Residential user carbon emission prediction method and system based on energy carbon accounting

Abstract

The invention belongs to the field of carbon emission prediction and calculation, and provides a residential user carbon emission prediction method and system based on energy carbon accounting, and the technical scheme is that the method comprises the steps: obtaining the historical energy consumption data of a residential user, the meteorological data of a corresponding time period, and the house area data of a corresponding energy consumption user; performing carbon emission accounting according to the historical energy consumption data of the residential users to obtain carbon emission sequence data; training the constructed carbon emission prediction model based on the carbon emission sequence data, the meteorological data of the corresponding time period and the house area data of the corresponding energy consumption user to obtain a trained carbon emission prediction model; and on the basis of the trained carbon emission prediction model, predicting the energy consumption carbon emission of the residential users in the future time period to obtain the energy consumption carbon emission of the residential users in the future time period. Individual prediction of energy consumption and carbon emission is realized, and making of a carbon emission reduction strategy can be supported to be more targeted and accurate.

Description

Technical Field

[0001] The present invention belongs to the field of carbon emission prediction and calculation, and in particular relates to a method and system for predicting carbon emissions of residential users based on energy-carbon accounting. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] With economic development and the improvement of people's living standards, residents' energy consumption is increasing, and energy use trends are becoming increasingly diversified, causing a rapid increase in carbon emissions. Residents' carbon emission control is a force that cannot be ignored in carbon reduction efforts.

[0004] In this context, many studies on carbon emission prediction based on residents' energy consumption have emerged, in order to provide a scientific data basis for residents to formulate carbon emission reduction strategies from the perspective of energy consumption. Among these studies, some are macroscopic studies on the prediction of carbon emissions of residents in a region, and the overall prediction data of residents' carbon emissions in a certain region in the future period are obtained. The problem with this type of research is that the basic energy data comes from regional statistical data, the accuracy and real-time performance of the data are poor, and there is a lack of ability to predict carbon emissions for individual residents, and it is impossible to provide feedback and early warning on the trend of residents' carbon emissions. In addition, some studies only focus on the prediction of carbon emissions related to electricity, with a single energy dimension, and the prediction results are even weaker in reflecting the carbon emissions of residents' energy consumption. Most studies build carbon emission prediction models based on neural networks and / or time series models, but they often ignore the energy consumption habits of residents, so that the model does not learn the periodic characteristics of individual residents' energy consumption, affecting the accuracy of residents' carbon emission prediction. Summary of the invention

[0005] In order to solve at least one technical problem existing in the above-mentioned background technology, the present invention provides a residential user carbon emission prediction method and system based on energy-carbon accounting, which constructs a model for predicting energy consumption-related carbon emissions for individual residents, and provides data basis for the formulation of residents' carbon emission control strategies.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a method for predicting carbon emissions of residential users based on energy-carbon accounting, comprising the following steps:

[0008] Obtain historical energy consumption data of residential users, meteorological data of the corresponding period, and house area data of corresponding energy consumption users;

[0009] Calculate carbon emissions based on historical energy consumption data of residential users to obtain carbon emissions series data;

[0010] The constructed carbon emission prediction model is trained based on the carbon emission series data, the meteorological data of the corresponding period, and the house area data of the corresponding energy consumption users to obtain the trained carbon emission prediction model; wherein, the construction process of the carbon emission prediction model includes:

[0011] The extracted residential carbon emission series data are processed in the frequency domain and time domain to obtain the frequency domain features containing the periodic information of the carbon emission series and the time domain features containing the short-term dependency of the carbon emission series;

[0012] Combine frequency domain characteristics, time domain characteristics, meteorological data and house area data to predict the energy consumption and carbon emission data of residential users in the next period;

[0013] Based on the trained carbon emission prediction model, the energy consumption carbon emission of residential users in the future period is predicted to obtain the energy consumption carbon emission of residential users in the future period.

