Method, medium and system for checking carbon emission above scale of chemical industry park

Abstract

The invention provides an accounting method, medium and system for carbon emission above a chemical industry park scale, and belongs to the technical field of carbon emission monitoring. Comprising the steps that S1, a yield prediction model is constructed, yield prediction is carried out based on electricity consumption and first external factors, a yield prediction result is obtained, and the first external factors comprise macroeconomic conditions, technical progress and energy prices; and S2, a carbon emission prediction model is constructed, carbon emission prediction is carried out based on the yield prediction result and second external factors, and the second external factors comprise a carbon emission policy, an energy-saving and carbon-reducing technology and market demands. According to the invention, the comprehensive influence of different links of the production process of the chemical industrial park and various external factors on carbon emission is deeply researched, and a carbon emission prediction technology is introduced to predict and check the carbon emission, so that the accuracy of the carbon emission check of the chemical industrial park is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission monitoring, and in particular to a method, medium and system for calculating carbon emissions above the scale of a chemical park. Background Art

[0002] Carrying out research on carbon emissions and electricity consumption forecasts in chemical parks is of great significance for systematically grasping carbon emissions, promoting the green transformation of the chemical industry, and achieving the "dual carbon" goals.

[0003] Traditional carbon emission accounting methods often only consider a single energy type or production process, ignoring the comprehensive impact of multiple factors on carbon emissions, such as energy structure, industrial structure, production process, and economic growth. As a result, the current carbon emission accounting for industries and enterprises in chemical parks is inaccurate and difficult to meet the needs of grid planning and management. Therefore, it is particularly important to develop a method that can comprehensively consider multiple influencing factors and accurately calculate the carbon emissions of chemical parks. Summary of the invention

[0004] In view of this, the present invention provides a method for calculating carbon emissions above the scale of a chemical park to solve the technical problem that traditional carbon emissions calculation methods often only consider a single energy type or production process, have poor calculation accuracy, and are difficult to meet the needs of grid planning and management.

[0005] The technical solution adopted by the present invention to solve its technical problem is: A method for calculating carbon emissions of chemical parks above a certain scale, comprising the following steps: S1. constructing a production forecasting model, and performing production forecasting based on electricity consumption and a first external factor to obtain a production forecasting result, wherein the first external factor includes macroeconomic conditions, technological progress, and energy prices; S2. Construct a carbon emission prediction model, and predict carbon emissions based on the output prediction result and a second external factor, wherein the second external factor includes carbon emission policies, energy-saving and carbon-reduction technologies, and market demand.

[0006] Preferably, in step S1, constructing a production forecasting model, and performing production forecasting based on power consumption and the first external factor specifically includes the following steps: S11. Collect and analyze electricity consumption information, production information and meteorological information, and clarify electricity consumption and production data; S12. Construct a VAR yield prediction model based on the VAR model and deep learning combined algorithm; S13, dividing the power consumption, output data and the first external factor data into a training set and a test set to train the VAR output prediction model; S14, evaluating the trained VAR yield prediction model; S15. Input the first external factor into the improved STIRPAT factor decomposition model to obtain the first external factor data, input the power consumption and the first external factor data into the evaluated VAR production forecasting model to perform production forecasting.

[0007] Preferably, in step S2, building a carbon emission prediction model, and predicting carbon emissions based on the output prediction result and the second external factor specifically includes the following steps: S21. Collect and analyze the company's monthly output, historical carbon emissions, production information and raw material information, and clarify the output data and carbon emissions data; S21. Construct a carbon emission prediction model based on a neural network and tree combination algorithm; S23, dividing the production data, carbon emission data and the second external factor data into a training set and a test set to train the carbon emission prediction model; S24. Evaluate the trained carbon emission prediction model; S25. Input the second external factor into the improved STIRPAT factor decomposition model to obtain the second external factor data, input the output forecast result and the second external factor data into the evaluated carbon emission forecast model to perform carbon emission forecasting.

[0008] Preferably, the method for constructing the improved STIRPAT factor decomposition model comprises the following steps: S31. Build a carbon emission indicator system; S32. Screen the characteristic indicators of the carbon emission indicator system and use the screening results as input indicators for improving the STIRPAT factor decomposition model; S33. Construct an improved STIRPAT factor decomposition model and solve the coefficients based on ridge regression.

