A power load forecasting method and system

Abstract

The application discloses a kind of electric power load prediction method and system, method includes: obtaining multivariate load data, carries out comprehensive correlation analysis to multivariate load data, to obtain high correlation degree characteristic;The comprehensive correlation analysis includes the correlation analysis between meteorological factor and multivariate load, coupling analysis between multivariate load and multivariate load respective correlation analysis;By convolutional neural network, bidirectional long short-term memory network and attention mechanism are fused to build hybrid deep learning prediction model;The high correlation degree characteristic is input into hybrid deep learning prediction model and is predicted to obtain predicted multivariate load value.The application reduces the dimension of multivariate load prediction model input layer, improves input data quality.Can effectively prevent the overfitting phenomenon of prediction algorithm, improve model prediction accuracy;More close to and actual engineering system scheduling demand.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid load forecasting, and in particular to a power load forecasting method and system. BACKGROUND

[0002] Load forecasting is a crucial link in integrated energy systems, and is essential for power system planning, operation and management. Accurate load forecasting is of great significance to the safe and stable operation of the system, the optimization of resource scheduling, and market transactions, etc. It can help system operators to make plans and preparations in advance, ensure the balance between supply and demand of energy, and avoid energy shortages or surpluses.

[0003] However, traditional forecasting methods such as time series analysis and regression analysis have limitations. The former assumes that load data is linear and stationary, which is difficult to deal with nonlinear influencing factors. The latter relies on manual feature engineering, which is prone to errors. Specifically, linear assumption and many statistical models assume that data is linear and follows a certain probability distribution, which is not flexible enough when dealing with nonlinear factors such as temperature fluctuations and holiday effects, and it is difficult to capture complex patterns in load demand. As for feature selection and engineering, methods such as regression analysis require manual selection of input variables and feature engineering, which is not only time-consuming and laborious, but also prone to prediction bias due to improper feature selection. As for adaptability and generalization ability, with the continuous development of the power market and changes in user behavior, load patterns will also change. Traditional models often lack sufficient adaptability and generalization ability to cope with these changes. Therefore, there is an urgent need for a method to improve the accuracy and efficiency of multi-element load data forecasting. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a power load forecasting method and system to solve the problem of low prediction accuracy and low efficiency caused by the difficulty of load forecasting, the difficulty of load forecasting caused by the dependence on manual feature engineering, the introduction of errors, the large load fluctuation, the numerous nonlinear influencing factors, and the difficulty of data set construction.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a power load forecasting method, comprising:

[0008] Obtaining multi-element load data, performing comprehensive correlation analysis on the multi-element load data to obtain high correlation features; the comprehensive correlation analysis includes correlation analysis between meteorological factors and multi-element load, coupling analysis between multi-element load, and self-correlation analysis of multi-element load;

[0009] The hybrid deep learning prediction model is constructed by fusing a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism.

[0010] The high-correlation-degree features are input into the hybrid deep learning prediction model for prediction to obtain predicted multivariate load values.

[0011] As a preferred scheme of the power load prediction method, the correlation between the meteorological factors and the multivariate load is analyzed by using a maximum mutual information coefficient, and is expressed as:

[0012]

[0013] wherein, represents the MIC value of the features and . , respectively represent the marginal probability distribution functions of the features , . represent the joint probability distribution functions of the features , . , represent the number of grids divided in the directions of , in the two-dimensional space. is a constant.

[0014] As a preferred scheme of the power load prediction method, the coupling between the multivariate loads is analyzed by using a Spearman correlation coefficient, and is expressed as:

[0015]

[0016] wherein, represents the correlation degree between the load and the input features, and the closer to 1 indicates the higher correlation degree; , are the values of two factors of the i-th data point; , are the mean values of the two factors; is the number of data points.

[0017] As a preferred scheme of the power load prediction method, the self-correlation of the multivariate load is analyzed by using a self-correlation function, and is expressed as:

[0018]

[0019] wherein,​ , respectively represent the order of the lag and the ACF value of the lag ; , respectively represent the load at time and the load at time ; represent the average value of all loads.

