Industrial source carbon pollution emission list dynamic accounting method and system, storage medium and equipment

Abstract

The invention discloses an industrial source carbon pollution discharge list dynamic accounting method and system, a storage medium and equipment, and belongs to the field of pollution monitoring, and the method comprises the steps: S1, obtaining the multi-source data of an industrial enterprise in a target region; s2, preprocessing the multi-source data; s3, constructing a first random forest fusion model based on the enterprise electricity consumption by using the preprocessed data, and predicting and verifying the industry pollutant discharge amount based on the first random forest fusion model; using the first random forest fusion model to predict the pollutant discharge amount of the enterprise in the industry passing the verification, and executing the step S4 for the industry not passing the verification; s4, constructing a second random forest fusion model based on the electricity consumption of the enterprise production equipment, and predicting the pollutant discharge amount of the enterprise in the industry based on the second random forest fusion model; and S5, completing construction of a day / hour scale industrial source carbon pollution dynamic emission list. According to the invention, the actual emission level of industrial source carbon pollution can be dynamically and comprehensively reflected in real time.

Description

Technical Field

[0001] This invention relates to the field of pollution monitoring, and in particular to a method, system, storage medium, and equipment for dynamic accounting of carbon pollution emission inventories from industrial sources. Background Technology

[0003] Air pollutant source emission inventory is used to conduct PM2.5 analysis. 2.5 Research on the synergistic control of air pollutants and oxygen (O3) is fundamental and plays a crucial role in studies on the formation mechanisms of air pollution, the evaluation of pollution control measures, and the formulation of pollution prevention and control management policies. Greenhouse gas emission inventories serve as the "basic data foundation" for climate governance, low-carbon transition, and policy formulation. Carbon pollution emission inventories, which include both air pollutants and greenhouse gases, are of paramount importance for environmental and climate management.

[0004] To accurately implement emergency response to heavy air pollution during autumn and winter and to ensure and control ambient air quality during major events, support ambient air quality forecasting and prediction, and promptly reflect the impact of economic fluctuations, policy interventions, or emergencies (such as pandemics and energy crises) on carbon emissions, carbon emission inventories need to achieve a certain degree of dynamism. Typically, carbon emission inventories are calculated on an annual basis, allocating annual emissions to hourly or monthly emissions using empirical parameters such as changes in production load. However, this temporal resolution cannot promptly reflect changes in the activity levels of carbon emission sources and the amount of pollutants emitted, nor can it reflect the differences between weekdays, weekends, and holidays, or the significant changes caused by special periods such as pandemics and economic downturns. Therefore, it is difficult to meet the needs of environmental authorities for atmospheric environmental management, which primarily relies on off-site supervision of enterprises, under the current socio-economic situation.

[0005] Currently, the main sources of air pollutants in my country are industrial and motor vehicle emissions. The industrial emission source inventory is primarily constructed based on data from Continuous Emission Monitoring Systems (CEMS). CEMS equipment is mainly installed in power plants, steel mills, and cement plants, etc., where NO... x And enterprises that contribute significantly to SO2 emissions. Existing dynamic accounting methods for industrial emission sources are mainly based on CEMS data, and are primarily applied to the dynamic emission characterization of industries that consume large amounts of fossil fuels, such as thermal power and steel, with a focus on NO. x While accounting for conventional pollutants like SO2 is common, there is limited application of dynamic emission accounting for greenhouse gases such as CO2. Currently, it is not possible to comprehensively and dynamically grasp the current status of carbon pollution emissions from industrial enterprises.

[0006] To compensate for the insufficient coverage of CEMS, dynamic accounting methods for industrial carbon pollution emissions inventories have been gradually established in recent years, based on enterprise electricity consumption and real-time pollutant emission data (mainly CEMS data). The principle behind this method is the assumption that enterprise production activity levels are strongly correlated with electricity consumption. However, this method has serious drawbacks: The method does not consider whether the electricity consumption of enterprises can reflect the actual level of their production activities. For example, key carbon-polluting enterprises such as cement clinker and steel manufacturing use fossil fuels such as coal and natural gas as their main energy sources and electricity as an auxiliary energy source. Their carbon emissions are not strongly correlated with their electricity consumption. The dynamic accounting method established by using only enterprise electricity consumption and real-time pollutant emission data is difficult to accurately reflect the actual level of carbon emissions.

[0007] In summary, traditional static industrial carbon emission inventory accounting methods that rely solely on CMES data and directly use enterprise electricity consumption and CEMS data are insufficient to meet the current management and research needs for air pollution prevention and control and climate change response. Summary of the Invention

[0008] The purpose of this invention is to overcome the problems existing in the prior art and to provide a method, system, storage medium and equipment for dynamic accounting of industrial source carbon pollution emission inventory.

