Source-load joint prediction method and device, electronic equipment and storage medium

Abstract

The invention discloses a source-load joint prediction method and device, electronic equipment and a storage medium, and relates to the technical field of power equipment. The method comprises the following steps: acquiring photovoltaic output, electric heating and cold multi-element load data and historical meteorological data in a target time period from a regional integrated energy system; determining candidate meteorological indexes influencing photovoltaic output and electric heating and cold multi-element load data; screening the candidate meteorological indexes according to the influence degree of the candidate meteorological indexes on the photovoltaic output and the electric heating and cooling multi-element load data to obtain a target meteorological index; extracting target meteorological data corresponding to the target meteorological index from the historical meteorological data; and determining a photovoltaic output predicted value, an electric load predicted value, a thermal load predicted value and a cold load predicted value in a future time period according to the target meteorological data, the photovoltaic output and the electric-thermal-cold multi-element load data in the target time period through the source-load combined prediction model. According to the invention, the accuracy of photovoltaic output and load prediction can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power equipment, and in particular to a source-load joint prediction method, device, electronic equipment and storage medium. Background Art

[0002] In the context of coping with the dual crises of energy and environment, the country attaches great importance to the transformation and upgrading of the energy industry. The traditional centralized power supply model has been unable to adapt to the complex and changeable energy forms and diversified utilization needs. As a result, the regional integrated energy system has emerged. By integrating cold, heat, electricity, gas and renewable energy, it has achieved efficient coupling and coordinated control of energy, significantly improving energy utilization efficiency and system flexibility. However, with the increase in the proportion of renewable energy such as wind and solar power connected to the grid and the widespread access to flexible loads, the regional integrated energy system faces the challenge of high uncertainty on both the source and load sides, which puts higher requirements on the prediction accuracy of new energy and load. Therefore, how to accurately predict the source and load has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a source-load joint prediction method, device, electronic equipment and storage medium.

[0004] According to another aspect of the present invention, a source-load joint prediction method is provided, comprising:

[0005] Obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within the target time period from the regional integrated energy system;

[0006] Determine candidate meteorological indicators that affect photovoltaic output and electric heating and cooling multi-load data;

[0007] According to the degree of influence of candidate meteorological indicators on photovoltaic output and electric heating and cooling multi-load data, candidate meteorological indicators are screened to obtain target meteorological indicators;

[0008] Extracting target meteorological data corresponding to target meteorological indicators from historical meteorological data;

[0009] Through the source-load joint prediction model, according to the target meteorological data, photovoltaic output and electric, heating and cooling multi-load data in the target time period, the photovoltaic output forecast value, electric load forecast value, thermal load forecast value and cooling load forecast value in the future time period are determined.

[0010] In some embodiments, the source-load joint prediction model is used to determine the photovoltaic output prediction value, the electric load prediction value, the heat load prediction value, and the cold load prediction value in the future time period according to the target meteorological data, photovoltaic output, and electric heat and cold multi-load data in the target time period, including:

[0011] Through the multi-channel feature extraction module of the source-load joint prediction model, feature extraction and processing are performed on the target meteorological data, photovoltaic output, and historical multi-load data in parallel to obtain channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics;

[0012] Through the attention mechanism of the source-load joint prediction model, the extracted channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are enhanced and integrated;

[0013] Through the multi-task prediction module of the source-load joint prediction model, the fused features are jointly predicted to obtain the photovoltaic output prediction value, electric load prediction value, thermal load prediction value and cooling load prediction value in the future time period.

[0014] In some embodiments, the attention mechanism is a CBAM attention mechanism, and the CBAM attention mechanism includes a channel attention module and a spatial attention module;

[0015] Accordingly, the extracted channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are enhanced and integrated through the attention mechanism of the source-load joint prediction model, including:

[0016] Through the channel attention module, the channel weights corresponding to the meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are determined, and the weighted enhancement processing of the channel dimension is performed according to the corresponding channel weights;

[0017] Determine a spatial attention map through a spatial attention module; wherein the spatial attention map records the weights corresponding to different spatial positions;

[0018] Based on the spatial attention map, the features after weighted enhancement are spatially enhanced and then fused.

[0019] In some embodiments, the multi-channel feature extraction module is composed of a multi-channel convolutional neural network; the multi-task prediction module is composed of a model architecture based on a bidirectional long short-term memory network and multi-task learning.

