Intelligent carbon emission prediction system based on neural network

Abstract

The invention discloses a carbon emission intelligent prediction system based on a neural network, and relates to the technical field of carbon emission prediction. The system obtains historical energy consumption and carbon emission data of a target enterprise, performs feature extraction and dimension reduction processing, and constructs an energy consumption data set and a carbon emission data set. A complex nonlinear relation between energy consumption data and carbon emission data is captured through a Siamese network, potential features are extracted through a double-branch structure, and an association degree matrix and nonlinear features are output. And performing time sequence modeling on the correlation degree matrix and the nonlinear characteristics through a long short-term memory network, capturing a time sequence dependency relationship between carbon emission and energy consumption, and predicting a future carbon emission value of the target enterprise. Through a gradient descent algorithm, according to an error between a predicted value and an actual value, a parameter weight of the model is optimized. Accurate carbon emission prediction is provided for a target enterprise, and the enterprise is assisted to make a more scientific decision in the aspects of carbon emission management, energy conservation and consumption reduction.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission prediction, and in particular to a carbon emission intelligent prediction system based on a neural network. Background Art

[0002] The intelligent carbon emission prediction system not only has important environmental significance, but also helps enterprises optimize energy management, improve resource utilization efficiency and reduce operating costs.

[0003] For example, a Chinese invention patent with announcement number CN117035167B discloses a method for predicting corporate carbon emissions, which collects historical energy consumption data associated with the corporate carbon emission data, and performs the following steps: processing the historical energy consumption data, calculating the degree of correlation DTW between the historical energy consumption data and the corporate carbon emission data, judging the strength of the correlation between the historical energy consumption data and the corporate carbon emission data, substituting the strongly correlated historical energy consumption data into a long short-term memory neural network for iterative training to obtain a corporate carbon emission prediction model, and performing corporate carbon emission prediction; because the present invention selects historical energy consumption data that is strongly correlated with corporate carbon emissions as training data for the long short-term memory neural network through different steps, it greatly reduces the types of correlations, enhances the mining efficiency of the long short-term memory neural network, and makes the prediction accuracy of the obtained corporate carbon emission prediction model higher.

[0004] However, existing methods usually select training data by calculating the correlation between historical energy consumption data and carbon emission data. Although they can screen out strongly correlated data, they may ignore the nonlinear relationships and complex features in the data, resulting in insufficient expression of the model. Secondly, although the long short-term memory neural network (LSTM) can process time series data, the existing methods do not fully consider the complex time series dependencies between carbon emissions and energy consumption, and fail to effectively capture long-term and short-term dependencies in time. Therefore, the existing methods still need to be improved when dealing with complex time series data and nonlinear relationships, and cannot accurately predict the future carbon emissions of enterprises. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a carbon emission intelligent prediction system based on a neural network. The system adopts a technical solution combining Siamese networks and long short-term memory networks, which can efficiently model complex nonlinear relationships and time series dependencies in historical data, thereby greatly improving the accuracy and stability of carbon emission prediction. In addition, by adjusting the model parameters through the gradient descent optimization algorithm, the accuracy of the prediction results and the generalization ability of the system are further improved. The above-mentioned background technology problems are solved.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a carbon emission intelligent prediction system based on a neural network, comprising the following modules: a data processing module, a nonlinear relationship modeling module, a prediction module, and an optimization module; the data processing module is used to obtain the historical energy consumption data and carbon emission data of the target enterprise and perform preprocessing, including extracting key features and dimensionality reduction processing, and constructing an energy consumption data set and a carbon emission data set; the nonlinear relationship modeling module is used to capture the complex nonlinear relationship between historical energy consumption data and carbon emissions through a Siamese network based on the energy consumption data set and the carbon emission data set, establish a mapping model between energy consumption data and carbon emission data, and output a correlation matrix and nonlinear features; the prediction module is used to perform time series modeling on the correlation matrix and nonlinear features through a long short-term memory network, capture the time series dependency between carbon emissions and energy consumption, and predict the future carbon emission value of the target enterprise; the optimization module is used to adjust the parameter weights of the mapping model through a gradient descent algorithm according to the error between the future carbon emission value of the target enterprise and the actual carbon emission value.

[0007] Furthermore, the specific process of obtaining the historical energy consumption data and historical carbon emission data of the target enterprise and preprocessing them is as follows: obtain the historical energy consumption data and historical carbon emission data of the target enterprise, clean the acquired data, and eliminate missing values, outliers, and duplicate data; extract the key features of the historical energy consumption data and historical carbon emission data of the target enterprise, including energy consumption, production activity type, equipment operating status, seasonal fluctuations, and perform feature dimensionality reduction processing.