[0014] Furthermore, the carbon emission accounting is performed based on the historical energy consumption data of residential users to obtain the carbon emission series data, including:

[0015] Multiply the electricity consumption data of residential users by the electricity carbon emission coefficient to obtain the carbon emissions generated by the household electricity consumption of the residential user;

[0016] Multiply the water consumption data of residential users by the carbon emission coefficient of tap water to obtain the carbon emissions generated by the household water consumption of the residential user;

[0017] Multiply the gas consumption data of residential users by the natural gas carbon emission coefficient to obtain the carbon emissions generated by the household gas consumption of the residential users;

[0018] Superimpose the time series data of the same dimension corresponding to the carbon emissions generated by electricity consumption, the carbon emissions generated by water consumption, and the carbon emissions generated by gas consumption to obtain the total carbon emissions from energy consumption of the residential user;

[0019] The total carbon emission series data of residential users’ energy consumption are normalized to obtain the residential users’ carbon emission series data.

[0020] Furthermore, a temporal convolutional network is used to extract the residential users’ carbon emission sequence data for frequency domain processing to obtain frequency domain features containing the periodic information of the carbon emission sequence.

[0021] Furthermore, the extracted residential carbon emission series data is processed in the time domain to obtain the time domain features containing the short-term dependency of the carbon emission series, including:

[0022] Calculate the degree of forgotten information f of the carbon emission sequence data of the current resident user in the SRU network t ;

[0023] Calculate the current newly formed cell state based on the cell state at the previous moment, the degree of forgotten information, and the required updated information;

[0024] According to the current cell state, update degree, and carbon emission sequence data of resident users at the current moment, the output at the current moment is calculated, that is, the time domain characteristics containing the short-term dependency of the carbon emission sequence.

[0025] Furthermore, the method also includes preprocessing the historical energy consumption data of residential users, the meteorological data of the corresponding time period, and the house area data of the corresponding energy consumption users, including the elimination of invalid data and the filling of missing data.

[0026] Furthermore, the BP neural network is used to predict the energy consumption and carbon emission data of residential users in the next period by combining frequency domain characteristics, time domain characteristics, meteorological data and house area data.

[0027] Furthermore, the meteorological information of a month t is defined as a tuple p n =(t,h n ),n=1,2,3 means, where p 1 ,h 1 Respectively represent the temperature information and average temperature data of a month t, p 2 ,h 2 They represent the sunshine duration information and average sunshine duration data of a month t, p 3 ,h 3 They respectively represent the precipitation information and average precipitation data of month t.

[0028] A second aspect of the present invention provides a residential user carbon emission prediction system based on energy-carbon accounting, comprising:

[0029] The data acquisition module is used to obtain the historical energy consumption data of residential users, the meteorological data of the corresponding time period, and the house area data of the corresponding energy consumption users;

[0030] A carbon emission sequence data acquisition module is used to calculate carbon emissions based on historical energy consumption data of residential users to obtain carbon emission sequence data;

[0031] The carbon emission prediction model training module is used to train the constructed carbon emission prediction model based on the carbon emission sequence data, the meteorological data of the corresponding period, and the house area data of the corresponding energy consumption user to obtain the trained carbon emission prediction model; wherein, the construction process of the carbon emission prediction model includes:

[0032] The extracted residential carbon emission series data are processed in the frequency domain and time domain to obtain the frequency domain features containing the periodic information of the carbon emission series and the time domain features containing the short-term dependency of the carbon emission series;

[0033] Combine frequency domain characteristics, time domain characteristics, meteorological data and house area data to predict the energy consumption and carbon emission data of residential users in the next period;

[0034] The carbon emission prediction module is used to predict the carbon emission of energy consumption of residential users in the future period based on the trained carbon emission prediction model to obtain the carbon emission of energy consumption of residential users in the future period.

[0035] A third aspect of the present invention provides a computer-readable storage medium.

[0036] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in the above-mentioned method for predicting carbon emissions of residential users based on energy-carbon accounting.

[0037] A fourth aspect of the present invention provides a computer device.

[0038] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for predicting carbon emissions of residential users based on energy-carbon accounting as described above are implemented.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. The present invention combines data from various dimensions of users with residents' energy consumption habits and individual residents' periodic energy consumption characteristics, including frequency domain characteristics of periodic information of carbon emission sequences and time domain characteristics including short-term dependencies of carbon emission sequences. It can better and more targetedly predict energy consumption and carbon emissions for residential users, realize individualized prediction of energy consumption and carbon emissions, and formulate carbon emission reduction strategies more accurately.