[0020] As a preferred scheme of the power load prediction method, the high-correlation features are input into a hybrid deep learning prediction model for prediction, which comprises:

[0021] The convolutional neural network layer extracts spatial features from the input high-correlation feature data to obtain a spatial feature set;

[0022] The spatial feature set is input into a BiLSTM layer, which trains multiple LSTM units from the front and back directions respectively. The hidden layer simultaneously reads the information features of the data before a certain time of the spatial feature set from the front and back directions, obtains the relationship between the current time data and the past and future time data, and fits the current time data from two directions to obtain the power load data at a certain future time.

[0023] As a preferred scheme of the power load prediction method, the high-correlation features are input into a hybrid deep learning prediction model for prediction, which further comprises:

[0024] The attention mechanism layer assigns different weights to each feature, and the weights are calculated by a softmax function on the attention score;

[0025] The features with weights are input into a fully connected layer for prediction, and the predicted multivariate load value is output.

[0026] As a preferred scheme of the power load prediction method, the attention score is obtained by calculating the similarity between the query vector and each input vector, and there are attention scores, which are represented as: ;

[0027] The softmax function converts the original attention score into a weight in the form of probability distribution. For any one attention score , the weight calculation formula after softmax is:

[0028]

[0029] wherein, represents the attention score Taking an exponential, is the sum of all attention scores after taking an exponential;

[0030] The output vector after attention weighting is represented as , the output is: = , wherein, is the corresponding weight.

[0031] In the second aspect, the present application provides a power load prediction system, comprising:

[0032] An acquisition module is configured to acquire multi-element load data, and perform comprehensive correlation analysis on the multi-element load data to obtain high-correlation-degree features; the comprehensive correlation analysis includes correlation analysis between meteorological factors and the multi-element load, coupling analysis between the multi-element load, and self-correlation analysis of the multi-element load,

[0033] A construction module is configured to construct a hybrid deep learning prediction model by fusing a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism;

[0034] A prediction module is configured to input the high-correlation-degree features into the hybrid deep learning prediction model for prediction to obtain predicted multi-element load values.

[0035] In the third aspect, the present application provides a computing device, comprising:

[0036] A memory and a processor;

[0037] The memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, so as to realize the steps of the power load prediction method.

[0038] In the fourth aspect, the present application provides a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions are executed by the processor to realize the steps of the power load prediction method.

[0039] Compared with the prior art, the present application has the beneficial effects: the present application adopts MIC, PCC, ACF and other data correlation analysis technologies to fuse into a feature comprehensive correlation analysis method, analyzes the correlation between meteorological factors and load, between load and load coupling, and between before and after load, proposes high correlation features, reduces the dimension of the input layer of the multivariate load prediction model, and improves the input data quality. The organic combination of CNN, BiLSTM and Attention is used to predict data, which can fully play the advantages of CNN in effectively extracting spatial features, BiLSTM network in bidirectional time sequence feature extraction of sequence data, and effectively reduce the loss of historical information and highlight the information of key historical time points to reduce the influence of redundant information on load prediction results. Thus, the overfitting phenomenon of the prediction algorithm can be effectively prevented, and the model prediction accuracy can be improved. It is more suitable for the scheduling needs of the system in actual engineering, and the prediction method can provide good support for system day-ahead scheduling and intra-day operation optimization. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0041] Figure 1 The overall flow logic diagram of the power load prediction method described in the first embodiment of the present application;

[0042] Figure 2 The prediction effect diagram before the electric load model is fused in the power load prediction method described in the second embodiment of the present application;

[0043] Figure 3 The prediction effect diagram after the electric load model is fused in the power load prediction method described in the second embodiment of the present application;

[0044] Figure 4 The prediction error diagram before the electric load model is fused in the power load prediction method described in the second embodiment of the present application;

