Zone area photovoltaic cluster output on-site prediction system based on photovoltaic side micrometeorological monitoring

Abstract

The invention discloses a zone area photovoltaic cluster output local prediction system based on photovoltaic side micro-meteorological monitoring, which comprises a plurality of micro-meteorological monitoring terminals arranged on a zone area photovoltaic side, the micro-meteorological monitoring terminals are connected with a data comprehensive processing unit through a communication unit, the data comprehensive processing unit is connected with a photovoltaic output prediction unit, and the photovoltaic output prediction unit is connected with the micro-meteorological monitoring terminals. The photovoltaic output prediction unit performs on-site prediction on photovoltaic output by calling data in the data comprehensive processing unit; according to the invention, the regional micro-meteorological data is collected, various influence factors are considered, and a photovoltaic on-site prediction system is established, so that accurate prediction of photovoltaic station processing based on meteorological data is realized, the power supply quality and the power supply reliability of a photovoltaic power station are improved, and stable power grid voltage and safe operation are ensured.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic power output prediction technology, specifically relating to a local prediction system for photovoltaic cluster power output based on photovoltaic side micro-meteorological monitoring. Background Technology

[0002] Residential distributed photovoltaic power generation refers to small-scale photovoltaic power generation systems developed in a decentralized manner at the user end. It has the characteristics of balancing and regulating the distribution network system. As the installed capacity of photovoltaic equipment in the power grid continues to increase, the intermittency and randomness of photovoltaic output have also brought severe challenges to the traditional power grid.

[0003] Photovoltaic power generation is greatly affected by weather conditions, especially during cloudy weather, when photovoltaic output fluctuates drastically, threatening the voltage stability and safe operation of the power grid in areas with high photovoltaic penetration. Therefore, with the large-scale penetration of photovoltaics, it is very important to anticipate weather conditions, accurately predict the output of photovoltaic power plants, and reduce peak and valley loads in advance to minimize grid impact.

[0004] Therefore, in order to solve the above problems, it is necessary to develop a local forecasting system for photovoltaic cluster output based on photovoltaic side micro-meteorological monitoring. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a local forecasting system for photovoltaic cluster output based on photovoltaic micro-meteorological monitoring. This system collects regional micro-meteorological data, considers various influencing factors, and builds a local forecasting system for photovoltaic power plants. It enables accurate forecasting of photovoltaic power plant processing based on meteorological data, improves the power supply quality and reliability of photovoltaic power plants, and ensures stable and safe operation of the power grid voltage.

[0006] The objective of this invention is achieved as follows: a photovoltaic cluster output prediction system based on photovoltaic side micro-meteorological monitoring includes multiple micro-meteorological monitoring terminals installed on the photovoltaic side of the photovoltaic area. The micro-meteorological monitoring terminals are connected to a data processing unit through a communication unit. The data processing unit is connected to a photovoltaic output prediction unit. The photovoltaic output prediction unit performs on-site prediction of photovoltaic output by retrieving data from the data processing unit.

[0007] The micro-meteorological monitoring terminal includes multiple meteorological sensors for monitoring meteorological data, a terminal data acquisition module connected to the multiple meteorological sensors, and a power supply module for supplying power to the terminal.

[0008] The micro-meteorological monitoring terminal transmits the monitored micro-meteorological data to the data processing unit through the communication unit;

[0009] The data processing unit includes a central control module, a data receiving module, a preprocessing module, a normalization module, and a data storage module. The central control module is connected to the data receiving module, the preprocessing module, the normalization module, and the data storage module and issues control commands. The data receiving module is connected to the micro-meteorological monitoring terminal through a communication unit. The preprocessing module is used to perform preliminary processing on the data, including cleaning, deduplication, and completion. The normalization module is used to normalize the preprocessed micro-meteorological data for neural network training. The data storage module is used to store the data.

[0010] The photovoltaic power output prediction unit includes a sample training module, a power output prediction module, a performance evaluation module, and a result output module. The sample training module is used to input processed micrometeorological data as training samples into an improved NARX nonlinear autoregressive neural network for training. The processing prediction module is used to input test samples into the trained NARX nonlinear autoregressive neural network to calculate the predicted value. The performance evaluation module is used to evaluate the generalization ability of the NARX nonlinear autoregressive neural network in the sample training module and the processing prediction module. The result output module is used to perform inverse normalization processing on the calculated photovoltaic power output prediction value to obtain the actual prediction value and output it.

[0011] Furthermore, the meteorological sensors in the micro-meteorological monitoring terminal include a temperature sensor, a humidity sensor, a light radiation intensity sensor, a wind transmitter, a wind direction transmitter, and a rainfall sensor.

