A carbon dioxide emission accounting method, device and system

Abstract

The application relates to a carbon dioxide emission accounting method, device and system. The method comprises the following steps: identifying at least one to-be-accounted emission node in a to-be-accounted petrochemical plant; for each to-be-accounted emission node, inputting the emission technical information of the to-be-accounted emission node into a pre-established confidence accounting calculation model, determining a reference emission node matched with the emission technical information of the to-be-accounted emission node through the confidence accounting calculation model, and calculating the target emission confidence interval of the to-be-accounted emission node according to the emission confidence interval of the reference emission node; wherein the confidence accounting calculation model is established based on the emission technical information and the emission confidence interval of each emission node of a petrochemical plant matched with the to-be-accounted petrochemical plant, and the reference emission node belongs to an emission node in the petrochemical plant; and determining the carbon dioxide emission accounting result of the to-be-accounted emission node according to the target emission confidence interval. The application can improve the accuracy and reliability of the carbon dioxide emission accounting result.

Description

Technical Field

[0001] This application relates to the field of accounting technology, and in particular to a method, apparatus, and system for calculating carbon dioxide emissions. Background Technology

[0002] The technology for calculating carbon dioxide emissions from petrochemical plants is a crucial foundation for each plant to fulfill its emission reduction responsibilities. The core of carbon dioxide emission accounting lies in obtaining accurate emission data at different production stages and equipment points within a petrochemical plant, and then making reasonable estimates and predictions.

[0003] In related technologies, carbon dioxide emission accounting techniques are mostly based on manual sampling and measurement, single-point monitoring and estimation, or calculation using fixed emission factors. Although these methods can estimate the total emissions, they are poorly adaptable to actual multi-node emission and multi-emission technology conditions, resulting in low accuracy and reliability of carbon dioxide emission accounting results. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, or system for calculating carbon dioxide emissions to address the aforementioned technical problems.

[0005] Firstly, this application provides a carbon dioxide emission accounting method, including:

[0006] Identify at least one emission point in the petrochemical plant to be accounted for;

[0007] For each emission node to be calculated, the emission technology information of the emission node is input into a pre-established confidence calculation model. The confidence calculation model determines a reference emission node that matches the emission technology information of the emission node to be calculated. The target emission placement confidence interval of the emission node to be calculated is calculated based on the emission placement confidence interval of the reference emission node. The confidence calculation model is established based on the emission technology information and emission placement confidence interval of each emission node in a petrochemical-like plant that matches the petrochemical plant to be calculated. The reference emission node is an emission node in the petrochemical-like plant.

[0008] Based on the target emission confidence interval, determine the carbon dioxide emission accounting results for the emission nodes to be accounted for.

[0009] In one embodiment, a reference emission node matching the emission technology information of the emission node to be accounted for is determined through a confidence calculation model. The target emission confidence interval for the emission node to be accounted for is then calculated based on the emission confidence interval of the reference emission node, including:

[0010] Based on the emission technology information of the emission node to be accounted for, a distance measurement is performed between the emission technology information of the reference emission node to obtain the characteristic distance between the emission technology information of the emission node to be accounted for and the emission technology information of at least one reference emission node.

[0011] Convert the distances of each feature into similarity weights;

[0012] Based on the similarity weights and the corresponding arrangement confidence intervals, the target arrangement confidence intervals for the emission nodes to be verified are calculated.

[0013] In one embodiment, the carbon dioxide emission calculation result of the emission node to be calculated is determined based on the target emission confidence interval, including:

[0014] Correlation analysis was performed on the emission change data of all emission nodes to be accounted for, and the correlation matrix between emission nodes was obtained.

[0015] The emission placement information intervals of all emission nodes to be accounted for are sampled and calculated to obtain the total emission placement information interval of the target petrochemical plant;

[0016] Based on the total emission confidence interval and the correlation matrix between emission nodes, carbon dioxide emission accounting is performed on the emission change data of each emission node to be accounted for, and the carbon dioxide emission accounting results are obtained.

[0017] In one embodiment, the construction of the confidence accounting calculation model includes:

[0018] Obtain emission change data and emission technology information for each reference emission node of the petrochemical-like plant that matches the petrochemical plant to be accounted for;

[0019] Based on the emission change data and emission technology information of each reference emission node of a petrochemical plant, the emission confidence interval of each reference emission node is determined.

[0020] Based on the emission technology information, emission change data and corresponding emission confidence intervals of all reference emission nodes of petrochemical plants, a confidence accounting calculation model is constructed.

[0021] In one embodiment, a confidence calculation model is established based on the emission technology information, emission change data, and corresponding emission confidence intervals of all reference emission nodes of a petrochemical plant, including:

[0022] Based on the emission change data of each reference emission node, construct the first prior probability distribution corresponding to each reference emission node;

[0023] Bayesian inference is performed on the first prior probability distribution and emission technology information of each reference emission node to obtain the first posterior probability distribution corresponding to each emission node.

[0024] The emission confidence interval corresponding to each reference emission node is determined based on the first posterior probability distribution of each reference emission node.

[0025] Each reference emission node, along with the corresponding emission technology information and emission confidence interval, is input into the initial accounting calculation model for fitting and training to obtain the confidence accounting calculation model.

[0026] In one embodiment, a confidence calculation model is established based on the emission technology information, emission change data, and corresponding emission confidence intervals of all reference emission nodes of a petrochemical plant, including:

[0027] For each emission node cluster, the mean of the emission change data of all reference emission nodes included in the emission node cluster is calculated to obtain the second prior probability distribution corresponding to the emission node cluster; wherein, the emission node cluster is obtained by clustering each reference emission node based on the emission change data and emission technology information of each reference emission node.

[0028] Based on the second prior probability distribution, emission node clusters, and technology information clusters, the posterior probability distribution is calculated to obtain the second posterior probability distribution corresponding to each emission node cluster; where the technology information cluster is the feature extraction result of the emission technology information of the emission node cluster.

[0029] The confidence interval for each emission node cluster is determined based on the second posterior probability distribution corresponding to each emission node cluster.

[0030] Each emission node cluster, technical information cluster, and emission confidence interval are input into a preset initial accounting calculation model for fitting and training to construct a confidence accounting calculation model.

[0031] In one embodiment, the method further includes:

[0032] Obtain the migration variance corresponding to each reference emission node. The migration variance characterizes the degree of fluctuation of emission change data under different operating conditions.

[0033] The migration variance was used as a hyperparameter for cross-validation to obtain the migration error distribution corresponding to each reference emission node.

[0034] The updated second posterior probability distribution is obtained by calculating the Bayesian posterior probability distribution based on the migration error distribution corresponding to each node, the second prior probability distribution corresponding to each emission node cluster, and the technical information cluster corresponding to each emission node cluster.

[0035] In one embodiment, determining the emission node cluster confidence interval corresponding to each emission node cluster based on the second posterior probability distribution of each emission node cluster includes:

[0036] The corresponding rank confidence interval is determined from the second posterior probability distribution according to the preset confidence level, wherein the representative value of the rank confidence interval is the statistical value of the posterior probability distribution.

