Coupling information separation and extraction multi-task learning method for multi-station photovoltaic power prediction

Abstract

The invention discloses a coupling information separation and extraction multi-task learning method for multi-station photovoltaic power prediction, and the method comprises the following steps: S1, calculating the power similarity, the meteorological similarity and the geographic position similarity according to the power information, the meteorological information and the geographic position information of a photovoltaic power station; the correlation between the photovoltaic power stations is comprehensively judged through the comprehensive similarity, and similar power station groups are recognized; s2, constructing a coupling information separation and extraction multi-task learning prediction model; 3, constructing a dynamic update loss function of the coupling information separation and extraction multi-task learning prediction model; and 4, training and selecting specific parameters of the coupling information separation and extraction multi-task learning prediction model. According to the method, the task expert network for specially extracting the specific information of each power station and the shared network for extracting the coupling information among the plurality of power stations are constructed, so that the specific task information can be separated from the coupling information, and the accuracy and robustness of the prediction model are improved.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic power generation prediction, and more specifically, to a coupled information separation and extraction multi-task learning method for multi-station photovoltaic power prediction. Background Art

[0002] Photovoltaic power generation forecasting is essential for optimizing energy management and grid stability. Accurate forecasting can enhance the integration of solar energy into the power system, reduce dependence on fossil fuels, and minimize operating costs. By forecasting photovoltaic power generation, power companies can better balance supply and demand and improve the utilization of renewable energy.

[0003] Most of the existing photovoltaic power prediction methods only consider the prediction of a single photovoltaic power station, ignoring the correlation between different photovoltaic power stations. Since different photovoltaic power stations in the same region have similar geographical conditions and weather patterns, there are spatiotemporal coupling characteristics between photovoltaic power stations. If this coupling relationship can be properly considered, it will help improve the prediction accuracy of photovoltaic power. In the field of power forecasting, multi-task learning is considered to be an effective method for mining the coupling relationship between different objects. However, traditional multi-task learning models cannot distinguish between the unique information of different prediction tasks and the coupling information between tasks, resulting in information interference in the prediction process and generally poor prediction results. Summary of the invention

[0004] In view of the problem that traditional multi-task learning models cannot distinguish between the unique information of different prediction tasks and the coupling information between tasks, the present invention provides a multi-task learning method for separating and extracting coupling information for multi-station photovoltaic power prediction. This method can simultaneously predict the output power of multiple related photovoltaic power stations. During the prediction process, the unique information of each photovoltaic power station can be separated from the coupling information between multiple photovoltaic power stations to avoid interference between information, thereby improving the accuracy and robustness of the prediction model.

[0005] To achieve the above object, the present invention provides a multi-task learning method for coupling information separation and extraction of multi-station photovoltaic power prediction, comprising the following steps:

[0006] Step S1: Calculate the power similarity, meteorological similarity, and geographical location similarity respectively according to the power information, meteorological information, and geographical location information of the photovoltaic power station, and comprehensively judge the correlation between photovoltaic power stations through the comprehensive similarity to identify similar power station groups;

[0007] Step S2: construct a coupled information separation and extraction multi-task learning prediction model, which consists of an input layer, a feature extraction network, a prediction network, a gating network, and an output layer;

[0008] The input layer inputs a prediction input feature set of a group of similar power stations;

[0009] The feature extraction network includes a task expert network for extracting specific information of each power station and a shared network for extracting coupling information between multiple power stations; the task expert network and the shared network are composed of multiple fully connected layer neural networks.

[0010] The gating network selectively fuses the outputs of the task expert network and the shared network; the output of the gating network for the kth task in the jth layer feature extraction network can be expressed as:

[0011] D k,j (x) = w D k,j (D k,j-1 (x))S k,j (x) (5);

[0012] Where: k represents the task number; x is the input data; w D k,j is the weight function of task k in the j-th layer feature extraction network, calculated by the Softmax function: D k,j =Softmax(W k,j x), where W k,j is the parameter matrix;

[0013] The prediction network uses the final output of the gated network as its input feature, which includes both the unique information of each photovoltaic power station and the coupling information between different power stations; the prediction network is composed of a long short-term memory neural network.