[0009] Preferably, in step S32, screening the characteristic index of the carbon emission system index includes the following steps: S321. Sort out the monthly indicator data list; S322. Based on the monthly indicator data list, select the filtering method, the wrapping method, and the embedding method to screen the characteristic indicators of the carbon emission system indicators respectively to obtain the first characteristic combination, the second characteristic combination, and the third characteristic combination; S323. Determine a final feature combination from the first feature combination, the second feature combination, and the third feature combination based on a voting method, and use the final feature combination as an input indicator for improving the STIRPAT factor decomposition model.

[0010] Preferably, in step S33, constructing an improved STIRPAT factor decomposition model, and solving the coefficients based on ridge regression specifically includes the following steps: The basic form of the improved STIRPAT factor decomposition model is: lnI=lna+blnP+clnA+dlnT+lnu; the coefficients are solved using the logarithmic sum form and ridge regression.

[0011] Preferably, in step S23, the historical carbon emissions are obtained by calculating the carbon emissions based on the inventory compilation method.

[0012] Preferably, the carbon emissions accounting based on the inventory compilation method specifically includes the following steps: Organize data requirements, clarify data reserve lists and energy types; Based on the data reserve list and energy types, combined with the authoritative documents on greenhouse gas accounting, reporting guidelines and energy consumption above the scale of chemical parks, the IPCC inventory compilation method is used to construct a carbon emission inventory; Collect default values ​​of carbon content from carbon emission sources to estimate carbon emission factors; Carbon emissions are calculated based on the carbon emissions inventory and carbon emission factors.

[0013] The present invention also provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the above-mentioned method for calculating carbon emissions above the scale of a chemical park.

[0014] The present invention also provides a system for calculating carbon emissions above a certain scale in a chemical park, comprising: the computer-readable storage medium as described above.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention first constructs a production forecasting model, and performs production forecasting based on power consumption and a first external factor to obtain a production forecasting result; then constructs a carbon emission forecasting model, and performs carbon emission forecasting based on the production forecasting result and a second external factor. Thus, by deeply studying the different links of the production process of the chemical park and the comprehensive impact of various external factors on carbon emissions, the carbon emission forecasting technology is introduced to predict and calculate carbon emissions, thereby improving the accuracy of carbon emission accounting in the chemical park. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The figure is a flow chart of the method of the present invention.

[0017] Figure 2 A flowchart of the production forecasting model construction process.

[0018] Figure 3 A flowchart of the construction process of the carbon emission prediction model.

[0019] Figure 4A flowchart of the construction process for improving the STIRPAT factor decomposition model.

[0020] Figure 5 This is a flowchart of the IPCC inventory construction process. DETAILED DESCRIPTION

[0021] The technical solutions and technical effects of the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings of the present invention.

[0022] Please see Figure 1 A method for calculating carbon emissions of chemical parks above a certain scale comprises the following steps: S1. construct a production forecasting model, and perform production forecasting based on electricity consumption and a first external factor to obtain a production forecasting result. The first external factor includes macroeconomic conditions, technological progress, and energy prices; S2. Construct a carbon emission prediction model and make a carbon emission prediction based on the output prediction results and the second external factor. The second external factor includes carbon emission policies, energy-saving and carbon reduction technologies and market demand.

[0023] The present invention first constructs a production forecasting model, and performs production forecasting based on power consumption and a first external factor to obtain a production forecasting result; then constructs a carbon emission forecasting model, and performs carbon emission forecasting based on the production forecasting result and a second external factor. Thus, by deeply studying the different links of the production process of the chemical park and the comprehensive impact of various external factors on carbon emissions, the carbon emission forecasting technology is introduced to predict and calculate carbon emissions, thereby improving the accuracy of carbon emission accounting in the chemical park.

[0024] Specifically, an "electricity-energy (output)-carbon" accounting model is constructed for chemical parks above the scale, and the output data is used to predict carbon emissions. However, the output data has the defect of lag, which has a certain impact on the prediction of carbon emissions. Therefore, firstly, the defect of timeliness of output data is compensated by the real-time nature of electricity, and an output prediction model is established to predict output using electricity consumption. Then, based on the predicted results of output, combined with the production information of the enterprise, a carbon emission prediction model is constructed to predict carbon emissions, and finally a "electricity-energy (output)-carbon" prediction mechanism is formed to improve the accuracy of carbon emission accounting.