[0012] Furthermore, the power supply module in the micro-meteorological monitoring terminal adopts a power supply method that combines photovoltaic panels and batteries.

[0013] Furthermore, the preprocessing module in the data integration processing unit uses interpolation to complete the data, specifically as follows: ,but time The interpolation value is: In the formula: The data value at time is , The data value at time is , The data value at time is .

[0014] Furthermore, the normalization module in the data processing unit normalizes the micrometeorological data using the following formula: In the formula: This represents the normalized data. This represents the data before normalization. Representing variables respectively The maximum and minimum values.

[0015] Furthermore, the sample training module in the photovoltaic output prediction unit uses an improved regularization algorithm to optimize the network weights of the neural network, wherein the neural network performance evaluation function is expressed as: In the formula: This represents the correction factor, which is 0.3 here. This represents the sum of squares of the neural network error. This represents the sum of squares of all values ​​in the neural network.

[0016] Furthermore, in the sample training module, the NARX nonlinear autoregressive neural network includes... Each delay layer node, One intermediate layer node and The output, then its i Sum of squared network errors for input and output samples Represented as: In the formula: Indicates the first The target output of each output layer node Indicates the first The network output of each output layer node Indicates the number of outputs; the first Sum of squared network weights for input and output sample data Represented as: In the formula: This indicates the number of adjustable weights in the network. Indicates the first From the delay layer node to the first Weights between intermediate layer nodes Indicates the first The intermediate layer node to the first The weights of each output layer node. Indicates the number of delay layer nodes. This indicates the number of intermediate layer nodes.

[0017] Furthermore, since the gradient descent method is used to optimize the weights in the sample training module, the activation function of the intermediate layer is expressed as: In the formula: The parameter representing the activation function of the intermediate layer has a value of 1. The weight adjustment of the intermediate layer is expressed as: In the formula: Indicates the first The intermediate layer weight adjustment amount for input and output sample data. Indicates the correction factor. Indicates the first For the input and output sample data at the th The target output of each output layer node Indicates the first For the input and output sample data at the th The network output of each output layer node , They represent the first , For the input and output sample data at the th The intermediate layer node to the first The weights of each output layer node. Indicates the first For the input and output sample data at the th The input of each output layer node, This indicates the number of adjustable weights in the network; therefore, the intermediate layer weights are adjusted as follows: , This represents the learning rate, with a value of 0.006.

[0018] Furthermore, since the gradient descent method is used to optimize the weights in the sample training module, the linear processing function of the output layer is expressed as: The weight adjustment of the output layer is expressed as: In the formula: Indicates the first For the input and output sample data at the th If there are several intermediate layer nodes as input, then the output layer weights are adjusted as follows: , This represents the learning rate, with a value of 0.006.

[0019] Furthermore, the performance evaluation module uses root mean square error to evaluate the generalization ability of the neural network, specifically expressed as follows: In the formula, This represents the root mean square error. This represents the target value in the test sample. This represents the predicted value of the model after inverse normalization. This indicates the number of data samples.

[0020] Due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0021] (1) By collecting regional micro-meteorological data in real time, considering the continuity and trend of meteorological data, and combining various influencing factors, a photovoltaic on-site prediction system is built to realize accurate prediction of photovoltaic power station processing based on meteorological data, improve the power supply quality and reliability of photovoltaic power station, and ensure the stable voltage and safe operation of the power grid.

[0022] (2) By using an improved NARX nonlinear autoregressive neural network to predict photovoltaic power output and by optimizing the network weights of the neural network through an improved regularization algorithm, the output of photovoltaic power plants can be accurately predicted, peak shaving and valley filling can be carried out in advance, and the impact on the power grid can be reduced. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the principle of the present invention. Detailed Implementation

[0024] The technical solution of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0025] like Figure 1 As shown, the photovoltaic cluster output prediction system based on photovoltaic side micro-meteorological monitoring includes multiple micro-meteorological monitoring terminals set up on the photovoltaic side of the photovoltaic area. The micro-meteorological monitoring terminals are connected to a data processing unit through a communication unit. The data processing unit is connected to a photovoltaic output prediction unit. The photovoltaic output prediction unit performs on-site prediction of photovoltaic output by retrieving data from the data processing unit.

[0026] The micro-meteorological monitoring terminal includes multiple meteorological sensors for monitoring meteorological data, a terminal data acquisition module connected to the multiple meteorological sensors, and a power supply module for supplying power to the terminal; the micro-meteorological monitoring terminal transmits the monitored micro-meteorological data to the data processing unit through a communication unit.

[0027] Preferably, the meteorological sensors in the micro-meteorological monitoring terminal include a temperature sensor, a humidity sensor, a light radiation intensity sensor, a wind transmitter, a wind direction transmitter, and a rainfall sensor; the power supply module in the micro-meteorological monitoring terminal adopts a power supply method combining photovoltaic panels and batteries.

