A carbon emission analysis method and device, electronic equipment and storage medium

Abstract

The application provides a carbon emission analysis method and device, electronic equipment and storage medium, and relates to the technical field of carbon emission data processing. The method comprises the following steps: acquiring multi-source real-time data and energy consumption data of each energy consumption node in the park within a current time period, and extracting multi-dimensional key features; determining initial energy structure data; according to the correlation relationship between the energy consumption node and other energy consumption nodes in the park, the initial energy structure data is corrected to obtain final energy structure data; then the dynamic emission factor matrix of the energy consumption node is determined; the carbon emission of the energy consumption node is simulated through a hybrid model to obtain the carbon emission result of the energy consumption node; and then the dynamic analysis and interpretation of the carbon emission are performed through an attention mechanism to obtain the carbon emission analysis result of the park within the current time period. The application realizes high-precision dynamic analysis of the carbon emission of the park.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of carbon emission data processing, in particular to a carbon emission analysis method and device, electronic equipment and storage medium. BACKGROUND

[0002] At present, with the emphasis on environmental protection and sustainable development, as an important carrier of industrial agglomeration and economic development, the carbon emission analysis of the park is extremely important in energy management, environmental policy making and enterprise carbon emission reduction strategy. Among them, for the carbon emission analysis of the park, different emission factors need to be combined with different energy structures in the park to calculate the carbon emission, so as to obtain the analysis result of carbon emission.

[0003] In the related art, since the existing emission factor is based on fixed data and assumption, the emission factor mostly adopts fixed value, and the carbon emission in a certain period of time is analyzed statically by using fixed emission factor, so it is difficult to reflect the dynamic change of carbon emission, and thus it is unable to adapt to the dynamic change of park energy structure, resulting in large deviation of carbon emission analysis result. SUMMARY

[0004] The problem solved by the present application is how to improve the dynamic analysis accuracy of carbon emission in the park.

[0005] To solve the above problems, the present application provides a carbon emission analysis method, device, electronic equipment and storage medium.

[0006] In a first aspect, the present application provides a carbon emission analysis method, comprising:

[0007] Obtaining multi-source real-time data and energy consumption data of each energy consumption node in the park in a current time period;

[0008] According to the carbon emission analysis requirement, the multi-dimensional key features of each energy consumption node are extracted from the multi-source real-time data of each energy consumption node;

[0009] According to the multi-dimensional key features of each energy consumption node, the initial energy structure data of each energy consumption node is determined;

[0010] According to the correlation relationship between the energy consumption node and other energy consumption nodes in the park, the initial energy structure data is corrected to obtain the final energy structure data;

[0011] According to the final energy structure data, the dynamic emission factor matrix of the energy consumption node is obtained by using a hierarchical adaptive calculation model;

[0012] The carbon emission results of the energy consumption nodes are obtained by simulating carbon emissions of the energy consumption nodes through a hybrid model combined with the dynamic emission factor matrix, the multi-source real-time data and the energy consumption data;

[0013] The carbon emission analysis results of the park in the current time period are obtained by performing dynamic analysis and interpretation of carbon emissions through an attention mechanism according to the carbon emission results of each energy consumption node.

[0014] Optionally, the multi-dimensional key features of the energy consumption nodes are extracted from the multi-source real-time data of each energy consumption node according to carbon emission analysis requirements, including:

[0015] The type of the carbon emission analysis requirement is determined according to the management scenario of the park, and the type of the carbon emission analysis requirement includes short-term emission reduction monitoring requirements, long-term policy making requirements and equipment energy efficiency optimization requirements;

[0016] The target data type corresponding to the multi-source real-time data of each energy consumption node is matched according to the type of the carbon emission analysis requirement;

[0017] The original features of the target data type corresponding dimension are extracted from the target data type, wherein the time dimension energy consumption peak-valley fluctuation feature and the device dimension real-time load feature are extracted for the short-term emission reduction monitoring requirement, the spatial dimension functional area energy consumption difference feature and the policy dimension clean energy proportion change feature are extracted for the long-term policy making requirement, and the device dimension aging loss feature and the working condition dimension energy production-energy consumption correlation feature are extracted for the equipment energy efficiency optimization requirement;

[0018] The original features of each extracted dimension are preprocessed to obtain the multi-dimensional key features of each energy consumption node.

[0019] Optionally, the initial energy structure data of the energy consumption node is determined according to the multi-dimensional key features of each energy consumption node, including:

[0020] The multi-dimensional key features are analyzed and classified through an energy structure identification model to determine a plurality of energy structure types and scene correlation features of the energy consumption node;

[0021] The energy type proportion, energy conversion efficiency and energy consumption fluctuation data of the energy consumption node are determined according to the energy structure type;

[0022] The energy consumption strongly related scene influence data is determined according to the scene correlation feature;

[0023] The energy type proportion, the energy conversion efficiency, the energy consumption fluctuation data, and the scenario influence data of the energy consumption node are taken as the initial energy structure data;

[0024] The initial energy structure data is associated and corrected according to the association relationship of the energy consumption node with other energy consumption nodes in the park, to obtain final energy structure data, including:

[0025] According to the association relationship of each energy consumption node in the park, the affected coefficient of the energy consumption node is quantified;

[0026] The energy type proportion, the energy conversion efficiency, the energy consumption fluctuation data, and the scenario influence data are corrected according to the affected coefficient, to obtain final energy structure data.

[0027] Optionally, the hierarchical adaptive calculation model includes a basic factor matching layer, an efficiency correction layer, a scenario dynamic adjustment layer, and a bias calibration layer.