[0037] Secondly, this application also provides a carbon dioxide emission accounting device, comprising:

[0038] The acquisition module is used to identify at least one emission node in the petrochemical plant to be accounted for;

[0039] The calculation module is used to input the emission technology information of each emission node to be calculated into a pre-established confidence calculation model. The confidence calculation model determines a reference emission node that matches the emission technology information of the emission node to be calculated, and calculates the target emission placement confidence interval of the emission node to be calculated based on the emission placement confidence interval of the reference emission node. The confidence calculation model is established based on the emission technology information and emission placement confidence interval of each emission node in a petrochemical-like plant that matches the petrochemical plant to be calculated. The reference emission node is an emission node in the petrochemical-like plant.

[0040] The generation module is used to determine the carbon dioxide emission calculation results of the emission nodes to be calculated based on the target emission confidence interval.

[0041] Thirdly, this application also provides a carbon dioxide emission accounting system, including a data collection unit and a calculation unit;

[0042] The data acquisition unit is used to collect emission change data and emission technology information of each emission node to be calculated in the target petrochemical plant, as well as the overall production data of petrochemical plants.

[0043] The calculation unit is used to execute any of the methods described above based on the emission change data and emission technology information of each emission node to be calculated in the target petrochemical plant, as well as the overall production data of the petrochemical plant.

[0044] The aforementioned carbon dioxide emission accounting method, apparatus, and system, for each emission node in a petrochemical plant to be accounted for, inputs the emission technology information of the emission node to be accounted for into a pre-established confidence accounting calculation model. The confidence accounting calculation model determines a reference emission node that matches the emission technology information of the emission node to be accounted for. Based on the emission confidence interval of the reference emission node, a target emission confidence interval for the emission node to be accounted for is calculated. The confidence accounting calculation model is established based on the emission technology information and emission confidence intervals of each emission node in a petrochemical-like plant that matches the petrochemical plant to be accounted for, and the reference emission node belongs to the emission nodes in the petrochemical-like plant. Based on the target emission confidence interval, the carbon dioxide emission accounting result for the emission node to be accounted for is determined. This application embodiment achieves the construction of a confidence accounting calculation model that integrates emission change data, technology information, and emission confidence intervals from multiple reference nodes. This allows for the output of high-precision, quantifiable confidence emission confidence intervals under different production conditions, facilitating high-precision and reliable carbon dioxide emission accounting. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a diagram illustrating the application environment of a carbon dioxide emission accounting method in one embodiment.

[0047] Figure 2 This is a flowchart illustrating a carbon dioxide emission accounting method in one embodiment;

[0048] Figure 3 This is a flowchart illustrating the process of calculating the target placement interval in one embodiment;

[0049] Figure 4 This is a flowchart illustrating the calculation of carbon dioxide emission accounting results in one embodiment;

[0050] Figure 5 This is a flowchart illustrating the process of constructing a confidence accounting calculation model in one embodiment;

[0051] Figure 6 This is a flowchart illustrating the process of constructing a confidence accounting calculation model in another embodiment;

[0052] Figure 7 This is a flowchart illustrating the process of constructing a confidence accounting calculation model in another embodiment;

[0053] Figure 8 This is a flowchart illustrating the calculation of the second posterior probability distribution in one embodiment;

[0054] Figure 9 This is a flowchart illustrating a preferred embodiment of the carbon dioxide emission accounting method;

[0055] Figure 10 This is a structural block diagram of a carbon dioxide emission accounting device in one embodiment;

[0056] Figure 11 This is a block diagram of a carbon dioxide emission accounting system in one embodiment. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] The technical background will be explained below:

[0059] In high-carbon dioxide-emitting industries such as petrochemicals, steel, and power, accurate and reliable carbon dioxide emission accounting technology has become an important foundation for enterprises to fulfill their emission reduction responsibilities. The core of carbon dioxide emission accounting lies in obtaining accurate emission data at different production stages and equipment nodes, and combining it with statistical models and technical parameters to make reasonable estimates and predictions.

[0060] For petrochemical plants, high-precision carbon dioxide emission accounting technology is crucial. Currently, most carbon dioxide emission accounting schemes rely on manual sampling and measurement, single-point monitoring estimation, or calculation using fixed emission factors. While these methods achieve an estimate of total emissions, they suffer from poor adaptability under multi-node and multi-technology conditions, failing to fully utilize inter-node correlation information and technological differences, thus limiting the accuracy and reliability of the results. Furthermore, existing technologies generally lack the quantification of emission uncertainties. For example, confidence interval calculations often rely on simple statistical inference, ignoring the impact of production fluctuations, equipment performance differences, and technological upgrades, thus failing to dynamically update emission results under different operating conditions.

[0061] In summary, the main problems with the relevant technologies are: poor adaptability in actual multi-node emission and multi-emission technology conditions, resulting in low accuracy and reliability of carbon dioxide emission calculation results.

[0062] Based on this, this application provides a method for calculating carbon dioxide emissions, including: identifying at least one emission node to be calculated in a petrochemical plant; for each emission node to be calculated, inputting the emission technology information of the emission node to be calculated into a pre-established confidence calculation model, determining a reference emission node that matches the emission technology information of the emission node to be calculated through the confidence calculation model, and calculating a target emission confidence interval for the emission node to be calculated based on the emission confidence interval of the reference emission node; wherein, the confidence calculation model is established based on the emission technology information and emission confidence intervals of each emission node in a petrochemical-like plant that matches the petrochemical plant to be calculated, and the reference emission node belongs to the emission nodes in the petrochemical-like plant; and determining the carbon dioxide emission calculation result of the emission node to be calculated based on the target emission confidence interval. This application can improve the accuracy and reliability of carbon dioxide emission calculation. See the following embodiments for details:

[0063] The carbon dioxide emission accounting method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, the data acquisition terminal 102 communicates with the server 104 via a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated onto the server 104 or placed on a cloud or other network server. The process involves identifying at least one emission node in the petrochemical plant to be accounted for; for each emission node, the emission technology information of the emission node is input into a pre-established confidence calculation model; the confidence calculation model determines a reference emission node that matches the emission technology information of the emission node to be accounted for; and the target emission confidence interval of the emission node to be accounted for is calculated based on the emission confidence interval of the reference emission node. The confidence calculation model is established based on the emission technology information and emission confidence intervals of each emission node in a petrochemical-like plant that matches the petrochemical plant to be accounted for, and the reference emission node belongs to the emission nodes in the petrochemical-like plant. Based on the target emission confidence interval, the carbon dioxide emission accounting result of the emission node to be accounted for is determined. The data acquisition terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle systems, and projection devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. The server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0064] In one exemplary embodiment, such as Figure 2 As shown, a method for calculating carbon dioxide emissions is provided, which is then applied to... Figure 1 Taking the server in the example of this, the explanation includes:

[0065] S210, Identify at least one emission node to be accounted for in the petrochemical plant to be accounted for.

[0066] Among them, the petrochemical plants to be accounted for are those that require carbon dioxide emission accounting. The aforementioned emission nodes to be accounted for are the key carbon dioxide emission nodes included in the petrochemical plants to be accounted for. The emission nodes to be accounted for refer to specific equipment, process links or areas that have independent carbon dioxide emission behavior in the production process, such as oil refining towers, heating furnaces, combustion boilers, etc. The emission characteristics of the target plant can be reflected based on the emission nodes to be accounted for.