[0008] Furthermore, the specific process of constructing the energy consumption dataset and the carbon emission dataset is as follows: according to the pre-processed energy consumption data and carbon emission data of the target enterprise, a representative time window is selected for data division; the historical energy consumption data and carbon emission data are aligned according to the time series, and each time period corresponds to a set of features and target values ​​to construct the energy consumption dataset and the carbon emission dataset.

[0009] Furthermore, the specific process of capturing the complex nonlinear relationship between historical energy consumption data and carbon emissions through the Siamese network is as follows: the energy consumption dataset and the carbon emission dataset are taken as input, and the data are passed into two identical sub-networks for processing through the dual-branch structure of the Siamese network; in each sub-network, the input data is feature extracted and embedded through the neural network layer to extract the potential patterns in the historical energy consumption data and carbon emission data; the similarity of the outputs of the two sub-networks is calculated to capture the nonlinear relationship between the historical energy consumption data and carbon emissions.

[0010] Furthermore, the specific process of establishing a mapping model between energy consumption data and carbon emission data is as follows: after capturing the nonlinear relationship between historical energy consumption data and carbon emission data through the Siamese network, the key features of the historical data are combined; the deep correlation between energy consumption and carbon emissions is learned through the fully connected layer; the network weights are adjusted through the regression algorithm to construct a mapping model between energy consumption data and carbon emission data.

[0011] Furthermore, the specific process of outputting the correlation matrix and nonlinear features is as follows: the similarity of the outputs of the two sub-networks is calculated through the dual-branch structure of the Siamese network. Each sub-network processes energy consumption data and carbon emission data respectively. After extracting the features, the strength of the relationship between each pair of data points is evaluated by calculating the cosine similarity to form a correlation matrix. Each element in the matrix represents the degree of correlation between two data points, capturing the nonlinear relationship between energy consumption and carbon emissions; the input energy consumption data and carbon emission data are nonlinearly transformed through the neural network layer. In each layer of the neural network, the nonlinear features are extracted after being processed by the activation function; the obtained correlation matrix and nonlinear features are used as output.

[0012] Furthermore, the association matrix and nonlinear features are modeled for time series through the long short-term memory network to capture the time series dependency between carbon emissions and energy consumption. The specific process is as follows: the association matrix and nonlinear features are organized into time series data in chronological order as the input of the long short-term memory network; the long short-term memory network is used for time series modeling, and the long-term and short-term dependencies in the historical data are captured through the forget gate, input gate and output gate mechanism of the LSTM unit, and the state information of the historical moment is selectively retained or forgotten; the mutual influence between energy consumption and carbon emissions in different time periods is gradually learned through multi-layer LSTM units, and the time series features in the data are extracted; the feedback mechanism of LSTM is used to capture the time correlation of historical data based on the output of the previous moment and the current input, and a time series dependency model between carbon emissions and energy consumption is established; the result of capturing the time series dependency between carbon emissions and energy consumption is obtained through the output time series characteristics of the long short-term memory network.

[0013] Furthermore, the specific process of predicting the future carbon emission values ​​of the target enterprise is as follows: the time series features captured by the long short-term memory network are taken as input and input into the fully connected layer for further processing, including fusing the time series features output by multiple layers of LSTM in the fully connected layer to extract the potential correlation information between energy consumption and carbon emissions; the time series features are mapped through the regression model to predict the future carbon emission values ​​of the target enterprise.

[0014] Furthermore, the specific process of adjusting the parameter weights of the mapping model through the gradient descent algorithm is as follows: define a loss function to measure the difference between the predicted value and the actual value, and calculate the gradient of the loss function relative to each parameter of the mapping model; determine the learning rate, decide the step size of each parameter update, and update the parameter weights of the model according to the gradient and learning rate; repeat the gradient update steps until the loss function converges.

[0015] The present invention has the following beneficial effects:

[0016] (1) The carbon emission intelligent prediction system based on neural network obtains the historical energy consumption data and carbon emission data of the target enterprise through the data processing module, and performs preprocessing, extracts key features and reduces dimensions to construct energy consumption data sets and carbon emission data sets. This process effectively converts complex energy consumption and carbon emission data into structured data that can be used for further modeling. In addition, the nonlinear relationship modeling module captures the complex nonlinear relationship between historical energy consumption data and carbon emissions through the Siamese network, establishes a mapping model between energy consumption data and carbon emission data, and outputs the correlation matrix and nonlinear features, further enhancing the system's ability to model the potential relationship of the data.