[0014] The output layer is the power prediction value of each power station;

[0015] Step 3: Construct a dynamically updated loss function for the coupled information separation extraction multi-task learning prediction model;

[0016] Step 4: Train and select specific parameters for the coupled information separation extraction multi-task learning prediction model.

[0017] Furthermore, the power similarity is calculated, and the power similarity between different photovoltaic power plants is calculated using, for example, the Pearson correlation coefficient:

[0018]

[0019] Among them, P i and Q i These are the power data of the two power stations. and is their mean.

[0020] Furthermore, the meteorological similarity is calculated and the multidimensional vector similarity metric is used to calculate the similarity of meteorological data:

[0021]

[0022] Among them, M i 、N i are the meteorological data vectors of the two power stations respectively.

[0023] Furthermore, when calculating the geographical location similarity, the Haversine formula is used to calculate the geographical distance between photovoltaic power stations:

[0024]

[0025] Among them, φ and λ are longitude and latitude respectively, and r is the radius of the earth.

[0026] Furthermore, the comprehensive similarity is calculated:

[0027] S=w1·S power +w2·S weather +w3·S location (4);

[0028] Among them, S power , S location and S weather are the similarities of power, weather and geographical location respectively, and w1, w2 and w3 are their weights.

[0029] Furthermore, in step S1, similar power station groups are identified, specifically by using the DBSCAN clustering algorithm to group the power stations.

[0030] Furthermore, the prediction input feature set includes: photovoltaic power historical data, prediction time, irradiance, temperature, humidity, cloud cover, and rainfall.

[0031] Furthermore, the dynamically updated loss function of the coupled information separation extraction multi-task learning prediction model is:

[0032]

[0033] Where: L k loss is the loss function of the kth task; a k is the weight of the kth loss function. The weight of the loss function is dynamically updated according to the rate of change:

[0034] a k (t) = a k,0 ×γ k (7);

[0035] Where: ak (t) is the weight of the kth loss function at the tth iteration; a k,0 for a k The initial value of γ k is the rate of change.

[0036] Furthermore, the Adam optimization algorithm is used to solve the model parameters, and the cross-validation method is used to optimize the selection of hyperparameters such as learning rate, number of expert networks, number of feature extraction network layers, and number of network layers.

[0037] The beneficial effects of the present invention are as follows:

[0038] 1. The present invention proposes a multi-task learning prediction method for coupling information separation and extraction for multi-station photovoltaic power prediction. This method constructs a task expert network that specifically extracts the unique information of each power station, and a shared network that extracts the coupling information between multiple power stations. It can separate the task-specific information from the coupling information to avoid interference between the information, thereby improving the accuracy and robustness of the prediction model.

[0039] 2. In order to solve the problem that it is difficult to determine the weights of different prediction tasks, the present invention proposes a loss function that is dynamically updated according to a certain rate of change, which helps to find the most appropriate task weights and further improve the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for describing the embodiments are briefly introduced below.

[0041] Figure 1 This is a flow chart of a multi-task learning method for coupling information separation and extraction for multi-station photovoltaic power prediction according to an embodiment of the present invention;

[0042] Figure 2 This is a diagram of the multi-task learning framework for coupling information separation and extraction proposed in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they cannot be understood as limitations on the present invention.

[0045] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0046] like Figure 1 As shown, an embodiment of the present invention provides a multi-task learning method for coupling information separation and extraction for multi-station photovoltaic power prediction, comprising the following steps:

[0047] Step S1: Calculate the power similarity, meteorological similarity, and geographical location similarity respectively according to the power information, meteorological information, and geographical location information of the photovoltaic power station, and comprehensively judge the correlation between photovoltaic power stations through the comprehensive similarity to identify similar power station groups;

[0048] Specifically, when calculating power similarity, the Pearson correlation coefficient is used to calculate the power similarity between different photovoltaic power stations:

[0049]

[0050] Among them, P i and Q i These are the power data of the two power stations. and is their mean.