[0067] In this embodiment of the application, each emission node to be accounted for can be identified through the production data of the petrochemical plant to be accounted for. The production data includes the operating status of each piece of equipment, fuel consumption, process parameters, etc. in the petrochemical plant to be accounted for. By analyzing the production data, the main carbon dioxide emission sources in the petrochemical plant to be accounted for can be determined and divided into n independent emission nodes, namely the emission nodes to be accounted for, where n represents the number of emission nodes to be accounted for, and n is an integer greater than or equal to 1.

[0068] S220, for each emission node to be calculated, the emission technology information of the emission node to be calculated is input into a pre-established confidence calculation model. The confidence calculation model determines a reference emission node that matches the emission technology information of the emission node to be calculated. The target emission placement confidence interval of the emission node to be calculated is calculated based on the emission placement confidence interval of the reference emission node. The confidence calculation model is established based on the emission technology information and emission placement confidence interval of each emission node of a petrochemical-like plant that matches the petrochemical plant to be calculated. The reference emission node is an emission node in the petrochemical-like plant.

[0069] In this context, "petrochemical-like plants" refers to other petrochemical plants that are similar to the petrochemical plant to be accounted for in various aspects. Similarly, "reference emission nodes" refers to the carbon dioxide emission nodes included in the "petrochemical-like plants." The aforementioned confidence calculation model is a pre-established model that characterizes the mapping relationship between the emission technology information of a node and its emission confidence interval. The emission technology information of the aforementioned nodes (whether they are emission nodes to be accounted for or reference emission nodes) includes, but is not limited to, process parameters and equipment characteristics that affect carbon dioxide emissions, such as fuel type, process flow, equipment model, operating load, and temperature and pressure conditions. The emission confidence interval of the aforementioned nodes (whether they are emission nodes to be accounted for or reference emission nodes) includes a range that can cover the actual emission value at a certain confidence level. For example, a node has a 95% probability that its daily carbon dioxide emission confidence interval is between 9.7 tons and 10.5 tons. In practical applications, the emission confidence interval can be determined from the fluctuation of the node's emissions over time.

[0070] In this embodiment of the application, a confidence accounting calculation model is first established based on the emission technology information of each reference emission node in the petrochemical plant and the corresponding emission confidence interval. In practical applications, this confidence accounting calculation model is established based on the Bayesian hierarchical model.

[0071] Then, for each emission node to be calculated, the emission technology information of the emission node to be calculated is input into the aforementioned confidence calculation model. The confidence calculation model determines a reference emission node that matches the emission technology information of the emission node to be calculated, that is, at least one reference emission node among petrochemical plants with similar emission technology information to the emission node to be calculated. Then, the target emission placement confidence interval of the emission node to be calculated is calculated based on the emission placement confidence interval corresponding to the reference emission node. The calculation of the target emission placement confidence interval of the emission node to be calculated based on the emission placement confidence interval of the reference emission node includes, but is not limited to, the following two calculation methods: 1. Determine the reference emission node among petrochemical plants that is most similar to the emission technology information of the emission node to be calculated, and directly assign the emission placement confidence interval of the most similar reference emission node to the emission node to be calculated, that is, directly use the emission placement confidence interval of the most similar reference emission node as the target emission placement confidence interval of the emission node to be calculated. 2. Identify multiple reference emission nodes in petrochemical plants whose emission technology information is similar to that of the emission node to be calculated. Calculate the similarity score between the emission technology information of each reference emission node and the emission technology information of the emission node to be calculated, and then calculate the corresponding weight based on the similarity score. For example, if the similarity between reference emission node 1 and the emission technology information of the emission node to be calculated is 90%, its weight is 0.9; if the similarity between reference emission node 2 and the emission technology information of the emission node to be calculated is 80%, its weight is 0.8; and if the similarity between reference emission node 3 and the emission technology information of the emission node to be calculated is 80%, its weight is 0.8. The similarity between the emission technology information of emission nodes is 60%, and its weight is 0.6. Similarly, the emission placement confidence intervals of all reference emission nodes corresponding to the emission node to be calculated are weighted and summed. The weight value is the weight determined based on similarity mentioned above. Thus, the target emission placement confidence interval of the emission node to be calculated is obtained by combining the emission placement confidence of multiple reference emission nodes. It can be understood that in order to obtain an accurate target emission placement confidence interval, the weights of each reference emission node can be calculated using softmax to ensure that the sum of the weights of multiple reference emission nodes is 1.

[0072] S230, based on the target emission confidence interval, determine the carbon dioxide emission accounting results for the emission nodes to be calculated.

[0073] Among them, carbon dioxide emission accounting refers to the calculation of the total or sub-node emissions by using the numerical range of confidence intervals and combining the operating characteristics of each node in the production system, so as to obtain a result that conforms to the actual uncertainty conditions. The carbon dioxide emission accounting result is the output of this calculation, including but not limited to the total emissions, the emissions of each node, and the fluctuation range of each node.

[0074] In this embodiment of the application, it can be understood that the target emission confidence interval is the range of possible fluctuations in the emission of the corresponding emission node to be calculated at a certain confidence level. Therefore, the above carbon dioxide emission calculation results (such as total emissions, emissions of sub-nodes, and fluctuation ranges of each node) can be obtained based on the target emission confidence interval of each emission node to be calculated.

[0075] In summary, a confidence calculation model based on a petrochemical plant-like structure can be used to quickly calculate the emission confidence intervals for the emission nodes to be calculated, and further, the calculation results for the emission nodes to be calculated can be completed based on the emission confidence intervals. The embodiments of this application realize the construction of a confidence calculation model that integrates emission change data, technical information, and emission confidence intervals from multiple reference nodes. This allows for the output of high-precision, quantifiable confidence intervals for emission emissions under different production conditions, facilitating accurate and reliable calculation of carbon dioxide emissions.

[0076] In one exemplary embodiment, such as Figure 3 As shown, a reference emission node matching the emission technology information of the emission node to be accounted for is determined through a confidence calculation model. The target emission confidence interval for the emission node to be accounted for is then calculated based on the emission confidence interval of the reference emission node, including:

[0077] S310, based on the emission technology information of the emission node to be calculated, a distance measurement is performed between the emission technology information of the reference emission node to obtain the characteristic distance between the emission technology information of the emission node to be calculated and the emission technology information of at least one reference emission node.

[0078] The calculation of the distance metric between the technical information of each emission node to be calculated and each reference emission node refers to comparing the differences in each feature dimension between the technical information of each emission node to be calculated in the petrochemical plant and the technical information of each reference emission node in the similar petrochemical plant. The aforementioned emission technical information includes, but is not limited to, fuel type, process type, equipment model, equipment load range, and equipment operating temperature and pressure. This emission technical information can be numerical, such as equipment load percentage, or categorical, such as whether the fuel is natural gas or coal. The distance metric is a method of quantifying differences; for example, Euclidean distance can be used for numerical features, while edit distance can be used for string features.

[0079] The aforementioned feature distance refers to the similarity between each emission node to be accounted for and at least one reference emission node after distance metric calculation. In one embodiment, each emission node to be accounted for can be compared with each reference emission node in the petrochemical plant (assuming there are m reference emission nodes), so that each emission node to be accounted for can have m feature distances. In another embodiment, each emission node to be accounted for can be compared with m reference emission nodes in the petrochemical plant, and finally retain the x feature distances with the highest similarity, that is, the closest similarity, where m and x are both positive integers greater than or equal to 1.