[0028] The dynamic emission factor matrix of the energy consumption node is obtained according to the final energy structure data through the hierarchical adaptive calculation model, including:

[0029] An initial emission factor vector is obtained according to the energy type proportion through the basic factor matching layer, in combination with a preset industry benchmark emission factor library.

[0030] The initial emission factor vector is corrected according to the energy conversion efficiency and the energy consumption fluctuation data through the efficiency correction layer, to obtain an efficiency-corrected factor vector.

[0031] The efficiency-corrected factor vector is dynamically adjusted in real time according to the scenario influence data through the scenario dynamic adjustment layer, to obtain a scenario-adjusted factor matrix.

[0032] The bias calibration layer is combined with historical carbon emission evaluation data to determine a bias value of the scenario-adjusted factor matrix and historical actual monitoring data, and the scenario-adjusted factor matrix is self-calibrated and optimized based on the bias value, to obtain the dynamic emission factor matrix.

[0033] Optionally, the hierarchical adaptive calculation model further includes a real-time feedback layer.

[0034] The dynamic emission factor matrix of the energy consumption node is obtained according to the final energy structure data through the hierarchical adaptive calculation model, further including:

[0035] The real-time carbon emission concentration data collected by the real-time carbon emission monitoring device arranged in the park are acquired through the real-time feedback layer;

[0036] A feedback compensation factor is constructed according to the deviation between the real-time carbon emission concentration data and the estimated carbon emission concentration data generated by the dynamic emission factor matrix;

[0037] The dynamic emission factor matrix is iteratively optimized online through the feedback compensation factor, and the emission factor values in the dynamic emission factor matrix are updated.

[0038] Optionally, the carbon emission of the energy consumption node is simulated through a hybrid model by combining the dynamic emission factor matrix, the multi-source real-time data and the energy consumption data, and the carbon emission result of the energy consumption node is obtained, including:

[0039] The energy consumption data is split according to energy types, and the sub-consumption amount of each energy type of the energy consumption node is obtained;

[0040] The sub-consumption amount is multiplied by the emission factor of the corresponding energy type in the dynamic emission factor matrix element by element, and the basic carbon emission amount of the energy consumption node is obtained;

[0041] According to the multi-source real-time data, a bidirectional long short-term memory network model is used for fitting, and a nonlinear correction amount of the energy consumption node is obtained;

[0042] The basic carbon emission amount and the nonlinear correction amount are weighted and fused to obtain a preliminary carbon emission result of the energy consumption node;

[0043] The preliminary carbon emission result is subjected to outlier detection and elimination, and is labeled in combination with the collection frequency of the energy consumption data, to generate the carbon emission result containing a timestamp, an energy type sub-carbon emission value and a total carbon emission value.

[0044] Optionally, according to the carbon emission result of each energy consumption node, a dynamic analysis and interpretation of carbon emission are performed through an attention mechanism, and a carbon emission analysis result of the park in the current time period is obtained, including:

[0045] The carbon emission result of the energy consumption node is input into an attention mechanism model;

[0046] The energy type sub-carbon emission value and the total carbon emission value in the carbon emission result are subjected to weight distribution through the attention mechanism model, and the contribution of each energy consumption individual in the energy consumption node to carbon emission is determined;

[0047] According to the contribution degree, the energy consumption individual is dynamically analyzed to obtain a carbon emission source and a key driving factor of the park;

[0048] According to the carbon emission source and the key driving factor of the park, the carbon emission analysis result of the park in the current time period is generated.

[0049] In a second aspect, a carbon emission analysis device is provided, which comprises:

[0050] A data acquisition unit is configured to acquire multi-source real-time data and energy consumption data of each energy consumption node in a park in a current time period;

[0051] A feature extraction unit is configured to extract multi-dimensional key features of each energy consumption node from the multi-source real-time data of the energy consumption node according to carbon emission analysis requirements;

[0052] An energy structure analysis unit is configured to determine initial energy structure data of each energy consumption node according to the multi-dimensional key features of the energy consumption node;

[0053] A correction unit is configured to correct the initial energy structure data according to the correlation between the energy consumption node and other energy consumption nodes in the park to obtain final energy structure data;

[0054] An emission factor matrix construction unit is configured to obtain a dynamic emission factor matrix of the energy consumption node according to the final energy structure data through a hierarchical adaptive calculation model;

[0055] An analog unit is configured to simulate carbon emission of the energy consumption node through a hybrid model in combination with the dynamic emission factor matrix, the multi-source real-time data and the energy consumption data to obtain a carbon emission result of the energy consumption node;

[0056] An analysis unit is configured to perform dynamic analysis and interpretation of carbon emission through an attention mechanism according to the carbon emission result of each energy consumption node to obtain a carbon emission analysis result of the park in the current time period.

[0057] In a third aspect, an electronic device is provided, which comprises a processor and a memory, and the memory is configured to store a computer program;

[0058] The computer program, when loaded by the processor, enables the processor to perform the carbon emission analysis method as described above.

[0059] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the carbon emission analysis method as described above.

[0060] The carbon emission analysis method, device, electronic equipment and storage medium of the present application can dynamically reflect the actual state of energy consumption in the park by obtaining multi-source real-time data and energy consumption data of each energy consumption node in the park, determining initial energy structure data from multiple dimensions of key features, and correcting in combination with the correlation between nodes. Furthermore, a dynamic emission factor matrix can be generated using a hierarchical adaptive calculation model, which can automatically adjust according to the real-time changes of the energy structure, thereby accurately capturing the dynamic changes of carbon emissions in the carbon emission simulation process and achieving high-precision dynamic analysis of carbon emissions in the park. The introduction of the attention mechanism can dynamically analyze and explain the carbon emission results of each energy consumption node, which can deeply mine the key information and dynamic change rules behind the carbon emission data. Not only can it help to more clearly understand the dynamic changes of carbon emissions in the park, but also can provide in-depth and practically meaningful analysis results for formulating targeted carbon emission reduction strategies.