[0025] In some embodiments, the first external factor includes macroeconomic conditions, technological progress and energy prices. Among them, changes in macroeconomic conditions will affect the production activities of the enterprise, thereby affecting electricity consumption and output. For example, economic growth will drive enterprises to expand their production scale, increase electricity consumption and output. Advances in production processes and technologies can improve energy efficiency and reduce energy consumption per unit of output, thereby affecting electricity consumption and output. Changes in energy prices will affect the production costs and electricity use decisions of enterprises, thereby affecting output. For example, rising electricity prices may prompt enterprises to optimize energy use and reduce electricity consumption, but they may also compensate for cost increases by improving production efficiency, thereby maintaining or increasing output. Changes in these external factors will affect the prediction results of the output prediction model, so when applying the output prediction model for prediction, it is necessary to fully consider the changes in these external factors.

[0026] In some embodiments, the second external factor includes carbon emission policies, energy-saving and carbon reduction technologies, and market demand. Among them, changes in carbon emission policies will directly affect the company's carbon emission management. For example, restrictions on carbon emission quotas and changes in trading prices will prompt companies to take emission reduction measures. The development and application of energy-saving and carbon reduction technologies can reduce the company's carbon emission intensity. These technologies may include more efficient energy utilization methods, more environmentally friendly production processes, etc. Changes in market demand will also affect the company's production activities and carbon emissions. For example, increased market demand for environmentally friendly products will prompt companies to adopt more environmentally friendly production methods and reduce carbon emissions. Changes in these external factors will affect the prediction results of the carbon emission prediction model. Therefore, when applying the carbon emission prediction model for prediction, it is necessary to fully consider the changes in these external factors, so as to achieve accurate prediction of carbon emissions.

[0027] In some implementations, in step S1, constructing a production forecasting model, and performing production forecasting based on power consumption and the first external factor specifically includes the following steps: S11. Collect and analyze electricity consumption information, production information and meteorological information, and clarify electricity consumption and production data; among them, production data mainly refers to the number of products produced by an enterprise in a certain period (such as a month, a quarter or a year). In the accounting of carbon emissions above a certain scale in chemical parks, production data is an important reference indicator because it has a direct relationship with energy consumption and carbon emissions. Production data usually covers the number of various products produced by an enterprise. Production data not only reflects the production scale of an enterprise, but also reflects its production efficiency and market demand. However, it takes a certain amount of time to collect and organize production data. Enterprises usually need to wait until the end of the production cycle before they can count and confirm the production. In addition, the data organization and review process may also cause time delays. This makes the production data have the defect of lag, which has a certain impact on the prediction of carbon emissions. Therefore, the real-time nature of electricity is used to make up for the timeliness of production data, establish a production prediction model, use electricity consumption to predict production, and then predict carbon emissions through production data.

[0028] Specifically, by querying the internal records of each enterprise in the chemical park and the upstream and downstream supply chain information, the historical electricity consumption information and historical production information in the chemical park are collected. When collecting information, the accuracy, completeness, consistency and timeliness of the data must be guaranteed, so as to ensure the accuracy and reliability of the subsequent production forecast model. Next, the historical electricity consumption information and historical production information are processed, the information is sorted and cleaned, errors and outliers in the data are removed or corrected, and duplicate data are processed to clarify the electricity consumption and production data, providing a data basis for the production forecast model. In some embodiments, when the weather is bad, it will affect the electricity consumption, so meteorological information can also be collected and analyzed, so that the electricity consumption can be analyzed more accurately.

[0029] S12. Construct a VAR yield prediction model based on the VAR model and deep learning combined algorithm; Specifically, constructing a VAR production forecasting model includes: first constructing a basic VAR model to capture the dynamic relationship between electricity consumption, production data and the first external factor data. Select a suitable deep learning model (such as LSTM, GRU, etc.) to capture the nonlinear relationship and time series characteristics in the data. Combine the VAR model with the deep learning model to form a VAR-deep learning combined model. This is achieved by using the prediction results of the VAR model as the input features of the deep learning model, or by weighted averaging the prediction results of the two. The VAR-deep learning combined model improves the prediction accuracy of the VAR production forecasting model.

[0030] S13, dividing the electricity consumption, output data and the first external factor data into a training set and a test set to train the VAR output prediction model; specifically, dividing the electricity consumption, output data and the first external factor data into a training set and a test set in proportion, inputting the training set into the VAR output prediction model to perform iterative training on it, testing the accuracy of the output prediction model through the test set, and obtaining a trained VAR output prediction model.