[0080] In this embodiment of the application, for each emission node to be calculated, the distance metric between each emission node to be calculated and the emission technology information of each reference emission node in the petrochemical plant is calculated to obtain the characteristic distance between each emission node to be calculated and the emission technology information of at least one reference emission node. This characteristic distance characterizes the degree of difference in the emission technology information of the two emission nodes being compared.

[0081] S320 converts the distances of each feature into similarity weights.

[0082] In this embodiment, each feature distance is converted into a similarity weight. The similarity weight is a proportional value used to reflect the degree of similarity in the technical information of two nodes. The smaller the distance, the higher the similarity weight. For example, for the emission node A to be calculated, the average distance between the reference node A and the emission node A to be calculated across all technical information features is only 0.1, while the average distance between the reference node B and the emission node A to be calculated across all technical information features is 0.5. Therefore, the similarity weight between the reference node A and the emission node A to be calculated can be set to 0.9, and the similarity weight between the reference node B and the emission node A to be calculated can be set to 0.5. In summary, the conversion of feature distance into similarity weight can be performed using methods commonly used in the art, as long as the feature distance and the similarity weight are inversely proportional.

[0083] S330, based on the similarity weight and the corresponding arrangement confidence interval, the target arrangement confidence interval of the emission node to be verified is calculated.

[0084] In this application embodiment, two methods are given for calculating the target emission placement confidence interval based on similarity weight and placement confidence interval: 1. Determine the reference emission node in the petrochemical plant that is most similar to the emission technology information of the emission node to be calculated, and directly assign the placement confidence interval of the most similar reference emission node to the emission node to be calculated, that is, directly use the placement confidence interval of the most similar reference emission node as the target placement confidence interval of the emission node to be calculated. 2. Identify multiple reference emission nodes in the petrochemical-like plant whose emission technology information is similar to that of the emission node to be calculated (this can be all reference emission nodes of the petrochemical-like plant, or a few of the most similar reference emission nodes selected; the specific number of reference emission nodes can be determined by relevant technical personnel according to actual needs). Calculate the similarity weight based on the characteristic distance between the emission technology information of each reference emission node and the emission technology information of the emission node to be calculated. Finally, sum the emission placement confidence intervals of all reference emission nodes corresponding to the emission node to be calculated using a weighted average. The weight value is the aforementioned similarity weight. By combining the emission placement confidence intervals of multiple reference emission nodes, the target emission placement confidence interval of the emission node to be calculated can be obtained. It is understandable that, in order to obtain an accurate target emission placement confidence interval, the weights of each reference emission node can be calculated using softmax to ensure that the sum of the weights of multiple reference emission nodes is 1.

[0085] In one exemplary embodiment, such as Figure 4 As shown, based on the target emission confidence interval, the carbon dioxide emission calculation results for the emission nodes to be calculated are determined, including:

[0086] S410, perform correlation analysis on the emission change data of all emission nodes to be accounted for, and obtain the correlation matrix between emission nodes.

[0087] Correlation analysis refers to the use of statistical methods to measure the degree of interrelation between emission changes at different emission nodes in a petrochemical plant, such as measuring whether two or more emission nodes change synchronously.

[0088] The aforementioned correlation matrix is ​​also known as the covariance matrix. The correlation matrix is ​​a set of rules in which each element represents the covariance value between two emission nodes to be accounted for, thereby reflecting whether the emission change data of the two emission nodes to be accounted for change synchronously. A positive covariance value indicates that the two nodes change in the same direction, a negative covariance value indicates that the two nodes change in opposite directions, and a covariance value close to zero indicates that the two nodes have no obvious linear relationship.

[0089] In this embodiment of the application, based on the emission change data of all emission nodes to be accounted for, a correlation analysis is performed on each emission node to be accounted for to obtain the covariance matrix between the emission nodes to be accounted for, i.e. the aforementioned correlation matrix.

[0090] S420, sample and calculate the emission placement information intervals of all emission nodes to be accounted for, and obtain the total emission placement information interval of the petrochemical plant to be accounted for.

[0091] The sampling calculation can be implemented using the Monte Carlo sampling algorithm.

[0092] In this embodiment, the Monte Carlo sampling algorithm can be used to perform multiple random samplings of the emission confidence intervals of all emission nodes to be accounted for in order to estimate the uncertainty range of the total emissions. For example, 10,000 random simulation samplings can be performed. Each time, emission data is generated and summed according to the probability distribution of each emission node to be accounted for. The total distribution can be used to calculate the total emission confidence interval, which represents the possible range of values ​​for the total emissions of all emission nodes to be accounted for at a certain confidence level (e.g., 95%).

[0093] S430, based on the total emission distribution interval and the correlation matrix between emission nodes, performs carbon dioxide emission accounting on the emission change data of each emission node to be accounted for, and obtains the carbon dioxide emission accounting results.

[0094] In this embodiment, within the obtained total emission placement confidence interval, the aforementioned correlation matrix is ​​used to perform carbon dioxide emission calculation on the emission change data of all emission nodes to be calculated, yielding the aforementioned carbon dioxide emission calculation result. This calculation combines correlation information with the interval range, ensuring that the estimation of total emissions considers not only the uncertainty of individual nodes but also their mutual influence, thus obtaining a more accurate carbon dioxide emission calculation result. For example, based on the previous total emission placement confidence interval calculation, there is a 95% probability that the total emission placement confidence interval for emission nodes A, B, and C to be calculated should be [32, 35] tons, with nodes A and B positively correlated, nodes A and C negatively correlated, and nodes B and C essentially uncorrelated. If, at this point, the total emissions of emission nodes A, B, and C to be calculated are detected to be 36 tons, exceeding the aforementioned total emission placement confidence interval, this indicates that there may be a random error in the instrument of a certain monitoring device, requiring calculation. Furthermore, based on the correlation between emission nodes A, B, and C to be calculated, each node can be re-detected, thereby obtaining a more accurate calculation result. In summary, the carbon dioxide emission accounting results include, but are not limited to, the summation, decomposition, or weighted calculation of overall and node emissions, and the calculation results basically conform to the theoretically established uncertainty conditions, that is, they have higher accuracy and precision.

[0095] Through the embodiments of this application, the emission change data of the emission nodes to be calculated can be calculated by using the correlation matrix between each emission node to be calculated and the total emission placement confidence interval. It can not only calculate the total emission of all emission nodes to be calculated and the emission of each node by summing and decomposition, but also verify the calculation results and the collection results of emission change data, and promptly detect and correct the errors of individual abnormal nodes, so that the calculation results have higher accuracy and precision.

[0096] In one exemplary embodiment, such as Figure 5 As shown, the construction of the confidence accounting calculation model includes:

[0097] S510: Obtain emission change data and emission technology information for each reference emission node of a petrochemical-like plant that matches the petrochemical plant to be accounted for.

[0098] Among them, the petrochemical plants that match the petrochemical plant to be accounted for are other petrochemical plants that are similar to the petrochemical plant to be accounted for in terms of production processes and other characteristics.

[0099] The emission change data for a node refers to the fluctuations in carbon dioxide emissions based on time series data for that node, including but not limited to instantaneous carbon emission rates, peak emissions, and cumulative emissions per unit time. Specifically, carbon dioxide emission change data refers to the changes in emissions over time recorded and analyzed for that node in a time series dimension, reflecting the dynamic fluctuation trend of emissions. Time series emissions mean that there is a corresponding emission record for each moment. The node's carbon dioxide emission change data includes instantaneous carbon emission rates, peak emissions, and cumulative emissions per unit time. Instantaneous carbon emission rate is the emission rate at an instant within a unit of time; peak emissions are the maximum emissions that occur during the monitoring period; and cumulative emissions per unit time refer to the cumulative value of total emissions over a fixed time interval.