[0009] The objective of this invention is achieved through the following technical solution: Firstly, a method for dynamically calculating industrial carbon pollution emission inventories is provided, including the following steps: S1. Obtain multi-source data of industrial enterprises within the target area. The multi-source data includes energy composition data, enterprise electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data of industrial carbon pollution emissions. S2. Preprocess the multi-source data; S3. Construct a first stochastic forest fusion model based on enterprise electricity consumption using the preprocessed data, predict and verify the industry pollutant emissions based on the first stochastic forest fusion model; for industries that pass the verification, use the first stochastic forest fusion model to predict the pollutant emissions of enterprises within the industry, and for industries that fail the verification, proceed to step S4. S4. Construct a second stochastic forest fusion model based on the electricity consumption of enterprise production equipment, and predict the pollutant emissions of enterprises in the industry based on the second stochastic forest fusion model; S5. Complete the construction of a dynamic carbon emission inventory of industrial sources on a daily / hourly scale: For enterprises with continuous emission monitoring systems, calculate carbon emissions directly using data from the continuous emission monitoring systems; for enterprises without continuous emission monitoring systems, calculate carbon emissions using the first and second random forest fusion models, combined with annual data on industrial carbon emissions; integrate the carbon emissions of all enterprises to form a dynamic carbon emission inventory of industrial sources for the target area.

[0010] In some embodiments, the energy composition data includes industry energy composition data and enterprise energy composition data.

[0011] In some embodiments, the preprocessing of the multi-source data includes: Use industry energy composition data to perform overall constraints and verification on enterprise energy composition data; Outlier removal and spatial and temporal scale matching were performed on the enterprise's electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data on industrial source carbon pollution emissions.

[0012] In some embodiments, constructing the first random forest fusion model based on enterprise electricity consumption specifically includes: A fusion framework of linear and random forest models was established, with enterprise electricity consumption, enterprise energy composition, meteorological data, and enterprise activity data as feature variables and air pollutant emissions as the target variable; and model training and cross-validation were carried out by national economic sector.

[0013] In some embodiments, the prediction and verification of industry pollutant emissions based on the first random forest fusion model includes: The Pearson correlation coefficient between the emissions predicted by the model and the emissions measured by the continuous emission monitoring system was calculated. For industries with a Pearson correlation coefficient ≥ the first threshold, or a Pearson correlation coefficient between the second and first thresholds and a corporate electricity share in the energy mix ≥ the third threshold, the verification is deemed successful; for industries with a Pearson correlation coefficient < the first threshold and a corporate electricity share in the energy mix < the third threshold, the verification is deemed unsuccessful.

[0014] In some embodiments, constructing a second random forest fusion model based on the electricity consumption of enterprise production equipment includes: Based on the first random forest fusion model, adjustments were made. First, the electricity consumption of production equipment was replaced with the emission of air pollutants as the target variable, while other feature variables remained unchanged. The model was then trained and the electricity consumption of production equipment was predicted. Then, the predicted electricity consumption of production equipment is used as a new feature variable, and the target variable is restored to air pollutant emissions, and the model is trained again.

[0015] In some embodiments, in step S5, for enterprises that do not have a continuous emission monitoring system for flue gas, the daily / hourly emissions of unmonitored air pollutants are calculated using the following formula: in, For corporate air pollutants i Daily / hourly emissions The daily / hourly NOx emissions predicted by the first or second random forest fusion model. This represents the company's NOx emissions from the previous year. For the pollutants of the enterprise in the previous year i Emissions, i For air pollutants not included in the continuous emission monitoring system for flue gas; Its daily / hourly CO2 emissions are calculated using the following formula: in, This refers to the company's daily / hourly CO2 emissions. The daily / hourly electricity consumption of the enterprise's electricity meter. The annual electricity consumption of the enterprise's electricity meter. This represents the company's CO2 emissions in the previous year.

[0016] Secondly, a dynamic accounting system for industrial source carbon pollution emission inventories is provided, including: The data acquisition module is used to acquire multi-source data of industrial enterprises within the target area. The multi-source data includes energy composition data, enterprise electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data of industrial carbon pollution emissions. The data preprocessing module is used to preprocess the multi-source data; The first stochastic forest fusion model construction module uses preprocessed data to construct a first stochastic forest fusion model based on enterprise electricity consumption, predicts and verifies industry pollutant emissions based on the first stochastic forest fusion model, and uses the first stochastic forest fusion model to predict the pollutant emissions of enterprises within the industry for industries that have passed verification. The second random forest fusion model construction module is used to construct a second random forest fusion model based on the electricity consumption of enterprise production equipment for industries that have not passed the validation, and to predict the pollutant emissions of enterprises in the industry based on the second random forest fusion model. The emission inventory construction module is used to construct a dynamic emission inventory of industrial carbon pollution on a daily / hourly scale: for enterprises with continuous emission monitoring systems, carbon pollution emissions are calculated directly using data from these systems; for enterprises without continuous emission monitoring systems, carbon pollution emissions are calculated using a first random forest fusion model and a second random forest fusion model, combined with annual data on industrial carbon pollution emissions; and the carbon pollution emissions of all enterprises are integrated to form a dynamic emission inventory of industrial carbon pollution in the target area.