[0020] In some embodiments, candidate meteorological indicators are screened according to their influence on photovoltaic output and electric heating and cooling multi-load data to obtain target meteorological indicators, including:

[0021] For any moment in the target time period, the meteorological data corresponding to all candidate meteorological indicators at that moment and the photovoltaic output at that moment are taken as a training sample; the meteorological data corresponding to all candidate meteorological indicators at that moment and the electric heating and cooling multi-load data at that moment are taken as a training sample;

[0022] Through the determined training samples, the constructed data classification model is classified and trained to obtain the influence of each candidate meteorological index on photovoltaic output and the influence of each candidate meteorological index on the multi-load data of electricity, heating and cooling;

[0023] The candidate meteorological indicators whose impact on photovoltaic output is greater than the threshold are taken as target meteorological indicators; the candidate meteorological indicators whose impact on electric, heating and cooling multi-load data is greater than the threshold are taken as target meteorological indicators.

[0024] In some embodiments, the candidate meteorological indicators include at least one of the following: irradiance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, and rainfall.

[0025] According to another aspect of the present invention, a source-load joint prediction device is provided, comprising:

[0026] The data acquisition module is used to obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within the target time period from the regional integrated energy system;

[0027] The first indicator determination module is used to determine candidate meteorological indicators that affect photovoltaic output and electric heating and cooling multi-load data;

[0028] The second indicator determination module is used to screen candidate meteorological indicators according to the degree of influence of candidate meteorological indicators on photovoltaic output and electric heating and cooling multi-load data to obtain target meteorological indicators;

[0029] A data screening module is used to extract target meteorological data corresponding to target meteorological indicators from historical meteorological data;

[0030] The prediction module is used to determine the photovoltaic output prediction value, electric load prediction value, thermal load prediction value, and cooling load prediction value in the future time period based on the target meteorological data, photovoltaic output, and electric, heating, and cooling multi-load data in the target time period through the source-load joint prediction model.

[0031] In some embodiments, the prediction module includes:

[0032] The feature extraction unit is used to perform feature extraction processing on the target meteorological data, photovoltaic output, and historical multi-load data through the multi-channel feature extraction module of the source-load joint prediction model, so as to obtain channel-level meteorological factor features, photovoltaic output features, and multi-load features;

[0033] The enhancement unit is used to enhance and fuse the extracted channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics through the attention mechanism of the source-load joint prediction model;

[0034] The prediction unit is used to jointly predict the fused features through the multi-task prediction module of the source-load joint prediction model to obtain the photovoltaic output prediction value, electric load prediction value, thermal load prediction value and cooling load prediction value in the future time period.

[0035] In some embodiments, the attention mechanism is a CBAM attention mechanism, and the CBAM attention mechanism includes a channel attention module and a spatial attention module;

[0036] Accordingly, the enhancement unit is specifically used for:

[0037] Through the channel attention module, the channel weights corresponding to the meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are determined, and the weighted enhancement processing of the channel dimension is performed according to the corresponding channel weights;

[0038] Determine a spatial attention map through a spatial attention module; wherein the spatial attention map records the weights corresponding to different spatial positions;

[0039] Based on the spatial attention map, the features after weighted enhancement are spatially enhanced and then fused.

[0040] In some embodiments, the multi-channel feature extraction module is composed of a multi-channel convolutional neural network; the multi-task prediction module is composed of a model architecture based on a bidirectional long short-term memory network and multi-task learning.

[0041] In some embodiments, the second indicator determination module is specifically used to:

[0042] For any moment in the target time period, the meteorological data corresponding to all candidate meteorological indicators at that moment and the photovoltaic output at that moment are taken as a training sample; the meteorological data corresponding to all candidate meteorological indicators at that moment and the electric heating and cooling multi-load data at that moment are taken as a training sample;

[0043] Through the determined training samples, the constructed data classification model is classified and trained to obtain the influence of each candidate meteorological index on photovoltaic output and the influence of each candidate meteorological index on the multi-load data of electricity, heating and cooling;

[0044] The candidate meteorological indicators whose impact on photovoltaic output is greater than the threshold are taken as target meteorological indicators; the candidate meteorological indicators whose impact on electric, heating and cooling multi-load data is greater than the threshold are taken as target meteorological indicators.