[0100] The emission technology information of a node refers to the technical parameters related to the emission process and conditions of that node, which affect the quantity and characteristics of emissions. Carbon dioxide emission technology information includes fuel type, process type, equipment model, equipment load range, and equipment operating temperature and pressure. Fuel type determines the amount of carbon dioxide produced after combustion; for example, natural gas, coal, and oil have different carbon contents. Process type refers to the technological process used in production, such as catalytic cracking or steam cracking. Equipment model involves the design and performance parameters of the equipment. Equipment load range indicates the power or capacity range of the equipment. Equipment operating temperature and pressure are the temperature and pressure of the equipment during production; for example, a reactor may operate at 500 degrees Celsius and 2 MPa. In practical applications, the actual acquired node carbon dioxide emission change data and node emission technology information can be subjected to feature engineering processing (this feature engineering processing includes, but is not limited to, processing and optimization to better suit subsequent analysis and calculation, outputting processed carbon dioxide emission change data and carbon dioxide emission technology information for each node. Feature engineering processing may include outlier removal, data standardization, normalization, feature selection, and feature combination, etc.) to obtain the emission change data and emission technology information for each of the aforementioned reference emission nodes. The table below is a schematic diagram of emission variation data and technical information data for a node in one embodiment, for reference only:

[0101]

[0102] S520 determines the emission confidence interval for each reference emission node based on emission change data and emission technology information of each reference emission node of the petrochemical plant.

[0103] In this embodiment of the application, the confidence interval is determined by statistical methods, which combine the calculation of the confidence interval for each reference emission node. That is, a range that can cover the actual emission value at a certain confidence level is calculated. Arranging the confidence interval can help quantify the uncertainty of emissions and improve the scientific nature of the accounting.

[0104] This application further provides a specific method for calculating emission confidence intervals: First, emission change data of each reference emission node within a certain time period (e.g., 24 hours) are collected, and the technical parameters of the node within the same time period are recorded to determine whether these emission change data were collected under normal operating conditions. Then, the collected emission change data and technical information are organized, and extreme values ​​caused by equipment start-up / shutdown, fault maintenance, or obviously abnormal operating conditions are removed, retaining only data from normal and stable operation. Finally, statistical calculations are performed on the cleaned emission change data to obtain its average value, and the standard deviation is calculated based on the data fluctuation range. Typically, a 95% confidence level is used, i.e., the average value is added to and subtracted from twice the standard deviation to obtain a numerical range, which is the aforementioned emission confidence interval. This interval represents that, under normal conditions, there is a 95% probability that the emission of the node falls within this interval. The above method for calculating emission confidence intervals is only an example for reference.

[0105] S530 constructs a confidence accounting calculation model based on the emission technology information, emission change data and corresponding emission confidence intervals of all reference emission nodes of petrochemical plants.

[0106] In this embodiment, a ternary confidence accounting calculation model is constructed based on the emission technology information, emission change data, and emission confidence intervals of all reference emission nodes. The aforementioned confidence accounting calculation model is ternary, which means that the emission change data (including the identity of the node), emission technology information, and confidence interval of each reference node are associated as a whole to establish a mapping relationship. This enables the comprehensive utilization of emission technology information and emission confidence intervals, that is, the establishment of a mapping relationship between emission technology information and emission confidence intervals, so as to realize the emission calculation and migration inference of different emission nodes, thereby providing accurate emission estimates for the petrochemical plant to be calculated.

[0107] In one exemplary embodiment, such as Figure 6 As shown, based on the emission technology information, emission change data, and corresponding emission confidence intervals of all reference emission nodes of petrochemical plants, a confidence calculation model is established, including:

[0108] S610, based on the emission change data of each reference emission node, construct the first prior probability distribution corresponding to each reference emission node.

[0109] This application presents a method for establishing a confidence accounting calculation model. First, based on the emission change data of each reference emission node, a first prior probability distribution is constructed. This involves converting emission fluctuation data at different times into a probability distribution. The prior distribution is a distribution form inferred based on historical data and existing experience. For example, if the emission rate of a reference emission node is mostly between 2 and 3 kg / min, occasionally reaching 3.5 kg / min, then the prior distribution will concentrate probability weights between 2,000 and 3 kg / min, and retain a small probability tail near 3.5 kg / min. In summary, the corresponding first prior probability distribution can be constructed based on the emission change data of each reference emission node.

[0110] S620, Bayesian inference is performed on the first prior probability distribution and emission technology information of each reference emission node to obtain the first posterior probability distribution corresponding to each emission node.

[0111] In this embodiment, Bayesian inference is performed based on the first prior probability distribution and emission technology information of each node. That is, based on the first prior probability distribution, emission technology information and newly collected observation data, the posterior distribution is calculated according to Bayes' theorem. The first prior probability distribution is combined with the newly collected observation data to calculate the updated first posterior probability distribution corresponding to each emission node. The newly collected observation data is the carbon dioxide emission data of each reference emission node that was actually collected, and the newly collected observation data does not belong to the emission change data of the corresponding reference emission node.

[0112] In practical applications, a predefined Bayesian hierarchical model, i.e., a predefined initial confidence calculation model, can be fitted and trained based on the first prior probability distribution and emission technology information of each reference emission node. Historical emission data and technological background information of the nodes are input into the model, and the model parameters are continuously adjusted through statistical calculations to maximize the Bayesian hierarchical model's interpretability of actual observation data. The result of the fitting and training is the Bayesian posterior probability distribution for each node. This posterior distribution integrates prior knowledge and new data, and can more accurately reflect the true emission characteristics of that node.

[0113] S630, determine the emission confidence interval corresponding to each reference emission node based on the first posterior probability distribution of each reference emission node.

[0114] In this embodiment of the application, for each reference emission node, a large number of samples can be drawn from the first posterior probability distribution. These samples represent various possible values ​​of the actual emission level of the node. Then, a confidence level is set, which can generally be set to the commonly used 95%, that is, the expected interval contains the actual emission value with a 95% probability. Finally, the samples are sorted, and all the sample values ​​are arranged in ascending order. The value located at the 2.5th percentile is used as the lower limit of the interval, and the value located at the 97.5th percentile is used as the upper limit of the interval. The range between these two endpoints is the aforementioned arrangement confidence interval.

[0115] The confidence intervals are calculated based on the first posterior probability distribution, which makes it highly probable that the confidence interval contains the true emission value. The final result is the confidence interval for each reference emission node, which is an intervalized estimate of the emission level of each reference emission node. This can reflect both the average level of emissions and its uncertainty.

[0116] S640, input each reference emission node, as well as the emission technology information and emission confidence interval corresponding to each reference emission node, into the initial accounting calculation model for fitting training to obtain the confidence accounting calculation model.