[0031] S14. Evaluate the trained VAR yield prediction model; specifically, comprehensively select MAE, MSE, RMSE, and MAPE indicators for model evaluation. Among them, the mean absolute error (MAE) is used to measure the average absolute error between the predicted value and the true value. It is a non-negative value. The smaller the MAE, the better the model. The mean square error (MSE) also measures the gap between the predicted value and the true value. The root mean square error (RMSE) is a typical indicator of the regression model, which is used to indicate how large the error will be in the model prediction. For larger errors, the weight is higher, and the smaller the RMSE, the better. The mean absolute percentage error (MAPE) is similar to MAE. It is standardized on the basis of MAE. The smaller the MAPE, the better the model.

[0032] These indicators measure the prediction performance of the yield prediction model from different angles, so that the prediction accuracy and stability of the model can be comprehensively evaluated. Different evaluation indicators can also be applied to different scenarios and data characteristics. Further, by using these indicators in combination, we can better understand the source of the model's prediction error, so as to carry out targeted optimization. For example, MSE and RMSE are sensitive to outliers, while MAE is more robust to outliers. In addition, we can find the performance differences of the model on different data sets, and better understand the performance of the model under different conditions, so as to optimize the generalization ability of the model and guide model tuning. For example, MSE and RMSE can be used to evaluate the stability of the model, while MAE and MAPE can be used to evaluate the relative error of the model. By adjusting model parameters or selecting different features, the values ​​of MSE and RMSE can be reduced, while keeping MAE and MAPE within an acceptable range. In summary, the comprehensive use of MSE, RMSE, MAE, MAPE and other indicators for model evaluation can comprehensively and accurately evaluate the prediction performance of the model, improve the interpretability, robustness and generalization ability of the model, and thus guide model tuning and optimization.

[0033] S15. Input the first external factor into the improved STIRPAT factor decomposition model to obtain the first external factor data, and input the power consumption and the first external factor data into the evaluated VAR production forecasting model to forecast production. Production data has a lag, while power consumption has outstanding advantages in reflecting enterprise production, such as real-time accuracy, wide coverage, and flexible application. There is a strong correlation between power and production, so a production forecasting model is constructed to accurately forecast production.

[0034] In the specific implementation process, in step S2, a carbon emission prediction model is constructed, and carbon emission prediction is performed based on the output prediction result and the second external factor, specifically including the following steps: S21. Collect and analyze the monthly output, historical carbon emissions, production information and raw material information of enterprises, and clarify the output data and carbon emission data; specifically, collect the monthly output, historical carbon emissions, production information and raw material information of enterprises in the chemical park by querying the internal records and upstream and downstream supply chain information of each enterprise in the chemical park. Then, sort out and clean up the collected information to clarify the output data and carbon emission data, and provide a data basis for the carbon emission prediction model.

[0035] S22. Construct a carbon emission prediction model based on a neural network and tree combination algorithm; Specifically, building a carbon emission prediction model includes: 1) Determine the model architecture: First, according to the characteristics of the carbon emission prediction problem, select the appropriate type of neural network, such as multi-layer perceptron (MLP), long short-term memory network (LSTM) or convolutional neural network (CNN). Considering that carbon emissions are affected by time series data, LSTM is preferred. Then, select a tree algorithm, such as decision tree, random forest or gradient boosted decision tree (GBDT), to integrate with the neural network. The integration method can be model fusion (such as weighted average, voting, etc.) or feature fusion (using the output of the tree algorithm as the input feature of the neural network).

[0036] 2) Build the model: First, according to the determined neural network type, a neural network model is constructed, and parameters such as the number of network layers, number of nodes, activation function, loss function, etc. are set. Then, the selected tree algorithm model is constructed, and relevant parameters such as the depth of the tree, splitting conditions, etc. are set. Thirdly, the neural network and the tree algorithm are integrated to form a combined model.

[0037] S23. Divide the production data, carbon emission data and the second external factor data into a training set and a test set to train the carbon emission prediction model; specifically, divide the production data, carbon emission data and the second external factor data into a training set and a test set in proportion, input the training set into the carbon emission prediction model to perform iterative training, test the accuracy of the prediction of the carbon emission prediction model through the test set, and obtain a trained carbon emission prediction model.

[0038] S24. Evaluate the trained carbon emission prediction model; specifically, comprehensively select MAE, MSE, RMSE, and MAPE indicators for model evaluation, which can comprehensively and accurately evaluate the prediction performance of the carbon emission prediction model, improve the model's interpretability, robustness, and generalization ability, thereby guiding model tuning and optimization and improving the prediction accuracy of the carbon emission prediction model.