[0051] When calculating meteorological similarity, multidimensional vector similarity metrics (such as Euclidean distance and cosine similarity) are used to calculate the similarity of meteorological data:

[0052]

[0053] Among them, M i 、N i are the meteorological data vectors of the two power stations respectively.

[0054] When calculating the geographical location similarity, the Haversine formula is used to calculate the geographical distance between photovoltaic power stations:

[0055]

[0056] Among them, φ and λ are longitude and latitude respectively, and r is the radius of the earth.

[0057] Calculate the comprehensive similarity:

[0058] S=w1·S power +w2·S weather +w3·S location (4);

[0059] Among them, S power , S location and S weather are the similarities of power, weather and geographical location respectively, and w1, w2 and w3 are their weights.

[0060] The DBSCAN clustering algorithm is used to group the power plants and identify similar power plant groups.

[0061] Step S2: Construct a coupled information separation extraction multi-task learning prediction model. The coupled information separation extraction multi-task learning prediction model consists of an input layer, a feature extraction network, a prediction network, a gating network, and an output layer, as shown in the attached figure. Figure 2 shown.

[0062] The input layer inputs the prediction input feature set of similar power station groups, including: historical data of photovoltaic power, prediction time, irradiance, temperature, humidity, cloud cover, rainfall, etc.

[0063] The feature extraction network includes a task expert network that extracts the unique information of each power station and a shared network that extracts the coupling information between multiple power stations. The task expert network and the shared network are composed of multiple fully connected layer neural networks.

[0064] The gating network selectively fuses the outputs of the task expert network and the shared network. The output of the gating network for the kth task in the jth layer feature extraction network can be expressed as:

[0065] D k,j (x) = w D k,j (D k,j-1 (x))S k,j (x) (5);

[0066] Where: k represents the task number; x is the input data; w D k,jis the weight function of task k in the j-th layer feature extraction network, calculated by the Softmax function: D k,j =Softmax(W k,j x), where W k,j is the parameter matrix.

[0067] The prediction network uses the final output of the gating network as its input features, which include both the unique information of each photovoltaic power station and the coupling information between different power stations. The prediction network is composed of a long short-term memory neural network.

[0068] The output layer is the power prediction value of each power station.

[0069] Step 3: Construct a dynamically updated loss function for the coupled information separation extraction multi-task learning prediction model.

[0070] Specifically, the total loss function of the model is:

[0071]

[0072] Where: L k loss is the loss function of the kth task; a k is the weight of the kth loss function. The weight of the loss function is dynamically updated according to the rate of change:

[0073] a k (t) = a k,0 ×γ k (7);

[0074] Where: a k (t) is the weight of the kth loss function at the tth iteration; a k,0 for a k The initial value of γ k is the rate of change.

[0075] Step 4: Train and select specific parameters of the coupled information separation extraction multi-task learning prediction model. The Adam optimization algorithm is used to solve the model parameters, and the cross-validation method is used to optimize the hyperparameters such as learning rate, number of expert networks, number of feature extraction network layers, and number of network layers.

[0076] Implementation comparison: A real photovoltaic data set from a certain place was used for example verification. Three similar photovoltaic power stations were selected through correlation analysis, and the method of the present invention was used to predict the three photovoltaic power stations. The LSTM algorithm was used as a comparison method. The mean absolute percentage error (MAPE) was used to evaluate the prediction accuracy of the method. The prediction results are: the prediction MAPE of the LSTM method for the three photovoltaic power stations are 8.83%, 9.02%, and 14.21%, respectively, while the prediction MAPE of the method of the present invention are 7.25%, 8.41%, and 6.98%, respectively, which is better than the comparison, indicating the effectiveness and advancement of the method of the present invention.

