Distribution network load prediction model based on artificial intelligence

Abstract

The invention belongs to the technical field of power systems, and particularly relates to a distribution network load prediction model based on artificial intelligence, which comprises a DM-LSTM neural network model, and a distribution network load prediction method of the DM-LSTM neural network model comprises the following steps: S1: firstly, collecting original data; comprise historical generating capacity, weather data, season information, geographic position parameters and equipment performance indexes; s2, establishing a Deep MLP deep network encoder, inputting the collected original data, and then carrying out initialization and normalization processing on data information; s3, a Sat iotempora l LSTM (Long Short Term Memory) model is built; s4, fusing the output of the Deep MLP deep network encoder and the output of the Sat iotempora l LSTM model to obtain a mixed weight so as to reduce a prediction difference; and S5, constructing a Softmax layer at the tail end of the Sat iotempora l LSTM (Long Short Term Memory) model. According to the method, static and dynamic factors are comprehensively considered, more background information, time sequence modeling capability and dynamic mode recognition capability are provided, prediction errors are effectively reduced, and prediction accuracy and reliability are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and in particular relates to a distribution network load prediction model based on artificial intelligence. Background Art

[0002] Load forecasting is the basic work in urban power grid planning and plays a key role in the quality of planning. Accurate load forecasting will provide scientific decision-making basis for the reasonable layout of power sources, timely power grid construction, optimal investment time, and maximum economic and social benefits, and provide guarantee for the safe, economical and reliable operation of the power grid.

[0003] With the surge in power grid business, load forecasting is becoming more and more important in business scenarios such as regulation, power grid planning, and power trading. The commonly used load forecasting models now include BP neural network, LSTM neural network, and CNN-LSTM neural network. However, in actual application, there are certain disadvantages: for BP neural network, it is easy to fall into local optimal solution, the training process may be more difficult, and it has high requirements for data quality and feature engineering, and it is not good at handling outliers and missing values. At the same time, it is difficult to handle long-term dependencies in time series data; for LSTM neural network, although it can make up for the defects of BP neural network to a certain extent, its training and parameter adjustment complexity is relatively high, it requires more computing resources, and it is not suitable for simple load In addition, when the amount of data is small, overfitting may occur. For the CNN-LSTM neural network, although it combines the advantages of CNN and LSTM, it can not only effectively extract spatial features in sequence data, but also capture time dependencies. However, the model is relatively complex and requires high computing resources and time costs, and it is difficult to adjust parameters and interpret the model. At the same time, it has high requirements for data quality and preprocessing, and requires appropriate feature engineering. Based on the above, we provide an effective use of policies, industries, weather, festivals and other factors on load forecasting, and better predict the node load of the distribution network based on artificial intelligence. Summary of the invention

[0004] In response to the above-mentioned technical problems, the present invention proposes a distribution network load forecasting model based on artificial intelligence, which has a reasonable design, simple structure, easy processing and can effectively use the important influence of policies, industries, weather, festivals and other factors on load forecasting to better predict the node load of the distribution network.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is a distribution network load prediction model based on artificial intelligence, including a DM-LSTM neural network model. The method for distribution network load prediction of the DM-LSTM neural network model includes the following steps:

[0006] S1: First, collect raw data, including historical power generation, weather data, seasonal information, geographic location parameters, and equipment performance indicators;

[0007] S2: Build a DeepMLP deep network encoder, input the collected raw data, initialize and normalize the data information, extract and fuse features from different data sources, and then perform the final output features of each node;

[0008] S3: Build a Spatiotemporal LSTM model to model spatiotemporal sequence data to capture long-term dependencies and spatiotemporal dynamics in time series data;

[0009] S4: Fuse the output of the DeepMLP deep network encoder and the output of the SpatiotemporalLSTM model to obtain a mixed weight and reduce the prediction difference;

[0010] S5: A Softmax layer is built at the end of the SpatiotemporalLSTM model to realize the weight allocation of different influencing factors, so as to measure the influence of different factors on the prediction of renewable energy output of the distribution network, and the prediction result of the SpatiotemporalLSTM model is multiplied by the output of the Softmax layer to obtain the prediction result under various factors.