[0017] (2) The carbon emission intelligent prediction system based on neural network uses the long short-term memory network (LSTM) to perform time series modeling on the correlation matrix and nonlinear features through the prediction module, capturing the time series dependency between carbon emissions and energy consumption, thereby improving the accuracy of carbon emission prediction. The optimization module uses the gradient descent algorithm to dynamically adjust the parameter weights of the mapping model according to the error between the future carbon emission value of the target enterprise and the actual carbon emission value, so that the prediction model can be continuously optimized to improve the accuracy and stability of the prediction. The collaborative work of this series of modules enables the system to accurately predict the future carbon emission values ​​of enterprises in different time periods, which helps enterprises to scientifically formulate carbon emission reduction strategies and energy management plans.

[0018] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of the carbon emission intelligent prediction system based on neural network of the present invention. DETAILED DESCRIPTION

[0020] The embodiment of the present application solves the problem of low accuracy of traditional carbon emission prediction methods and the inability to effectively capture the complex nonlinear relationship and time dependence between energy consumption and carbon emissions through a carbon emission intelligent prediction system based on a neural network.

[0021] The overall idea of ​​the solution in the embodiments of this application is as follows:

[0022] Obtain the historical energy consumption data and carbon emission data of the target enterprise and perform preprocessing, including extracting key features and dimensionality reduction, and constructing energy consumption data sets and carbon emission data sets.

[0023] According to the energy consumption data set and the carbon emission data set, the complex nonlinear relationship between historical energy consumption data and carbon emissions is captured through the Siamese network, a mapping model between energy consumption data and carbon emission data is established, and the correlation matrix and nonlinear features are output.

[0024] The correlation matrix and nonlinear features are modeled in time series through long short-term memory networks to capture the time series dependency between carbon emissions and energy consumption and predict the future carbon emissions of target enterprises.

[0025] According to the error between the future carbon emission value of the target enterprise and the actual carbon emission value, the parameter weights of the mapping model are adjusted through the gradient descent algorithm.

[0026] See also Figure 1 The embodiment of the present invention provides a technical solution: a carbon emission intelligent prediction system based on a neural network, comprising the following modules: a data processing module, a nonlinear relationship modeling module, a prediction module, and an optimization module; the data processing module is used to obtain the historical energy consumption data and carbon emission data of the target enterprise and perform preprocessing, including extracting key features and dimensionality reduction processing, and constructing an energy consumption data set and a carbon emission data set; the nonlinear relationship modeling module is used to capture the complex nonlinear relationship between the historical energy consumption data and carbon emissions through a Siamese network according to the energy consumption data set and the carbon emission data set, establish a mapping model between the energy consumption data and the carbon emission data, and output a correlation matrix and nonlinear features; the prediction module is used to perform time series modeling on the correlation matrix and the nonlinear features through a long short-term memory network, capture the time series dependency between carbon emissions and energy consumption, and predict the future carbon emission value of the target enterprise; the optimization module is used to adjust the parameter weights of the mapping model through a gradient descent algorithm according to the error between the future carbon emission value of the target enterprise and the actual carbon emission value.

[0027] In this implementation scheme, data processing module: The main task of this module is to obtain the historical energy consumption data and carbon emission data of the target enterprise and preprocess them. Preprocessing includes two aspects: Extract key features: Extract features related to carbon emission prediction from the original data to ensure that the model can focus on the data part that has a higher impact on carbon emission prediction. Dimensionality reduction processing: In order to avoid overfitting and low computational efficiency caused by too high a dimension of the data, a dimensionality reduction method is used to reduce the number of features of the data while retaining the key information in the data. Finally, the data processed by the data processing module will be constructed into an "energy consumption data set" and a "carbon emission data set", which will serve as the basis for subsequent modeling. Nonlinear relationship modeling module: The function of this module is to use Siamese networks (twin neural networks) to model the complex nonlinear relationship between energy consumption and carbon emissions. The Siamese network uses a shared weight structure to enable the model to efficiently learn and capture the similarities and differences between input data sets. The output of this module includes an association matrix and nonlinear features, where the association matrix reflects the strength of the relationship between energy consumption data and carbon emission data, while the nonlinear features describe the complex interaction pattern between them. Prediction module: The prediction module uses the long short-term memory network (LSTM) to perform time series modeling on the correlation matrix and nonlinear features output by the nonlinear relationship modeling module. LSTM is a neural network that can capture long-term and short-term dependencies in time series data, and is particularly suitable for processing time series data. Through the training of the LSTM network, this module can effectively capture the time series dependency between carbon emissions and energy consumption, thereby predicting future carbon emission values. The prediction results will provide a quantitative basis for the future carbon emission levels of the target enterprise. Optimization module: The purpose of this module is to further improve the accuracy of carbon emission prediction by optimizing the prediction results. The specific method is to calculate the error between the predicted carbon emission value and the actual carbon emission value of the target enterprise, and adjust the parameter weights of the model through the gradient descent algorithm. The gradient descent algorithm will optimize the parameters of the mapping model based on the error back propagation, so that the model can more accurately predict future carbon emission values.