[0045] Figure 5 The prediction error diagram after the electric load model is fused in the power load prediction method described in the second embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0047] Embodiment 1

[0048] With reference to Figure 1 For an embodiment of the present application, a power load prediction method is provided, comprising:

[0049] S100: Obtain multi-element load data, and perform comprehensive correlation analysis on the multi-element load data to obtain high correlation characteristics; the comprehensive correlation analysis includes correlation analysis between meteorological factors and multi-element load, coupling analysis between multi-element loads, and self-correlation analysis of multi-element load;

[0050] In the embodiments of the present application, the correlation analysis between the meteorological factors and the multi-element load is performed by using the maximum mutual information coefficient, which is expressed as:

[0051]

[0052] wherein, denotes the MIC value of the characteristics and . , denote the marginal probability distribution functions of the characteristics , . denote the joint probability distribution functions of the characteristics , . , denote the number of grids divided in the , direction in the two-dimensional space. is a constant.

[0053] Specifically, Generally, it is about 0.6 power of the data amount.

[0054] It should be noted that the maximum mutual information coefficient can deeply mine the evaluation index of the correlation between variables in big data. In the feature selection process, the maximum normalized information under different interval divisions is used to deeply mine the correlation between the input characteristics and the output characteristics, so as to accurately measure the correlation between the feature sequence and the photovoltaic sequence.

[0055] In the embodiment of the present application, the coupling analysis between the multivariate loads is performed using the Spearman correlation coefficient, which is expressed as:

[0056]

[0057] in, represents the correlation between the load and the input feature, , the closer it is to 1, the higher the correlation; 、 Called The values ​​of the two factors for each data point; 、 It is called the mean of the two factors; is the number of data points.

[0058] In the embodiment of the present application, the multivariate load autocorrelation analysis is performed using the autocorrelation function, which is expressed as:

[0059]

[0060] in, 、 denote the order of lag and lag respectively ACF value of the order; 、 Respectively Moment and The load of the moment; Represents the average value of all loads.

[0061] It should be noted that multivariate load forecasting requires the simultaneous use of multiple loads as input variables for prediction. However, different loads exhibit varying characteristics at different time scales. This, coupled with the variability of meteorological factors, can lead to a sharp increase in the input requirements of the forecast model, significantly increasing the computational complexity. To address this issue, a single feature correlation analysis method is insufficient for accurate forecasting. Therefore, the aforementioned approach employs a comprehensive feature analysis to streamline the model's inputs and improve its forecast accuracy.

[0062] S200: Constructing a hybrid deep learning prediction model by integrating convolutional neural networks, bidirectional long short-term memory networks, and attention mechanisms;

[0063] It should be noted that the convolutional neural network (CNN) is a deep learning model with efficient feature extraction capabilities. It consists of two core layers: the convolution layer and the pooling layer. The convolution layer effectively extracts the nonlinear local features of multivariate loads through convolution operations and outputs feature maps; the pooling layer is used to compress the size of the feature map, that is, to reduce the dimensionality of the input data while retaining important features and improving the generalization ability of the features.

[0064] Among them, the long short-term memory network (LSTM) is a recurrent neural network structure, and the bidirectional long short-term memory network (BiLSTM) is a further expansion based on it, which is composed of forward and backward LSTMs.

[0065] Among them, the attention mechanism (Attention) can accurately capture the importance of the input features, that is, different weights are assigned to different features, and greater weights are assigned to key factors and smaller weights are assigned to weakly related features, so as to distinguish the importance of the input features and improve the processing efficiency of the model for the input features.

[0066] The organic combination of CNN, BiLSTM and Attention is used to predict data, which can fully exert the advantages of CNN in effectively extracting spatial features, BiLSTM network in extracting bidirectional time sequence features of sequence data, and Attention mechanism in effectively reducing the loss of historical information and highlighting the information of key historical time points to reduce the influence of redundant information on load prediction results. Thus, the overfitting phenomenon of the prediction algorithm can be effectively prevented, and the prediction accuracy of the model can be improved.