[0061] For the problem of dynamic change of energy structure in the park, the application captures the dynamic change information of the energy structure in time by acquiring multi-source real-time data and extracting multi-dimensional key features. At the same time, a hierarchical adaptive calculation model is used to generate a dynamic emission factor matrix, which can automatically adjust the calculation of the emission factor according to the real-time change of the energy structure, so as to accurately reflect the carbon emission under the dynamic change of the energy structure, and solve the problem of large deviation of the analysis result caused by the fixed emission factor in the traditional method. The previous carbon emission analysis may only determine the energy structure according to a single data source or limited data dimensions, resulting in insufficient and inaccurate data. The application acquires multi-source real-time data of each energy consumption node in the park, and extracts multi-dimensional key features therefrom, fully integrates data of different types and different sources, and makes the determination of the energy structure data more comprehensive and accurate. In addition, the correlation between the energy consumption nodes is considered to correct the initial energy structure data, further improving the accuracy and reliability of the data, providing a solid data foundation for subsequent carbon emission analysis, and solving the problem of deviation of the analysis result caused by insufficient and inaccurate data integration in the past. At the same time, the previous carbon emission analysis result is often general, lacking in-depth analysis and explanation of the carbon emission of each energy consumption node, and it is difficult to effectively guide the carbon emission reduction work of the park. The application dynamically analyzes and explains the carbon emission result of each node through the attention mechanism, deeply mines the information behind the carbon emission data, and clearly defines the position and role of each node in the park carbon emission, as well as its dynamic change trend and influencing factors, further providing strong support for the park managers to formulate specific and effective carbon emission reduction strategies, and solving the problem of lack of depth and targeted guidance of the previous carbon emission analysis result. In summary, the application effectively solves the problems of dynamic deficiency and precision deviation caused by fixed emission factor in the traditional park carbon emission analysis by combining multi-source real-time data, dynamic emission factor matrix and attention mechanism, and realizes high-precision dynamic analysis of the park carbon emission. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 A flowchart of a carbon emission analysis method in an embodiment of the application is shown.

[0063] Figure 2 A structural diagram of a carbon emission analysis system in an embodiment of the application is shown.

[0064] Figure 3 A structural diagram of an electronic device in an embodiment of the application is shown. DETAILED DESCRIPTION

[0065] In order to make the above objectives, characteristics and advantages of the present application more apparent, concrete embodiments of the present application will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms, and should not be interpreted as being limited to the embodiments set forth herein, but rather, these embodiments are provided in order to more thoroughly and completely understand the present application. It should be understood that the drawings and embodiments of the present application are for illustrative purposes only, and are not intended to limit the scope of protection of the present application.

[0066] It should be understood that each of the steps recited in the method embodiments of the present application can be performed in different orders, and / or in parallel. In addition, the method embodiments can include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.

[0067] The term "comprising" and variations thereof as used herein are open-ended, that is "including, but not limited to"; the term "based on" is "based, at least in part, on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optional" means "optional in at least some embodiments". Related definitions are given throughout the description. It should be noted that the concepts mentioned in the present application using "first", "second", etc. are only used to distinguish different devices, modules or units, and are not intended to limit the order or interdependence of the functions performed by these devices, modules or units.

[0068] It should be noted that the modification of "one" or "multiple" mentioned in the present application is illustrative rather than limiting, and those skilled in the art should understand that, unless otherwise explicitly indicated in the context, it should be understood as "one or more".

[0069] The names of the messages or information exchanged between the devices in the embodiments of the present application are only for illustrative purposes, and are not intended to limit the scope of the messages or information.

[0070] In combination Figure 1 As shown, the carbon emission analysis method provided by the embodiments of the present application comprises:

[0071] Obtaining multi-source real-time data and energy consumption data of each energy consumption node in the park within a current time period.

[0072] Specifically, when acquiring multi-source real-time data and energy consumption data of each energy consumption node in the park within the current time period, first, the energy consumption nodes in the park are determined, such as production workshops, office buildings, charging piles, boiler rooms, etc. Through the deployment of Internet of Things sensors such as power sensors, gas flow meters, water temperature sensors, etc. at each node, real-time collection of equipment operating parameters such as voltage, current, flow, pressure, environmental data such as temperature, humidity, illumination is performed, and the park energy management system, equipment operation and maintenance log system, etc. are connected to acquire the energy consumption raw data of each node such as hourly power consumption, daily natural gas consumption, monthly coal consumption, etc. and set the current time period such as the past 24 hours, the first week of the month. The collected data is time-stamped and aligned, outliers such as jump data caused by sensor failure are removed, missing values are completed by interpolation with adjacent period means, and finally structured multi-source real-time data sets and energy consumption data sets are formed.

[0073] According to the carbon emission analysis requirements, multi-dimensional key features of each energy consumption node are extracted from the multi-source real-time data of each energy consumption node.