[0045] In some embodiments, the candidate meteorological indicators include at least one of the following: irradiance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, and rainfall.

[0046] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0047] at least one processor; and

[0048] a memory communicatively connected to at least one processor; wherein,

[0049] The memory stores a computer program that can be executed by at least one processor. The computer program is executed by at least one processor so that the at least one processor can execute the source-load joint prediction method of the embodiment of the present invention.

[0050] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the source-load joint prediction method according to an embodiment of the present invention when executed.

[0051] The technical solution of the embodiment of the present invention can improve the accuracy of predicting photovoltaic output and load, and further can effectively support the optimized scheduling and energy management of the regional integrated energy system based on the prediction results.

[0052] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0054] Figure 1 It is a flow chart of a source-load joint prediction method provided by an embodiment of the present invention;

[0055] Figure 2a It is a flow chart of another source-load joint prediction method provided by an embodiment of the present invention;

[0056] Figure 2b It is a structural diagram of the MCNN model provided by an embodiment of the present invention;

[0057] Figure 2c It is a logical schematic diagram of a source-load joint prediction method provided by an embodiment of the present invention;

[0058] Figure 3 It is a structural schematic diagram of a source-load joint prediction device provided by an embodiment of the present invention;

[0059] Figure 4 It is a structural schematic diagram of an electronic device for implementing the source-load joint prediction method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0060] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0061] Embodiment 1

[0062] Figure 1 A flow chart of a source-load joint prediction method is provided for an embodiment of the present invention. This embodiment is applicable to scenarios in which the photovoltaic output and multiple loads of a regional integrated energy system are jointly predicted. The method can be executed by a source-load joint prediction device, which can be implemented in the form of hardware and / or software. The source-load joint prediction device can be configured in an electronic device.

[0063] like Figure 1 As shown, the source-load joint prediction method includes:

[0064] S101. Obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within a target time period from the regional integrated energy system.

[0065] In the embodiment of the present invention, the regional integrated energy system refers to a system that realizes efficient coupling and coordinated control of energy by integrating cold, heat, electricity, gas and renewable energy; the target time period refers to a period of time before the current moment, and the specific duration can be set according to actual needs.

[0066] In the regional integrated energy system, photovoltaic output refers to the ability of the solar photovoltaic power generation system in the region to actually generate electricity or the output power of electricity under specific time and conditions. Therefore, the photovoltaic output obtained in the target time period includes information such as the actual photovoltaic power generation or power generation at different times in the target time period.

[0067] In the regional integrated energy system, the electric, heat and cold multi-load data refers to the various types of data recorded for the consumption demand of three different forms of energy, namely electricity, heat and cold, within the specific area. It reflects the usage of electricity, heat and cold energy by different users in the area at different times. Therefore, the acquired electric, heat and cold multi-load data can include electric load data at different times (such as residential, commercial and industrial power and electricity consumption, etc.), heat load data (such as heating, hot water supply and other related parameters), and cold load data (such as air conditioning cooling power and operation time, etc.).

[0068] Historical meteorological data refers to data describing meteorological conditions within a target time period. The historical meteorological data within the target time period may be recorded by a device constituting the regional energy management system, or may be obtained from an external system, and is not specifically limited here.

[0069] In the embodiment of the present invention, to perform source-load prediction, it is necessary to first obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within the target time period from the regional energy management system; and then implement the prediction based on the three types of data obtained according to the following steps.

[0070] S102. Determine candidate meteorological indicators that affect photovoltaic output and electric heating and cooling multi-load data.

[0071] In an embodiment of the present invention, the correlation analysis method can be used to calculate the correlation between photovoltaic output and different meteorological indicators, as well as the correlation between electric, heating and cooling multi-load data and different meteorological indicators; and then, based on the correlation calculation results, the candidate meteorological indicators that affect photovoltaic output and electric, heating and cooling multi-load data are determined. For example, if the Pearson correlation coefficient between solar radiation illuminance and photovoltaic output is calculated to be 0.85, it means that there is a strong positive correlation between the two, that is, the higher the solar radiation illuminance, the higher the photovoltaic output is generally. For another example, the Pearson correlation coefficient between humidity and residential air-conditioning refrigeration and cooling load is calculated. If it is 0.6, it indicates that there is a certain positive correlation between humidity and residential air-conditioning refrigeration and cooling load. The higher the humidity, the greater the residential air-conditioning refrigeration and cooling load may be. In this way, radiation illuminance and humidity can be used as candidate climate indicators.