[0041] 2. The present invention adopts a time convolutional network combined with a simple cyclic unit network to obtain the periodic characteristics and short-term sequence dependencies of the historical carbon emission data of residential users, so that the expression of the sequence characteristics of carbon emission data is richer, which can improve the accuracy of carbon emission prediction of residential users. The time convolutional network used is more suitable for frequency domain analysis of time series data. Combined with the characteristics of parallel calculation of simple cyclic units, the model calculation rate is improved, which can meet the actual prediction tasks with different time frequencies and high real-time requirements.

[0042] 3. The present invention collects data on electricity, tap water and gas from residential users, covers the energy consumption of residential users in all dimensions, and provides a more realistic reflection of the carbon emissions of residential energy consumption.

[0043] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0045] Figure 1 This is a flow chart of a method for predicting carbon emissions of residential users based on energy-carbon accounting provided by an embodiment of the present invention;

[0046] Figure 2 is a data preprocessing flow chart provided by an embodiment of the present invention;

[0047] Figure 3 It is an energy carbon accounting framework diagram provided by an embodiment of the present invention;

[0048] Figure 4 It is an overall framework diagram of the carbon emission prediction model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0050] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0051] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0052] In view of the lack of carbon emission prediction for individual residents mentioned in the background technology, the single energy dimension, and the fact that residents' energy consumption habits are often ignored, the model fails to learn the periodic characteristics of individual residents' energy consumption, affecting the accuracy of residents' carbon emission prediction; the purpose of the present invention is to truly reflect the carbon emission situation of energy consumption of residential users, to make more accurate and rapid carbon emission predictions for individual residents, and to propose a residential user carbon emission prediction method based on energy-carbon accounting. By performing energy-carbon accounting on various energy consumption data, the historical carbon emission data of residential users is obtained, and the carbon emission time series data features are captured based on the time convolution network and the simple cyclic unit network. Finally, the features are input into the BP neural network model, and the meteorological data and house area data are integrated to achieve accurate prediction of the carbon emissions of residential users in the future period, so as to provide a data basis for formulating residential carbon emission reduction strategies, and provide intuitive data reflection for residents to carry out low-carbon life.

[0053] Embodiment 1

[0054] like Figure 1 As shown, this embodiment provides a method for predicting carbon emissions of residential users based on energy-carbon accounting, comprising the following steps:

[0055] Step 1: Obtain the monthly energy consumption data of the residential users to be analyzed, the meteorological data of the corresponding period, the house area data of the corresponding energy consumption users, etc.;

[0056] In this embodiment, the energy consumption information of a residential user in a certain month t is defined as a tuple q m =(t,l m ), m = 1, 2, 3 means, where q 1 ,l 1 Respectively represent the electricity consumption information and electricity consumption data of residential users, q 2 ,l 2 Respectively represent the water consumption information and water consumption data of residential users, q 3 ,l 3 Respectively represent the gas usage information and gas usage data of residential users;

[0057] Given a residential user u, the energy consumption sequence E of the user is represented by a tuple sequence, namely

[0058] The weather information of a month t is defined by a tuple p n =(t,h n ),n=1,2,3 means, where p 1 ,h 1 Respectively represent the temperature information and average temperature data of a month t, p 2 ,h 2 They represent the sunshine duration information and average sunshine duration data of a month t, p3 ,h 3 Respectively represent the precipitation information and average precipitation data of a month t;

[0059] Given a residential user area v, the meteorological information sequence W of the area is represented by a tuple sequence, namely

[0060] Define the building area of ​​a residential user u as c, then the house area sequence S of the user is expressed as: S u =c1c2…c n , where c1=c2=c n =c.

[0061] Step 2: Preprocess the acquired data, including removing invalid data and filling in missing data;

[0062] like Figure 2 As shown in Figure 2, if the recent monthly energy consumption data of residential user u are all zero, that is: The residential user u is judged to be in an idle state, and the carbon emissions of the residential user are not predicted. The sequence data E is removed. u , W v , S u ;

[0063] If the sequence data E u , W v There are missing values ​​in Then use Fill in missing values. If the sequence data E u If there are more than 5 missing values ​​NA in the , the carbon emissions of the residential user u will not be predicted, and the sequence data E will be removed. u , W v , S u .