[0074] Specifically, according to the carbon emission analysis requirements, such as identifying key emission nodes, analyzing the impact of energy types on emissions, and evaluating periodical emission differences, when extracting multi-dimensional key features from the multi-source real-time data of each energy consumption node, first, the feature dimensions are determined based on the requirements. For example, if the requirement is to analyze the impact of device operating status on emissions, the dimensions can include device operating parameters such as motor power peak, boiler operating time, load characteristics such as production load rate, power load fluctuation coefficient; if the requirement is environmental factor influence analysis, the dimensions can include environmental parameters such as outdoor temperature mean, humidity change rate; then through statistical methods such as calculating the mean, variance, maximum value of each hour data to extract basic features, through time series analysis such as calculating the change amount, trend slope of adjacent period data to extract dynamic features, and then combining feature selection algorithms such as mutual information-based screening, variance threshold method to remove redundant features irrelevant to the analysis requirements, such as removing illumination data which has no significant impact on boiler house emission analysis, and finally obtaining the multi-dimensional key feature set of each node.

[0075] According to the multi-dimensional key features of each energy consumption node, the initial energy structure data of the energy consumption node is determined.

[0076] Specifically, when determining the initial energy structure data based on the multi-dimensional key features of each energy consumption node, a mapping relationship between key features and energy structure is first established. For example, equipment type features, such as an electric forklift, correspond to its energy consumption being electricity. Operating parameter features, such as the gas flow sensor data of a gas boiler, are directly related to natural gas consumption. For composite nodes, such as workshops that use both electricity and natural gas, the consumption ratio of various energy types is calculated using energy-related parameters in the key features, such as the cumulative value of the electricity sensor and the cumulative value of the gas flow meter. This is then calibrated by combining the energy structure template of similar nodes in history with the average electricity / gas consumption ratio of similar workshops. If the features of a node show that its production load has increased by 20% compared to the historical template, the initial consumption of various energy types is adjusted proportionally. Finally, initial energy structure data containing the energy types consumed by each node, such as electricity, natural gas, coal, and biomass energy, and their corresponding consumption amounts are formed.

[0077] Based on the correlation between the energy consumption nodes and other energy consumption nodes in the park, the initial energy structure data is corrected to obtain the final energy structure data.

[0078] Specifically, when obtaining the final energy structure data by correlating and correcting the initial energy structure data based on the correlation between energy consumption nodes and other nodes in the park, a node correlation graph is first constructed using the park's energy network topology and equipment connection table to clarify the correlation types, such as power supply correlation (the substation supplies power to workshops A and B) and energy sharing correlation (multiple offices share a central air conditioning system). Then, the initial data is verified based on the correlation rules. For example, the initial data of the total power supply of the substation should be equal to the sum of the initial power consumption data of each downstream node, such as workshops A and B. If there is a discrepancy, such as the substation recording a power supply of 1000 kWh, while the total power consumption of downstream nodes is 900 kWh, then according to the correlation strength, such as workshop A's historical power consumption ratio being 60% and workshop B's being 40%, a correction amount of 100 kWh is allocated to make up the difference, workshop A is supplemented with 60 kWh, and workshop B is supplemented with 40 kWh. For correlations involving energy conversion, such as the boiler room converting coal into steam to supply the workshops, the matching of steam consumption and coal consumption is verified based on the conversion efficiency, such as the coal-to-steam conversion rate being 80%, to correct unreasonable initial data and finally obtain the final energy structure data that conforms to the correlation logic.

[0079] The dynamic emission factor matrix of the energy consumption node is obtained by using a hierarchical adaptive calculation model based on the final energy structure data.

[0080] Specifically, when obtaining the dynamic emission factor matrix of energy consumption nodes based on the final energy structure data through the hierarchical adaptive calculation model, the hierarchical model consists of three layers: the bottom layer is the basic emission factor layer, which stores the baseline emission factors of various energy sources, such as the emission factor per kWh of electricity and the emission factor per cubic meter of natural gas published by the state; the middle layer is the node type correction layer, which sets correction coefficients according to node attributes, such as industrial nodes and residential nodes. For example, the emission factor of industrial electricity is 5% higher than that of residential electricity due to high grid load; the top layer is the time dynamic layer, which adjusts the emission factors based on external factors of the current time period, such as the real-time energy composition of the grid and seasonal factors. For example, the electricity emission factor is 10% lower than that of winter due to high photovoltaic power generation in summer; the adaptive mechanism monitors the rate of change of the proportion of energy types in the final energy structure data. For example, if the proportion of natural gas in a certain node suddenly increases by 30%, it automatically increases the weight of the correction coefficient of the corresponding node type in the middle layer and triggers the time dynamic layer of the top layer to recalculate. The final output is a dynamic matrix containing the emission factors of each node, each energy type, and each time period, such as the electricity emission factor of node A in time period t1 and the natural gas emission factor in time period t2.

[0081] By combining the dynamic emission factor matrix, the multi-source real-time data, and the energy consumption data, a hybrid model is used to simulate carbon emissions at the energy consumption node, and the carbon emission results of the energy consumption node are obtained.

[0082] Specifically, when simulating carbon emissions at energy consumption nodes using a hybrid model by combining dynamic emission factor matrices, multi-source real-time data, and energy consumption data, the hybrid model integrates a physical model and a data-driven model: the physical model part calculates the basic emissions based on the formula: carbon emissions from a certain energy source = energy consumption × corresponding dynamic emission factor. For example, node B consumes 500 m³ of natural gas. 3 The corresponding dynamic emission factor is 2.0 kg CO2 / m³. 3 The baseline emission is 1000 kg CO2. The data-driven model, such as the LSTM neural network, uses multi-source real-time data, such as equipment operating temperature and load rate, to correct the baseline value. For example, when the equipment load rate exceeds 80%, the emission efficiency decreases, and the model outputs a correction coefficient of 1.05 to adjust the baseline value to 1050 kg CO2. The hybrid model dynamically adjusts the weights of the two parts through historical simulation errors. For example, the weight of the physical model error is increased to 0.7 when the error is small. Finally, it outputs the hourly carbon emission data and cumulative carbon emission results of each node in the current time period.