[0039] S25. Input the second external factor into the improved STIRPAT factor decomposition model to obtain the second external factor data, input the output forecast result and the second external factor data into the evaluated carbon emission forecast model, and forecast carbon emissions. There is a correlation between the output of chemical parks above a certain scale and carbon emissions, and the output based on electricity forecast has a stronger timeliness, providing a more valuable indicator for the forecast of carbon emissions. Therefore, a carbon emission forecast model is constructed to accurately forecast carbon emissions.

[0040] Among them, the improved STIRPAT factor decomposition model is used to analyze the influence of each influencing factor on carbon emissions, and the second external factor is input into the improved STIRPAT factor decomposition model to obtain the second external factor data, which includes the influence direction and influence degree of the second external factor on carbon emissions. By inputting the output forecast results and the second external factor data into the evaluated carbon emission forecast model to predict carbon emissions, the impact of external factors on carbon emissions can be effectively incorporated into the carbon emission forecast model, thereby achieving accurate prediction of carbon emissions.

[0041] Furthermore, the construction method of improving the STIRPAT factor decomposition model includes the following steps: S31. Build a carbon emission index system; the carbon emissions of chemical parks above a certain scale are affected by electricity consumption, energy structure, economic growth, industrial structure, production technology and other aspects, so it is necessary to build a carbon emission index system to support subsequent analysis. In order to quantify the impact of various aspects on carbon emissions, the carbon emission index system includes enterprise electricity consumption level indicators (such as monthly electricity consumption of enterprises), enterprise energy structure indicators (such as energy output rate, proportion of renewable energy use, clean energy use rate, etc.), regional economic growth level indicators (such as regional GDP, industrial added value, fixed asset investment, etc.), enterprise structure indicators (such as the proportion of leading industries, the proportion of high value-added industries, output value scale, comprehensive utilization rate of industrial solid waste, etc.) and enterprise production process level indicators (such as raw material utilization rate).

[0042] S32. Feature index screening is performed on the carbon emission index system, and the screening results are used as input indicators for improving the STIRPAT factor decomposition model. The characteristics of the carbon emission influencing factors corresponding to each indicator are screened out by performing feature screening on each indicator in the carbon emission index system, and the screened characteristics are used as input indicators for improving the STIRPAT factor decomposition model. These indicators are screened from the constructed carbon emission index system and are used to quantify the degree of influence of each influencing factor on carbon emissions.

[0043] S33. Construct an improved STIRPAT factor decomposition model and solve the coefficients based on ridge regression. Carbon emissions are closely related to external factors. Based on the screened features, an improved STIRPAT factor decomposition model is constructed, and the coefficients are solved based on ridge regression. Through the improved STIRPAT factor decomposition model, each influencing factor is decomposed, and a functional relationship between chemical park elements and carbon emission accounting is established, realizing the quantitative analysis of carbon emission influencing factors, including the direction and degree of influence of each influencing factor on carbon emissions.

[0044] In some embodiments, in step S32, screening the characteristic index of the carbon emission system index includes the following steps: S321. Sort out the monthly indicator data list; specifically, external influencing factors are closely related to carbon emissions. Based on the established carbon emission indicator system, collect external factor data through materials, sort out the monthly indicator data list, and clarify the historical carbon emission data. Among them, the monthly indicator data list mainly includes various monthly data related to carbon emissions, such as corporate electricity consumption, energy use, economic growth data, corporate structure data, and production process data, which are used to support subsequent analysis and model construction.

[0045] S322. Based on the monthly indicator data list, the filtering method, the wrapping method, and the embedding method are selected to screen the characteristic indicators of the carbon emission system indicators respectively, and the first characteristic combination, the second characteristic combination, and the third characteristic combination are obtained; specifically, the filtering method evaluates the importance of the characteristics through the statistical characteristics of the characteristics themselves, thereby screening out the first characteristic combination, and can efficiently and quickly eliminate obviously irrelevant features. The wrapping method evaluates the effectiveness of different feature subsets by training and testing specific machine learning algorithms, and selects the optimal feature subset by iteratively adding or removing features to obtain the second characteristic combination. The embedding method embeds the feature selection process into the model construction process, and uses some features of the model itself for automatic feature selection to obtain the third characteristic combination.

[0046] Among them, the feature combination includes indicators with obvious statistical correlation such as monthly electricity consumption of enterprises, regional GDP, proportion of leading industries, indicators such as clean energy utilization rate, industrial added value, proportion of high value-added industries, etc. that can significantly improve the accuracy of model prediction, as well as energy output rate, raw material utilization rate, etc. that can reflect the energy utilization efficiency and production process level of enterprises.