[0017] Thirdly, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the dynamic accounting method for industrial source carbon pollution emission inventory described in the first aspect.

[0018] Fourthly, an electronic device is provided, including a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, characterized in that the processor executes the dynamic accounting method for industrial source carbon pollution emission inventory described in the first aspect when executing the computer instructions.

[0019] It should be further noted that the technical features corresponding to the above-mentioned options and embodiments can be combined or substituted with each other to form new technical solutions without conflict.

[0020] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention utilizes multi-source enterprise-level data, including enterprise energy composition, electricity consumption, electricity consumption of production equipment, and CEMS data, combined with meteorological reanalysis data. It employs a random forest algorithm to establish a random forest fusion model (combining a random forest model with a linear model). This model predicts pollutant emissions based on enterprise electricity consumption across different sectors of the national economy. For sectors with poor prediction results, pollutant emissions are predicted again based on their production equipment electricity consumption. This achieves dynamic emission accounting covering all enterprises involved in carbon pollution emissions, establishing a dynamic accounting method for industrial source carbon pollution emissions, which has significant practical value. This invention can reflect the actual level of industrial source carbon pollution emissions in real time, dynamically, and comprehensively, meeting the current management and research needs for air pollution prevention and control and climate change response.

[0021] 2. This invention establishes a fusion framework based on a linear model and a random forest model using multi-source electricity consumption data from enterprises. The linear model can capture the overall, macro-level average relationship and trend of "electricity-emissions" across the industry; the random forest model excels at learning complex nonlinear relationships and capturing the fluctuations and characteristics of individual enterprises. This fusion effectively balances the robustness of overall prediction with the sensitivity of individual predictions, improving the model's generalization ability and prediction accuracy. Simultaneously, the introduction of meteorological data as a feature variable considers the impact of environmental factors on non-production electricity consumption and emission diffusion conditions, further enhancing the model's scientific rigor. The fusion framework balances overall industrial level prediction with capturing individual enterprise fluctuations, while also addressing the issue that electricity consumption in some energy-intensive industries (steel, cement, etc.) does not reflect the actual production activity levels of these enterprises, demonstrating high practicality.

[0022] 3. This invention creatively combines CEMS measured data with machine learning model prediction data. For key enterprises that have installed CEMS, high-precision measured data is used directly; for enterprises that account for the vast majority of industrial emissions and have not installed CEMS, the fusion model of this invention is used for prediction. This method breaks through the physical limitations of CEMS equipment coverage and, for the first time, achieves dynamic emission accounting for all industrial enterprises involved in carbon pollution emissions within a region, constructing a truly comprehensive dynamic inventory.

[0023] 4. This invention introduces enterprise energy composition as a key criterion and designs a dual modeling path for different industries. For industries with a high proportion of electricity consumption and a strong correlation between electricity consumption and production activities (such as some machinery manufacturing and electronics industries), the total electricity consumption of enterprises is directly used for modeling, which is highly efficient. For industries dominated by fossil fuels, such as steel and cement, and with a weak correlation to electricity, the modeling path is intelligently switched to one based on the electricity consumption of production equipment. The electricity consumption of production equipment is directly linked to the core process, which can more accurately reflect the actual production load, thereby significantly improving the accuracy of dynamic emission prediction for these key emission industries.

[0024] 5. This invention ultimately outputs a high spatiotemporal resolution carbon emissions inventory at the daily or even hourly level. This can reflect the impact of weekdays, weekends, holidays, seasonal production adjustments, and sudden economic and policy events on emissions in real time, completely changing the lag of traditional annual inventories. It provides unprecedented data support for environmental departments to carry out precise emergency response to heavy pollution weather, air quality assurance for major events, real-time perception of emissions trends, and precise "non-site" enforcement, effectively serving the current modern environmental governance model of "non-interference during non-event periods and non-site supervision."

[0025] 6. The method of this invention simultaneously calculates the dynamic characterization of emissions of air pollutants (SO2, NOx, etc.) and greenhouse gases (CO2), directly generating a dynamic inventory of integrated carbon and pollution. This provides a unified and dynamic data foundation for analyzing the correlation between pollution and emissions, assessing the synergistic effect of pollution reduction measures on carbon reduction, and formulating synergistic control strategies. It has significant practical value for promoting the realization of the national strategic goal of "synergistic efficiency improvement in pollution and carbon reduction". Attached Figure Description

[0026] Figure 1 This is a flowchart of a dynamic accounting method for industrial carbon pollution emission inventory according to the present invention; Figure 2 This is a schematic diagram of the operation process of the first random forest fusion model based on enterprise electricity consumption in this invention; Figure 3 This is a schematic diagram of the operation process of the second random forest fusion model based on the electricity consumption of enterprise production equipment in this invention. Detailed Implementation

[0027] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] It should be noted that the defects in the solutions in the prior art are all the results of the inventors' practice and careful research. Therefore, the discovery process of the above problems and the solutions proposed by the embodiments of this application in the following text should be the inventors' contributions to this application in the process of invention and creation, and should not be understood as technical content known to those skilled in the art.