[0064] Step 3: Obtain the energy consumption data E of resident u u Carry out carbon emission accounting, that is, energy consumption data E u Convert to carbon emission data M mu , m = 1, 2, 3, and after superposition, the total amount of energy consumption and carbon emissions of residents during this period is obtained through normalization processing M u ;

[0065] like Figure 3 As shown, the specific steps include:

[0066] Step 301: The power consumption data l of the residential user u obtained in step 2 is 1 Multiply them by the electricity carbon emission coefficient α to get the carbon emissions m generated by the electricity consumption of the household u 1u= l 1u ·α, aggregates the electricity-carbon emissions data of t months and stores them in the form of one-dimensional time series data as M 1u =[m1 1u ,m2 1u ,m3 1u ,...,m t 1u ] T ,In this embodiment, according to the data published by the Ministry of Science and Technology of China, the electricity carbon emission coefficient α=0.97kg CO2 / kwh;

[0067] Step 302: The water consumption data l of the residential user u obtained in step 2 is 2 Multiply them by the tap water carbon emission coefficient β to get the carbon emission m generated by the water consumption of household u 2u = l 2u β, aggregates the tap water-carbon emission data of t months and stores it in the form of one-dimensional time series data as M 2u =[m1 2u ,m2 2u ,m3 2u ,...,m t 2u ] T In this embodiment, according to the data published by the Ministry of Science and Technology of China, the carbon emission coefficient of tap water is β = 0.30 kg CO2 / m 3 ;

[0068] Step 303: The gas consumption data l of the residential user u obtained in step 2 is 3 Multiply them by the natural gas carbon emission coefficient γ to obtain the carbon emissions m generated by the gas consumption of the residential user household u 3u = l 3u γ, aggregates the natural gas-carbon emissions data for t months and stores them in the form of one-dimensional time series data as M 3u =[m1 3u ,m2 3u ,m3 3u ,...,m t 3u ] T In this embodiment, according to the data published by the Ministry of Science and Technology of China, the carbon emission coefficient of natural gas is γ = 2.2 kg CO2 / m 3 ;

[0069] Step 304: superimpose the time series data of the same dimension obtained in steps 301 to 303 to obtain the monthly total energy consumption and carbon emissions of the residential user:

[0070]

[0071] Step 305: Calculate the monthly energy consumption and carbon emission total series data M u Perform normalization processing and set Sequence M u The minimum and maximum values ​​in the normalized residential user carbon emission series data are It is also stored in the form of one-dimensional time series data.

[0072] Step 4: Train the constructed carbon emission prediction model based on the carbon emission series data, the meteorological data of the corresponding period, and the house area data of the corresponding energy consumption users to obtain the trained carbon emission prediction model;

[0073] The specific steps include:

[0074] Step 401, extracting frequency domain features containing periodic information of carbon emission sequences and time domain features containing short-term dependency of carbon emission sequences from carbon emission sequence data of residential users;

[0075] like Figure 4 As shown, the specific steps include:

[0076] Step 4011: Extracting residential user carbon emission sequence data M′ based on the time convolutional network u Perform frequency domain processing to obtain frequency domain features containing periodic information of carbon emission sequences;

[0077] In this embodiment, the temporal convolutional network (TCN) model includes a causal convolutional layer, a dilated convolutional layer, and a residual block network structure;

[0078] Specifically, the causal convolution layer ensures that the network output depends only on the input of current and historical data, that is, for the time series data of residents' energy consumption and carbon emissions obtained in step 3 The value of time t is determined only by the current and previous inputs Decisions are made, thus achieving strict time constraints.

[0079] To solve the problem of limited receptive field of the causal convolutional layer, a dilation factor d is applied to the filter feature f through dilated convolution, so that the network can scan the input sequence with a larger stride, thereby expanding the receptive field.

[0080] Specifically, for a one-dimensional time series input feature , dilated convolution Operations on time series elements The above definition is: , In the formula, Represents a time series element The convolution operation in represents the filter size, represents the expansion factor, By inputting time series data Apply the dilation factor To achieve, Indicates the past direction, by adjusting and , which can increase the network's receptive field, accept a wider input range, and improve the model's operating efficiency through parallelism.

[0082] Where H(T) represents the convolution operation in the time series element T, n represents the filter size, d represents the dilation factor, and M′ u * d By inputting the timing data M′ u The expansion factor d is applied above to achieve this. Td·i represents the past direction. By adjusting n and d, the network receptive field can be increased to accept a wider input range, and the model operation efficiency can be improved through parallelization.