[0067] S300: inputting the high-correlation-degree features into the hybrid deep learning prediction model for prediction to obtain predicted multivariate load values.

[0068] In the embodiments of the present application, the high-correlation-degree features are inputted into the hybrid deep learning prediction model for prediction, which includes:

[0069] The convolutional neural network layer extracts spatial features from the input high-correlation-degree feature data to obtain a spatial feature set;

[0070] The spatial feature set is inputted into the BiLSTM layer, and the BiLSTM layer trains multiple LSTM units from the front and back directions respectively. The hidden layer reads the information features of the data before a certain moment of the spatial feature set from the front and back directions at the same time, obtains the relationship between the current moment data and the past and future moment data, and fits the current moment data from two directions to obtain the power load data at a certain future moment.

[0071] In the embodiments of the present application, the high-correlation-degree features are inputted into the hybrid deep learning prediction model for prediction, which further includes:

[0072] The attention mechanism layer assigns different weights to each feature, and the weights are calculated by a softmax function on the attention score;

[0073] Each feature with a weight is inputted into a fully connected layer for prediction, and the predicted multivariate load values are outputted.

[0074] In the embodiment of the present application, the attention score is obtained by calculating the similarity between the query vector and each input vector, and the attention score is represented as:

[0075] The Softmax function converts the original attention score into a weight in the form of a probability distribution. For any one attention score , the weight calculation formula after the softmax is:

[0076]

[0077] wherein, represents the attention score taking the exponential, is the sum of the exponentials of all attention scores;

[0078] The calculated weight is used to weight and sum the corresponding input elements, and the output vector after attention weighting is represented as , and the output is: = , wherein, is the corresponding weight.

[0079] The above is a schematic scheme of the power load prediction method of the embodiment. It should be noted that the technical scheme of the power load prediction system belongs to the same concept as the technical scheme of the power load prediction method described above. The technical scheme of the power load prediction system in the embodiment is not described in detail, and the description of the technical scheme of the power load prediction method can be referred to.

[0080] The power load prediction system in the embodiment comprises:

[0081] The acquisition module is configured to acquire multi-element load data, and perform comprehensive correlation analysis on the multi-element load data to obtain high-correlation-degree features. The comprehensive correlation analysis includes correlation analysis between meteorological factors and multi-element loads, coupling analysis between multi-element loads, and self-correlation analysis of multi-element loads,

[0082] The construction module is configured to construct a hybrid deep learning prediction model by fusing a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism.

[0083] The prediction module is configured to input the high-correlation-degree features into the hybrid deep learning prediction model for prediction to obtain predicted multi-element load values.

[0084] The embodiment also provides a computing device suitable for power load prediction, comprising:

[0085] ​​The memory is configured to store computer-executable instructions, and the processor is configured to execute the computer-executable instructions to implement the power load prediction method.

[0086] The embodiment also provides a storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the power load prediction method.

[0087] The storage medium provided by the embodiment belongs to the same inventive concept as the power load prediction method provided by the above embodiment, and the technical details not described in the embodiment can be referred to the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0088] Through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary universal hardware, and of course can also be realized by hardware. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, and the computer software product can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH, a hard disk or an optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods of various embodiments of the present application.

[0089] Embodiment 2

[0090] Referring to Tables 1-2 and Figures 2-5 Based on the previous embodiment, the present embodiment provides an application case of the power load prediction method to illustrate the implementability and beneficial effects of the present scheme.

[0091] In order to verify the performance of the prediction model proposed in the present application, actual data sets are used for experimental test analysis. The data is derived from the measured data of cold, heat, electricity and gas load provided by a certain comprehensive energy system in a certain place in southern China from January 1, 2020 to December 30, 2020, with a sampling frequency of 1h and 24 sampling points per day. Each set of original data contains 13 external environmental factors such as month, hour, day, holiday type, wind speed, ambient temperature, air pressure, radiation, rainfall, dry bulb temperature, dew point temperature, wet bulb temperature and humidity, as well as load power data, a total of 8760 groups of data (6130 groups of training, 2628 groups of testing).

