A carbon emission calculation method and device based on big data

Abstract

The application provides a carbon emission calculation method and device based on big data, the method comprises the following steps: obtaining load power information, power grid voltage information and power grid current information of a power grid; setting configuration parameters for the power information, power grid voltage information and power grid current information according to the attribute characteristics of input power, and generating a prediction set; inputting the prediction set into a convolutional neural network training, including encoding data, inputting the encoded data into a convolutional layer, the convolutional layer extracts feature data according to the time period continuity of the data, and generates a time power graph and a time load graph according to the feature data, inputs the time power graph and the time load graph into a fusion layer for fusion to generate a fusion graph; and calculating carbon emission according to the fusion graph. The application directly calculates the carbon emission of the power grid voltage, current and load data by using the convolutional neural network model, the data acquisition is simple, the calculation is relatively accurate, and the applicability is strong.

Description

Technical Field

[0001] This application claims protection for a big data prediction method, and more particularly relates to a big data-based carbon emission calculation method. This application also relates to a big data-based carbon emission calculation device. Background Technology

[0002] Electricity is mainly obtained through thermal power, which is generated by burning fossil fuels. This process emits carbon dioxide and other carbon-containing gases, known as carbon emissions.

[0003] Currently, green development is considered the main direction of social development, and controlling carbon emissions is one of the important aspects of green development. The first step in controlling carbon emissions is to conduct carbon emission calculations.

[0004] In existing technologies, carbon emissions are calculated through a series of detections, which requires extremely precise measurement methods and a large amount of calculation. This is very difficult for ordinary enterprises to obtain, so there is an urgent need for a simple and direct method to calculate carbon emissions. Summary of the Invention

[0005] To address one or more of the technical problems described above, this application provides a carbon emission calculation method based on big data. This application also relates to a carbon emission calculation device based on big data.

[0006] This application provides a carbon emission measurement method based on big data, including:

[0007] Acquire load power information, grid voltage information, and grid current information of the power grid;

[0008] Based on the attribute characteristics of the input power, configuration parameters are set for the power information, grid voltage information, and grid current information, and a prediction set is generated;

[0009] The prediction set is input into a convolutional neural network for training, including encoding the data, inputting the encoded data into a convolutional layer, the convolutional layer extracting feature data based on the time period continuity of the data, generating a time-based power consumption map and a time-based load map based on the feature data, and inputting the time-based power consumption map and the time-based load map into a fusion layer for fusion to generate a fused map;

[0010] Carbon emissions are calculated based on the fusion diagram.

[0011] Optionally, the power grid includes a public power grid, a local power grid, and a residential power grid;

[0012] Optionally, the training method for the convolutional neural network includes:

[0013] Extract historical data on grid load power, grid voltage, and grid current;

[0014] A sample set is generated based on the historical data and configured parameters.

[0015] The data from the sample set is input into the convolutional neural network model to train and obtain the fusion map;

[0016] Adjust the parameters of the convolutional neural network based on the fused graph, and repeat the above steps until the fused graph meets expectations.

[0017] Optionally, the configuration parameter is the ratio of power generation to electricity consumption in the power grid.

[0018] Optionally, the configuration parameter is the ratio of fossil fuel power to clean energy power in the power grid.

[0019] This application also provides a carbon emission measurement device based on big data, comprising:

[0020] The acquisition module is used to acquire load power information, grid voltage information, and grid current information of the power grid;

[0021] The preprocessing module is used to set configuration parameters for the power information, grid voltage information and grid current information based on the attribute characteristics of the input power, and generate a prediction set;

[0022] An analysis module is used to input the prediction set into a convolutional neural network for training, including encoding the data, inputting the encoded data into a convolutional layer, the convolutional layer extracting feature data based on the time period continuity of the data, generating a time-based power consumption map and a time-based load map based on the feature data, and inputting the time-based power consumption map and the time-based load map into a fusion layer for fusion to generate a fused map;

[0023] The calculation module is used to calculate carbon emissions based on the fusion graph.

[0024] Optionally, the power grid includes a public power grid, a local power grid, and a residential power grid;

[0025] Optionally, a training module may also be included;

[0026] The training module includes:

[0027] The extraction unit is used to extract historical data of the grid load power, grid voltage, and grid current.

[0028] A sample unit is used to generate a sample set after configuring parameters based on the historical data.