[0083] In order to solve the problem of gradient vanishing and training instability that occurs as the number of network layers increases, a residual block network structure is adopted. Each residual block includes a main path for the input value M′. u (h-1) performs the transformation operation, and a 1×1 convolution bypass to maintain the consistency of the number of features. The output M′ of the h-th residual block u (h) The calculation formula is:

[0084] M′ u (h) = δ(F(M′) u(h-1) )+M′ u(h-1) ),

[0085] In the formula, M′ u (h) Frequency domain features of the h-th residual block, M′ u(h-1) is the frequency domain feature of the h-1th residual block, δ is the activation function, which is used to introduce nonlinear features. δ activates the output of the residual block to ensure the nonlinear representation of the features;

[0086] Step 4012: extracting the time domain features of the residential user carbon emission sequence data including the short-term dependency of the carbon emission sequence based on the simple cyclic unit network;

[0087] In this embodiment, the carbon emission series data M′ of residential users is analyzed by using a simple cyclic unit network model constructed by SRU. u The time domain processing was carried out to obtain the time domain characteristics with short-term dependence of carbon emission series;

[0088] Build an SRU network, set the number of network layers to 3, the number of SRU neurons in the first layer to 8, and the number of neurons in subsequent layers to decrease by integer multiples of 2. The network initialization method is selected as uniform distribution initialization, and the activation function is set to tanh function and sigmoid function. The function expression is:

[0089]

[0090] The number of neural iterations is 100, the batch processing parameter is 8, and the number of neuron jitters is 5;

[0091] Specifically, for the current input data m t ' u , calculate the degree of forgotten information in the network f t , the calculation formula is:

[0092]

[0093] Determine the information to be updated

[0094] Determine the degree of update information

[0095] Among them, W f , W r 、b f 、b r are the parameter matrices of the network;

[0096] According to the cell state C at the last moment t-1 , the degree of forgotten information f t , Required update information Calculate the current newly formed cell state C t , the calculation formula is as follows:

[0097]

[0098] According to the current cell state C t , update degree t , Current Input Calculate the current output h t , the calculation formula is as follows:

[0099]

[0100] Therefore, the internal information transmission of SRU only has the current input Participate in the operation and output h at the current moment t The calculation does not depend on the output h at the previous moment t-1This structure enables SRU to complete independent parallel operations, greatly improving the computational efficiency, and finally obtaining the time domain feature h t .

[0101] Step 402: The frequency domain feature M′ of the sequence obtained in step 401 u (h), time domain features h t , combined with the meteorological data W processed in step 2 v And the house area data S u , output the prediction result;

[0102] This embodiment constructs a 3-layer BP neural network. Figure 4 , we can determine that the number of neurons in the input layer is 6, that is, the frequency domain feature M′ of the input u (h), time domain features h t , and the mean temperature series Average sunshine duration series Average precipitation series And the house area sequence S u , the energy consumption and carbon emission data of residential users in the next period As the output result, the number of neurons in the output layer is 1;

[0103] Specifically, according to Kolmogorov's theorem, the formula for determining the number of neurons in the hidden layer is: Where n is the number of neurons in the input layer, m is the number of neurons in the output layer, and the value range of α is 1 to 10. In this embodiment, the number of input neurons n=6, and the number of output neurons m=1. It can be seen that the value range of the number of neurons in the hidden layer is 4 to 13. Through experiments, when the number of neurons in the hidden layer is 5, the prediction error value of the BP neural network is the smallest, that is, this embodiment confirms that the number of neurons in the hidden layer is 5;

[0104] Specifically, this embodiment determines that the hidden layer of the BP neural network uses the Sigmoid function as the transfer function, that is:

[0105]

[0106] The output layer uses the linear transfer function Purelin as the transfer function, that is:

[0107] f(x)=ax+b.

[0108] Step 5: Train the carbon emission prediction model. Train the end-to-end carbon emission prediction model constructed in step 4 to obtain model hyperparameters.