[0047] S323, based on the voting method, determine the final feature combination from the first feature combination, the second feature combination and the third feature combination, and use the final feature combination as the input indicator of the improved STIRPAT factor decomposition model. Specifically, the voting method adopts the principle of "the minority obeys the majority", and each feature combination includes multiple features. By checking the number of times each feature appears in the three feature combinations, the one with the most appearances will be selected into the final feature combination. In this way, the respective advantages of the filtering method, the wrapping method and the embedding method are fully utilized, and the effect of feature selection is improved while taking into account the computational efficiency and the accuracy of feature selection, thereby improving the accuracy of the results of the improved STIRPAT factor decomposition model.

[0048] Further, in step S33, constructing an improved STIRPAT factor decomposition model, and solving the coefficients based on ridge regression specifically includes the following steps: The basic form of the improved STIRPAT factor decomposition model is: lnI=lna+blnP+clnA+dlnT+lnu; the coefficients are solved by ridge regression in the form of logarithmic sum. Among them, I, P, A, T represent environmental impact, population, affluence and technological level; a is the model coefficient; b, c, d are the indexes to be estimated; u is the error term.

[0049] Since variables such as population size, economic growth and technological progress restrict each other, they may lead to multicollinearity problems. Therefore, ridge regression is used to solve the STIRPAT model. First, the traditional STIRPAT factor decomposition model may cause the regression coefficients to be unstable or difficult to interpret when processing multivariate linear regression data. However, adding the L2 norm regularization term in the loss function of ridge regression can effectively alleviate the multicollinearity problem and make the regression coefficients more stable and reliable. Secondly, the traditional STIRPAT factor decomposition model uses the least squares method to solve the coefficients, which may cause overfitting problems. The STIRPAT factor decomposition model based on ridge regression can more accurately capture the true relationship between the independent variables and the dependent variables, thereby improving the model's prediction accuracy for future environmental pressures. Finally, the accuracy and applicability of the traditional STIRPAT factor decomposition model are highly dependent on the integrity and accuracy of the data, while the regularization parameter (usually denoted as λ) in ridge regression can be flexibly adjusted according to the characteristics of the data, and can show good adaptability and stability in different application scenarios.

[0050] The decomposition of carbon emission influencing factors is achieved by improving the STIRPAT factor decomposition model. First, a carbon emission index system is built; second, the filtering method, wrapping method, and embedding method are used to screen features respectively, and the final feature selection is completed based on the voting method; finally, based on the relevant characteristic indicators of carbon emissions, the ridge regression calculation method is used to improve the traditional STIRPAT factor decomposition model, and the carbon emission influencing factors are decomposed based on the improved STIRPAT factor decomposition model to achieve quantitative analysis of carbon emission influencing factors.

[0051] Furthermore, in step S23, the historical carbon emissions are obtained by calculating the carbon emissions based on the inventory compilation method.

[0052] In the specific implementation process, the carbon emissions accounting based on the inventory compilation method includes the following steps: First, sort out the data needs, clarify the data reserve list and energy types. Specifically, through data collection and analysis of carbon accounting data needs, according to internal and external data access needs, and based on the data such as carbon emissions, fuel consumption, energy carbon emission factors, electricity, etc. in the monitoring center database, clarify the data reserve list and energy types. Among them, energy types mainly include two categories: fossil energy and renewable energy. Fossil energy includes coal, oil, and natural gas. Renewable energy includes wind energy, solar energy, biomass energy, and hydropower: using water resources such as water flow, water level or tide to generate electricity.

[0053] In some embodiments, it can be based on the enterprise's energy consumption records or heat purchase and sale settlement vouchers / grid settlement meter readings, the enterprise's original production records, statistical ledgers, statistical reports, sales reports, financial details, etc. For example, the amount of natural gas used in a certain year is counted through the production ledger, and the amount of natural gas purchased is checked through the financial details. For the same activity level (physical quantity), the enterprise may have multiple data source options to ensure that the data source must be consistent throughout the time series. The same data can be collected in multiple ways and the statistical data can be cross-checked. The purpose of this step is to collect and organize the basic data required for accounting, including carbon emissions, fuel consumption, energy carbon emission factors, electricity data, etc., and determine the type of energy used to ensure the accuracy and reliability of the accounting results.