[0029] The training unit is used to train the convolutional neural network model based on the data input from the sample set and obtain the fusion map.

[0030] An adjustment unit is used to adjust the parameters of the convolutional neural network based on the fusion graph, and repeat the above steps until the fusion graph meets expectations.

[0031] Optionally, the configuration parameter is the ratio of power generation to electricity consumption in the power grid.

[0032] Optionally, the configuration parameter is the ratio of fossil fuel power to clean energy power in the power grid.

[0033] The advantages of this application compared to the prior art are:

[0034] This application provides a carbon emission calculation method based on big data, comprising: acquiring grid load power information, grid voltage information, and grid current information; setting configuration parameters for the power information, grid voltage information, and grid current information according to the attribute characteristics of the input electricity, and generating a prediction set; inputting the prediction set into a convolutional neural network for training, including encoding the data, inputting the encoded data into a convolutional layer, the convolutional layer extracting feature data based on the time continuity of the data, and generating a time-based electricity graph and a time-based load graph based on the feature data, inputting the time-based electricity graph and the time-based load graph into a fusion layer for fusion, generating a fused graph; and calculating carbon emissions based on the fused graph. This application uses a convolutional neural network model to directly calculate carbon emissions from grid voltage, current, and load data, which is simple to acquire, relatively accurate, and highly adaptable. Attached Figure Description

[0035] Figure 1 This is a flowchart of the carbon emission calculation based on big data in this application.

[0036] Figure 2 This is a flowchart of the training process for the convolutional neural network in this application.

[0037] Figure 3 This is a schematic diagram of the carbon emission measurement device based on big data in this application. Detailed Implementation

[0038] The following are examples of specific implementation processes provided to illustrate the technical solutions to be protected in this application. However, this application may also be implemented in other ways different from those described herein. Those skilled in the art can implement this application by different technical means under the guidance of the concept of this application. Therefore, this application is not limited to the specific embodiments below.

[0039] This application provides a carbon emission calculation method based on big data, comprising: acquiring grid load power information, grid voltage information, and grid current information; setting configuration parameters for the power information, grid voltage information, and grid current information according to the attribute characteristics of the input electricity, and generating a prediction set; inputting the prediction set into a convolutional neural network for training, including encoding the data, inputting the encoded data into a convolutional layer, the convolutional layer extracting feature data based on the time continuity of the data, and generating a time-based electricity graph and a time-based load graph based on the feature data, inputting the time-based electricity graph and the time-based load graph into a fusion layer for fusion, generating a fused graph; and calculating carbon emissions based on the fused graph. This application uses a convolutional neural network model to directly calculate carbon emissions from grid voltage, current, and load data, which is simple to acquire, relatively accurate, and highly adaptable.

[0040] Figure 1 This is a flowchart of the carbon emission calculation based on big data in this application.

[0041] Please refer to Figure 1 As shown, S101 acquires the load power information, grid voltage information, and grid current information of the power grid.

[0042] The load power information, grid voltage information, and grid current information shown are primarily data values. For example, the load power, grid voltage, and grid current are shown. The load power information, grid voltage information, and grid current information can be obtained through direct reading; for example, to read the load's power information, simply read the load's power parameter. The grid voltage information and grid current information can be read using voltage transformers and current transformers, and then converted into digital information via an analog-to-digital converter (AD converter).

[0043] Preferably, the power grid includes public power grids, local power grids, and household power grids.

[0044] Please refer to Figure 1 As shown, S102 sets configuration parameters for the power information, grid voltage information and grid current information based on the attribute characteristics of the input power, and generates a prediction set.

[0045] Specifically, the attribute characteristic refers to the carbon content of electricity in the power grid. In particular, the electricity in the power grid may be generated by multiple power sources, including thermal power and new energy sources such as wind power and hydropower. Only thermal power will produce carbon emissions, while new energy sources will not produce carbon emissions. Therefore, the attribute characteristic refers to a ratio, namely the carbon content of electricity.

[0046] Specifically, the ratio of new energy to thermal power is the attribute characteristic, and carbon emission analysis can be more accurate based on this attribute characteristic.

[0047] The carbon content of the electricity in the power grid is used as the parameter for calculation. The power grid voltage, power grid current and load power are reduced proportionally to form prediction data. The set of all prediction data is used as the prediction set.