[0072] It is understandable that the above-mentioned radiance and humidity are merely examples, and the determined candidate meteorological indicators may include at least one of the following: radiance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, and rainfall.

[0073] S103. Screen candidate meteorological indicators according to their influence on photovoltaic output and electric heating and cooling multi-load data to obtain target meteorological indicators.

[0074] In an embodiment of the present invention, different model frameworks can be selected to evaluate the degree of influence of candidate meteorological indicators on photovoltaic output and electric heating and cooling multi-load data. Optional models include linear regression models, decision tree models (such as CART decision trees), random forest models, XGBoost models, etc. Then, some candidate meteorological indicators with a greater degree of influence on photovoltaic output and electric heating and cooling multi-load data are selected as target meteorological indicators. For example, the candidate meteorological indicators whose influence is greater than a preset threshold are used as target meteorological indicators; for example, the candidate meteorological indicators whose influence ranks in the top n are used as target meteorological indicators, where n is a preset constant.

[0075] S104. Extract target meteorological data corresponding to the target meteorological index from the historical meteorological data.

[0076] Optionally, historical meteorological data within a target time period is filtered according to target meteorological indicators to obtain required target meteorological data.

[0077] S105. Determine the photovoltaic output forecast value, electric load forecast value, heat load forecast value, and cooling load forecast value in the future time period through the source-load joint forecast model according to the target meteorological data, photovoltaic output, and electric, heat, and cooling multi-load data in the target time period.

[0078] In an embodiment of the present invention, the source-load joint prediction model is a model architecture for collaboratively predicting energy supply sources (sources) and energy consumption loads (loads) in a regional integrated energy system. The model may include a multi-channel feature extraction module (e.g., composed of a multi-channel convolutional neural network MCNN), an attention mechanism (e.g., a CBAM attention mechanism), and a multi-task prediction module (e.g., a model architecture based on a bidirectional long short-term memory network and multi-task learning).

[0079] After introducing the structure of the source-load joint prediction model, the source-load joint prediction model is used to determine the photovoltaic output prediction value, electric load prediction value, heat load prediction value, and cold load prediction value in the future time period according to the target meteorological data, photovoltaic output, and electric, heating, and cooling multi-load data in the target time period, including: through the multi-channel feature extraction module of the source-load joint prediction model, feature extraction and processing are performed on the target meteorological data, photovoltaic output, and historical multi-load data in parallel to obtain channel-level meteorological factor features, photovoltaic output features, and multi-load features; through the attention mechanism of the source-load joint prediction model, the extracted channel-level meteorological factor features, photovoltaic output features, and multi-load features are enhanced and fused; through the multi-task prediction module of the source-load joint prediction model, the fused features are jointly predicted to obtain the photovoltaic output prediction value, electric load prediction value, heat load prediction value, and cooling load prediction value in the future time period.

[0080] In the embodiment of the present invention, by screening meteorological indicators, the meteorological data with the greatest impact on photovoltaic output and multiple loads are selected, and then combined with the photovoltaic output and multiple loads in the target time period, the photovoltaic output, electric load, thermal load, and cooling load in the future time period are predicted. Compared with the isolated source-load prediction, the prediction accuracy can be improved; at the same time, according to the prediction results, a solid technical support can be provided for the optimized scheduling and efficient operation of the regional energy system.

[0081] Embodiment 2

[0082] Figure 2a A flow chart of a source-load joint prediction method is provided for an embodiment of the present invention. Figure 2a , the method comprises the following steps:

[0083] S201. Obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within a target time period from the regional integrated energy system.

[0084] S202. Determine candidate meteorological indicators that affect photovoltaic output and electric heating and cooling multi-load data.

[0085] The candidate meteorological indicators may include at least one of the following: irradiance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, and rainfall.

[0086] In some embodiments, candidate meteorological indicators are screened according to their influence on photovoltaic output and electric heating and cooling multi-load data, and the process of obtaining target meteorological indicators can refer to steps S203-S205.

[0087] S203. For any moment in the target time period, the meteorological data corresponding to all candidate meteorological indicators at that moment and the photovoltaic output at that moment are taken as a training sample; the meteorological data corresponding to all candidate meteorological indicators at that moment and the electric heating and cooling multi-load data at that moment are taken as a training sample.