[0083] Based on the carbon emission results of each energy consumption node, the carbon emissions are dynamically analyzed and interpreted through an attention mechanism to obtain the carbon emission analysis results of the park in the current time period.

[0084] Specifically, based on the carbon emission results of each energy consumption node, the carbon emission analysis results for the current time period in the park are obtained through dynamic analysis and interpretation of carbon emissions using an attention mechanism. First, the carbon emission results of each node are input into the attention model to calculate the attention weight of each node. For example, if the carbon emission of node C accounts for 30% of the total emissions in the park and the Pearson correlation coefficient with the total emissions is 0.9, then the weight is 0.3. The top 20% of key nodes with the highest weights are selected. Then, the emission driving factors of the key nodes are analyzed. The attention mechanism is used to locate specific characteristics, such as 80% of the carbon emissions of node C come from coal consumption, and emissions surge during the peak load period from 8 to 10 am every day. The analysis results are dynamically generated, including the total carbon emissions of the park, a pie chart of the emission proportion of each node, a list of key emission sources including node names and contributions, emission change curves over time, and explanations of the main driving factors, such as coal consumption being the primary source of emissions in the park, accounting for 65%, etc., forming a complete carbon emission analysis report.

[0085] The carbon emission analysis method in this embodiment acquires multi-source real-time data and energy consumption data from various energy consumption nodes within the park. It determines the initial energy structure data based on key features from multiple dimensions and corrects it by incorporating the correlations between nodes, enabling the energy structure data to dynamically reflect the actual state of energy consumption in the park. Then, a hierarchical adaptive computation model is used to generate a dynamic emission factor matrix. This matrix automatically adjusts according to real-time changes in the energy structure, accurately capturing dynamic changes in carbon emissions during the carbon emission simulation process and achieving high-precision dynamic analysis of carbon emissions in the park. Introducing an attention mechanism for dynamic analysis and interpretation of carbon emission results from each energy consumption node allows for in-depth mining of key information and dynamic change patterns behind the carbon emission data. This not only helps to more clearly understand the dynamic changes in carbon emissions in the park but also provides in-depth and practically guiding analytical results for developing targeted carbon reduction strategies.

[0086] To address the issue of dynamic changes in the energy structure of industrial parks, this embodiment acquires multi-source real-time data and extracts multi-dimensional key features to promptly capture information on these dynamic changes. Simultaneously, a hierarchical adaptive calculation model is used to generate a dynamic emission factor matrix. This model automatically adjusts the calculation of emission factors based on real-time changes in the energy structure, accurately reflecting carbon emissions under dynamic energy structure changes. This solves the problem of significant bias in analysis results caused by fixed emission factors in traditional methods. Previous carbon emission analyses may have relied on a single data source or limited data dimensions to determine the energy structure, resulting in incomplete and inaccurate data. This invention acquires multi-source real-time data from various energy consumption nodes within the park and extracts multi-dimensional key features, fully integrating data of different types and sources, making the determination of energy structure data more comprehensive and accurate. Furthermore, considering the correlation between energy consumption nodes, the initial energy structure data is corrected, further improving the accuracy and reliability of the data. This provides a solid data foundation for subsequent carbon emission analysis, solving the problem of biased analysis results caused by insufficient data integration and low accuracy in the past. Moreover, previous carbon emission analyses were often too general, lacking in-depth analysis and interpretation of carbon emissions at each energy consumption node, making it difficult to effectively guide the park's carbon reduction efforts. This embodiment utilizes an attention mechanism to dynamically analyze and interpret the carbon emission results of each node, deeply mining the information behind the carbon emission data. It clarifies the position and role of each node in the park's carbon emissions, as well as its dynamic trends and influencing factors. This provides strong support for park managers to formulate specific and effective carbon reduction strategies, addressing the problem of previous carbon emission analysis results lacking depth and targeted guidance. In summary, this embodiment, by combining multi-source real-time data, a dynamic emission factor matrix, and an attention mechanism, effectively solves the problems of insufficient dynamism and accuracy deviation caused by fixed emission factors in traditional park carbon emission analysis, achieving high-precision dynamic analysis of park carbon emissions.

[0087] Optionally, the step of extracting multi-dimensional key features of each energy consumption node from the multi-source real-time data of each energy consumption node according to carbon emission analysis requirements includes:

[0088] Based on the management scenario of the park, the types of carbon emission analysis needs are determined, including short-term emission reduction monitoring needs, long-term policy formulation needs, and equipment energy efficiency optimization needs.

[0089] For each type of carbon emission analysis requirement, match the target data type corresponding to the multi-source real-time data of each energy consumption node;

[0090] Extract the original features of the corresponding dimensions of the target data type; specifically, for the short-term emission reduction monitoring needs, extract the energy consumption peak and valley fluctuation features of the time dimension and the real-time load features of the equipment dimension; for the long-term policy formulation needs, extract the functional area energy consumption difference features of the spatial dimension and the clean energy proportion change features of the policy dimension; for the equipment energy efficiency optimization needs, extract the aging and loss features of the equipment dimension and the capacity-energy consumption correlation features of the operating condition dimension.

[0091] Data preprocessing is performed on the original features of each extracted dimension to obtain multi-dimensional key features for each energy consumption node.