[0028] Specifically, the specific process of obtaining the historical energy consumption data and historical carbon emission data of the target enterprise and performing preprocessing is as follows: obtain the historical energy consumption data and historical carbon emission data of the target enterprise, clean the acquired data, and eliminate missing values, outliers, and duplicate data; extract the key features of the historical energy consumption data and historical carbon emission data of the target enterprise, including energy consumption, production activity type, equipment operating status, seasonal fluctuations, and perform feature dimensionality reduction processing.

[0029] In this implementation plan, historical data of the target enterprise is obtained: first, historical energy consumption data and carbon emission data are extracted from the system or database of the target enterprise. Energy consumption data includes the consumption of various types of energy such as electricity, natural gas, and steam; carbon emission data is the carbon emission calculated based on energy consumption data. These data are usually time series, recording the energy use and carbon emissions of enterprises in different periods. Data cleaning: Clean the acquired data to ensure the quality and accuracy of the data. Specific operations include: Eliminate missing values: Remove records containing null values ​​or missing data to avoid these missing data from having adverse effects on subsequent analysis and modeling processes. Eliminate outliers: Identify and eliminate obviously unreasonable outliers through statistical methods or business rules. For example, some values ​​are far higher or far lower than the reasonable range of energy consumption or carbon emissions. Eliminate duplicate data: Check and delete duplicate records in the data set to ensure the uniqueness of each data and avoid data redundancy affecting the analysis results. Feature extraction: Extract key features from the cleaned historical data. These features include: Energy consumption: The energy usage of the target enterprise in different time periods is an important basis for predicting carbon emissions. Type of production activities: Different production activities have different impacts on energy consumption and carbon emissions, so it is necessary to extract relevant activity type information, such as the production of high-energy consumption products or low-energy consumption products. Equipment operating status: The startup, shutdown status, and failure of the equipment will also affect energy consumption and carbon emissions, so it is necessary to record and extract the operating status of the equipment. Seasonal fluctuations: Energy consumption and carbon emissions are usually affected by seasonal changes. Especially in areas with large temperature changes, the use of equipment such as heating and air conditioning will cause seasonal fluctuations in energy consumption and carbon emissions. Therefore, it is very important to extract seasonal change characteristics. Feature dimensionality reduction processing: The extracted multiple features are reduced in dimensionality through technical means to reduce the data dimension and improve the training efficiency of subsequent models. Common dimensionality reduction methods include principal component analysis (PCA), etc., which retains the most representative features in the data and removes redundant information, making the data more concise and able to more effectively support model training.

[0030] Specifically, the specific process of constructing the energy consumption dataset and the carbon emission dataset is as follows: according to the pre-processed energy consumption data and carbon emission data of the target enterprise, a representative time window is selected for data division; the historical energy consumption data and carbon emission data are aligned according to the time series, and each time period corresponds to a set of features and target values ​​to construct the energy consumption dataset and the carbon emission dataset.

[0031] In this implementation scheme, time windows are selected for data division: based on the pre-processed historical energy consumption data and carbon emission data, representative time windows are selected, such as divided by month, quarter or year. These time windows help capture the cyclical changes in energy consumption and carbon emissions. Data alignment: Align energy consumption data and carbon emission data in time series. That is, the energy consumption and corresponding carbon emissions in each time period must match to ensure that the characteristics of each time period (energy consumption data) correspond to the target value (carbon emission data). Construct a data set: Based on the aligned data, the energy consumption data of each time period is used as the characteristic value, and the carbon emission data is used as the target value to construct an energy consumption data set and a carbon emission data set respectively. In this way, each set of data contains the relevant characteristics and target output of a time period, which is convenient for subsequent modeling and analysis.

[0032] Specifically, the specific process of capturing the complex nonlinear relationship between historical energy consumption data and carbon emissions through the Siamese network is as follows: the energy consumption dataset and the carbon emission dataset are taken as input, and the data are passed into two identical sub-networks for processing through the dual-branch structure of the Siamese network; in each sub-network, the input data is feature extracted and embedded through the neural network layer to extract the potential patterns in the historical energy consumption data and carbon emission data; the similarity of the outputs of the two sub-networks is calculated to capture the nonlinear relationship between the historical energy consumption data and carbon emissions.