[0011] Preferably, a multilayer perceptron is constructed in the DeepMLP deep network encoder of step S2, and the multilayer perceptron includes an input layer, a hidden layer and an output layer, and includes the following steps:

[0012] S2.1: The collected raw data is used as input features and fed into the multi-layer perceptron. The data in the input layer is processed by the hidden layer.

[0013] S2.2: In the hidden layer, each neuron is connected to all neurons in the previous layer, and nonlinearity is introduced through the activation function, so that the network can capture complex data patterns;

[0014] S2.3: The output of the hidden layer is transmitted to one or more output neurons of the output layer, and a linear activation function is set up in the output layer to achieve a real-valued output;

[0015] S2.4: Apply regularization technology to the DeepMLP deep network encoder and optimize the model training process by adjusting hyperparameters such as learning rate and batch size.

[0016] Preferably, the activation function from the input layer to the hidden layer is expressed as: h1=σ(W1·x+b1), and the activation function from the hidden layer to the output layer is expressed as: y=sigma(W2·h+b2), wherein x is the input feature vector, W1 is the weight matrix from the input layer to the hidden layer, b1 is the bias vector of the hidden layer, sigma is usually a nonlinear activation function, h is the output of the hidden layer, W2 is the weight matrix from the hidden layer to the output layer, b2 is the bias vector of the output layer, and y is the final output, which is the predicted value of the model.

[0017] Preferably, a gating mechanism is built in the Spatiotemporal LSTM model of step S3, wherein the gating mechanism includes an input gate, a forget gate, a cell state update and an output gate, and includes the following steps:

[0018] S3.1: Use the input gate to view the current input data and the hidden state of the previous time step, and determine the information to be stored. The formula for determining the information to be stored by the input gate is:

[0019] i t =σ(W xi ·[x t ;h t-1 ]+b i )

[0020] Among them, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, W xi is the weight matrix corresponding to the input gate, b i is the bias term, σ represents the sigmoid activation function;

[0021] S3.2: Use the forget gate to determine the unimportant information that is allowed to be discarded in the model. The formula of the forget gate that allows the model to discard information is: t =σ(W xf ·[x t ;h t-1 ]+b f ), where x t is the input of the current time step, h t-1 is the hidden state of the previous time step, W xf is the weight matrix of the forget gate, bf is the bias term;

[0022] S3.3: According to the signal input in S3.1 and S3.2, the cell state is updated. The formula for updating the cell state is:

[0023] c t =f t ·c t-1 +i t tanh(W xc ·[x t ;h t-1 ]+b c )

[0024] Among them, c t is the cell state at the current time step, c t-1 is the cell state at the previous time step, f t is the output of the forget gate, i t is the output of the input gate, W xc is the associated weight matrix, b c is the bias term;

[0025] S3.4: Use the output gate to determine how much information is output in the cell state update to the final result. The formula for determining how much information is output by the output gate is:

[0026] o t =σ(W xo ·[x t ;h t-1 ]+b o )

[0027] Among them, W xo is the weight matrix of the output gate, b o is the bias term.

[0028] Preferably, step S5 comprises the following steps:

[0029] S5.1: Assuming that there are n influencing factors, a probability distribution is generated for them using the constructed Softmax layer. The probability distribution determination formula of the Softmax layer is:

[0030] P=(p1,p2,...,p n )

[0031] Among them, p i Represents the weight of the ith factor.

[0032] S5.2: Assume that the LSTM layer makes a prediction based on each factor, denoted as:

[0033] L=(l1,l2,...,l n );

[0034] S5.3: Use the output in step S5.1 as a weight to perform weighted summation on the predicted values ​​in S5.2. The formula for determining the weighted prediction result is expressed as: Among them, p i ·l i is the weighted estimate of the (i)th factor.

[0035] Compared with the prior art, the advantages and positive effects of the present invention are:

[0036] 1. The invention provides a distribution network load forecasting model based on artificial intelligence. Through the established DM-LSTM neural network model, its advantages in load forecasting are that it comprehensively considers static and dynamic factors, provides more background information, time series modeling capabilities and dynamic pattern recognition capabilities. The advantages of DeepMLP and LSTM are comprehensively utilized to reduce the prediction error and improve the accuracy and reliability of the prediction; the invention can effectively use the important influence of policies, industries, weather, festivals and other factors on load forecasting, and better predict the node load of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 This is an architecture diagram of the DM-LSTM neural network model provided by the present invention;

[0039] Figure 2 This is a network architecture diagram of the Spatiotemporal LSTM model provided by the present invention. DETAILED DESCRIPTION

[0040] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0041] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways than those described herein. Therefore, the present invention is not limited to the specific embodiments of the following disclosure.