[0092] Specifically, when determining the type of carbon emission analysis needs based on the park's management scenario, the core management objectives of the park should be clarified first. For example, if the management scenario is daily operation monitoring, such as the dispatch center monitoring the implementation of emission reduction measures in real time, then the corresponding short-term emission reduction monitoring needs are needed, focusing on emission fluctuations and immediate emission reduction effects in the past 1-7 days; if the management scenario is the formulation of annual / five-year plans, such as the environmental protection department formulating the park's carbon peaking path, then the corresponding long-term policy formulation needs are needed, focusing on emission trends and structural optimization space over a year or more; if the management scenario is equipment operation and maintenance upgrades, such as workshops carrying out energy efficiency transformations for high-energy-consuming equipment, then the corresponding equipment energy efficiency optimization needs are needed, focusing on the matching of energy consumption and output of a single unit / type of equipment.

[0093] When matching target data types to the types of carbon emission analysis needs, data sources are selected based on the time scale of the need and the objects of interest. Short-term emission reduction monitoring needs require matching high-frequency real-time data, such as equipment current / power data every 5 minutes, hourly energy consumption statistics, and equipment status data, such as operating / shutdown status and load switching records. Long-term policy-making needs require matching periodic cumulative data, such as monthly total energy consumption of each functional area, quarterly clean energy photovoltaic / wind power generation, and policy-related data, such as new energy equipment installed capacity and carbon quota execution records. Equipment energy efficiency optimization needs require matching equipment lifecycle data, such as operating time, maintenance records, factory parameters, and production data, such as unit time capacity and product qualification rate. Specifically, when extracting raw features from target data types, it is necessary to focus on key dimensions in conjunction with the type of demand: For short-term emission reduction monitoring needs, the energy consumption peak-valley fluctuation characteristics in the time dimension can be extracted by calculating the deviation rate between hourly energy consumption and the daily average energy consumption, the duration of peak-valley periods such as 9:00-11:00 as the peak and 0:00-6:00 as the valley, and the energy consumption difference; the real-time load characteristics in the equipment dimension can be extracted by the current load rate, actual power / rated power, and load change rate, the load difference between adjacent 10-minute intervals / baseline load; for long-term policy formulation needs, the functional area energy consumption difference characteristics in the spatial dimension can be extracted by calculating the energy consumption difference of each functional area such as production area / office area / living area. Energy consumption per unit area of ​​the active zone, energy density, and standard deviation of total energy consumption / functional area can be extracted. The characteristics of changes in the proportion of clean energy in the policy dimension can be extracted by the ratio of monthly clean energy consumption to total energy consumption and the month-on-month growth rate of this ratio. For equipment energy efficiency optimization needs, the aging and wear characteristics of equipment can be extracted by the linear regression coefficient of running time and energy consumption growth, such as the proportion of energy consumption increase per 1000 hours of operation, and the Pearson correlation coefficient of monthly failures and energy consumption. The capacity-energy consumption correlation characteristics in the operating condition dimension can be extracted by the energy consumption per unit product, total energy consumption / total output, and the synchronization rate of daily capacity standard deviation and energy consumption fluctuation.

[0094] When preprocessing the raw features of each extracted dimension, the following operations are performed sequentially: cleaning, standardization, and dimensionality reduction. In the data cleaning stage, the 3σ rule is used to remove outliers, such as energy consumption fluctuations exceeding the mean ± 3 standard deviations, and linear interpolation is used to fill in missing values, such as 1-2 missing data points caused by brief equipment offline. In the standardization stage, min-max normalization is used to convert features of different dimensions, such as energy consumption values ​​in kWh and load rates as percentages, to the [0,1] interval to eliminate the influence of dimensional differences. In the dimensionality reduction stage, principal component analysis (PCA) is used to retain principal components with a cumulative contribution rate ≥ 90% and to remove redundant features, such as highly correlated load rates and power fluctuations.

[0095] In this embodiment of the invention, based on the park management scenario, three types of needs are divided: short-term emission reduction monitoring, long-term policy formulation, and equipment energy efficiency optimization. This clarifies the target boundaries of feature extraction from the outset, avoiding the information redundancy problem caused by indiscriminate feature collection in traditional methods. This ensures that subsequent analysis always revolves around actual business needs, laying the foundation for targeted features. Secondly, corresponding target data types are matched to different need types. For example, high-frequency real-time data is matched for short-term needs, and periodic cumulative data is matched for long-term needs. Through precise alignment of data and needs, interference from irrelevant data is reduced, lowering the computational cost of data processing and improving feature extraction. The efficiency is improved; secondly, based on demand-oriented extraction of original features, such as short-term demand focusing on peak and valley fluctuations and equipment load in the time dimension, and long-term demand focusing on differences in spatial functional areas and changes in the proportion of clean energy, the extracted features are directly related to the core analysis objectives, strengthening the correlation between features and carbon emission influencing factors, and providing high-information-density input variables for subsequent models; finally, through preprocessing operations such as cleaning, standardization, and dimensionality reduction, interference caused by outliers, dimensional differences, and feature redundancy is eliminated, making the features more accurately reflect the essential attributes of energy consumption nodes, and avoiding the disturbance of noisy data to subsequent energy structure analysis and carbon emission simulation.

[0096] Optionally, determining the initial energy structure data of each energy consumption node based on its multi-dimensional key features includes:

[0097] The multi-dimensional key features are analyzed and classified using an energy structure identification model to determine multiple energy structure types and scenario-related features of the energy consumption nodes.