[0033] In this implementation, input data: energy consumption dataset X energy and carbon emission dataset X carbon , respectively represent the energy consumption and carbon emission data of the target enterprise in different time periods. Each data set is in the form of X energy ,X carbon ∈R N ×D , where N is the number of data points and D is the feature dimension of each data point. Dual-branch structure: The Siamese network adopts a dual-branch structure, that is, the two data sets are input into two identical sub-networks respectively, and the processing method is consistent. Each sub-network has the same architecture and weights to ensure that the learned features are comparable between the two. Feature extraction and embedding: The sub-network extracts and embeds features of the input data through the neural network layer. The purpose is to convert the original energy consumption data and carbon emission data into low-dimensional, representative feature vectors, which is usually accomplished through structures such as convolutional layers, fully connected layers, and pooling layers. Calculate similarity metrics: Similarity metrics: Capture the nonlinear relationship between data by calculating the similarity or distance between the feature vectors output by the two sub-networks. Euclidean distance: Where m is the dimension of the embedding space, and Respectively represent the i-th eigenvalue. By calculating similarity or distance, the Siamese network can reveal the nonlinear relationship between historical energy consumption data and carbon emission data. During the network training process, by minimizing the loss function (such as the contrast loss function), the network weights are optimized so that similar data points are brought closer and dissimilar data points are pushed away, thereby capturing the complex nonlinear associations between data.

[0034] Specifically, the specific process of establishing a mapping model between energy consumption data and carbon emission data is as follows: after capturing the nonlinear relationship between historical energy consumption data and carbon emission data through the Siamese network, the key features of the historical data are combined; the deep correlation between energy consumption and carbon emissions is learned through the fully connected layer; the network weights are adjusted through the regression algorithm to construct a mapping model between energy consumption data and carbon emission data.

[0035] In this implementation, Siamese network model nonlinear relationship extraction: The complex nonlinear relationship between historical energy consumption data and carbon emission data is captured through the Siamese network. The feature representation of the network output (i.e., and H carbon ) contains the potential patterns and nonlinear dependencies of the two data sets. Combined with the key features extracted from energy consumption data and carbon emission data, such as energy consumption, type of production activity, equipment operating status, and seasonal fluctuations, these features help enrich the representation of the data and provide more contextual information that is helpful for prediction. The fully connected layer learns deep associations. These key features can be further improved by connecting and to the fully connected layer. Mathematical description of the fully connected layer: The fully connected layer performs deep association learning on the features obtained from the Siamese network. The fully connected layers play an important role at this stage. They further explore the implicit relationship between energy consumption and carbon emissions through linear combinations and nonlinear activation functions. Application of activation function: The input passes through a series of fully connected layers, and the output of each layer is transformed by a nonlinear activation function (such as ReLU, Sigmoid, etc.): F c =σ(W c H energy +b c ), where F c represents the output of layer c, W c and b care the weight and bias of the layer respectively, and σ is the activation function. The regression algorithm adjusts the network weights, and the network weights are adjusted by the regression algorithm to construct a mapping model between energy consumption data and carbon emission data. This regression process can use the mean square error (MSE) loss function to optimize the model. The goal is to update the network weights by minimizing the error between the predicted carbon emission value and the actual carbon emission value. Mathematical description of the regression algorithm. The goal of the regression algorithm is to learn a suitable mapping function that can predict the corresponding carbon emission data based on the given energy consumption data. The form of the mapping model is: in, is the predicted carbon emission value, θ is the weight parameter of the network, H energy and are the characteristics of the input data.

[0036] Specifically, the specific process of outputting the correlation matrix and nonlinear features is as follows: the similarity of the outputs of the two sub-networks is calculated through the dual-branch structure of the Siamese network. Each sub-network processes energy consumption data and carbon emission data respectively. After extracting the features, the strength of the relationship between each pair of data points is evaluated by calculating the cosine similarity to form a correlation matrix. Each element in the matrix represents the degree of correlation between two data points, capturing the nonlinear relationship between energy consumption and carbon emissions; the input energy consumption data and carbon emission data are nonlinearly transformed through the neural network layer. In each layer of the neural network, the nonlinear features are extracted after being processed by the activation function; the obtained correlation matrix and nonlinear features are used as output.

[0037] In this embodiment, the Siamese network consists of two identical subnetworks, each of which processes energy consumption data and carbon emission data respectively. The two subnetworks share the same weights and structure in order to learn the similarities and differences between the two data sets. The specific process is: Input processing: Input energy consumption data and carbon emission data into the two subnetworks respectively. Feature extraction: Each subnetwork extracts the features of the input data through a multi-layer neural network. The output of each subnetwork contains the potential pattern and feature representation in the input data. In each subnetwork of the Siamese network, the input data (i.e., energy consumption data and carbon emission data) is subjected to feature extraction through a multi-layer neural network. The output of each layer is nonlinearly transformed by an activation function, and the role of the activation function is to increase the nonlinear expression ability of the model, thereby capturing the complex patterns in the data. In each layer of the neural network, the input is weighted and summed, and the activation function converts it into more complex nonlinear features, thereby better capturing the nonlinear relationship between source consumption data and carbon emission data. Output association matrix and nonlinear features: Association matrix: The association matrix obtained by calculating the cosine similarity, each element represents the similarity or association between a feature pair, and the matrix shows the relationship between energy consumption and carbon emission data. Non-linear features: The features extracted by each sub-network after being processed by the neural network layer represent the complex patterns in the data. These features reflect the deep non-linear relationships in the data.