[0028] The data processing unit includes a central control module, a data receiving module, a preprocessing module, a normalization module, and a data storage module. The central control module is connected to the data receiving module, the preprocessing module, the normalization module, and the data storage module and issues control commands. The data receiving module is connected to the micro-meteorological monitoring terminal through a communication unit. The preprocessing module performs preliminary processing on the data, including cleaning, deduplication, and completion. The normalization module normalizes the preprocessed micro-meteorological data for neural network training. The data storage module stores the data.

[0029] Preferably, the preprocessing module in the data integration processing unit uses interpolation to complete the data, specifically as follows: ,but time The interpolation value is: In the formula: The data value at time is , The data value at time is , The data value at time is .

[0030] Preferably, the normalization module in the data processing unit normalizes the micrometeorological data using the following formula: In the formula: This represents the normalized data. This represents the data before normalization. Representing variables respectively The maximum and minimum values.

[0031] The photovoltaic power output prediction unit includes a sample training module, a power output prediction module, a performance evaluation module, and a result output module. The sample training module is used to input processed micrometeorological data as training samples into an improved NARX nonlinear autoregressive neural network for training. The processing prediction module is used to input test samples into the trained NARX nonlinear autoregressive neural network to calculate the predicted value. The performance evaluation module is used to evaluate the generalization ability of the NARX nonlinear autoregressive neural network in the sample training module and the processing prediction module. The result output module is used to perform inverse normalization processing on the calculated photovoltaic power output prediction value to obtain the actual prediction value and output it.

[0032] Preferably, the sample training module in the photovoltaic output prediction unit uses an improved regularization algorithm to optimize the network weights of the neural network, wherein the neural network performance evaluation function is expressed as: In the formula: This represents the correction factor, which is 0.3 here. This represents the sum of squares of the neural network error. This represents the sum of squares of all values ​​in the neural network.

[0033] Preferably, in the sample training module, the NARX nonlinear autoregressive neural network includes Each delay layer node, One intermediate layer node and The output, then its i Sum of squared network errors for input and output samples Represented as: In the formula: Indicates the first The target output of each output layer node Indicates the first The network output of each output layer node Indicates the number of outputs; the first Sum of squared network weights for input and output sample data Represented as: In the formula: This indicates the number of adjustable weights in the network. Indicates the first From the delay layer node to the first Weights between intermediate layer nodes Indicates the first The intermediate layer node to the first The weights of each output layer node. Indicates the number of delay layer nodes. This indicates the number of intermediate layer nodes.

[0034] Preferably, in the sample training module, gradient descent is used to optimize the weights, then the activation function of the intermediate layer is expressed as: In the formula: The parameter representing the activation function of the intermediate layer has a value of 1. The weight adjustment of the intermediate layer is expressed as: In the formula: Indicates the first The intermediate layer weight adjustment amount for input and output sample data. Indicates the correction factor. Indicates the first For the input and output sample data at the th The target output of each output layer node Indicates the first For the input and output sample data at the th The network output of each output layer node , They represent the first , For the input and output sample data at the th The intermediate layer node to the first The weights of each output layer node. Indicates the first For the input and output sample data at the th The input of each output layer node, This indicates the number of adjustable weights in the network; therefore, the intermediate layer weights are adjusted as follows: , This represents the learning rate, with a value of 0.006.

[0035] Preferably, if gradient descent is used to optimize the weights in the sample training module, then the linear processing function of the output layer is expressed as: The weight adjustment of the output layer is expressed as: In the formula: Indicates the first For the input and output sample data at the th If there are several intermediate layer nodes as input, then the output layer weights are adjusted as follows: , This represents the learning rate, with a value of 0.006.

[0036] Preferably, the performance evaluation module uses root mean square error to evaluate the generalization ability of the neural network, specifically expressed as follows: In the formula, This represents the root mean square error. This represents the target value in the test sample. This represents the predicted value of the model after inverse normalization. This indicates the number of data samples.