[0109] The specific steps include:

[0110] Step 501: train the carbon emission prediction model, and transform the meteorological data W in step 2 into v , Housing area data S u , and the time series data M of residents' energy consumption and carbon emission obtained in step 3 u 90% of the samples are divided into model training sets, and 10% of the samples are divided into test sets. The constructed end-to-end carbon emission prediction model is trained to obtain model hyperparameters;

[0111] Step 502: Optimize network hyperparameters through a large number of experiments, and use early stopping method to prevent model overfitting;

[0112] Step 503: Select the square loss function as the learning criterion, and its formula is:

[0113]

[0114] Among them, y represents the actual energy consumption and carbon emissions of residential users, f(x; θ) represents the energy consumption and carbon emissions of residential users predicted by the model, x is the input of the model, θ is the model parameter, N represents the total number of training samples, and in this embodiment, the Adam optimization algorithm is used to train the model.

[0115] Step 6: Based on the carbon emission prediction model trained in step 5, the energy consumption carbon emissions of residential user u in the future period t+1 are predicted to provide a data basis for the subsequent formulation of residential carbon emission control strategies, and the predicted carbon emission data can be pushed to residential users.

[0116] Step 601: Based on the carbon emission prediction model trained in step 5, the monthly energy consumption carbon emission time series data M′ of the residential user u obtained in step 3 is input. u , predict the energy consumption and carbon emissions of residential users in the future period t+1

[0117] Step 602: Gather energy consumption and carbon emission forecast data of multiple residential users Provide data basis for formulating carbon emission reduction strategies for residential users;

[0118] Step 603: The predicted carbon emission data By pushing the information to residential users, the residents who generate carbon emissions can have a more intuitive feeling and be more aware of the carbon emissions generated by their energy consumption activities, so as to impose appropriate constraints on the carbon emissions generated by energy consumption and guide the public to live a low-carbon life.

[0119] Embodiment 2

[0120] This embodiment provides a residential user carbon emission prediction system based on energy-carbon accounting, including:

[0121] The data acquisition module is used to obtain the historical energy consumption data of residential users, the meteorological data of the corresponding time period, and the house area data of the corresponding energy consumption users;

[0122] A carbon emission sequence data acquisition module is used to calculate carbon emissions based on historical energy consumption data of residential users to obtain carbon emission sequence data;

[0123] The carbon emission prediction model training module is used to train the constructed carbon emission prediction model based on the carbon emission sequence data, the meteorological data of the corresponding period, and the house area data of the corresponding energy consumption user to obtain the trained carbon emission prediction model; wherein, the construction process of the carbon emission prediction model includes:

[0124] The extracted residential carbon emission series data are processed in the frequency domain and time domain to obtain the frequency domain features containing the periodic information of the carbon emission series and the time domain features containing the short-term dependency of the carbon emission series;

[0125] Combine frequency domain characteristics, time domain characteristics, meteorological data and house area data to predict the energy consumption and carbon emission data of residential users in the next period;

[0126] The carbon emission prediction module is used to predict the carbon emission of energy consumption of residential users in the future period based on the trained carbon emission prediction model to obtain the carbon emission of energy consumption of residential users in the future period.

[0127] Embodiment 3

[0128] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps in the method for predicting carbon emissions of residential users based on energy-carbon accounting as described above are implemented.

[0129] Embodiment 4

[0130] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for predicting carbon emissions of residential users based on energy-carbon accounting as described above are implemented.

[0131] 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 hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, 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 and optical storage, etc.) containing computer-usable program code.

[0132] The present invention is described with reference to flowcharts and / or block diagrams of methods, 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.

[0133] 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.

[0134] 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.

[0135] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0136] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting carbon emissions from residential users based on energy-carbon accounting, characterized in that: The steps include: Obtain historical energy consumption data of residential users, meteorological data of the corresponding period, and house area data of corresponding energy consumption users; Calculate carbon emissions based on historical energy consumption data of residential users to obtain carbon emissions series data; The constructed carbon emission prediction model is trained based on the carbon emission series data, the meteorological data of the corresponding period, and the house area data of the corresponding energy consumption users to obtain the trained carbon emission prediction model; wherein, the construction process of the carbon emission prediction model includes: The extracted residential carbon emission series data are processed in the frequency domain and time domain to obtain the frequency domain features containing the periodic information of the carbon emission series and the time domain features containing the short-term dependency of the carbon emission series; Combine frequency domain characteristics, time domain characteristics, meteorological data and house area data to predict the energy consumption and carbon emission data of residential users in the next period; Based on the trained carbon emission prediction model, the energy consumption carbon emission of residential users in the future period is predicted to obtain the energy consumption carbon emission of residential users in the future period.