[0038] Specifically, the association matrix and nonlinear features are modeled for time series through the long short-term memory network to capture the time series dependency between carbon emissions and energy consumption. The specific process is as follows: the association matrix and nonlinear features are organized into time series data in chronological order as the input of the long short-term memory network; the long short-term memory network is used for time series modeling, and the long-term and short-term dependencies in the historical data are captured through the forget gate, input gate and output gate mechanism of the LSTM unit, and the state information of the historical moment is selectively retained or forgotten; the mutual influence between energy consumption and carbon emissions in different time periods is gradually learned through multi-layer LSTM units, and the time series features in the data are extracted; the feedback mechanism of LSTM is used to capture the time correlation of historical data based on the output of the previous moment and the current input, and a time series dependency model between carbon emissions and energy consumption is established; the result of capturing the time series dependency between carbon emissions and energy consumption is obtained through the output time series characteristics of the long short-term memory network.

[0039] In this implementation scheme, the correlation matrix and nonlinear features are organized into time series data in chronological order. Each time point corresponds to a data input, which contains feature values ​​related to the relationship between energy consumption and carbon emissions. These data will be input into the LSTM network for further processing. The purpose of this process is to enable the network to learn the law of data changes over time and capture the time series dynamics between carbon emissions and energy consumption. The LSTM network captures long-term and short-term dependencies in time series data through its internal forget gate, input gate and output gate mechanism: Forget gate: determines how much information of the current state should be forgotten. The forget gate generates a value between 0 and 1 based on the current input and the output of the previous time step to determine how much historical information is forgotten. Input gate: determines how much new information should be stored in the memory cell of the LSTM unit at the current moment. The input gate generates a value to control the input amount of new information by processing the current input and the output of the previous moment. Output gate: determines what information the LSTM unit outputs at the current moment. The output gate generates an output based on the memory cell and input signal of the current state as the input of the next moment and the final output. Capturing long-term and short-term dependencies: Through the above mechanism, LSTM is able to selectively retain or forget information based on long-term dependencies and short-term changes in historical data, ensuring that the model can capture important dependencies in long-term historical data while avoiding forgetting key information in the short term. This is crucial for capturing the complex time series relationship between carbon emissions and energy consumption, especially when the system is subject to periodic or sudden changes, LSTM can effectively maintain memory and make appropriate adjustments. Multi-layer LSTM learning time series features: Multi-layer LSTM units: By stacking multiple LSTM layers, the model is able to learn different levels of time series features in the data layer by layer. At each layer, LSTM extracts deeper time series information based on the output of its previous layer and the current input data. This multi-layer structure enables the network to learn dependencies from different time scales, thereby better capturing the long-term and short-term effects between carbon emissions and energy consumption. Building a time series dependency model: The LSTM network uses the output of the previous moment and the input of the current moment to capture the time correlation of historical data through its feedback mechanism. This mechanism enables LSTM to handle complex time dependencies in time series data. Through these mechanisms, LSTM gradually learns and builds a temporal dependency model between carbon emissions and energy consumption, that is, how to predict future changes based on past energy consumption and carbon emissions data. The output of LSTM is the temporal features obtained after multi-layer neural network processing. These features contain the temporal dependency relationship between carbon emissions and energy consumption. Through the output of LSTM, we can get a prediction of future carbon emissions and the relationship between carbon emissions and energy consumption.

[0040] Specifically, the specific process of predicting the future carbon emission values ​​of the target enterprise is as follows: the time series features captured by the long short-term memory network are used as input and input into the fully connected layer for further processing, including fusing the time series features output by multiple layers of LSTM in the fully connected layer to extract the potential correlation information between energy consumption and carbon emissions; the time series features are mapped through the regression model to predict the future carbon emission values ​​of the target enterprise.

[0041] In this implementation, the time series features extracted by LSTM can be represented as a vector: in, is the time series feature output by the ath LSTM unit at the time, and n is the number of LSTM layers. This vector aggregates the time series information of the target enterprise's energy consumption and carbon emissions over the past period of time. Next, the time series features will be input into the fully connected layer for further processing. The output of the fully connected layer can be expressed as: Z t =σ(W·H t +b); where: W is the weight matrix of the fully connected layer, m is the output dimension of the fully connected layer. b is the bias term. σ is the activation function, which is used to introduce nonlinear transformation. The output of the fully connected layer of the regression model will be passed to the regression model for final prediction. Assuming that the regression model uses linear mapping, the predicted carbon emission value can be expressed as: Where: Wr is the weight matrix of the regression model, 1 means that a single carbon emission value is predicted. br is the bias term of the regression model. It is the predicted carbon emission value of the target enterprise at a certain moment in time.