[0088] S204. Classification training is performed on the constructed data classification model using the determined training samples to obtain the influence of each candidate meteorological indicator on photovoltaic output and the influence of each candidate meteorological indicator on the electric, heating and cooling multi-load data.

[0089] In an embodiment of the present invention, the data classification model may be an XGBoost model. When analyzing the degree of influence of photovoltaic output, because we want to understand the degree of influence of candidate meteorological indicators on it, the photovoltaic output data is defined as the target variable (i.e., dependent variable), and the candidate meteorological indicator data (radiant illuminance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, rainfall, etc.) is defined as a characteristic variable (i.e., independent variable). When analyzing the degree of influence of electric, thermal, and cold multivariate load data, the electric load, thermal load, and cold load data are defined as target variables, and the candidate meteorological indicator data are also defined as characteristic variables. Further, using the training set data, according to the target variables and characteristic variables defined above, the data is input into the configured XGBoost model for training. During the training process, the XGBoost model adjusts the parameters of the model (such as the weight of the decision tree, etc.) according to the set loss function (such as mean square error, mean absolute error, etc.) so that the model can fit the relationship between the data as accurately as possible, that is, minimize the gap between the predicted value and the true value.

[0090] During the training process, the XGBoost model automatically calculates the importance of each feature (i.e., candidate meteorological indicator) to the model prediction results and presents it in the form of a score, which is the feature importance score. The higher the feature importance score, the greater the impact of the candidate meteorological indicator on the corresponding data. Therefore, by looking at these feature importance scores, you can directly understand the impact of each candidate meteorological indicator on photovoltaic output and on the multi-load data of electric heating and cooling.

[0091] S205. The candidate meteorological index having an impact on photovoltaic output greater than a threshold is used as the target meteorological index; the candidate meteorological index having an impact on electric, heating and cooling multi-load data greater than a threshold is used as the target meteorological index.

[0092] In this embodiment, in addition to screening meteorological indicators according to the threshold value, a constant n can be preset to select the top n meteorological indicators in terms of impact as the target meteorological parameters.

[0093] In an optional implementation, in order to ensure the accuracy of the screened target meteorological indicators, some candidate meteorological indicators whose impact on photovoltaic output is greater than a threshold and some candidate meteorological indicators whose impact on electric, heating and cooling multi-load data is greater than a threshold are screened out, and then the training samples are reconstructed using the screened candidate meteorological indicators. Any newly constructed training sample includes the meteorological data of some of the screened candidate meteorological indicators at a certain moment and the photovoltaic output at that moment, or a training sample includes the meteorological data of some of the screened candidate meteorological indicators at a certain moment and the electric, heating and cooling multi-load data at that moment; according to the newly constructed training samples, the training and influence calculation are performed according to the steps of step S204, and the indicator screening is performed according to step S205. The above process can be repeated multiple times to determine the final target meteorological indicator.

[0094] S206: Extract target meteorological data corresponding to the target meteorological index from the historical meteorological data.

[0095] S207. Determine the photovoltaic output forecast value, electric load forecast value, heat load forecast value, and cooling load forecast value in the future time period through the source-load joint forecast model according to the target meteorological data, photovoltaic output, and electric, heat, and cooling multi-load data in the target time period.

[0096] In the embodiment of the present invention, the source-load joint prediction model includes a multi-channel feature extraction module composed of a multi-channel convolutional neural network MCNN, a CBAM attention mechanism, a multi-task prediction module composed of a model architecture based on a bidirectional long short-term memory network BiLSTM and multi-task learning MTL. On this basis, the prediction process includes the following steps:

[0097] Step 1: Through the multi-channel feature extraction module of the source-load joint prediction model, feature extraction and processing are performed on the target meteorological data, photovoltaic output, and historical multi-load data in parallel to obtain channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics.

[0098] For example, see Figure 2b , which shows a schematic diagram of the structure of the multi-channel convolutional neural network MCNN. The multi-channel convolutional neural network MCNN includes 3 channels, each of which is composed of a convolution layer, a batch normalization layer, an activation function layer, a pooling layer, and a fully connected layer. Channel 1 can extract features from the input target meteorological data to obtain meteorological factor features; channel 2 can extract features from the input photovoltaic output to obtain photovoltaic output features; channel 3 can extract features from the input electric heating and cooling multi-loads to obtain multi-load features.

