoltaic power plant, the above normalization method cannot be directly applied to normalize the output power of the photovoltaic power plant. The following mathematical expression is used to normalize the output power of the photovoltaic power plant:
[0048]
[0049] Where P i P′ represents the actual power output of the photovoltaic power station. i P represents the normalized power output of a photovoltaic power station. N This represents the rated installed capacity of the photovoltaic power station. The spatial characteristic difference sequences and photovoltaic power station power sequences used later are all normalized data.
[0050] The spatial feature difference sequence between the baseline photovoltaic (PV0) and neighboring distributed photovoltaic (PV) sites PV1-PVm, along with the historical power sequences of PV1-PVm, are used as inputs. The power value of the baseline PV0 at the corresponding time point is output. A locally connected neural network is employed to achieve dimensionality reduction and feature extraction for the data from each PV site. Furthermore, the locally connected approach reduces the complexity of the network model, decreases the number of training parameters, and improves training efficiency. The spatial correlation model of the output power of distributed PV sites is as follows: Figure 3 As shown.
[0051] Figure 3 The diagram shows the spatial correlation analysis model using PV0 as the benchmark photovoltaic (PV) power station. Similarly, PV1 can also be used as the benchmark power station. In this case, the model training input data consists of normalized historical power data of PV0, PV2-PVm, and the spatial feature difference sequence between PV0, PV2-PVm, and PV1. To fully explore the spatial correlation of the output power of distributed photovoltaic (PV) power stations, neighboring distributed PV power stations (PV0-PVm) are used as benchmark power stations for model training. Therefore, when applying the established spatial correlation model of the output power of distributed PV power stations to the target PV power station PVx, the target PV power station PVx is considered as the benchmark power station. The distributed PV power station furthest from the target PV power station is removed, and the corresponding data of the remaining m neighboring distributed PV power stations (normalized predicted power values of neighboring power stations, and the installation location difference sequence between neighboring power stations and the target power station) are used as model input to estimate the power value of the target power station. Figure 4 As shown. Extensive research has been conducted on power prediction models for benchmark photovoltaic power plants based on time-series power sequences, and any prediction model can be applied to this invention; therefore, further details are omitted here.
[0052] This embodiment addresses the spatial correlation of output power among neighboring distributed photovoltaic (PV) power plants. It utilizes historical data from monitored PV plants to study the correlation between the output power of a benchmark PV plant and the output power and relative spatial characteristic parameters of adjacent PV plants, and applies this correlation to target PV plant power estimation. A power estimation model for target PV plants without historical data is established based on historical power data from neighboring monitored PV plants, solving the problem of difficulty in power estimation for newly added or newly constructed PV plants lacking meteorological information and historical data.
[0053] Example 2
[0054] like Figure 5As shown, a distributed photovoltaic (PV) power estimation method mainly includes two steps: first, spatial correlation analysis of the output power of distributed PV power plants; and second, power estimation of target power plants based on the spatial correlation model of the output power of distributed PV power plants. The specific steps of the spatial correlation analysis are as follows: One neighboring distributed PV power plant is selected as a benchmark PV power plant. The spatial characteristic difference sequence between the benchmark PV power plant and other neighboring distributed PV power plants is obtained, and the data is normalized. The normalized spatial characteristic difference sequence and the historical power sequences of each PV power plant are used as a set of training data for the PV output spatial correlation model. The installation characteristic difference sequence and the power data of the remaining neighboring distributed PV power plants are the model input, and the power data of the benchmark PV power plant is the model output. Using m+1 neighboring distributed PV power plants as benchmark PV power plants, m+1 sets of training data can be obtained. The spatial correlation model is trained sequentially, and the correlation between the output power of the benchmark PV power plant and the output power of neighboring PV power plants, as well as the relative spatial characteristic parameters, is analyzed to obtain the spatial correlation model of the output power of distributed PV power plants.
[0055] The specific steps for estimating the power of the target photovoltaic (PV) power station are as follows: From the m+1 neighboring distributed PV power stations, the PV power station furthest from the target PV power station is removed. The normalized power prediction values of the remaining m neighboring distributed PV power stations and the difference in installation characteristics between the neighboring distributed PV power stations and the target PV power station are used as inputs to achieve power estimation to the target PV power station. The obtained power estimation value is normalized data, and the actual estimated power is calculated by combining it with the rated installed capacity of the target PV power station.
[0056] This embodiment relies on historical power data from nearby monitored distributed photovoltaic (PV) power plants, but does not require power data or meteorological information from the target PV power plant. When building the model, nearby monitored PV power plants are used as target power plants for training, fully exploring the spatial correlation of PV power plant output. It is suitable for power estimation of newly built PV power plants without any historical power or meteorological data, and can be applied to power and economic forecasting before the construction of PV power plants.