[0042] Specifically, the specific process of adjusting the parameter weights of the mapping model through the gradient descent algorithm is as follows: define a loss function to measure the difference between the predicted value and the actual value, and calculate the gradient of the loss function relative to each parameter of the mapping model; determine the learning rate, determine the step size of each parameter update, and update the parameter weights of the model according to the gradient and learning rate; repeat the gradient update steps until the loss function converges.

[0043] In this implementation, a loss function is defined to measure the difference between the predicted value and the actual value. In regression problems, mean square error (MSE) is usually used as the loss function. Given a predicted value and an actual value y t , the loss function L is in the form of: Where: T is the total number of training samples. is the model's prediction at time y. tis the actual carbon emission value of the target enterprise at the time. Calculate the gradient of the loss function with respect to each parameter of the mapping model. The gradient is the partial derivative of the loss function with respect to each parameter, which represents the sensitivity of the loss function to the change of the parameter. Specifically, assuming that the parameters of the mapping model are and b r , we need to calculate the gradient of the loss function with respect to these parameters: Gradient with respect to weights: Gradient with respect to the bias: Determine the learning rate. The learning rate η is a key hyperparameter in the gradient descent method, which determines the step size of each parameter update. A learning rate that is too small may cause the training process to be too slow, while a learning rate that is too large may cause the parameters to be updated too quickly and miss the optimal solution. The ideal learning rate needs to be selected through tuning in actual applications. Update the parameter weights of the model. According to the gradient and learning rate, the model parameters W r and b r Will be updated. The update rule using gradient descent is as follows: Where: η is the learning rate. and They are the loss functions with respect to W r and b r The gradient of the loss function is repeated until the loss function converges. The update step will continue, and each update will make the loss function L approach the minimum value. The convergence condition of the gradient descent algorithm can be judged based on the change of the loss function. If the change of the loss function is less than the preset threshold or reaches the maximum number of iterations, the model is considered to have converged. Specifically, the convergence condition can be expressed as: |L( k )-L( k-1 )|<∈where: L( k ) is the loss function value of the kth iteration. L( k-1 ) is the loss function value of the previous iteration. ∈ is a preset very small value, indicating the tolerance range of loss function convergence.

[0044] In summary, this application has at least the following effects:

[0045] The carbon emission intelligent prediction system based on neural network can efficiently capture the deep temporal dependency between carbon emission and energy consumption by using Siamese network to capture the complex nonlinear relationship between historical energy consumption data and carbon emission, and combine the long short-term memory network (LSTM) to model the time series characteristics, thereby improving the accuracy of the prediction. The system can perform personalized modeling and prediction based on the historical data characteristics of the target enterprise, adapt to the energy consumption patterns and production activities of different enterprises, and realize customized carbon emission prediction and management solutions. The application of Siamese network can deeply explore the complex nonlinear relationship between energy consumption data and carbon emission data, overcome the limitations of traditional linear modeling methods, and provide more accurate feature mapping for carbon emission prediction. Dynamic optimization of mapping model parameters through gradient descent algorithm can continuously improve the prediction performance of the model and ensure high stability and low error of carbon emission prediction. Long short-term memory network (LSTM) can effectively capture the long-term and short-term dependencies in historical data through its unique gating mechanism, so as to accurately model the temporal changes of carbon emissions and predict future trends.

[0046] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0047] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0048] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0049] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0050] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0051] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. The carbon emission intelligent prediction system based on neural network is characterized by: It includes the following modules: data processing module, nonlinear relationship modeling module, prediction module, and optimization module; The data processing module is used to obtain the historical energy consumption data and carbon emission data of the target enterprise and perform preprocessing, including extracting key features and dimensionality reduction processing, and constructing energy consumption data sets and carbon emission data sets; The nonlinear relationship modeling module is used to capture the complex nonlinear relationship between historical energy consumption data and carbon emissions through the Siamese network based on the energy consumption data set and the carbon emission data set, establish a mapping model between the energy consumption data and the carbon emission data, and output the correlation matrix and nonlinear features; The prediction module is used to perform time series modeling on the correlation matrix and nonlinear features through a long short-term memory network, capture the time series dependency between carbon emissions and energy consumption, and predict the future carbon emissions value of the target enterprise; The optimization module is used to adjust the parameter weights of the mapping model through a gradient descent algorithm according to the error between the future carbon emission value of the target enterprise and the actual carbon emission value.