[0092] The cross-correlations between the multivariate loads and meteorological factors in the meteorological factor correlation results vary. A larger MIC value indicates a greater correlation between the feature and the output. Based on empirical analysis, if the MIC threshold for the feature and the electric load is set to 0.5 and the MIC value for the ambient temperature and the electric load is 0.5172, then the ambient temperature is selected as the input for the electric load prediction dataset. If the heating load MIC is set to > 0.5, then the ambient temperature and air pressure are selected. If the cooling load MIC is set to > 0.6, then the ambient temperature, air pressure, global radiation, dry-bulb temperature, wet-bulb temperature, and dew point temperature are selected. Similarly, if the air load MIC is set to > 0.6, then the ambient temperature, air pressure, dry-bulb temperature, wet-bulb temperature, and dew point temperature are selected.

[0093] The correlations between the electric load and the heating load, cooling load, and gas load are 0.5566, 0.7407, and 0.7188, respectively. The correlations between the heating load and the cooling load and gas load are 0.7867 and 0.7671, respectively. The correlation between the cooling load and the gas load is 0.8967. Setting the multi-load correlation threshold to 0.7 allows you to select cooling and gas loads as inputs for the electric and heating load forecasting datasets, while selecting heating, gas, and electric loads as inputs for the cooling and gas load forecasting datasets.

[0094] The multivariate load autocorrelation results show a cyclical pattern in the correlation between predicted and historical values ​​as the forecast time moves forward (historical time), decreasing sequentially while the load values ​​at that moment are biased upward. The ACF thresholds for electricity, heating, cooling, and gas loads were set at 0.85, 0.9, 0.95, and 0.9, respectively. The selected historical load values ​​served as the input for the forecast dataset.

[0095] A prediction data set is constructed based on the features screened out by the comprehensive feature analysis in the early stage and the output load, and a prediction data set without feature screening is also set up.

[0096] A day-ahead forecast comparison experiment was conducted under the same forecast model, focusing only on the electricity load. The experimental comparison results are as follows: Figures 2~5 shown.

[0097] The specific prediction performance indicators of other loads are shown in Tables 1 and 2.

[0098] Table 1 Day-ahead forecast accuracy without feature selection

[0099]

[0100] Table 2 Day-ahead prediction accuracy after feature screening

[0101]

[0102] For electrical load, Figure 2 、3 From Table 1-2, it can be seen that before the feature comprehensive analysis method is combined with the hybrid deep learning model, the predicted value of the model is not well fitted with the true value, especially the error at the wave peak is large, and after the feature comprehensive analysis method is combined with the hybrid deep learning model, the prediction effect is better and the error is smaller.

[0103] From Table 1-2, it can be seen that before the feature comprehensive analysis method is combined with the hybrid deep learning model, the predicted value of the model is not well fitted with the true value, especially the error at the wave peak is large, and after the feature comprehensive analysis method is combined with the hybrid deep learning model, the prediction effect is better and the error is smaller.

[0104] From the above experimental results, it can be seen that the feature comprehensive correlation analysis and the hybrid deep learning prediction method described in the present application have the best prediction performance in the aspect of multivariate load prediction in IES.