[0029] In view of the technical problems pointed out in the background art, the present invention provides the following embodiments: In one exemplary embodiment, a method for dynamic accounting of industrial source carbon pollution emission inventories is provided, such as... Figure 1 As shown, it includes the following steps: S1. Obtain multi-source data of industrial enterprises within the target area. The multi-source data includes energy composition data, enterprise electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data of industrial carbon pollution emissions. S2. Preprocess the multi-source data; S3. Construct a first stochastic forest fusion model based on enterprise electricity consumption using the preprocessed data, predict and verify the industry pollutant emissions based on the first stochastic forest fusion model; for industries that pass the verification, use the first stochastic forest fusion model to predict the pollutant emissions of enterprises within the industry, and for industries that fail the verification, proceed to step S4. S4. Construct a second stochastic forest fusion model based on the electricity consumption of enterprise production equipment, and predict the pollutant emissions of enterprises in the industry based on the second stochastic forest fusion model; S5. Complete the construction of a dynamic carbon emission inventory of industrial sources on a daily / hourly scale: For enterprises with continuous emission monitoring systems, calculate carbon emissions directly using data from the continuous emission monitoring systems; for enterprises without continuous emission monitoring systems, calculate carbon emissions using the first and second random forest fusion models, combined with annual data on industrial carbon emissions; integrate the carbon emissions of all enterprises to form a dynamic carbon emission inventory of industrial sources for the target area.

[0030] Specifically, this method mainly consists of three parts: data preprocessing, construction of a random forest fusion model based on enterprise electricity consumption data, and establishment of a dynamic emission inventory of industrial carbon pollution. The first part aims to optimize the input data of the random forest fusion model by performing verification, outlier removal, and matching on the input data. The second part focuses on model construction, establishing an atmospheric pollutant prediction model based on enterprise-level data. The third part establishes a dynamic accounting method based on CEMS data, combined with the random forest fusion model method, ultimately establishing a comprehensive dynamic accounting method for industrial carbon pollution. This invention can reflect the actual emission level of industrial carbon pollution in real time, dynamically, and comprehensively, meeting the current management and research needs for air pollution prevention and control and climate change response.

[0031] In step S1, key data such as energy, air pollutant emissions, and meteorological environment required for the input of the random forest fusion model are obtained through various channels, including publicly available statistical and research data, environmental authorities, and power supply companies. The acquisition of multi-source data is described in detail below: 1. Obtain energy composition data. Energy composition includes two dimensions: first, industry energy composition, which serves as an overall constraint on enterprise energy composition; second, enterprise energy composition, which serves as a key input parameter for the random forest fusion model and is also a core basis for judging the model's emission prediction performance.

[0032] Industry energy composition data is obtained directly from publicly available national, provincial, and municipal statistical data. This data is categorized according to the major categories in the National Economic Industry Classification Standard, consistent with the industry classification method for industrial enterprises in this methodology. Enterprise energy composition data needs to be collected from the bottom up, including but not limited to enterprise name, latitude and longitude, industry type (categorized into major, medium, and minor categories according to the National Economic Industry Classification Standard), main products and output, raw material information, and energy information such as fuel type, fuel name, fuel unit, annual consumption, and lower heating value of fuel. Energy information is collected separately for boilers, kilns, production lines, and end-of-pipe treatment facilities, and the logical consistency of the energy consumption data for the entire plant and each stage is simultaneously verified. The time resolution is monthly, quarterly, or annual.

[0033] 2. Obtain enterprise electricity consumption data, which consists of enterprise electricity meter data and represents the total electricity consumption of facilities and equipment such as production, auxiliary facilities, and pollution control facilities. The data is collected from the power grid company and includes, but is not limited to, enterprise name, time, and anonymized enterprise electricity consumption. The time resolution is daily or hourly. This data serves as a feature variable in the random forest fusion model for predicting air pollutant emissions.

[0034] 3. Obtain electricity consumption data for enterprise production equipment. Data is collected from the electricity meters installed on the equipment side of the enterprise's main process facilities, and includes, but is not limited to, enterprise name, monitoring point name, equipment type, monitoring time, and monitoring value. The time resolution is 15 minutes. This data serves as the target variable for a random forest fusion model to predict air pollutant emissions.

[0035] 4. Obtain CEMS data. Data is collected from air pollution control authorities and includes, but is not limited to, company name, emission outlet name, air pollutant name, time, and air pollutant emission amount. The time resolution is daily or hourly. This data serves as the target variable in a random forest fusion model to predict air pollutant emissions.

[0036] 5. Obtain meteorological reanalysis data. Data was collected from publicly available information from the European Centre for Medium-Range Weather Forecasts (ECMWF-ERA5) and included, but was not limited to, temperature, atmospheric pressure, east-west wind speed, north-south wind speed, dew point temperature, and planetary boundary layer height. The temporal resolution was daily or hourly. This data served as a feature variable in the random forest fusion model for predicting atmospheric pollutant emissions.