[0054] Secondly, based on the data reserve list and energy types, combined with authoritative documents on greenhouse gas accounting, reporting guidelines, and energy consumption above a certain scale in chemical parks, the IPCC inventory compilation method is used to construct a carbon emission inventory. Specifically, based on the data reserve list and energy types, combined with authoritative documents such as "Greenhouse Gas Accounting Methods for 24 Industries", "China's Chemical Production Enterprise Greenhouse Gas Emissions Accounting Methods and Reporting Guidelines", and energy consumption above a certain scale in chemical parks, the IPCC inventory compilation method is used to construct a carbon emission inventory based on the carbon emission sources of fossil fuel combustion emissions, industrial production process emissions, net purchased electricity and heat consumption, and recycling; the constructed carbon emission inventory lists in detail the various energy uses and emissions, providing basic data for subsequent analysis and prediction.

[0055] Next, collect the default values ​​of carbon content of carbon emission sources to estimate carbon emission factors. Emission factors are coefficients corresponding to activity level data, which are used to quantify greenhouse gas emissions per unit activity level. It represents the representative emission rate of a certain activity level under given operating conditions. In the absence of conditions for actual measurement, the default values ​​of common fossil fuel low calorific value, carbon oxidation rate, carbon content, etc. listed in the relevant accounting guidelines can be used to estimate emission factors. Specifically, refer to the "Guidelines for Accounting Methods and Reporting of Greenhouse Gas Emissions of Chinese Chemical Production Enterprises" to collect the default values ​​of carbon content of carbon emission sources to estimate carbon emission factors and provide accurate parameters for subsequent calculations; carbon emission factors are key parameters for accounting carbon emissions, which reflect the carbon emissions generated by unit energy or material use. Among them, the calculation formula of carbon emission factors is: emission factor = carbon content per unit calorific value of fossil fuel × carbon oxidation rate of fossil fuel × (44 / 12), and the carbon content per unit calorific value of fossil fuel can be estimated based on the default value of carbon content.

[0056] Finally, carbon emissions are calculated based on the carbon emission inventory and carbon emission factors. Specifically, referring to the "Guidelines for the Calculation and Reporting of Greenhouse Gas Emissions of Chemical Production Enterprises in China", carbon accounting for chemical parks above scale is carried out based on carbon emission inventory and carbon emission factor data, that is, carbon emissions generated by fossil fuel combustion, industrial production process emissions, emissions generated by net purchased electricity and heat, and recycling, to provide a clear direction and basis for the subsequent accounting process. Among them, fossil fuel combustion emissions include CO2 emissions generated by the full combustion of fossil fuels with oxygen in various types of fixed or mobile combustion equipment (such as boilers, burners, furnaces, ovens, internal combustion engines, etc.), which should be included in the total carbon emissions of the enterprise. Emissions from industrial production processes include: CO2 emissions from fossil fuels and other hydrocarbons used as raw materials, including CO2 emissions from vented exhaust gas after flaring, which should be included in the company's total carbon emissions; CO2 emissions from carbonate use (such as limestone, dolomite, etc. used as raw materials, flux or desulfurization agent); N2O emissions from nitric acid or adipic acid production processes, whose CO2 emission equivalents should be included in the company's total carbon emissions. The net purchase of electricity and heat by the company will indirectly lead to CO2 emissions from power / heat plants, and this part of CO2 emissions should be included in the company's total carbon emissions. The amount of recycling is the amount of CO2 that the company recycles into the atmosphere within the legal entity boundary and supplies it to other units as a product. This part of CO2 should be deducted from the company's total carbon emissions.

[0057] In some embodiments, the total carbon emissions should be equal to the CO2 emissions from fossil fuel combustion plus the CO2 emissions from industrial production processes, plus the CO2 emissions caused by the company's net purchased electricity and heat consumption, and finally minus the CO2 emissions recovered and supplied by the company. The formula is as follows: EGHG=ECO2-combustion+EGHG-process+EGHG-net electricity+EGHG-net heat-RCO2-recovery Among them, the calculation method for CO2 emissions generated by fossil fuel combustion, CO2 emissions during production and use, and CO2 emissions generated by net purchased electricity and heat is the product of the consumption of each type of energy and the carbon emission factor.

[0058] The present invention also provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the above-mentioned method for calculating carbon emissions above the scale of a chemical park.

[0059] The present invention also provides a system for calculating carbon emissions above a certain scale in a chemical park, comprising: the computer-readable storage medium as described above.