[0099] Step 2: Through the attention mechanism of the source-load joint prediction model, the extracted channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are enhanced and integrated.

[0100] Optionally, the attention mechanism is a CBAM attention mechanism, which includes a channel attention module and a spatial attention module; accordingly, the extracted channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are enhanced and fused through the attention mechanism of the source-load joint prediction model, including: determining the channel weights corresponding to the meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics through the channel attention module, and performing weighted enhancement processing on the channel dimension according to the corresponding channel weights; determining a spatial attention map through the spatial attention module; wherein the weights corresponding to different spatial positions are recorded in the spatial attention map; based on the spatial attention map, performing spatial enhancement processing on the features after weighted enhancement processing, and fusing the processed features.

[0101] Step 3: Through the multi-task prediction module of the source-load joint prediction model, the fused features are jointly predicted to obtain the photovoltaic output prediction value, electric load prediction value, thermal load prediction value and cooling load prediction value in the future time period.

[0102] Optionally, the multi-task prediction module is composed of a model architecture based on a bidirectional long short-term memory network BiLSTM and multi-task learning MTL, and the multi-task prediction module consists of an input layer, a shared BiLSTM layer, a fully connected layer, and an output layer. The multi-task prediction module adopts the hard parameter sharing mechanism in MTL, and inputs the fused meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics into the shared BiLSTM layer for feature extraction and sharing, and then generates independent outputs for different source-load prediction tasks through their respective fully connected layers. It can be understood that the BiLSTM network in the multi-task prediction module connects the forward and reverse LSTM networks to realize a bidirectional feature extraction mechanism for time series data. This mechanism not only covers the forward evolution characteristics of the time series, but also deeply captures its hidden information of reverse backtracking, thereby ensuring that the neural network is more global and complete in the feature extraction process.

[0103] In the embodiment of the present invention, by screening meteorological indicators, selecting the meteorological data with the greatest rigidity for photovoltaic output and multiple loads, and then combining the photovoltaic output and multiple loads in the target time period, the photovoltaic output, electric load, thermal load, and cooling load in the future time period are predicted. Compared with isolated source-load prediction, the prediction accuracy can be improved; at the same time, it provides solid technical support for the optimized scheduling and efficient operation of the regional energy system.

[0104] On the basis of the above, see Figure 2c, which shows the logical flow chart of source-load prediction. The specific process is as follows: First, determine the input data (including the photovoltaic output, electric heating and cooling multi-load data and historical meteorological data collected in the target time period), and then based on the input data, select the target meteorological indicators that meet the conditions for the influence of photovoltaic output and electric heating and cooling multi-load data through the XGBoost model; then extract the target meteorological data, photovoltaic output, and electric heating and cooling multi-load data corresponding to the target meteorological indicators in the target time period through the MCNN model; further, enhance and fuse the features extracted by MCNN through the attention CBAM layer, and the specific process can be found in the description of the above embodiment; further, input the enhanced and fused features into the BiLSTM shared layer; then, according to the output of the BiLSTM shared layer, determine the photovoltaic output prediction value, electric load prediction value, heat load prediction value, and cold load prediction value of the future time period through different fully connected layers.

[0105] Embodiment 3

[0106] Figure 3 The schematic diagram of the structure of a source-load joint prediction device provided by an embodiment of the present invention is shown in FIG. This embodiment is applicable to scenarios where photovoltaic output and multiple loads of a regional integrated energy system are predicted. Figure 3 As shown, the device comprises:

[0107] The data acquisition module 301 is used to acquire photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within a target time period from the regional integrated energy system;

[0108] The first indicator determination module 302 is used to determine candidate meteorological indicators that affect photovoltaic output and electric heating and cooling multi-load data;

[0109] The second indicator determination module 303 is used to screen candidate meteorological indicators according to the degree of influence of candidate meteorological indicators on photovoltaic output and electric heating and cooling multi-load data to obtain target meteorological indicators;

[0110] The data screening module 304 is used to extract target meteorological data corresponding to the target meteorological index from the historical meteorological data;

[0111] The prediction module 305 is used to determine the photovoltaic output prediction value, electric load prediction value, thermal load prediction value, and cooling load prediction value in the future time period based on the target meteorological data, photovoltaic output, and electric, heating, and cooling multi-load data in the target time period through the source-load joint prediction model.