[0117] The initial accounting calculation model is an initial Bayesian hierarchical model. This model can capture the commonalities of all nodes while retaining the individual characteristics of each node in probabilistic inference. Specifically, the parameters of the Bayesian hierarchical model include global regression coefficients, node-specific biases, and time series terms. Global regression coefficients are parameters shared by all nodes and are used to describe global influencing factors, such as the average increase in emission rates across all nodes due to rising temperatures. Node-specific biases are independent correction values ​​for individual nodes, representing the difference between that node and the global trend. Time series terms are parameters that describe the pattern of node emissions over time, such as periodic fluctuations or long-term trends.

[0118] In this embodiment, constructing the confidence calculation model includes: forming a triplet of emission technology information, confidence interval, and unique node identifier for each reference emission node; and using the triplet data of all reference emission nodes to train the model for fitting; continuously adjusting the model parameters through statistical calculations to ensure the Bayesian hierarchical model has the strongest interpretability of actual observation data. Specifically, the emission technology information and confidence interval of each node are used as input, and the posterior distribution of all unknown parameters in the model is calculated using the Bayesian inference method. After training, the model internally forms a mapping function from technology information to emission confidence intervals, thus completing the construction of the confidence calculation model. This model has the ability to infer the emission confidence interval of any node based on its emission technology information, and can quantify the uncertainty of the prediction. In other words, this step mainly integrates the global parameters of all reference emission nodes. After fitting training, the model has transfer capability, thus enabling direct input of the technology information of the node to be calculated and output of the corresponding emission confidence interval.

[0119] In one exemplary embodiment, such as Figure 7 As shown, based on the emission technology information, emission change data, and corresponding emission confidence intervals of all reference emission nodes of petrochemical plants, a confidence calculation model is established, including:

[0120] S710, for each emission node cluster, the mean of the emission change data of all reference emission nodes included in the emission node cluster is calculated to obtain the second prior probability distribution corresponding to the emission node cluster; wherein, the emission node cluster is obtained by clustering each reference emission node based on the emission change data and emission technology information of each reference emission node.

[0121] This application provides another method for establishing a confidence accounting calculation model. First, multiple reference emission nodes are clustered. Specifically, all reference emission nodes are clustered according to their emission change data and emission technology information. Nodes that are similar across multiple levels are grouped together to form multiple emission node clusters. Clustering is an unsupervised learning method that groups essentially similar nodes together by calculating the similarity between them, making the data structure simpler. Furthermore, the number of emission node clusters is less than the number of reference emission nodes, meaning that multiple emission nodes are merged into fewer groups, which facilitates subsequent analysis and modeling.

[0122] Then, the mean value of the emission change data of all reference emission nodes included in the emission node cluster is calculated to generate the Bayesian prior probability distribution corresponding to each emission node cluster, namely the second prior probability distribution mentioned above. That is, the emission data of each node in the emission node cluster are averaged to form the typical emission characteristics of the cluster. Then, the second prior probability distribution is constructed based on the mean value to express the preliminary characteristics of the emission characteristics of the cluster.

[0123] S720, posterior probability distribution is calculated based on the second prior probability distribution, emission node clusters, and technology information clusters to obtain the second posterior probability distribution corresponding to each emission node cluster; wherein, the technology information cluster is the feature extraction result of the emission technology information of the emission node cluster.

[0124] Within this cluster, each technical information cluster corresponds one-to-one with an emission node cluster. A technical information cluster is formed by summarizing or refining the technical information of each reference node within the corresponding emission node cluster, such as fuel type, process flow, and equipment model, to create a representative technical information description for that cluster. Technical information clusters can reflect the commonalities among nodes within a cluster. For example, if a cluster's equipment primarily uses natural gas fuel, its process type is catalytic cracking, and its equipment models are similar, it facilitates the unified application of subsequent models.

[0125] In this embodiment of the application, the posterior probability distribution is calculated based on the second prior probability distribution, the emission node cluster, and the technical information cluster. That is, the second prior probability distribution is combined with the actual acquired observation data to update the probability distribution of emissions for each cluster, thereby obtaining a new posterior probability distribution, namely the second posterior probability distribution. The observation data is newly collected carbon dioxide emission data and does not belong to the emission change data mentioned above.

[0126] S730, determine the emission confidence interval corresponding to each emission node cluster based on the second posterior probability distribution corresponding to each emission node cluster.

[0127] In this embodiment, the emissions included in the second posterior probability distribution are first arranged in ascending order. Assuming a preset confidence level of 95%, the value at the 2.5% position is taken as the lower limit of the interval, and the value at the 97.5% position is taken as the upper limit of the interval. The range between these two data represents the range within which the emissions of the emission node cluster have a 95% probability of falling.

[0128] The same operation was repeated for each emission node cluster to obtain the emission placement information interval for each emission node cluster.

[0129] S740 constructs a confidence accounting calculation model based on each emission node cluster, technical information cluster, and emission confidence interval.

[0130] In this embodiment, the identifier of each emission node cluster, the corresponding technical information cluster, and the corresponding emission confidence interval are combined to form a triple. The triples of all emission node clusters are then combined and input into the initial calculation model for fitting and training, thereby obtaining the final confidence calculation model. In this embodiment, technical features are associated with confidence intervals on a cluster-by-cluster basis, making the overall calculation more concise and statistically significant, and facilitating its application to emission estimation at petrochemical plants.

[0131] In one exemplary embodiment, such as Figure 8 As shown, the method also includes:

[0132] S810, obtain the migration variance corresponding to each reference emission node. The migration variance characterizes the degree of fluctuation of the emission change data under different operating conditions.

[0133] Migration variance is a statistical indicator of the degree of fluctuation of component data under different conditions. A larger variance indicates more drastic changes in the emission characteristics of that node during migration; a smaller variance indicates more stable characteristics after migration. Methods for obtaining the migration variance of each reference emission node include, but are not limited to: for each reference emission node, obtaining the emission characteristic values ​​of its various emission features (such as baseline emission levels, sensitivity to load changes, emission fluctuation amplitude, etc.) under multiple different conditions (i.e., different production conditions, different application scenarios, etc.). Then, for each emission feature's characteristic value under different conditions, calculating the variance of that characteristic value under all conditions (if there are only two different conditions, the difference between the two characteristic values ​​can be calculated). This yields the variance of one emission feature for that reference emission node. If multiple emission features need to be considered for that node, a weighted average or the maximum value can be taken from the variances of multiple emission features to obtain a comprehensive value, which is the migration variance of that node. Repeating the above steps for each reference emission node yields the migration variance for each reference emission node.

[0134] S820 uses migration variance as a hyperparameter for cross-validation to obtain the migration error distribution corresponding to each reference emission node.

[0135] In this embodiment, using transfer variance as a hyperparameter for cross-validation means that during model training, transfer variance is treated as an external parameter controlling the model's fit. Multiple rounds of cross-validation are used to evaluate the model's generalization ability under different parameter combinations. Cross-validation is a method that involves dividing data into multiple training and validation sets for repeated testing, effectively reducing the risk of overfitting. Transfer variance is used to influence the model's regularization or weight allocation, thereby indirectly controlling the model's sensitivity to nodes with unstable transfer behavior.

[0136] Subsequently, cross-validation was used to obtain multiple transfer error distributions, each corresponding one-to-one with the transfer error itself; that is, the probability distribution of the model prediction error for each node under different transfer variance conditions. The error distribution reflects the range of uncertainty in the prediction results during node migration.