[0042] Examples, such as Figure 1As shown, a distribution network load forecasting model based on artificial intelligence is designed to better predict the node load of the distribution network and effectively use the important influence of policies, industries, weather, festivals and other factors on load forecasting. The DM-LSTM neural network model is specially designed. The DeepMLP part is responsible for processing factors such as policies, industries, and weather, and the LSTM part is responsible for processing historical time series information. Then, the results of the two parts are weighted to obtain a more accurate load forecast value, including the DM-LSTM neural network model. The method for distribution network load forecasting using the DM-LSTM neural network model includes the following steps:

[0043] S1: First, collect raw data, including historical power generation, weather data, seasonal information, geographic location parameters, and equipment performance indicators;

[0044] The specific description is: This step is used in conjunction with the data processing and feature engineering modules to collect raw data, including historical power generation, weather data (such as temperature, wind speed, sunshine, etc.), seasonal information, geographic location parameters, and equipment performance indicators. The data needs to be cleaned, standardized, and normalized for model processing. Different data sources are extracted and fused for features. For example, weather and seasonal factors can be used to identify patterns through historical data analysis, geographic factors may include information such as the site's latitude and longitude, altitude, and equipment performance may include equipment model, efficiency, maintenance records, etc.;

[0045] S2: Build a DeepMLP deep network encoder, input the collected raw data, initialize and normalize the data information, extract and fuse features from different data sources, and then perform the final output features of each node;

[0046] The specific description is as follows: This model is one of the core parts, and its advantage is that it can extract and learn the feature representation of these static information, so as to better understand and explain the impact of environmental factors on load;

[0047] S3: Build a Spatiotemporal LSTM model to model spatiotemporal sequence data to capture long-term dependencies and spatiotemporal dynamics in time series data;

[0048] The specific description is: This model is also one of the core parts. The LSTM neural network is good at processing time series data and can capture time dependencies, long-term memory and short-term changes. The advantage of this part is that it can effectively use historical time series data to learn the change rules and trends of load over time;

[0049] S4: Fuse the output of the DeepMLP deep network encoder and the output of the SpatiotemporalLSTM model to obtain a mixed weight and reduce the prediction difference;

[0050] Specifically described as follows: The output of the DeepMLP module will be fused with the output of the LSTM module. This fusion can be a simple weighted average or a more complex fusion technique, such as feature concatenation or learned mixed weights; the training process of the entire DM-LSTM model includes adjusting the parameters of the two parts of DeepMLP and LSTM, as well as the weight of the fusion strategy, which is usually achieved through the back propagation algorithm and the gradient descent method, with the goal of minimizing the difference between the predicted load and the actual load; this step is a weighted processing, that is, combining the prediction results of the DeepMLP part and the LSTM part, and comprehensively considering static information and dynamic time series information. Because more information and factors are comprehensively considered, the prediction error can be reduced, and it can better adapt to load changes in different situations, and improve the accuracy and reliability of the prediction;

[0051] S5: A Softmax layer is built at the end of the SpatiotemporalLSTM model to allocate weights to different influencing factors, so as to measure the influence of different factors on the prediction of the output of renewable energy in the distribution network. The output of the Softmax layer is multiplied by the prediction result of the SpatiotemporalLSTM model to obtain the prediction result under various factors.

[0052] The specific description is: The role of the SoftMax layer after the LSTM network is to further process the sequence features generated by the LSTM layer to output the final weights and prediction results. Adding a Softmax layer at the end of the LSTM model can assign weights to different influencing factors (such as weather, policies, etc.). These weights reflect the contribution of each factor to the final load prediction; by multiplying the weight output by the Softmax layer with the prediction result of the LSTM, the weighted prediction value can be obtained to more accurately estimate the distribution network load;

[0053] In the above process: the established DM-LSTM neural network model has the advantage of comprehensive consideration of static and dynamic factors, providing more background information, time series modeling capabilities and dynamic pattern recognition capabilities in load forecasting. The advantages of DeepMLP and LSTM are comprehensively utilized to reduce the prediction error and improve the accuracy and reliability of the prediction; the present invention can effectively use the important influence of policies, industries, weather, festivals and other factors on load forecasting, and better predict the node load of the distribution network.