2. The carbon emission intelligent prediction system based on neural network according to claim 1 is characterized by: The specific process of obtaining the target enterprise's historical energy consumption data and historical carbon emission data and preprocessing them is as follows: Obtain the target company's historical energy consumption data and carbon emission data, clean the acquired data, and remove missing values, outliers, and duplicate data; Extract key features from the target enterprise’s historical energy consumption data and historical carbon emission data, including energy consumption, production activity type, equipment operating status, seasonal fluctuations, and perform feature dimensionality reduction processing.

3. The carbon emission intelligent prediction system based on neural network according to claim 2 is characterized by: The specific process of constructing the energy consumption dataset and carbon emission dataset is as follows: According to the pre-processed energy consumption data and carbon emission data of the target enterprise, a representative time window is selected for data segmentation; The historical energy consumption data and carbon emission data are aligned in time series, with each time period corresponding to a set of features and target values, to construct energy consumption datasets and carbon emission datasets.

4. The carbon emission intelligent prediction system based on neural network according to claim 3 is characterized by: The specific process of capturing the complex nonlinear relationship between historical energy consumption data and carbon emissions through the Siamese network is as follows: The energy consumption dataset and carbon emission dataset are taken as input, and the data are passed into two identical sub-networks for processing through the dual-branch structure of the Siamese network; In each sub-network, the input data is feature extracted and embedded through the neural network layer to extract the potential patterns in the historical energy consumption data and carbon emission data; The similarity of the outputs of the two sub-networks is calculated to capture the nonlinear relationship between historical energy consumption data and carbon emissions.

5. The carbon emission intelligent prediction system based on neural network according to claim 4 is characterized in that: The specific process of establishing a mapping model between energy consumption data and carbon emission data is as follows: After capturing the nonlinear relationship between historical energy consumption data and carbon emission data through Siamese network, the key features of historical data are combined; Learn the deep relationship between energy consumption and carbon emissions through fully connected layers; The network weights are adjusted through the regression algorithm to construct a mapping model between energy consumption data and carbon emission data.

6. The carbon emission intelligent prediction system based on neural network according to claim 5 is characterized in that: The specific process of outputting the correlation matrix and nonlinear features is as follows: The similarity of the outputs of the two sub-networks is calculated through the dual-branch structure of the Siamese network. Each sub-network processes energy consumption data and carbon emission data respectively. After extracting features, the strength of the relationship between each pair of data points is evaluated by calculating the cosine similarity to form a correlation matrix. Each element in the matrix represents the degree of correlation between two data points, capturing the nonlinear relationship between energy consumption and carbon emissions. The input energy consumption data and carbon emission data are transformed nonlinearly through the neural network layer. In each layer of the neural network, the activation function is processed to extract nonlinear features. The obtained correlation matrix and nonlinear features are taken as output.

7. The carbon emission intelligent prediction system based on neural network according to claim 6 is characterized by: The specific process of using the long short-term memory network to perform temporal modeling of the correlation matrix and nonlinear features to capture the temporal dependency between carbon emissions and energy consumption is as follows: Organize the correlation matrix and nonlinear features into time series data in chronological order as the input of the long short-term memory network; Use long short-term memory networks for time series modeling. Through the forget gate, input gate, and output gate mechanism of LSTM units, it captures long-term and short-term dependencies in historical data and selectively retains or forgets the state information of historical moments. Through multi-layer LSTM units, we gradually learn the mutual influence between energy consumption and carbon emissions in different time periods and extract the time series features in the data; Using the LSTM feedback mechanism, based on the output of the previous moment and the current input, the temporal correlation of historical data is captured to establish a temporal dependency model between carbon emissions and energy consumption; Through the output timing characteristics of the long short-term memory network, we obtain the result of capturing the temporal dependency between carbon emissions and energy consumption.

8. The carbon emission intelligent prediction system based on neural network according to claim 7 is characterized by: The specific process of predicting the future carbon emissions of the target enterprise is as follows: The time series features captured by the long short-term memory network are used as input to the fully connected layer for further processing, including fusing the time series features output by multiple layers of LSTM in the fully connected layer to extract the potential correlation information between energy consumption and carbon emissions; The time series characteristics are mapped through the regression model to predict the future carbon emission values ​​of the target enterprise.

9. The carbon emission intelligent prediction system based on neural network according to claim 8 is characterized by: The specific process of adjusting the parameter weights of the mapping model through the gradient descent algorithm is as follows: Define the loss function to measure the difference between the predicted value and the actual value, and calculate the gradient of the loss function with respect to each parameter of the mapping model; Determine the learning rate, determine the step size of each parameter update, and update the model's parameter weights based on the gradient and learning rate; The gradient update step is repeated until the loss function converges.

Images