[0112] In some embodiments, the prediction module 305 includes:

[0113] The feature extraction unit is used to perform feature extraction processing on the target meteorological data, photovoltaic output, and historical multi-load data through the multi-channel feature extraction module of the source-load joint prediction model, so as to obtain channel-level meteorological factor features, photovoltaic output features, and multi-load features;

[0114] The enhancement unit is used to enhance and fuse the extracted channel-level meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics through the attention mechanism of the source-load joint prediction model;

[0115] The prediction unit is used to jointly predict the fused features through the multi-task prediction module of the source-load joint prediction model to obtain the photovoltaic output prediction value, electric load prediction value, thermal load prediction value and cooling load prediction value in the future time period.

[0116] In some embodiments, the attention mechanism is a CBAM attention mechanism, and the CBAM attention mechanism includes a channel attention module and a spatial attention module;

[0117] Accordingly, the enhancement unit is specifically used for:

[0118] Through the channel attention module, the channel weights corresponding to the meteorological factor characteristics, photovoltaic output characteristics, and multi-load characteristics are determined, and the weighted enhancement processing of the channel dimension is performed according to the corresponding channel weights;

[0119] Determine a spatial attention map through a spatial attention module; wherein the spatial attention map records the weights corresponding to different spatial positions;

[0120] Based on the spatial attention map, the features after weighted enhancement are spatially enhanced and then fused.

[0121] In some embodiments, the multi-channel feature extraction module is composed of a multi-channel convolutional neural network; the multi-task prediction module is composed of a model architecture based on a bidirectional long short-term memory network and multi-task learning.

[0122] In some embodiments, the second indicator determination module 303 is specifically used to:

[0123] For any moment in the target time period, the meteorological data corresponding to all candidate meteorological indicators at that moment and the photovoltaic output at that moment are taken as a training sample; the meteorological data corresponding to all candidate meteorological indicators at that moment and the electric heating and cooling multi-load data at that moment are taken as a training sample;

[0124] Through the determined training samples, the constructed data classification model is classified and trained to obtain the influence of each candidate meteorological index on photovoltaic output and the influence of each candidate meteorological index on the multi-load data of electricity, heating and cooling;

[0125] The candidate meteorological indicators whose impact on photovoltaic output is greater than the threshold are taken as target meteorological indicators; the candidate meteorological indicators whose impact on electric, heating and cooling multi-load data is greater than the threshold are taken as target meteorological indicators.

[0126] In some embodiments, the candidate meteorological indicators include at least one of the following: irradiance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, and rainfall.

[0127] The source-load joint prediction device provided in the embodiment of the present invention can execute the source-load joint prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0128] Embodiment 4

[0129] Figure 4 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to be a controller device in a new energy vehicle. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0130] like Figure 4 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0131] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0132] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as executing a source-load joint prediction method.

[0133] In some embodiments, the source-load joint prediction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the source-load joint prediction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the source-load joint prediction method in any other appropriate manner (e.g., by means of firmware).

[0134] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0135] The computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable source-load joint prediction device, so that when the computer programs are executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer programs may be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0136] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0137] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0138] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0139] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0140] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0141] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A source-load joint prediction method, characterized in that: include: Obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within the target time period from the regional integrated energy system; Determine candidate meteorological indicators that affect the photovoltaic output and the electric, heating and cooling multi-load data; According to the influence degree of the candidate meteorological indicators on the photovoltaic output and the electric, heating and cooling multi-load data, the candidate meteorological indicators are screened to obtain target meteorological indicators; Extracting target meteorological data corresponding to the target meteorological index from the historical meteorological data; Through the source-load joint prediction model, according to the target meteorological data, the photovoltaic output and the electric, thermal and cooling multi-load data in the target time period, the photovoltaic output prediction value, the electric load prediction value, the thermal load prediction value and the cooling load prediction value in the future time period are determined.

2. The method according to claim 1, characterized in that The source-load joint prediction model determines the photovoltaic output prediction value, the electric load prediction value, the heat load prediction value, and the cold load prediction value in the future time period according to the target meteorological data, the photovoltaic output, and the electric, heat, and cold multi-load data in the target time period, including: Through the multi-channel feature extraction module of the source-load joint prediction model, feature extraction processing is performed on the target meteorological data, the photovoltaic output, and the historical multi-load data to obtain channel-level meteorological factor features, photovoltaic output features, and multi-load features; The extracted channel-level meteorological factor characteristics, the photovoltaic output characteristics, and the multi-load characteristics are enhanced and integrated through the attention mechanism of the source-load joint prediction model; Through the multi-task prediction module of the source-load joint prediction model, the fused features are jointly predicted to obtain the photovoltaic output prediction value, electric load prediction value, thermal load prediction value and cooling load prediction value in the future time period.