[0057] The present invention also provides the following product examples:
[0058] A distributed photovoltaic power estimation system, comprising:
[0059] The module for constructing a spatial correlation model of the output power of distributed photovoltaic power plants is configured to extract spatial characteristic parameters of distributed photovoltaic power plants, use historical power data of neighboring distributed photovoltaic power plants to analyze the correlation between the output power of the benchmark photovoltaic power plant and the output power of neighboring photovoltaic power plants, as well as the relative spatial characteristic parameters, and establish a spatial correlation model of the output power of distributed photovoltaic power plants that takes into account spatial distribution characteristics.
[0060] The neighboring distributed photovoltaic power prediction module is configured to establish a power prediction model for the neighboring distributed photovoltaic power station based on historical power data, and obtain the predicted power data of the neighboring distributed photovoltaic power station.
[0061] The estimation module is configured to obtain the power estimation result of the target distributed photovoltaic power station based on the spatial correlation model of the output power of the distributed photovoltaic power station and the power prediction data of neighboring photovoltaic power stations.
[0062] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the steps in the above method.
[0063] An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the steps in the method described above.
[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., that can be made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for estimating distributed photovoltaic power, characterized in that, Includes the following steps: Spatial characteristic parameters of distributed photovoltaic (PV) power plants are extracted. Historical power data of neighboring distributed PV power plants are used to analyze the correlation between the output power of the benchmark PV power plant and the output power and relative spatial characteristic parameters of neighboring PV power plants. A spatial correlation model of the output power of distributed PV power plants that takes into account spatial distribution characteristics is established. A power prediction model is established based on historical power data of nearby distributed photovoltaic power stations to obtain predicted power data of nearby distributed photovoltaic power stations; Based on the spatial correlation model of the output power of distributed photovoltaic power plants and the power prediction data of neighboring photovoltaic power plants, the power estimation results of the target distributed photovoltaic power plant are obtained.
2. The distributed photovoltaic power estimation method as described in claim 1, characterized in that, The spatial correlation model of the output power of the distributed photovoltaic power station shows that there is a numerical correlation between the output trajectories of distributed photovoltaic power stations with different spatial characteristics under the same time dimension.
3. The distributed photovoltaic power estimation method as described in claim 1, characterized in that, The specific process of establishing a correlation model between the output power of a distributed photovoltaic power station and its spatial distribution includes: One of the neighboring distributed photovoltaic (PV) power plants is selected as the benchmark PV power plant. The spatial characteristic difference sequence between the benchmark PV power plant and other neighboring distributed PV power plants is obtained. The data is normalized. The normalized spatial characteristic difference sequence and the historical power sequence of each PV power plant are used as a set of training data for the spatial correlation model of PV output. Each of the m+1 neighboring distributed PV power plants is used as the benchmark PV power plant, and m+1 sets of training data are obtained. The model is trained in sequence. The correlation between the output power of the benchmark PV power plant and the output power of the neighboring PV power plants and the relative spatial characteristic parameters is analyzed to obtain the spatial correlation model of the output power of distributed PV power plants.
4. The distributed photovoltaic power estimation method as described in claim 3, characterized in that, The normalized spatial feature difference sequence and the power data of other nearby distributed photovoltaic power plants are used as model inputs, and the normalized benchmark photovoltaic power plant power data are used as model outputs.
5. The distributed photovoltaic power estimation method as described in claim 1, characterized in that, The specific process of obtaining the power estimation result of the target distributed photovoltaic power station includes applying the established spatial correlation model of the output power of the distributed photovoltaic power station to the target photovoltaic power station. The target photovoltaic power station is regarded as the benchmark photovoltaic power station. The distributed photovoltaic power station that is farthest away from the target photovoltaic power station is eliminated. The corresponding data of the remaining m neighboring distributed photovoltaic power stations are used as the model input to estimate the power value of the target power station.
6. The distributed photovoltaic power estimation method as described in claim 5, characterized in that, The corresponding data includes normalized power prediction values of neighboring power stations and a sequence of installation location differences between neighboring power stations and the target power station.
7. The distributed photovoltaic power estimation method as described in claim 5, characterized in that, The estimated power value of the target power plant is normalized data. After combining it with the rated installed capacity of the target photovoltaic power plant, the actual estimated power is calculated.
8. A distributed photovoltaic power estimation system, characterized in that it comprises: The distributed photovoltaic (PV) power output power spatial correlation model construction module is configured to extract spatial characteristic parameters of distributed PV power plants, analyze the correlation between the output power of the benchmark PV power plant and the output power of adjacent PV power plants and relative spatial characteristic parameters using historical power data of neighboring distributed PV power plants, and establish a distributed PV power output power spatial correlation model that takes into account spatial distribution characteristics; the neighboring distributed PV power prediction module is configured to establish a power prediction model for neighboring distributed PV power plants based on historical power data, and obtain predicted power data for neighboring distributed PV power plants; The estimation module is configured to obtain the power estimation result of the target distributed photovoltaic power station based on the spatial correlation model of the output power of the distributed photovoltaic power station and the power prediction data of neighboring photovoltaic power stations.
9. A computer-readable storage medium, characterized in that, Used to store computer instructions, which, when executed by a processor, complete the steps of the method according to any one of claims 1-7.
10. An electronic device, characterized in that, It includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the steps of the method according to any one of claims 1-7.