[0037] 6. Obtain annual data on industrial carbon emissions (annual emissions of industrial air pollutants or greenhouse gases). Industrial activity level data are collected from industrial enterprises, and the annual data on industrial carbon emissions are calculated using the methods outlined in the "Technical Guidelines for Compiling Integrated Emission Inventories of Air Pollutants and Greenhouse Gases (Trial)" issued by the Ministry of Ecology and Environment. This data serves as the baseline for carbon emissions in the previous year and is used to construct a dynamic carbon emission inventory.

[0038] In step S2, on the one hand, constraint analysis and preprocessing are performed on the energy composition data. Specifically, based on publicly available statistical data on industrial energy consumption, overall constraints and verification are performed on the enterprise-level energy consumption data obtained by the bottom-up approach. The verification formula is as shown in equation (1).

[0039] (1) This is the difference (in tons of standard coal) between the energy consumption data in the statistical data and the total energy consumption data of enterprises. Energy consumption (tons of standard coal) in the statistical data. This represents enterprise-level energy consumption data (tons of standard coal); i represents the energy type; j represents the national economic sector; and k represents a specific enterprise within that national economic sector.

[0040] The enterprise-level energy composition is the proportion of the enterprise's electricity consumption in all energy types. It is calculated based on all energy types and consumption of the enterprise, and the calculation formula is as shown in equation (2).

[0041] (2) PEC represents the composition of a company's electricity consumption within its total energy consumption; it is dimensionless. ECP i ECi,j represents the enterprise's electricity consumption (tons of standard coal); ECi,j represents the enterprise's consumption of j types of energy (tons of standard coal) at stage i; here i represents the energy use stage, including production and manufacturing, auxiliary facilities, pollution control facilities, etc.; here j represents the energy type, including electricity, coal, natural gas, fuel oil, diesel, etc.

[0042] On the other hand, preprocessing is performed on enterprise electricity consumption, enterprise production equipment electricity consumption, data from continuous emission monitoring systems for flue gas, meteorological reanalysis data, and annual data on industrial carbon emissions. Specifically, this includes: 1. Based on company name and time, match company electricity consumption, electricity consumption of company production equipment, and company CEMS. In addition, remove missing items and outlier data with negative values ​​from the above data, and use the interquartile range (IQR) method to filter and remove data with large deviations.

[0043] 2. The meteorological reanalysis data consisted of hourly near-surface meteorological data, including six variables such as temperature. The spatial resolution of the meteorological data was 0.25°, and bilinear interpolation was used to spatially interpolate the meteorological data to a 1 km × 1 km grid. Based on the latitude and longitude of each enterprise, the meteorological data was spatially matched to obtain the meteorological information for each enterprise.

[0044] In step S3, a fusion framework of a linear model and a random forest model based on enterprise electricity consumption is established to construct a machine learning model for "electricity-air pollutant emissions," which takes into account both overall industrial level prediction and individual enterprise fluctuation capture. The operation process is as follows: Figure 2 As shown, the verification and estimation process for a single industry category of pollutants is presented. Meteorological data is a crucial variable in the prediction of electricity-air pollutant emissions. Since non-production electricity consumption by enterprises is highly correlated with ambient temperature, incorporating meteorological data can significantly improve prediction accuracy. Therefore, the characteristic variables of the first random forest fusion model established in this method include meteorological data such as lateral wind speed, longitudinal wind speed, dew point temperature, temperature, boundary layer height, and atmospheric pressure, as well as key data on enterprise activities such as electricity consumption, energy composition, types of air pollutants, national economic industry categories, annual product output, annual product capacity, types of raw materials, and raw material consumption. The target variable of the model is air pollutant emissions.

[0045] During model training, for each national economic sector, one enterprise is selected as the validation set, and the other enterprises in the same sector are selected as the training set for cross-validation. A linear regression model is established using the average and standard deviation of the electricity consumption and pollutant emissions of the enterprises in the training set. Based on the linear regression model, the pollutant emissions and standard deviation of the validation enterprises are predicted. The pollutant emissions and electricity consumption input into the random forest model are normalized using the formula (3). The normalized pollutant emissions are predicted using the random forest model, and this result is combined with the linear regression model for inverse normalization calculation using the formula (4) to obtain the predicted pollutant emissions of the enterprises.

[0046] (3) ZSD represents the normalized values ​​of enterprise electricity consumption and pollutant emissions, which are dimensionless; x represents enterprise electricity consumption and pollutant emissions. is the average of the enterprise's electricity consumption and pollutant emissions; SD is the standard deviation of the enterprise's electricity consumption and pollutant emissions.