[0054] In order to extract and learn the feature representation of these static information, so as to better understand and explain the impact of environmental factors on load, a multi-layer perceptron is built in the DeepMLP deep network encoder of step S2. The multi-layer perceptron includes an input layer, a hidden layer and an output layer, and includes the following steps:

[0055] S2.1: The collected raw data is used as input features and fed into the multi-layer perceptron. The data in the input layer is processed by the hidden layer.

[0056] Among them, the original data includes influencing factors such as policies, industries, and weather, which are used as input features. These input data may have been preprocessed, such as normalization or standardization, to facilitate better processing by the neural network.

[0057] S2.2: The data in the input layer is processed through a series of fully connected layers (hidden layers), in which each neuron is connected to all neurons in the previous layer, and nonlinearity is introduced through activation functions, so that the network can capture complex data patterns.

[0058] S2.3: The output of the hidden layer is transmitted to one or more output neurons of the output layer, and a linear activation function is set up in the output layer to achieve a real-valued output to represent the influence intensity of the corresponding factor.

[0059] S2.4: Apply regularization techniques to the DeepMLP deep network encoder and optimize the model training process by adjusting hyperparameters such as learning rate and batch size.

[0060] The specific description is as follows: Step S2.4 is to prevent overfitting by using regularization techniques such as L2 regularization or dropout in DeepMLP. At the same time, the training process of the model is optimized by adjusting hyperparameters such as learning rate and batch size.

[0061] Further, MLP can be described as a combination of a series of stacked linear transformations and nonlinear activation functions, each layer consists of a weight matrix, a bias vector and an activation function. For a three-layer MLP, it includes an input layer, a hidden layer and an output layer, which is further described as:, the activation function from the input layer to the hidden layer is expressed as: h1 = σ(W1·x+b1), and the activation function from the hidden layer to the output layer is expressed as: y = sigma(W2·h+b2), where x is the input feature vector, W1 is the weight matrix from the input layer to the hidden layer, b1 is the bias vector of the hidden layer, sigma is usually a nonlinear activation function, h is the output of the hidden layer, W2 is the weight matrix from the hidden layer to the output layer, b2 is the bias vector of the output layer, and y is the final output, which is the predicted value of the model.

[0062] In order to fully guarantee the specific extraction of spatiotemporal features by the Spatiotemporal LSTM model, a gating mechanism is built in the Spatiotemporal LSTM model in step S3. The gating mechanism includes an input gate, a forget gate, a cell state update, and an output gate, and includes the following steps:

[0063] S3.1: Use the input gate to view the current input data and the hidden state of the previous time step to determine the information to be stored. The formula for determining the information to be stored by the input gate is:

[0064] i t =σ(W xi ·[x t ;h t-1 ]+b i )

[0065] Among them, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, W xi is the weight matrix corresponding to the input gate, b i is the bias term, σ represents the sigmoid activation function;

[0066] S3.2: Use the forget gate to determine the unimportant information that is allowed to be discarded in the model. The formula for the forget gate to allow the model to discard information is: t =σ(W xf ·[x t ;h t-1 ]+b f ), where x t is the input of the current time step, h t-1 is the hidden state of the previous time step, W xf is the weight matrix of the forget gate, bf is the bias term;

[0067] S3.3: Update the cell state according to the signal input in S3.1 and S3.2. The formula for updating the cell state is:

[0068] c t =f t ·c t-1 +i t tanh(W xc ·[x t ;h t-1 ]+b c )

[0069] Among them, c t is the cell state at the current time step, c t-1 is the cell state at the previous time step, f t is the output of the forget gate, it is the output of the input gate, W xc is the associated weight matrix, b c is the bias term;

[0070] S3.4: Use the output gate to determine how much information is output in the cell state update to the final result. The formula for determining how much information is output by the output gate is:

[0071] o t =σ(W xo ·[x t ;h t-1 ]+b o )

[0072] Among them, W xo is the weight matrix of the output gate, b o is the bias term;

[0073] The specific description is: Figure 2 As shown in the figure, SpatiotemporalLSTM adopts a hierarchical structure, in which each layer extracts specific spatiotemporal features. This structure allows the network to capture different levels of abstract information at different levels, and this information can be further utilized and refined in subsequent layers. Key information is retained by updating the temporal and spatial states. The temporal state is responsible for maintaining the continuity of the time series, while the spatial state is responsible for recording changes in spatial information. This separate state update mechanism enables the network to process complex spatiotemporal dependencies more effectively. The parallel structural design of SpatiotemporalLSTM enables the model to process information independently between different layers. Although this design may lead to insufficient information transfer between layers, it also allows the model to focus on specific types of feature extraction at different levels, thereby improving the model's ability to process complex spatiotemporal data.