[0048] Please refer to Figure 1 As shown, S103 inputs the prediction set into the convolutional neural network for training, including encoding the data, inputting the encoded data into the convolutional layer, the convolutional layer extracting feature data based on the time period continuity of the data, generating a time-based power consumption map and a time-based load map based on the feature data, and inputting the time-based power consumption map and the time-based load map into the fusion layer for fusion to generate a fused map.

[0049] Figure 2 This is a flowchart of the training process for the convolutional neural network in this application.

[0050] Please refer to Figure 2 As shown, S201 extracts historical data of the grid load power, grid voltage, and grid current;

[0051] S202 generates a sample set based on the historical data configuration parameters;

[0052] S203 Input the data from the sample set into the convolutional neural network model, train it, and obtain the fusion map;

[0053] S204 adjusts the parameters of the convolutional neural network based on the fusion graph, and repeats the above steps until the fusion graph meets expectations.

[0054] The configuration parameter is the ratio of power generation to electricity generation in the power grid, or the configuration parameter is the ratio of fossil fuel electricity to clean energy electricity in the power grid. In this application, the configuration parameter can also be the product of the above two ratios.

[0055] The convolutional neural network has a first convolutional layer, a second convolutional layer, and a third convolutional layer. After the prediction set is input into the first convolutional layer, power information of the power grid is generated based on the grid voltage and current information. The power information is then arranged in chronological order, and key points are extracted. The expression for determining these key points is as follows:

[0056] A(x, y) = (t i p i ), p i -p i -1 <B,p i -p i +1

[0057] Where A represents a key point, and t​i p represents the time points taken at preset intervals. i Indicates t i The power information at a given time point, where B represents the power difference threshold, which is the power difference between two adjacent time points.

[0058] The keypoints selected using the above expression represent the most representative power information, and these keypoints are used as the first input data for the second convolutional layer. In practice, the first convolutional layer also needs to select keypoints for load power in the same way, and these selected load power keypoints are then used as the second input data for the second convolutional layer.

[0059] The second convolutional layer further filters the key points of the first and second input data to generate the third and fourth input data, which are then input into the third convolutional layer.

[0060] The third convolutional layer further filters the key points of the third and fourth input data, and generates power key points and load key points.

[0061] The fusion layer takes the power key points and load key points as input, and then fuses them based on time sequence. The load key points are numerically adjusted according to the power loss of the power grid itself. Power key points and load key points that overlap at the same time point are merged, and non-overlapping key points are deleted to generate a set of points representing carbon emission trends.

[0062] Please refer to Figure 1 As shown, S104 calculates carbon emissions based on the fusion diagram.

[0063] Specifically, a carbon emission trend map is generated based on the coordinates of the carbon emission trend points set on the time axis. The power grid energy consumption can be determined based on the trend map. Then, the energy consumption is multiplied by the carbon emission factor to obtain the carbon emission amount.

[0064] This application also provides a carbon emission measurement device based on big data, including an acquisition module 301, a preprocessing module 302, an analysis module 303, and a calculation module 304.

[0065] Figure 3 This is a schematic diagram of the carbon emission measurement device based on big data in this application.

[0066] The acquisition module 301 is used to acquire the load power information, grid voltage information and grid current information of the power grid.

[0067] The load power information, grid voltage information, and grid current information shown are primarily data values. For example, the load power, grid voltage, and grid current are shown. The load power information, grid voltage information, and grid current information can be obtained through direct reading; for example, to read the load's power information, simply read the load's power parameter. The grid voltage information and grid current information can be read using voltage transformers and current transformers, and then converted into digital information via an analog-to-digital converter (AD converter).

[0068] Preferably, the power grid includes public power grids, local power grids, and household power grids.

[0069] Please refer to Figure 3 As shown, the preprocessing module 302 is used to set configuration parameters for the power information, grid voltage information and grid current information according to the attribute characteristics of the input power, and generate a prediction set.

[0070] Specifically, the attribute characteristic refers to the carbon content of electricity in the power grid. In particular, the electricity in the power grid may be generated by multiple power sources, including thermal power and new energy sources such as wind power and hydropower. Only thermal power will produce carbon emissions, while new energy sources will not produce carbon emissions. Therefore, the attribute characteristic refers to a ratio, namely the carbon content of electricity.

[0071] Specifically, the ratio of new energy to thermal power is the attribute characteristic, and carbon emission analysis can be more accurate based on this attribute characteristic.