[0047] (4) EMI represents the predicted pollutant emissions (tons) of the enterprise; ZSDpol represents the normalized predicted pollutant emissions calculated by the random forest model, which is dimensionless; SDpol represents the standard deviation of the predicted pollutant emissions calculated by the linear model, which is dimensionless. The average value (in tons) of the predicted pollutant emissions calculated for the linear model.

[0048] The model's predicted emissions results are summarized and analyzed by industry: For industries where the Pearson correlation coefficient between predicted emissions and CEMS measured emissions is ≥0.7, or where the correlation coefficient is ≥0.6 but electricity accounts for a high proportion of the enterprise's energy mix (≥ the third threshold), the validation is considered successful, and the random forest fusion model is used to predict the enterprise's pollutant emissions; For industries where the Pearson correlation coefficient between predicted emissions and CEMS measured pollutant emissions is <0.7 and electricity accounts for a low proportion of the energy mix (< the third threshold), the validation is considered unsuccessful, and the next step is taken to continue predicting pollutant emissions.

[0049] In step S4, for industries that failed the verification in the previous step, a second random forest fusion model based on the electricity consumption of enterprise production equipment is constructed to further predict emissions. Specifically, adjustments are made to the first random forest fusion model to establish a linear model + random forest model based on the electricity consumption of enterprise production equipment. The operation process is as follows: Figure 3 As shown, the verification and estimation process for a single pollutant in an industry is presented. For industries where electricity accounts for a small proportion of the enterprise's energy mix, enterprise electricity consumption is no longer suitable as a key feature variable for the model.

[0050] In the first part of this step, we replace air pollutant emissions with electricity consumption from enterprise production equipment, which is more closely related to the enterprise's production process, as the target variable of the model, while keeping other feature variables unchanged. The random forest fusion model remains consistent with the previous step, using the random forest model to predict normalized electricity consumption from enterprise production equipment, then combining it with a linear regression model for inverse normalization calculation, and finally predicting electricity consumption from production equipment for enterprises without such consumption. In the second part of this step, we change the target variable of the model to air pollutant emissions, using the electricity consumption from production equipment predicted in the first part of this step as the feature variable, while keeping other feature variables unchanged, and continue using the model from the previous step (…). Figure 2 The method was used to determine the air pollutant emissions of companies in industries with a low proportion of electricity consumption, based on a Pearson correlation coefficient of <0.7.

[0051] In step S5, considering that enterprises involved in carbon pollution emissions are divided into two categories—those with CEMS data and those without—establishing a dynamic carbon pollution emission inventory covering all industrial sources consists of the following two parts: 1. Companies with CEMS data can directly use CEMS data to calculate carbon emissions. Using CEMS data, daily / hourly time profiles of flue gas emission variations are established for key enterprises with CEMS data. For air pollutants (such as CO and PM2.5) not included in CEMS data but required for industrial source carbon emission inventories, [further details are needed]. 2.5 SO2 and CO2 are allocated to daily / hourly emissions based on annual emissions, as shown in equation (5). For CEMS data including SO2 and NO... x And VOCs, etc., calculate the daily / hourly emissions based on CEMS data, as shown in equation (6).

[0052] (5) Hourly emissions of air pollutants (tons); Annual emissions of air pollutants (tons); Q i The hourly average value (m) of the flue gas volume in the i-th hour 3 / h).

[0053] (6) Here Q represents hourly emissions (tons) of air pollutants or CO2. i The hourly average value (m) of the flue gas volume in the i-th hour 3 / h); C i The average concentration (mg / m³) for the i-th hour 3 ).

[0054] 2. For companies without CEMS data, use the established first / second stochastic forest fusion model to predict carbon emissions. Using step 3 (running process) Figure 2 ) and step 4 (running process) Figure 3 The model in this paper predicts the air pollutant emissions of enterprises that do not have CEMS data, that is, it obtains the SO2 and NO emissions contained in the CEMS data. x The daily / hourly emissions of air pollutants such as VOCs. For air pollutants not included in the CEMS data, the daily / hourly emissions are allocated based on the enterprise's air pollutant emission composition in the previous year, as shown in Equation (7); for greenhouse gases not included in the CEMS, the daily / hourly emissions are allocated based on the enterprise's CO2 emissions in the previous year, as shown in Equation (8).

[0055] (7) in, For corporate air pollutants i Daily / hourly emissions (tons). The daily / hourly NOx emissions (tons) predicted by the first or second random forest fusion model. This represents the company's NOx emissions (in tons) for the previous year. For the pollutants of the enterprise in the previous year i Emissions (tons). i For air pollutants not included in the continuous emission monitoring system, including CO and PM2.5. 2.5 wait; Its daily / hourly CO2 emissions are calculated using the following formula: (8) in, This represents the company's daily / hourly CO2 emissions (tons). This refers to the daily / hourly electricity consumption of the enterprise's electricity meter, dimensionless. This refers to the company's annual electricity consumption, dimensionless. This represents the company's CO2 emissions (in tons) for the previous year.