[0074] The node connections and weights output by the Deep MLP deep network encoder can be used as features to input into the Spatiotemporal LSTM model. This information reflects the interdependence and spatial correlation between different energy sites. Using the structured data provided by the graph neural network, time series data can be constructed, which contains information such as historical power generation, weather conditions, geographical factors, etc. arranged in chronological order. In LSTM, the input gate determines what information will be stored in the cell state. This gating mechanism is determined by looking at the current input data and the hidden state of the previous time step. The forget gate is responsible for determining which information should be discarded from the cell state. This allows the model to discard information that is no longer important, allowing the model to better handle long-term dependencies. Based on the signals of the input gate and the forget gate, the cell state is updated. This state is a core feature of LSTM, which can maintain important information in long sequences. The output gate determines how much information is output from the cell state to the final result. In this way, the model can generate a prediction result for the next time step. Through the above-mentioned gating mechanism, Spatial Temporal LSTM can not only capture the dependency of time series data, but also combine the spatial characteristics provided by the graph neural network to more comprehensively understand and predict the changes in renewable energy power generation. The model is trained using historical data, and the prediction error is minimized by continuously adjusting the weights. During the training process, the network parameters can be optimized through the back-propagation algorithm to obtain more accurate prediction results. After sufficient training, the model will be able to predict future renewable energy power generation, taking into account multiple factors such as historical data, weather changes, seasonal effects, geographical environment, and equipment performance.

[0075] In order to output the final weights and prediction results, the role of the SoftMax layer after the LSTM network is to further process the sequence features generated by the LSTM layer to output the final weights and prediction results; further speaking: adding a Softmax layer at the end of the LSTM model can assign weights to different influencing factors (such as weather, policies, etc.), and these weights reflect the contribution of each factor to the final load prediction; by multiplying the weight output by the Softmax layer with the prediction result of the LSTM, we can get the weighted prediction value and estimate the distribution network load more accurately. Step S5 includes the following steps:

[0076] S5.1: Assuming that there are n influencing factors, a probability distribution is generated for them using the constructed Softmax layer. The probability distribution determination formula of the Softmax layer is:

[0077] P=(p1,p2,...,p n )

[0078] Among them, pi Represents the weight of the ith factor.

[0079] S5.2: Assume that the LSTM layer makes a prediction based on each factor, denoted as:

[0080] L=(l1,l2,...,l n );

[0081] S5.3: Use the output in step S5.1 as a weight to perform weighted summation on the predicted values ​​in S5.2. The formula for determining the weighted prediction result is expressed as: Among them, p i ·l i is the weighted estimate of the (i)th factor;

[0082] Specifically described as follows: Steps S4.2 and S4.3 are weighted prediction results. The role of the Softmax layer is to learn a probability distribution that can measure the degree of influence of different factors on load forecasting. For example, if the probability of weather factors is high, it means that weather has a greater impact on load forecasting. By multiplying the output of the Softmax layer with the prediction result of LSTM, we get a prediction result that comprehensively considers the importance of various factors. This processing method allows the model to more flexibly adjust the proportion of each factor in the final prediction.

[0083] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A distribution network load prediction model based on artificial intelligence, characterized in that: The method includes a DM-LSTM neural network model, wherein the method for predicting the load of a distribution network using the DM-LSTM neural network model includes the following steps: S1: First, collect raw data, including historical power generation, weather data, seasonal information, geographic location parameters, and equipment performance indicators; S2: Build a Deep MLP network encoder, input the collected raw data, initialize and normalize the data information, extract and fuse features from different data sources, and then perform the final output features of each node; S3: Build a Spatiotemporal LSTM model to model spatiotemporal sequence data to capture long-term dependencies and spatiotemporal dynamics in time series data; S4: Fuse the output of the Deep MLP deep network encoder and the output of the Spatiotemporal LSTM model to obtain a mixed weight and reduce the prediction difference; S5: A Softmax layer is built at the end of the Spatiotemporal LSTM model to allocate weights to different influencing factors, so as to measure the influence of different factors on the prediction of renewable energy output of the distribution network. The output of the Softmax layer is multiplied by the prediction result of the Spatiotemporal LSTM model to obtain the prediction result under various comprehensive factors.