[0137] S830, based on the migration error distribution corresponding to each node, the second prior probability distribution corresponding to each emission node cluster, and the technical information cluster corresponding to each emission node cluster, Bayesian posterior probability distribution is calculated to obtain the updated second posterior probability distribution.

[0138] In this embodiment, the Bayesian posterior probability distribution is recalculated based on the migration error distribution, the second prior probability distribution, the emission node cluster, and the emission technology information cluster. The error distribution is used as an additional weight or correction factor, combined with the prior information and technical characteristics of the cluster, to update the posterior distribution of each node, making the model more accurately reflect the true emission characteristics of the nodes after migration. After comprehensive correction using migration variance, error distribution, and cluster information, the updated second posterior probability distribution is output, providing an emission probability estimate that is closer to the actual migration scenario than the previous second posterior probability distribution.

[0139] In one exemplary embodiment, determining the arrangement confidence interval corresponding to each emission node cluster based on the second posterior probability distribution corresponding to each emission node cluster includes:

[0140] The corresponding rank confidence interval is determined from the second posterior probability distribution according to the preset confidence level, wherein the representative value of the rank confidence interval is the statistical value of the posterior probability distribution.

[0141] In this embodiment of the application, calculating the confidence interval based on the second posterior probability distribution means using the posterior probability distribution of each node under Bayesian statistics to determine a range that includes the true emission value at a given confidence level. The posterior probability distribution is a probability model derived from prior knowledge and observational data, reflecting the probability distribution of a certain emission value under known conditions.

[0142] Then, the corresponding emission confidence interval is calculated based on the second posterior probability distribution (the specific calculation method has been explained above and will not be repeated here). The emission confidence interval represents a credible numerical range corresponding to each reference emission node, which is used to measure the stability and reliability of the prediction. If the interval is narrow, it means that the emission prediction results of the node are relatively concentrated and have small fluctuations; if the interval is wide, it means that the uncertainty of the node is high.

[0143] Finally, the representative value for each emission confidence interval is taken as the posterior expectation of the corresponding second posterior probability distribution, which is the central value obtained by weighted averaging of all possible emission values ​​under the distribution. The posterior expectation is a single-point estimate that is most reasonable in a probabilistic sense and can be used as a representative figure for emissions in practical applications.

[0144] This application also provides a preferred embodiment of a carbon dioxide emission accounting method, such as... Figure 9 The figure shows a preferred embodiment of a carbon dioxide emission accounting method.

[0145] First, identify petrochemical-like plants similar to the petrochemical plant to be accounted for, and collect full-area production data of these petrochemical-like plants.

[0146] Then, based on the overall production data, m reference emission nodes in the petrochemical plant are identified, and the emission change data and emission technology information of the m reference emission nodes are determined. Based on the emission change data and emission technology information, the emission confidence intervals corresponding to the m reference emission nodes are determined. The methods for calculating the emission confidence intervals are as described above, including but not limited to: directly using statistical methods to define the emission confidence interval for each reference emission node based on the emission change data and emission technology information; or, as described above, first calculating the first prior probability distribution of the reference emission nodes, then performing Bayesian inference based on the first prior probability distribution and emission technology information to obtain the first posterior probability distribution corresponding to each emission node, and then determining the corresponding emission confidence interval based on the first posterior probability distribution; or, as described above, first clustering multiple reference emission nodes, then calculating the second prior probability distribution corresponding to each node cluster, then calculating the posterior probability distribution based on the second prior probability distribution, emission node clusters, and technology information clusters to obtain the second posterior probability distribution corresponding to each emission node cluster, and then determining the emission confidence interval corresponding to each emission node cluster based on the second posterior probability distribution.

[0147] After the confidence intervals for emission are calibrated, the confidence intervals for emission are input into the initial accounting calculation model, along with the emission change information, emission technology information, and corresponding reference emission nodes / emission node clusters (understandably, if the confidence interval for emission is the confidence interval for an emission node, then subsequent data will be the change information, technology information, and node identity of the node; similarly, if the confidence interval for emission is the confidence interval for an emission node cluster, then subsequent data will be the change information, technology information, and node identity of the cluster). This process is then used for fitting and training to form the confidence accounting calculation model.

[0148] Then, the petrochemical plants to be accounted for are identified, their production data is collected, and the emission nodes to be accounted for are identified based on the production data. The emission technology information of the emission nodes to be accounted for is input into the aforementioned confidence calculation model. Through similarity comparison, the target emission confidence interval corresponding to the emission node to be accounted for can be obtained. Finally, the corresponding carbon dioxide emission accounting result is determined based on the target emission confidence interval. The similarity comparison refers to matching the technical information of the emission node to be accounted for with the technical information of various reference emission nodes through mathematical distance measurement or similarity calculation methods, finding the closest reference sample in the feature space, thereby inferring the carbon dioxide emission confidence interval corresponding to the emission node to be accounted for, reflecting the possible fluctuation range of the node's emissions at a certain confidence level.

[0149] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0150] Based on the same inventive concept, this application also provides a carbon dioxide emission accounting apparatus for implementing the carbon dioxide emission accounting method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more embodiments of the carbon dioxide emission accounting apparatus provided below can be found in the limitations of the carbon dioxide emission accounting method described above, and will not be repeated here.

[0151] In one exemplary embodiment, such as Figure 10 As shown, a carbon dioxide emission accounting device is provided, comprising:

[0152] The acquisition module 1001 is used to identify at least one emission node to be accounted for in the petrochemical plant to be accounted for;

[0153] The calculation module 1002 is used to input the emission technology information of each emission node to be calculated into a pre-established confidence calculation model, determine a reference emission node that matches the emission technology information of the emission node to be calculated through the confidence calculation model, and calculate the target emission placement confidence interval of the emission node to be calculated based on the emission placement confidence interval of the reference emission node. The confidence calculation model is established based on the emission technology information and emission placement confidence interval of each emission node of a petrochemical-like plant that matches the petrochemical plant to be calculated, and the reference emission node is an emission node in the petrochemical-like plant.

[0154] The generation module 1003 is used to determine the carbon dioxide emission calculation results of the emission nodes to be calculated based on the target emission confidence interval.

[0155] Each module in the aforementioned carbon dioxide emission accounting device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module. The aforementioned carbon dioxide emission accounting device is also used to implement the carbon dioxide emission accounting method described in any of the preceding claims.

[0156] In one exemplary embodiment, such as Figure 11 As shown, a carbon dioxide emission accounting system is provided, including a data acquisition unit and a calculation unit:

[0157] The data acquisition unit 1101 is used to collect emission change data and emission technology information of each emission node to be calculated in the target petrochemical plant, as well as the overall production data of petrochemical plants.

[0158] The calculation unit 1102 is used to execute the carbon dioxide emission accounting method described above based on the emission change data and emission technology information of each emission node to be accounted for in the target petrochemical plant, as well as the overall production data of the petrochemical plant.

[0159] In one exemplary embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, implements any of the carbon dioxide emission accounting methods described above.

[0160] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements any of the carbon dioxide emission accounting methods described above.

[0161] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements any of the carbon dioxide emission accounting methods described above.