[0072] The carbon content of the electricity in the power grid is used as the parameter for calculation. The power grid voltage, power grid current and load power are reduced proportionally to form prediction data. The set of all prediction data is used as the prediction set.

[0073] Please refer to Figure 3 As shown, the analysis module 303 is used to input the prediction set into the training of the convolutional neural network, including encoding the data, inputting the encoded data into the convolutional layer, the convolutional layer extracting feature data according to the time period continuity of the data, generating a time-based power consumption map and a time-based load map according to the feature data, and inputting the time-based power consumption map and the time-based load map into the fusion layer for fusion to generate a fused map.

[0074] The convolutional neural network in this application is pre-trained using a training module, which includes:

[0075] The extraction unit is used to extract historical data of the grid load power, grid voltage, and grid current.

[0076] A sample unit is used to generate a sample set after configuring parameters based on the historical data.

[0077] The training unit is used to train the convolutional neural network model based on the data input from the sample set and obtain the fusion map.

[0078] An adjustment unit is used to adjust the parameters of the convolutional neural network based on the fusion graph, and repeat the above steps until the fusion graph meets expectations.

[0079] The configuration parameter is the ratio of power generation to electricity generation in the power grid, or the configuration parameter is the ratio of fossil fuel electricity to clean energy electricity in the power grid. In this application, the configuration parameter can also be the product of the above two ratios.

[0080] The convolutional neural network has a first convolutional layer, a second convolutional layer, and a third convolutional layer. After the prediction set is input into the first convolutional layer, power information of the power grid is generated based on the grid voltage and current information. The power information is then arranged in chronological order, and key points are extracted. The expression for determining these key points is as follows:

[0081] A(x, y) = (t i p i ), p i -p i -1 <B,p i -p i +1

[0082] Where A represents a key point, and t i p represents the time points taken at preset intervals. i Indicates t i The power information at a given time point, where B represents the power difference threshold, which is the power difference between two adjacent time points.

[0083] The keypoints selected using the above expression represent the most representative power information, and these keypoints are used as the first input data for the second convolutional layer. In practice, the first convolutional layer also needs to select keypoints for load power in the same way, and these selected load power keypoints are then used as the second input data for the second convolutional layer.

[0084] The second convolutional layer further filters the key points of the first and second input data to generate the third and fourth input data, which are then input into the third convolutional layer.

[0085] The third convolutional layer further filters the key points of the third and fourth input data, and generates power key points and load key points.

[0086] ​The fusion layer takes the power key points and load key points as input, and then fuses them based on time sequence. The load key points are numerically adjusted according to the power loss of the power grid itself. Power key points and load key points that overlap at the same time point are merged, and non-overlapping key points are deleted to generate a set of points representing carbon emission trends.

[0087] Please refer to Figure 3 As shown, the calculation module 303 is used to calculate carbon emissions based on the fusion graph.

[0088] Specifically, a carbon emission trend map is generated based on the coordinates of the carbon emission trend points set on the time axis. The power grid energy consumption can be determined based on the trend map. Then, the energy consumption is multiplied by the carbon emission factor to obtain the carbon emission amount.

Claims

1. A big data-based carbon emission calculation method, characterized in that, include: Acquire load power information, grid voltage information, and grid current information of the power grid; Based on the attribute characteristics of the input power, configuration parameters are set for the power information, grid voltage information, and grid current information, and a prediction set is generated; The prediction set is input into a convolutional neural network for training, including encoding the data, inputting the encoded data into a convolutional layer, the convolutional layer extracting feature data based on the time period continuity of the data, generating a time-based power consumption map and a time-based load map based on the feature data, and inputting the time-based power consumption map and the time-based load map into a fusion layer for fusion to generate a fused map; Calculate carbon emissions based on the fusion diagram; The convolutional neural network has a first convolutional layer, a second convolutional layer, and a third convolutional layer. After the prediction set is input into the first convolutional layer, the power information of the power grid is generated based on the grid voltage information and the grid current information. The power information is arranged in chronological order, and key points are extracted. The expression for determining the key points is as follows: Where A represents a key point. This indicates the time points taken at preset intervals. express Power information at a given time point, where B represents the power difference threshold, i.e., the power difference between two adjacent time points; The key points selected by the above expression are the most representative power information. The selected key points are used as the first input data of the second convolutional layer. In the actual execution, the first convolutional layer also needs to select the key points of the load power. After selecting the key points in the same way, the selected load power key points are used as the second input data of the second convolutional layer. The second convolutional layer further filters the key points of the first and second input data, generating the third and fourth input data, which are then fed into the third convolutional layer. The third convolutional layer further filters the key points of the third and fourth input data and generates power key points and load key points. The fusion layer takes the power key points and load key points as input, and then merges the power key points and load key points based on time sequence. The load key points are numerically adjusted according to the power loss of the power grid itself. The power key points and load key points that overlap at the same time point are merged, and the non-overlapping key points are deleted to generate a set of points for carbon emission trends.