[0105] It should be noted that the above examples are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A method for predicting power load, characterized in that: include: Obtain multivariate load data and conduct comprehensive correlation analysis on the multivariate load data to obtain high correlation features; The comprehensive correlation analysis includes correlation analysis between meteorological factors and multivariate loads, coupling analysis between multivariate loads, and autocorrelation analysis of multivariate loads; The correlation analysis between meteorological factors and multivariate loads was performed using the maximum mutual information coefficient, which is expressed as: ； in, Representation characteristics and MIC value of 、 Represents characteristics 、 The marginal probability distribution function of Representation characteristics 、 The joint probability distribution function of 、 Represented in two-dimensional space 、 The number of grids divided in the direction; is a constant; The coupling analysis between multiple loads was performed using the Spearman correlation coefficient, which is expressed as: ； in, represents the correlation between the load and the input feature, , the closer it is to 1, the higher the correlation; 、 Called The values ​​of the two factors for each data point; 、 It is called the mean of the two factors; is the number of data points; The multivariate load autocorrelation analysis is performed through the autocorrelation function, which is expressed as: ； in, 、 denote the order of lag and lag respectively ACF value of the order; 、 Respectively Moment and The load of the moment; Indicates the average value of all loads; A hybrid deep learning prediction model is constructed by integrating convolutional neural networks, bidirectional long short-term memory networks, and attention mechanisms; The highly correlated features are input into a hybrid deep learning prediction model for prediction to obtain predicted multivariate load values.

2. The power load forecasting method according to claim 1, wherein: Inputting the highly correlated features into a hybrid deep learning prediction model for prediction includes: The convolutional neural network layer extracts spatial features from the input high-correlation feature data to obtain a spatial feature set; The spatial feature set is input into the BiLSTM layer. The BiLSTM layer trains multiple LSTM units from the forward and backward directions respectively. The forward and backward information features of the data before a certain moment in the spatial feature set are read simultaneously through the hidden layer to obtain the relationship between the current moment data and the past and future moment data. The current moment data is fitted from two directions to obtain the power load data at a certain moment in the future.

3. The power load forecasting method according to claim 2, wherein: Inputting the highly correlated features into a hybrid deep learning prediction model for prediction, further comprising: The attention mechanism layer assigns different weights to each feature, and the weights are used to calculate the attention score through the softmax function; The weighted features are input into the fully connected layer for prediction, and the predicted multivariate load value is output.

4. The power load forecasting method according to claim 3, wherein: The attention score is obtained by calculating the similarity between the query vector and each input vector, The attention score is expressed as: ; The Softmax function converts the original attention score into a weight in the form of probability distribution. For any attention score , and the weight calculation formula after softmax is: ； in, Indicates the attention score Take the index, is the sum of all attention scores after taking the index; The calculated weights are used to perform weighted summation on the corresponding input elements. The output vector after attention weighting is expressed as , the output is: = ,in, is the corresponding weight.

5. A system for power load forecasting, applying the method according to claim 1, characterized in that: include: An acquisition module is used to acquire multivariate load data and perform comprehensive correlation analysis on the multivariate load data to obtain high correlation features; The comprehensive correlation analysis includes correlation analysis between meteorological factors and multivariate loads, coupling analysis between multivariate loads, and autocorrelation analysis of multivariate loads; The correlation analysis between meteorological factors and multivariate loads was performed using the maximum mutual information coefficient, which is expressed as: ； in, Representation characteristics and MIC value of 、 Represents characteristics 、 The marginal probability distribution function of Representation characteristics 、 The joint probability distribution function of 、 Represented in two-dimensional space 、 The number of grids divided in the direction; is a constant; The coupling analysis between multiple loads was performed using the Spearman correlation coefficient, which is expressed as: ； in, represents the correlation between the load and the input feature, , the closer it is to 1, the higher the correlation; 、 Called The values ​​of the two factors for each data point; 、 It is called the mean of the two factors; is the number of data points; The multivariate load autocorrelation analysis is performed through the autocorrelation function, which is expressed as: ； in, 、 denote the order of lag and lag respectively ACF value of the order; 、 Respectively Moment and The load of the moment; Indicates the average value of all loads; Building blocks for constructing hybrid deep learning prediction models by integrating convolutional neural networks, bidirectional long short-term memory networks, and attention mechanisms; The prediction module is used to input the high-correlation features into the hybrid deep learning prediction model for prediction to obtain a predicted multivariate load value.

6. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the power load forecasting method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the power load forecasting method according to any one of claims 1 to 4.

Images