2. A residential user carbon emission prediction method based on energy-carbon accounting as claimed in claim 1, characterized in that: The carbon emission accounting is performed based on the historical energy consumption data of residential users to obtain the carbon emission series data, including: Multiply the electricity consumption data of residential users by the electricity carbon emission coefficient to obtain the carbon emissions generated by the household electricity consumption of the residential user; Multiply the water consumption data of residential users by the carbon emission coefficient of tap water to obtain the carbon emissions generated by the household water consumption of the residential user; Multiply the gas consumption data of residential users by the natural gas carbon emission coefficient to obtain the carbon emissions generated by the household gas consumption of the residential users; Superimpose the time series data of the same dimension corresponding to the carbon emissions generated by electricity consumption, the carbon emissions generated by water consumption, and the carbon emissions generated by gas consumption to obtain the total carbon emissions from energy consumption of the residential user; The total carbon emission series data of residential users’ energy consumption are normalized to obtain the residential users’ carbon emission series data.

3. A residential user carbon emission prediction method based on energy-carbon accounting as claimed in claim 1, characterized in that: A temporal convolutional network is used to extract the residential users' carbon emission series data for frequency domain processing to obtain the frequency domain features containing the periodic information of the carbon emission series.

4. A residential user carbon emission prediction method based on energy-carbon accounting as claimed in claim 1, characterized in that: The extracted residential carbon emission series data is processed in the time domain to obtain the time domain features containing the short-term dependency of the carbon emission series, including: Calculate the degree of forgotten information f of the carbon emission sequence data of the current resident user in the SRU network t ; Calculate the current newly formed cell state based on the cell state at the previous moment, the degree of forgotten information, and the required updated information; According to the current cell state, update degree, and carbon emission sequence data of resident users at the current moment, the output at the current moment is calculated, that is, the time domain characteristics containing the short-term dependency of the carbon emission sequence.

5. A residential user carbon emission prediction method based on energy-carbon accounting as claimed in claim 1, characterized in that: The method also includes preprocessing the historical energy consumption data of residential users, the meteorological data of the corresponding time period, and the house area data of the corresponding energy consumption users, including the elimination of invalid data and the filling of missing data.

6. A residential user carbon emission prediction method based on energy-carbon accounting as claimed in claim 1, characterized in that: Combining frequency domain characteristics, time domain characteristics, meteorological data and house area data, BP neural network is used to predict the energy consumption and carbon emission data of residential users in the next period.

7. A residential user carbon emission prediction method based on energy-carbon accounting as claimed in claim 1, characterized in that: The weather information of a month t is defined by a tuple p n =(t,h n ),n=1,2,3 means, where p 1 ,h 1 Respectively represent the temperature information and average temperature data of a month t, p 2 ,h 2 They represent the sunshine duration information and average sunshine duration data of a month t, p 3 ,h 3 They respectively represent the precipitation information and average precipitation data of month t.

8. A residential user carbon emission prediction system based on energy-carbon accounting, characterized in that: include: The data acquisition module is used to obtain the historical energy consumption data of residential users, the meteorological data of the corresponding time period, and the house area data of the corresponding energy consumption users; A carbon emission sequence data acquisition module is used to calculate carbon emissions based on historical energy consumption data of residential users to obtain carbon emission sequence data; The carbon emission prediction model training module is used to train the constructed carbon emission prediction model based on the carbon emission sequence data, the meteorological data of the corresponding period, and the house area data of the corresponding energy consumption user to obtain the trained carbon emission prediction model; wherein, the construction process of the carbon emission prediction model includes: The extracted residential carbon emission series data are processed in the frequency domain and time domain to obtain the frequency domain features containing the periodic information of the carbon emission series and the time domain features containing the short-term dependency of the carbon emission series; Combine frequency domain characteristics, time domain characteristics, meteorological data and house area data to predict the energy consumption and carbon emission data of residential users in the next period; The carbon emission prediction module is used to predict the carbon emission of energy consumption of residential users in the future period based on the trained carbon emission prediction model to obtain the carbon emission of energy consumption of residential users in the future period.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a residential user carbon emission prediction method based on energy-carbon accounting as described in any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in a residential user carbon emission prediction method based on energy-carbon accounting as described in any one of claims 1-7 are implemented.

Images