[0037] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A system for on-site prediction of photovoltaic cluster output in a photovoltaic power station based on photovoltaic-side micro-meteorological monitoring, characterized in that: It includes multiple micro-meteorological monitoring terminals installed on the photovoltaic side of the distribution area. The micro-meteorological monitoring terminals are connected to a data processing unit through a communication unit. The data processing unit is connected to a photovoltaic output prediction unit. The photovoltaic output prediction unit makes on-site predictions of photovoltaic output by retrieving data from the data processing unit. The micro-meteorological monitoring terminal includes multiple meteorological sensors for monitoring meteorological data, a terminal data acquisition module connected to the multiple meteorological sensors, and a power supply module for supplying power to the terminal. The micro-meteorological monitoring terminal transmits the monitored micro-meteorological data to the data processing unit through the communication unit; The data processing unit includes a central control module, a data receiving module, a preprocessing module, a normalization module, and a data storage module. The central control module is connected to the data receiving module, the preprocessing module, the normalization module, and the data storage module and issues control commands. The data receiving module is connected to the micro-meteorological monitoring terminal through a communication unit. The preprocessing module is used to perform preliminary processing on the data, including cleaning, deduplication, and completion. The normalization module is used to normalize the preprocessed micro-meteorological data for neural network training. The data storage module is used to store the data. The photovoltaic power output prediction unit includes a sample training module, a power output prediction module, a performance evaluation module, and a result output module. The sample training module is used to input processed micrometeorological data as training samples into an improved NARX nonlinear autoregressive neural network for training. The processing prediction module is used to input test samples into the trained NARX nonlinear autoregressive neural network to calculate the predicted value. The performance evaluation module is used to evaluate the generalization ability of the NARX nonlinear autoregressive neural network in the sample training module and the processing prediction module. The result output module is used to perform inverse normalization processing on the calculated photovoltaic power output prediction value to obtain the actual prediction value and output it.

2. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 1, characterized in that: The meteorological sensors in the micro-meteorological monitoring terminal include a temperature sensor, a humidity sensor, a light radiation intensity sensor, a wind transmitter, a wind direction transmitter, and a rainfall sensor.

3. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 1, characterized in that: The power supply module in the micro-meteorological monitoring terminal adopts a power supply method that combines photovoltaic panels and batteries.

4. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 1, characterized in that: The preprocessing module in the data integration and processing unit uses interpolation to complete the data, specifically as follows: ,but time The interpolation value is: In the formula: The data value at time is , The data value at time is , The data value at time is .

5. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 1, characterized in that: The normalization module in the data processing unit uses the following formula to normalize the micro-meteorological data: In the formula: This represents the normalized data. This represents the data before normalization. Representing variables respectively The maximum and minimum values.

6. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 1, characterized in that: The sample training module in the photovoltaic output prediction unit uses an improved regularization algorithm to optimize the network weights of the neural network. The neural network performance evaluation function is expressed as: In the formula: This represents the correction factor, which is 0.3 here. This represents the sum of squares of the neural network error. This represents the sum of squares of all values ​​in the neural network.

7. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 6, characterized in that: The sample training module includes the NARX nonlinear autoregressive neural network. Each delay layer node, One intermediate layer node and The output, then its i Sum of squared network errors for input and output samples Represented as: In the formula: Indicates the first The target output of each output layer node Indicates the first The network output of each output layer node Indicates the number of outputs; the first Sum of squared network weights for input and output sample data Represented as: In the formula: This indicates the number of adjustable weights in the network. Indicates the first From the delay layer node to the first Weights between intermediate layer nodes Indicates the first The intermediate layer node to the first The weights of each output layer node. Indicates the number of delay layer nodes. This indicates the number of intermediate layer nodes.

8. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 7, characterized in that: In the sample training module, gradient descent is used to optimize the weights. Therefore, the activation function of the intermediate layer is expressed as: In the formula: The parameter representing the activation function of the intermediate layer has a value of 1. The weight adjustment of the intermediate layer is expressed as: In the formula: Indicates the first The intermediate layer weight adjustment amount for input and output sample data. Indicates the correction factor. Indicates the first For the input and output sample data at the th The target output of each output layer node Indicates the first For the input and output sample data at the th The network output of each output layer node , They represent the first , For the input and output sample data at the th The intermediate layer node to the first The weights of each output layer node. Indicates the first For the input and output sample data at the th The input of each output layer node, This indicates the number of adjustable weights in the network; therefore, the intermediate layer weights are adjusted as follows: , This represents the learning rate, with a value of 0.

006.

9. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 7, characterized in that: In the sample training module, gradient descent is used to optimize the weights. Therefore, the linear processing function of the output layer is expressed as: The weight adjustment of the output layer is expressed as: In the formula: Indicates the first For the input and output sample data at the th If there are several intermediate layer nodes as input, then the output layer weights are adjusted as follows: , This represents the learning rate, with a value of 0.

006.

10. The photovoltaic cluster output prediction system based on photovoltaic-side micro-meteorological monitoring according to claim 1, characterized in that: The performance evaluation module uses root mean square error to evaluate the generalization ability of the neural network, specifically as follows: In the formula, This represents the root mean square error. This represents the target value in the test sample. This represents the predicted value of the model after inverse normalization. This indicates the number of data samples.

Images