2. The distribution network load prediction model based on artificial intelligence according to claim 1 is characterized in that: The Deep MLP deep network encoder of step S2 is constructed with a multi-layer perceptron, which includes an input layer, a hidden layer and an output layer, and includes the following steps: S2.1: The collected raw data is used as input features and fed into the multi-layer perceptron. The data in the input layer is processed by the hidden layer. S2.2: In the hidden layer, each neuron is connected to all neurons in the previous layer, and nonlinearity is introduced through the activation function, so that the network can capture complex data patterns; S2.3: The output of the hidden layer is transmitted to one or more output neurons of the output layer, and a linear activation function is set up in the output layer to achieve a real-valued output; S2.4: Apply regularization technology to the Deep MLP deep network encoder and optimize the model training process by adjusting hyperparameters such as learning rate and batch size.

3. The distribution network load prediction model based on artificial intelligence according to claim 2 is characterized in that: The activation function from the input layer to the hidden layer is expressed as: h1=σ(W1·x+b1), and the activation function from the hidden layer to the output layer is expressed as: y=sigma(W2·h+b2), where x is the input feature vector, W1 is the weight matrix from the input layer to the hidden layer, b1 is the bias vector of the hidden layer, sigma is usually a nonlinear activation function, h is the output of the hidden layer, W2 is the weight matrix from the hidden layer to the output layer, b2 is the bias vector of the output layer, and y is the final output, which is the predicted value of the model.

4. The distribution network load prediction model based on artificial intelligence according to claim 3 is characterized in that: A gating mechanism is built in the Spatiotemporal LSTM model of step S3, wherein the gating mechanism includes an input gate, a forget gate, a cell state update and an output gate, and includes the following steps: S3.1: Use the input gate to view the current input data and the hidden state of the previous time step, and determine the information to be stored. The formula for determining the information to be stored by the input gate is: i t =σ(W xi ·[x t ;h t-1 ]+b i ) Among them, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, W xi is the weight matrix corresponding to the input gate, b i is the bias term, σ represents the sigmoid activation function; S3.2: Use the forget gate to determine the unimportant information that is allowed to be discarded in the model. The formula of the forget gate that allows the model to discard information is: t =σ(W xf ·[x t ;h t-1 ]+b f ), where x t is the input of the current time step, h t-1 is the hidden state of the previous time step, W xf is the weight matrix of the forget gate, bf is the bias term; S3.3: According to the signal input in S3.1 and S3.2, the cell state is updated. The formula for updating the cell state is: c t =f t ·c t-1 +i t ·tanh(W xc ·[x t ;h t-1 ]+b c ) Among them, c t is the cell state at the current time step, c t-1 is the cell state at the previous time step, f t is the output of the forget gate, i t is the output of the input gate, W xc is the associated weight matrix, b c is the bias term; S3.4: Use the output gate to determine how much information is output in the cell state update to the final result. The formula for determining how much information is output by the output gate is: the t =σ(W xo ·[x t ;h t-1 ]+b o ) Among them, W xo is the weight matrix of the output gate, b o is the bias term.

5. The distribution network load prediction model based on artificial intelligence according to claim 4 is characterized in that: The step S5 comprises the following steps: S5.1: Assuming that there are n influencing factors, a probability distribution is generated for them using the constructed Softmax layer. The probability distribution determination formula of the Softmax layer is: P=(p1,p2,...,p n ) Among them, p i Represents the weight of the ith factor. S5.2: Assume that the LSTM layer makes a prediction based on each factor, denoted as: L＝(l1,l2,...,l n )； S5.3: Use the output in step S5.1 as a weight to perform weighted summation on the predicted values ​​in S5.

2. The formula for determining the weighted prediction result is expressed as: Among them, p i ·l i is the weighted estimate of the (i)th factor.

Images