[0077] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A multi-task learning method for coupling information separation and extraction for multi-station photovoltaic power prediction, characterized in that: The following steps are involved: Step S1: Calculate the power similarity, meteorological similarity, and geographical location similarity respectively according to the power information, meteorological information, and geographical location information of the photovoltaic power station, and comprehensively judge the correlation between photovoltaic power stations through the comprehensive similarity to identify similar power station groups; Step S2: construct a coupled information separation and extraction multi-task learning prediction model, which consists of an input layer, a feature extraction network, a prediction network, a gating network, and an output layer; The input layer inputs a prediction input feature set of a group of similar power stations; The feature extraction network includes a task expert network for extracting specific information of each power station and a shared network for extracting coupling information between multiple power stations; the task expert network and the shared network are composed of multiple fully connected layer neural networks. The gating network selectively fuses the outputs of the task expert network and the shared network; the output of the gating network for the kth task in the jth layer feature extraction network can be expressed as: D k,j (x)=w D k,j (D k,j-1 (x))S k,j (x) (5) Where: k represents the task number; x is the input data; w D k,j is the weight function of task k in the j-th layer feature extraction network, calculated by the Softmax function: D k,j =Softmax(W k,j x), where W k,j is the parameter matrix; The prediction network uses the final output of the gated network as its input feature, which includes both the unique information of each photovoltaic power station and the coupling information between different power stations; the prediction network is composed of a long short-term memory neural network. The output layer is the power prediction value of each power station; Step 3: Construct a dynamically updated loss function for the coupled information separation extraction multi-task learning prediction model; Step 4: Train and select specific parameters for the coupled information separation extraction multi-task learning prediction model.

2. According to claim 1, a multi-task learning method for coupling information separation and extraction for multi-station photovoltaic power prediction is characterized by: Calculate the power similarity, such as using the Pearson correlation coefficient to calculate the power similarity between different photovoltaic power plants: Among them, P i and Q i These are the power data of the two power stations. and is their mean.

3. According to claim 1, a multi-task learning method for coupling information separation and extraction for multi-station photovoltaic power prediction is characterized by: Calculate meteorological similarity and use multidimensional vector similarity metric to calculate the similarity of meteorological data: Among them, M i 、N i are the meteorological data vectors of the two power stations respectively.

4. According to claim 1, a multi-task learning method for coupling information separation and extraction for multi-station photovoltaic power prediction is characterized by: When calculating the geographical location similarity, the Haversine formula is used to calculate the geographical distance between photovoltaic power stations: Among them, φ and λ are longitude and latitude respectively, and r is the radius of the earth.

5. The method for separating and extracting coupled information for multi-station photovoltaic power prediction according to claim 1 is characterized by: Calculate the comprehensive similarity: S=w1·S power +w2·S weather +w3·S location (4); Among them, S power , S location and S weather are the similarities of power, weather and geographical location respectively, and w1, w2 and w3 are their weights.

6. The method for separating and extracting coupled information for multi-station photovoltaic power prediction according to claim 1 is characterized by: In step S1, similar power station groups are identified, specifically by using the DBSCAN clustering algorithm to group the power stations.

7. The method for separating and extracting coupled information for multi-station photovoltaic power prediction according to claim 1 is characterized by: The prediction input feature set includes: photovoltaic power historical data, prediction time, irradiance, temperature, humidity, cloud cover, and rainfall.

8. The coupled information separation and extraction multi-task learning method for multi-station photovoltaic power prediction according to claim 1 is characterized by: The dynamically updated loss function of the coupled information separation extraction multi-task learning prediction model is: Where: L k loss is the loss function of the kth task; a k is the weight of the kth loss function. The weight of the loss function is dynamically updated according to the rate of change: a k (t)=a k,0 ×γ k (7); Where: a k (t) is the weight of the kth loss function at the tth iteration; a k,0 for a k The initial value of γ k is the rate of change.

9. The method for separating and extracting coupled information for multi-station photovoltaic power prediction according to claim 1 is characterized by: The Adam optimization algorithm is used to solve the model parameters, and the cross-validation method is used to optimize the hyperparameters such as learning rate, number of expert networks, number of feature extraction network layers, and number of network layers.

Images