[0098] Based on the energy structure type, determine the energy type proportion, energy conversion efficiency, and energy consumption fluctuation data of the energy consumption node;

[0099] Based on the aforementioned scenario-related characteristics, scenario-based impact data that are strongly correlated with energy consumption are determined;

[0100] The energy type proportion, energy conversion efficiency, energy consumption fluctuation data, and scenario-based impact data of the energy consumption nodes are used as the initial energy structure data.

[0101] The step of revising the initial energy structure data based on the correlation between the energy consumption nodes and other energy consumption nodes in the park to obtain the final energy structure data includes:

[0102] Based on the correlation between each energy consumption node in the park, the impact coefficient of the energy consumption node is quantified;

[0103] The energy type proportion, energy conversion efficiency, energy consumption fluctuation data, and scenario-based impact data are corrected based on the impact coefficient to obtain the final energy structure data.

[0104] Optionally, the hierarchical adaptive calculation model includes a basic factor matching layer, an efficiency correction layer, a scenario dynamic adjustment layer, and a deviation calibration layer;

[0105] The process of obtaining the dynamic emission factor matrix of the energy consumption nodes through a hierarchical adaptive calculation model based on the final energy structure data includes:

[0106] Through the basic factor matching layer, an initial emission factor vector is obtained based on the proportion of the energy type and in combination with a preset industry benchmark emission factor library;

[0107] The efficiency correction layer corrects the initial emission factor vector based on the energy conversion efficiency and the energy consumption fluctuation data to obtain the efficiency-corrected factor vector.

[0108] Through the scene dynamic adjustment layer, the efficiency-corrected factor vector is dynamically adjusted in real time according to the scene-based impact data to obtain the scene-adjusted factor matrix.

[0109] By using the deviation calibration layer and combining historical carbon emission assessment data, the deviation value between the scenario-adjusted factor matrix and the historical actual monitoring data is determined. Based on the deviation value, the scenario-adjusted factor matrix is ​​self-calibrated and optimized to obtain the dynamic emission factor matrix.

[0110] Specifically, firstly, an energy structure identification model is constructed. This model can adopt a classification model based on random forest, taking multi-dimensional key features (such as energy type-related features, equipment operation features, environmental features, etc.) as input, and training the model through training samples (including node feature data of known energy structure types) to enable the model to identify energy structure types. Then, the multi-dimensional key features of the target node are input into the trained model, and the energy structure type of the node is output, such as electricity-natural gas hybrid, coal-dominated, renewable energy-assisted, etc. At the same time, through the feature importance analysis module of the model, scenario features strongly related to the energy structure (such as production shift features, seasonal features, park activity arrangement features, etc.) are selected as scenario association features. For example, the energy structure of a production workshop is strongly related to the three-shift production scenario, and the energy structure of an office building is strongly related to the weekday / holiday scenario. Regarding the proportion of energy types, the consumption data of various energy sources in the multi-dimensional key characteristics of the node (such as cumulative values ​​of power sensors, gas flow meter data, etc.) are combined to calculate the proportion of each energy source in the total energy consumption. For example, in a hybrid electricity-natural gas node, if the electricity consumption is 600 kWh and the natural gas consumption is approximately 400 kWh (converted by calorific value), then the electricity accounts for 60% and the natural gas accounts for 40%. Regarding energy conversion efficiency, the actual conversion efficiency is calculated by referring to the equipment parameters (such as the rated conversion efficiency of the boiler) and operating characteristics (such as the load rate) corresponding to the energy structure type, using the formula: actual conversion efficiency = rated efficiency × load rate correction coefficient. For example, if the rated efficiency of the gas boiler is 90%, the current load rate is 80%, and the correction coefficient is 0.95, then the actual conversion efficiency is 90% × 0.95 = 85.5%. Regarding energy consumption fluctuation data, the standard deviation and peak-valley difference of energy consumption within a certain period of time (such as 1 day) are calculated to quantify the degree of fluctuation. For example, if the hourly energy consumption of a certain node is 100, 120, 90, and 110 kWh, the standard deviation is 12.9, and the peak-valley difference is 30 kWh.