[0162] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0163] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0164] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for calculating carbon dioxide emissions, characterized in that, The method includes: Identify at least one emission point in the petrochemical plant to be accounted for; For each emission node to be calculated, the emission technology information of the emission node is input into a pre-established confidence calculation model. The confidence calculation model determines a reference emission node that matches the emission technology information of the emission node to be calculated. Based on the emission confidence interval of the reference emission node, the target emission confidence interval of the emission node to be calculated is calculated. The confidence calculation model is established based on the emission technology information and emission confidence intervals of each emission node in a petrochemical-like plant that matches the petrochemical plant to be calculated. The reference emission node is an emission node within the petrochemical-like plant. The calculation model includes, for each emission node cluster, calculating the mean of emission change data of all reference emission nodes included in the emission node cluster to obtain a second prior probability distribution corresponding to the emission node cluster; wherein, the emission node cluster is obtained by clustering each reference emission node based on the emission change data and emission technology information of each reference emission node; and calculating a posterior probability distribution based on the second prior probability distribution, the emission node cluster, and the technology information cluster to obtain a second posterior probability distribution corresponding to each emission node cluster; wherein, the technology information cluster is the feature extraction result of the emission technology information of the emission node cluster. Obtain the migration variance corresponding to each reference emission node, wherein the migration variance characterizes the degree of fluctuation of the emission change data under different operating conditions; The migration variance was used as a hyperparameter for cross-validation to obtain the migration error distribution corresponding to each reference emission node. The updated second posterior probability distribution is obtained by calculating the Bayesian posterior probability distribution based on the migration error distribution corresponding to each node, the second prior probability distribution corresponding to each emission node cluster, and the technical information cluster corresponding to each emission node cluster. The confidence intervals for each emission node cluster are determined based on the updated second posterior probability distribution corresponding to each emission node cluster; each emission node cluster, technical information cluster, and confidence interval for each emission node cluster are input into a preset initial accounting calculation model for fitting training to construct the confidence accounting calculation model. Based on the target emission confidence interval, determine the carbon dioxide emission calculation result of the emission node to be calculated.

2. The method according to claim 1, characterized in that, The step of determining a reference emission node that matches the emission technology information of the emission node to be accounted for through the confidence calculation model, and calculating the target emission confidence interval of the emission node to be accounted for based on the emission confidence interval of the reference emission node, includes: Based on the emission technology information of the emission node to be calculated, a distance measurement is performed between it and the emission technology information of the reference emission node to obtain the characteristic distance between the emission technology information of the emission node to be calculated and the emission technology information of at least one of the reference emission nodes; Convert the distances of each feature into similarity weights; Based on the similarity weights and the corresponding arrangement confidence intervals, the target arrangement confidence intervals for the emission nodes to be verified are calculated.

3. The method according to claim 1, characterized in that, The step of determining the carbon dioxide emission calculation result of the emission node to be calculated based on the target emission confidence interval includes: Correlation analysis was performed on the emission change data of all the emission nodes to be accounted for, and the correlation matrix between emission nodes was obtained. The emission placement information intervals of all the emission nodes to be accounted for are sampled and calculated to obtain the total emission placement information interval of the petrochemical plant to be accounted for. Based on the total emission distribution interval and the correlation matrix between emission nodes, carbon dioxide emission calculation is performed on the emission change data of each emission node to be calculated, and the carbon dioxide emission calculation result is obtained.

4. The method according to any one of claims 1 to 3, characterized in that, The construction of the confidence accounting calculation model includes: Obtain emission change data and emission technology information for each reference emission node of the petrochemical-like plant that matches the petrochemical plant to be calculated; Based on the emission change data and emission technology information of each reference emission node of the petrochemical plant, the emission confidence interval of each reference emission node is determined; The confidence calculation model is constructed based on the emission technology information, emission change data and corresponding emission confidence intervals of all reference emission nodes of the petrochemical plant.

5. The method according to claim 1, characterized in that, The step of determining the arrangement confidence interval corresponding to each emission node cluster based on the second posterior probability distribution corresponding to each emission node cluster includes: The corresponding confidence interval is determined from the second posterior probability distribution according to the preset confidence level, wherein the representative value of the confidence interval is the statistical value of the posterior probability distribution.

6. The method according to claim 1, characterized in that, The emission technology information includes fuel type, process type, equipment model, equipment compliance range, and equipment operating temperature and pressure.

7. The method according to claim 1, characterized in that, The emission change data includes instantaneous carbon emission rate, peak carbon emission, and cumulative amount per unit time. The instantaneous carbon emission rate is the emission rate at any moment within a unit time. The peak carbon emission is the maximum emission amount that occurs during the monitoring period. The cumulative amount per unit time refers to the sum of the total emissions within a preset fixed time interval.

8. The method according to claim 1, characterized in that, The initial accounting calculation model is a Bayesian hierarchical model, and the parameters of the Bayesian hierarchical model include global regression coefficients, node-specific offsets, and time series terms.

9. A carbon dioxide emission accounting device, characterized in that, The device includes: The acquisition module is used to identify at least one emission node in the petrochemical plant to be accounted for; The calculation module is used to input the emission technology information of each emission node to be calculated into a pre-established confidence calculation model, determine a reference emission node that matches the emission technology information of the emission node to be calculated through the confidence calculation model, and calculate the target emission placement confidence interval of the emission node to be calculated based on the emission placement confidence interval of the reference emission node; wherein, the confidence calculation model is established based on the emission technology information and emission placement confidence intervals of each emission node of a petrochemical-like plant that matches the petrochemical plant to be calculated, and the reference emission node belongs to the emission node of the petrochemical-like plant; wherein, the establishment of the confidence calculation model is used to input the emission technology information and emission placement confidence intervals of each emission node of a petrochemical-like plant that matches the petrochemical plant to be calculated. The credit calculation model includes, for each emission node cluster, calculating the mean of emission change data of all reference emission nodes included in the emission node cluster to obtain a second prior probability distribution corresponding to the emission node cluster; wherein, the emission node cluster is obtained by clustering each reference emission node based on the emission change data and emission technology information of each reference emission node; and calculating a posterior probability distribution based on the second prior probability distribution, the emission node cluster, and the technology information cluster to obtain a second posterior probability distribution corresponding to each emission node cluster; wherein, the technology information cluster is the feature extraction result of the emission technology information of the emission node cluster. Obtain the migration variance corresponding to each reference emission node, wherein the migration variance characterizes the degree of fluctuation of the emission change data under different operating conditions; The migration variance was used as a hyperparameter for cross-validation to obtain the migration error distribution corresponding to each reference emission node. The updated second posterior probability distribution is obtained by calculating the Bayesian posterior probability distribution based on the migration error distribution corresponding to each node, the second prior probability distribution corresponding to each emission node cluster, and the technical information cluster corresponding to each emission node cluster. The confidence intervals for each emission node cluster are determined based on the updated second posterior probability distribution corresponding to each emission node cluster; each emission node cluster, technical information cluster, and confidence interval for each emission node cluster are input into a preset initial accounting calculation model for fitting training to construct the confidence accounting calculation model. The generation module is used to determine the carbon dioxide emission calculation result of the emission node to be calculated based on the target emission confidence interval.

10. A carbon dioxide emission accounting system, characterized in that, The system includes an acquisition unit and a computing unit; The data acquisition unit is used to collect emission change data and emission technology information of each emission node of the petrochemical plant to be accounted for, as well as the overall production data of petrochemical plants. The calculation unit is used to execute the method described in any one of claims 1 to 8 based on the emission change data and emission technology information of each emission node of the petrochemical plant to be calculated, as well as the overall production data of the petrochemical plant.

Images