[0056] In summary, by using formulas (5) to (8), based on enterprise electricity consumption and CEMS and other relevant data, we can obtain the daily / hourly emission data of industrial carbon pollution and complete the dynamic emission inventory of industrial carbon pollution at the daily / hour scale.

[0057] In another exemplary embodiment, based on the same inventive concept as the method embodiment, a dynamic accounting system for industrial source carbon pollution emissions is provided, comprising: The data acquisition module is used to acquire multi-source data of industrial enterprises within the target area. The multi-source data includes energy composition data, enterprise electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data of industrial carbon pollution emissions. The data preprocessing module is used to preprocess the multi-source data; The first stochastic forest fusion model construction module uses preprocessed data to construct a first stochastic forest fusion model based on enterprise electricity consumption, predicts and verifies industry pollutant emissions based on the first stochastic forest fusion model, and uses the first stochastic forest fusion model to predict the pollutant emissions of enterprises within the industry for industries that have passed verification. The second random forest fusion model construction module is used to construct a second random forest fusion model based on the electricity consumption of enterprise production equipment for industries that have not passed the validation, and to predict the pollutant emissions of enterprises in the industry based on the second random forest fusion model. The emission inventory construction module is used to construct a dynamic emission inventory of industrial carbon pollution on a daily / hourly scale: for enterprises with continuous emission monitoring systems, carbon pollution emissions are calculated directly using data from these systems; for enterprises without continuous emission monitoring systems, carbon pollution emissions are calculated using a first random forest fusion model and a second random forest fusion model, combined with annual data on industrial carbon pollution emissions; and the carbon pollution emissions of all enterprises are integrated to form a dynamic emission inventory of industrial carbon pollution in the target area.

[0058] In another exemplary embodiment, based on the same inventive concept as the method embodiment, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which, when executed by a processor, implements the dynamic accounting method for industrial source carbon pollution emission inventories provided in this embodiment of the invention. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0059] In another exemplary embodiment, based on the same inventive concept as the method embodiment, an electronic device is provided, including a memory and a processor. The memory stores computer instructions that can be executed on the processor. When the processor executes the computer instructions, it performs the dynamic accounting method for industrial source carbon pollution emission inventory provided in the embodiment of the present invention.

[0060] The processor may be a single-core or multi-core central processing unit or a specific integrated circuit, or one or more integrated circuits configured to implement the present invention.

[0061] The embodiments of the subject matter and functional operation described in this specification can be implemented in: tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or combinations thereof. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible, non-transitory program carrier for execution by a data processing device or for controlling the operation of a data processing device. Alternatively or additionally, the program instructions may be encoded on artificially generated propagation signals, such as machine-generated electrical, optical, or electromagnetic signals, which are generated to encode information and transmit it to a suitable receiving device for execution by the data processing device.

[0062] The processing and logic flow described in this specification can be executed by one or more programmable computers that execute one or more computer programs to perform corresponding functions by operating on input data and generating output. The processing and logic flow can also be executed by dedicated logic circuitry—such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits), and the device can also be implemented as dedicated logic circuitry.

[0063] Suitable processors for executing computer programs include, for example, general-purpose and / or special-purpose microprocessors, or any other type of central processing unit. Typically, the central processing unit receives instructions and data from read-only memory and / or random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data, such as disks, magneto-optical disks, or optical disks, or the computer will be operatively coupled to such mass storage devices to receive data from or transfer data to them, or both. However, a computer is not required to have such devices. Furthermore, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name a few.

[0064] It should be understood that each block in a flowchart or block diagram can represent a module, segment, or portion of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the figures. For example, two consecutive blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0065] The above detailed embodiments are a description of the present invention. It should not be considered that the specific embodiments of the present invention are limited to these descriptions. For those skilled in the art, several simple deductions and substitutions can be made without departing from the concept of the present invention, and all of these should be considered to fall within the protection scope of the present invention.

Claims

1. A method for dynamic accounting of industrial source carbon pollution emission inventories, characterized in that, Includes the following steps: S1. Obtain multi-source data of industrial enterprises within the target area. The multi-source data includes energy composition data, enterprise electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data of industrial carbon pollution emissions. S2. Preprocess the multi-source data; S3. Construct a first stochastic forest fusion model based on enterprise electricity consumption using the preprocessed data, predict and verify the industry pollutant emissions based on the first stochastic forest fusion model; for industries that pass the verification, use the first stochastic forest fusion model to predict the pollutant emissions of enterprises within the industry, and for industries that fail the verification, proceed to step S4. S4. Construct a second stochastic forest fusion model based on the electricity consumption of enterprise production equipment, and predict the pollutant emissions of enterprises in the industry based on the second stochastic forest fusion model; S5. Complete the construction of a dynamic carbon emission inventory of industrial sources on a daily / hourly scale: For enterprises with continuous emission monitoring systems, directly use the data from the continuous emission monitoring system to calculate carbon emissions; For enterprises without a continuous emission monitoring system, the carbon emissions are calculated using the first and second random forest fusion models, combined with annual data on industrial carbon emissions. The carbon emissions of all enterprises are then integrated to form a dynamic carbon emission inventory of industrial sources in the target area.