2. The carbon emission calculation method based on big data according to claim 1, characterized in that, The power grid includes public power grids, local power grids, and household power grids.

3. The carbon emission calculation method based on big data according to claim 1, characterized in that, The training method for the convolutional neural network includes: Extract historical data on grid load power, grid voltage, and grid current; A sample set is generated based on the historical data and configured parameters. The data from the sample set is input into the convolutional neural network model to train and obtain the fusion map; Adjust the parameters of the convolutional neural network based on the fused graph, and repeat the above steps until the fused graph meets expectations.

4. The carbon emission calculation method based on big data according to claim 1, characterized in that, The configuration parameter is the ratio of power generation to electricity consumption in the power grid.

5. The carbon emission calculation method based on big data according to claim 1, characterized in that, The configuration parameter is the ratio of fossil fuel power to clean energy power in the power grid.

6. A carbon emission measurement device based on big data, characterized in that, include: The acquisition module is used to acquire load power information, grid voltage information, and grid current information of the power grid; The preprocessing module is used to set configuration parameters for the power information, grid voltage information and grid current information based on the attribute characteristics of the input power, and generate a prediction set; An analysis module is used to input the prediction set into a convolutional neural network for training, including encoding the data, inputting the encoded data into a convolutional layer, the convolutional layer extracting feature data based on the time period continuity of the data, generating a time-based power consumption map and a time-based load map based on the feature data, and inputting the time-based power consumption map and the time-based load map into a fusion layer for fusion to generate a fused map; The calculation module is used to calculate carbon emissions based on the fusion graph; The convolutional neural network has a first convolutional layer, a second convolutional layer, and a third convolutional layer. After the prediction set is input into the first convolutional layer, the power information of the power grid is generated based on the grid voltage information and the grid current information. The power information is arranged in chronological order, and key points are extracted. The expression for determining the key points is as follows: Where A represents a key point. This indicates the time points taken at preset intervals. express Power information at a given time point, where B represents the power difference threshold, i.e., the power difference between two adjacent time points; The key points selected by the above expression are the most representative power information. The selected key points are used as the first input data of the second convolutional layer. In the actual execution, the first convolutional layer also needs to select the key points of the load power. After selecting the key points in the same way, the selected load power key points are used as the second input data of the second convolutional layer. The second convolutional layer further filters the key points of the first and second input data, generating the third and fourth input data, which are then fed into the third convolutional layer. The third convolutional layer further filters the key points of the third and fourth input data and generates power key points and load key points. The fusion layer takes the power key points and load key points as input, and then merges the power key points and load key points based on time sequence. The load key points are numerically adjusted according to the power loss of the power grid itself. The power key points and load key points that overlap at the same time point are merged, and the non-overlapping key points are deleted to generate a set of points for carbon emission trends.

7. The carbon emission measurement device based on big data according to claim 6, characterized in that, The power grid includes public power grids, local power grids, and household power grids.

8. The carbon emission measurement device based on big data according to claim 6, characterized in that, It also includes a training module; The training module includes: The extraction unit is used to extract historical data of the grid load power, grid voltage, and grid current. A sample unit is used to generate a sample set after configuring parameters based on the historical data. The training unit is used to train the convolutional neural network model based on the data input from the sample set and obtain the fusion map. An adjustment unit is used to adjust the parameters of the convolutional neural network based on the fusion graph, and repeat the above steps until the fusion graph meets expectations.

9. The carbon emission measurement device based on big data according to claim 6, characterized in that, The configuration parameter is the ratio of power generation to electricity consumption in the power grid.

10. The carbon emission measurement device based on big data according to claim 6, characterized in that, The configuration parameter is the ratio of fossil fuel power to clean energy power in the power grid.

Images