[0060] What is disclosed above is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. A person skilled in the art can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for calculating carbon emissions in chemical parks above a certain scale, characterized in that: The following steps are involved: S1. constructing a production forecasting model, and performing production forecasting based on electricity consumption and a first external factor to obtain a production forecasting result, wherein the first external factor includes macroeconomic conditions, technological progress, and energy prices; S2. Construct a carbon emission prediction model, and predict carbon emissions based on the output prediction result and a second external factor, wherein the second external factor includes carbon emission policies, energy-saving and carbon-reduction technologies, and market demand.

2. The method for calculating carbon emissions of chemical parks above scale according to claim 1, characterized in that: In step S1, constructing a production forecasting model, and performing production forecasting based on power consumption and a first external factor specifically includes the following steps: S11. Collect and analyze electricity consumption information, production information and meteorological information, and clarify electricity consumption and production data; S12. Construct a VAR yield prediction model based on the VAR model and deep learning combined algorithm; S13, dividing the power consumption, output data and the first external factor data into a training set and a test set to train the VAR output prediction model; S14, evaluating the trained VAR yield prediction model; S15. Input the first external factor into the improved STIRPAT factor decomposition model to obtain the first external factor data, input the power consumption and the first external factor data into the evaluated VAR production forecasting model to perform production forecasting.

3. The method for calculating carbon emissions of chemical parks above scale according to claim 1, characterized in that: In step S2, a carbon emission prediction model is constructed, and carbon emission prediction is performed based on the output prediction result and the second external factor, specifically including the following steps: S21. Collect and analyze the company's monthly output, historical carbon emissions, production information and raw material information, and clarify the output data and carbon emissions data; S21. Construct a carbon emission prediction model based on a neural network and tree combination algorithm; S23, dividing the production data, carbon emission data and the second external factor data into a training set and a test set to train the carbon emission prediction model; S24. Evaluate the trained carbon emission prediction model; S25. Input the second external factor into the improved STIRPAT factor decomposition model to obtain the second external factor data, input the output forecast result and the second external factor data into the evaluated carbon emission forecast model to perform carbon emission forecasting.

4. The method for calculating carbon emissions of chemical parks above scale according to claim 2 or 3, characterized in that: The method for constructing the improved STIRPAT factor decomposition model comprises the following steps: S31. Build a carbon emission indicator system; S32. Screen the characteristic indicators of the carbon emission indicator system and use the screening results as input indicators for improving the STIRPAT factor decomposition model; S33. Construct an improved STIRPAT factor decomposition model and solve the coefficients based on ridge regression.

5. The method for calculating carbon emissions of chemical parks above scale according to claim 4, characterized in that: In step S32, the characteristic index screening of the carbon emission system index includes the following steps: S321. Sort out the monthly indicator data list; S322. Based on the monthly indicator data list, select the filtering method, the wrapping method, and the embedding method to screen the characteristic indicators of the carbon emission system indicators respectively to obtain the first characteristic combination, the second characteristic combination, and the third characteristic combination; S323. Determine a final feature combination from the first feature combination, the second feature combination, and the third feature combination based on a voting method, and use the final feature combination as an input indicator for improving the STIRPAT factor decomposition model.

6. The method for calculating carbon emissions of chemical parks above scale according to claim 4, characterized in that: In step S33, constructing an improved STIRPAT factor decomposition model and solving the coefficients based on ridge regression specifically includes the following steps: The basic form of the improved STIRPAT factor decomposition model is: lnI=lna+blnP+clnA+dlnT+lnu; the coefficients are solved using the logarithmic sum form and ridge regression.

7. The method for calculating carbon emissions of chemical parks above a certain scale according to claim 3, characterized in that: In step S23, the historical carbon emissions are calculated based on the carbon emissions inventory compilation method.

8. The method for calculating carbon emissions of chemical parks above a certain scale according to claim 7, characterized in that: Carbon emissions accounting based on the inventory compilation method specifically includes the following steps: Organize data requirements, clarify data reserve lists and energy types; Based on the data reserve list and energy types, combined with the authoritative documents on greenhouse gas accounting, reporting guidelines and energy consumption above the scale of chemical parks, the IPCC inventory compilation method is used to construct a carbon emission inventory; Collect default values ​​of carbon content from carbon emission sources to estimate carbon emission factors; Carbon emissions are calculated based on the carbon emissions inventory and carbon emission factors.

9. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the method for calculating carbon emissions above the scale of a chemical park as described in any one of claims 1 to 8.

10. A carbon emission accounting system for chemical parks above scale, characterized in that: include: The computer readable storage medium of claim 9.

Images