2. The method for dynamic accounting of industrial source carbon pollution emission inventory according to claim 1, characterized in that, The energy composition data includes industry energy composition data and enterprise energy composition data.

3. The method for dynamic accounting of industrial source carbon pollution emission inventory according to claim 2, characterized in that, The preprocessing of the multi-source data includes: Use industry energy composition data to perform overall constraints and verification on enterprise energy composition data; Outlier removal and spatial and temporal scale matching were performed on the enterprise's electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data on industrial source carbon pollution emissions.

4. The method for dynamic accounting of industrial source carbon pollution emission inventory according to claim 1, characterized in that, The construction of the first random forest fusion model based on enterprise electricity consumption specifically includes: A fusion framework of linear and random forest models was established, with enterprise electricity consumption, enterprise energy composition, meteorological data, and enterprise activity data as feature variables and air pollutant emissions as the target variable; and model training and cross-validation were carried out by national economic sector.

5. The method for dynamic accounting of industrial source carbon pollution emission inventory according to claim 1, characterized in that, The prediction and verification of industry pollutant emissions based on the first random forest fusion model includes: The Pearson correlation coefficient between the emissions predicted by the model and the emissions measured by the continuous emission monitoring system was calculated. For industries with a Pearson correlation coefficient ≥ the first threshold, or a Pearson correlation coefficient between the second and first thresholds and a corporate electricity share in the energy mix ≥ the third threshold, the verification is deemed successful; for industries with a Pearson correlation coefficient < the first threshold and a corporate electricity share in the energy mix < the third threshold, the verification is deemed unsuccessful.

6. The method for dynamic accounting of industrial source carbon pollution emission inventory according to claim 1, characterized in that, The construction of the second random forest fusion model based on the electricity consumption of enterprise production equipment includes: Based on the first random forest fusion model, adjustments were made. First, the electricity consumption of production equipment was replaced with the emission of air pollutants as the target variable, while other feature variables remained unchanged. The model was then trained and the electricity consumption of production equipment was predicted. Then, the predicted electricity consumption of production equipment is used as a new feature variable, and the target variable is restored to air pollutant emissions, and the model is trained again.

7. The method for dynamic accounting of industrial source carbon pollution emission inventories according to claim 1, characterized in that, In step S5, for enterprises without a continuous emission monitoring system, the daily / hourly emissions of unmonitored air pollutants are calculated using the following formula: in, For corporate air pollutants i Daily / hourly emissions The daily / hourly NOx emissions predicted by the first or second random forest fusion model. This represents the company's NOx emissions from the previous year. For the pollutants of the enterprise in the previous year i Emissions, i For air pollutants not included in the continuous emission monitoring system for flue gas; Its daily / hourly CO2 emissions are calculated using the following formula: in, This refers to the company's daily / hourly CO2 emissions. The daily / hourly electricity consumption of the enterprise's electricity meter. The annual electricity consumption of the enterprise's electricity meter. This represents the company's CO2 emissions in the previous year.

8. A dynamic accounting system for industrial source carbon pollution emission inventories, characterized in that, include: The data acquisition module is used to acquire multi-source data of industrial enterprises within the target area. The multi-source data includes energy composition data, enterprise electricity consumption, enterprise production equipment electricity consumption, flue gas emission continuous monitoring system data, meteorological reanalysis data, and annual data of industrial carbon pollution emissions. The data preprocessing module is used to preprocess the multi-source data; The first stochastic forest fusion model construction module uses preprocessed data to construct a first stochastic forest fusion model based on enterprise electricity consumption, and predicts and verifies industry pollutant emissions based on the first stochastic forest fusion model. For industries that have passed the validation, the first random forest fusion model is used to predict the pollutant emissions of enterprises within that industry; The second random forest fusion model construction module is used to construct a second random forest fusion model based on the electricity consumption of enterprise production equipment for industries that have not passed the validation, and to predict the pollutant emissions of enterprises in the industry based on the second random forest fusion model. The emission inventory construction module is used to complete the construction of a dynamic emission inventory of industrial carbon pollution on a daily / hourly scale: For enterprises with a continuous emission monitoring system, carbon pollution emissions are calculated directly using data from the continuous emission monitoring system. For enterprises without a continuous emission monitoring system, the carbon emissions are calculated using the first and second random forest fusion models, combined with annual data on industrial carbon emissions. The carbon emissions of all enterprises are then integrated to form a dynamic carbon emission inventory of industrial sources in the target area.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the dynamic accounting method for industrial source carbon pollution emission inventory as described in any one of claims 1-7.

10. An electronic device comprising a memory and a processor, wherein the memory stores computer instructions executable by the processor, characterized in that, When the processor executes computer instructions, it performs the dynamic accounting method for industrial source carbon pollution emission inventory as described in any one of claims 1-7.

Images