3. The method according to claim 2, characterized in that The attention mechanism is a CBAM attention mechanism, and the CBAM attention mechanism includes a channel attention module and a spatial attention module; Accordingly, the extracted channel-level meteorological factor features, photovoltaic output features, and multi-load features are enhanced and integrated through the attention mechanism of the source-load joint prediction model, including: Determine the channel weights corresponding to the meteorological factor characteristics, the photovoltaic output characteristics, and the multi-load characteristics through the channel attention module, and perform weighted enhancement processing of the channel dimension according to the corresponding channel weights; Determine a spatial attention map by using the spatial attention module; wherein the spatial attention map records weights corresponding to different spatial positions; Based on the spatial attention map, spatial enhancement processing is performed on the features after weighted enhancement processing, and the processed features are fused.

4. The method according to claim 2, characterized in that: The multi-channel feature extraction module is composed of a multi-channel convolutional neural network; the multi-task prediction module is composed of a model architecture based on a bidirectional long short-term memory network and multi-task learning.

5. The method according to claim 1, characterized in that: The candidate meteorological indicators are screened according to the influence of the candidate meteorological indicators on the photovoltaic output and the electric, heating and cooling multi-load data to obtain the target meteorological indicators, including: For any moment within the target time period, the meteorological data corresponding to all candidate meteorological indicators at that moment and the photovoltaic output at that moment are taken as a training sample; the meteorological data corresponding to all candidate meteorological indicators at that moment and the electric heating and cooling multi-load data at that moment are taken as a training sample; Through the determined training samples, the constructed data classification model is classified and trained to obtain the influence of each candidate meteorological index on photovoltaic output and the influence of each candidate meteorological index on the multi-load data of electricity, heating and cooling; The candidate meteorological index whose influence on the photovoltaic output is greater than a threshold is used as the target meteorological index; the candidate meteorological index whose influence on the electric, heating and cooling multi-load data is greater than a threshold is used as the target meteorological index.

6. The method according to claim 1 or 5, characterized in that: The candidate meteorological indicators include at least one of the following: irradiance, wind speed, dew point temperature, humidity, atmospheric pressure, horizontal visibility, and rainfall.

7. A source-load joint prediction device, characterized in that: include: The data acquisition module is used to obtain photovoltaic output, electric heating and cooling multi-load data and historical meteorological data within the target time period from the regional integrated energy system; A first indicator determination module is used to determine candidate meteorological indicators that affect the photovoltaic output and the electric, heating and cooling multi-load data; A second indicator determination module is used to screen the candidate meteorological indicators according to the influence of the candidate meteorological indicators on the photovoltaic output and the electric, heating and cooling multi-load data to obtain the target meteorological indicators; A data screening module, used for extracting target meteorological data corresponding to the target meteorological index from the historical meteorological data; The prediction module is used to determine the photovoltaic output prediction value, electric load prediction value, thermal load prediction value, and cooling load prediction value in a future time period based on the target meteorological data, the photovoltaic output, and the electric, thermal, and cooling multi-load data in the target time period through a source-load joint prediction model.

8. The device according to claim 7, characterized in that The prediction module comprises: A feature extraction unit, configured to concurrently perform feature extraction processing on the target meteorological data, the photovoltaic output, and the historical multi-element load data through a multi-channel feature extraction module of the source-load joint prediction model, so as to obtain channel-level meteorological factor features, photovoltaic output features, and multi-element load features; An enhancement unit, configured to enhance and fuse the extracted channel-level meteorological factor characteristics, the photovoltaic output characteristics, and the multi-load characteristics through the attention mechanism of the source-load joint prediction model; The prediction unit is used to jointly predict the fused features through the multi-task prediction module of the source-load joint prediction model to obtain the photovoltaic output prediction value, electric load prediction value, thermal load prediction value and cooling load prediction value in the future time period.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method according to any one of claims 1 to 6 when executed.

Images