[0111] When determining scenario-based impact data strongly correlated with energy consumption based on scenario-related characteristics, the mapping relationship between scenario-related characteristics and energy consumption data is first established. The correlation between scenario characteristics and energy consumption is calculated using the Pearson correlation coefficient, and scenario characteristics with an absolute correlation coefficient value ≥ 0.6 are selected. Then, for these scenario characteristics, their impact on energy consumption is quantified. For example, in a three-shift production scenario, the energy consumption of the middle shift (16:00-24:00) is on average 20% higher than that of the early shift (8:00-16:00), and this 20% increase is the scenario-based impact data. In a low-temperature winter scenario, when the outdoor temperature is below 5℃, the heating energy consumption increases by 50kWh / day compared to normal temperature, and this 50kWh / day increase also belongs to the scenario-based impact data. Finally, the data obtained from the above calculations are integrated to form a structured data set. For example, the initial energy structure data of a certain node can be represented as: energy type ratio (electricity 70%, natural gas 30%), energy conversion efficiency (electric equipment 92%, natural gas equipment 88%), energy consumption fluctuation data (daily standard deviation 15kWh, peak-valley difference 40kWh), and scenario-based impact data (weekday energy consumption is 30% higher than weekend energy consumption, and cooling energy consumption increases by 25kWh / day in hot weather). The relationship types between nodes are clarified by using the park's energy network topology map (e.g., power supply relationship: node A supplies power to nodes B and C; heating relationship: node D provides heating to node E). Then, the relationship strength is calculated based on historical data. For example, for power supply relationships, the ratio of the change in power supply of node A to the change in energy consumption of node B is calculated to obtain the dependence coefficient of node B on node A. If the energy consumption of node B increases by an average of 80 kWh when the power supply of node A increases by 100 kWh, then the dependence coefficient is 0.8. The dependence coefficient and the relationship type weight (e.g., power supply relationship weight 0.6, heating relationship weight 0.4) are combined and quantified by the formula influence coefficient = dependence coefficient × relationship type weight. For example, the influence coefficient of node B is 0.8 × 0.6 = 0.48. Regarding the proportion of energy types, if node B is affected by upstream power supply node A (affect coefficient 0.48), and node A's power supply proportion decreases by 5% due to grid adjustments, then node B's power proportion is simultaneously corrected to the original proportion × (1 - 5% × 0.48). Regarding energy conversion efficiency, if node E is affected by heating node D (affect coefficient 0.3), and node D's heat conversion efficiency increases by 3%, then node E's heating-related conversion efficiency is corrected to the original efficiency × (1 + 3% × 0.3). Regarding energy consumption fluctuation data, if node C is strongly correlated with node A (affect coefficient 0.5), and node A's energy consumption fluctuation standard deviation increases by 10 kWh, then node C's fluctuation standard deviation is corrected to the original standard deviation + 10 × 0.5. Regarding scenario-based impact data, if node E is affected by node D's maintenance scenario (affect coefficient 0.4), and node D's heating scenario impact data increases by 20 kWh during maintenance, then node E's corresponding scenario impact data is corrected to the original data + 20 × 0.4, ultimately forming the final energy structure data that conforms to the correlation logic.

[0112] In this embodiment of the invention, the analysis and classification of multi-dimensional key features using an energy structure identification model overcomes the limitations of traditional methods that rely on human experience to judge energy structure. The model automatically identifies energy structure types and extracts scenario-related features, making the initial energy structure data more closely aligned with the actual energy consumption characteristics of nodes, avoiding subjective bias. Simultaneously, the introduction of scenario-based impact data compensates for the neglect of external scenario factors such as environment and production in simple energy consumption data, enhancing the comprehensiveness of the data. Secondly, by quantifying the impact coefficient based on node correlation and correcting the initial data, the problem of single-node data being difficult to reflect the power supply and heating in the park's energy network when calculated independently is solved. This process addresses issues of interconnectedness and interaction, such as correcting energy proportions and conversion efficiency by adjusting the influence coefficient of upstream nodes on downstream nodes. This ensures that the final energy structure data accurately reflects the interactions between nodes, eliminating data silos caused by ignoring connections. Ultimately, the final energy structure data obtained through this process retains the details of individual node energy consumption characteristics while conforming to the operational logic of the park's overall energy network. This provides a high-quality data foundation for subsequent calculations of dynamic emission factor matrices and carbon emission simulations, improving the reliability and practicality of carbon emission analysis results from the source and ensuring that the analysis conclusions truly reflect the intrinsic relationship between energy consumption and carbon emissions in the park.

[0113] Optionally, the hierarchical adaptive computing model further includes a real-time feedback layer;

[0114] The step of obtaining the dynamic emission factor matrix of the energy consumption node through a hierarchical adaptive calculation model based on the final energy structure data also includes:

[0115] The real-time feedback layer obtains real-time carbon emission concentration data collected by the real-time carbon emission monitoring equipment set up in the park.

[0116] Based on the deviation between the real-time carbon emission concentration data and the estimated carbon emission concentration data generated by the dynamic emission factor matrix, a feedback compensation factor is constructed.

[0117] The dynamic emission factor matrix is ​​iteratively optimized online using the feedback compensation factor to update the emission factor values ​​in the dynamic emission factor matrix.

[0118] Specifically, high-precision carbon emission monitoring equipment (such as non-dispersive infrared gas analyzers and laser spectrometers) are deployed around key energy consumption nodes in the park (such as boiler room exhaust outlets, ventilation points of high-energy-consuming workshops, and the park's main exhaust outlet). The data collection frequency is set to once every 10 minutes to collect data such as CO2 concentration, temperature, and air pressure in real time. The raw data is sent to the data receiving end of the real-time feedback layer through wireless transmission modules (such as LoRa and 5G). The receiving end performs preliminary verification on the data (such as removing outliers that exceed the equipment's range and converting the concentration data to values ​​under standard conditions through air pressure calibration), and finally forms a real-time carbon emission concentration dataset with timestamp alignment.

[0119] Based on the dynamic emission factor matrix and final energy structure data, estimated carbon emission concentration can be generated using the formula: Estimated carbon emission concentration = (Σ energy consumption × corresponding emission factor) / park space volume × diffusion coefficient (the diffusion coefficient is adjusted in real time according to meteorological data). Then, the deviation value is calculated using the relative deviation formula: Deviation = (Real-time concentration - Estimated concentration) / Estimated concentration × 100%. If the real-time concentration is 800 ppm and the estimated concentration is 750 ppm, the deviation is 6.67%. Subsequently, a compensation factor is constructed based on the magnitude and trend of the deviation, and the deviation is set accordingly. The threshold (e.g., ±5%) is set as follows: when the absolute value of the deviation is ≤5%, the compensation factor is 1 (no adjustment required); when the deviation is >5% (real-time deviation is higher than the estimate), the compensation factor = 1 + deviation × 0.3 (amplifying the emission factor); when the deviation is <-5% (real-time deviation is lower than the estimate), the compensation factor = 1 + deviation × 0.3 (reducing the emission factor). For example, the compensation factor corresponding to the above 6.67% deviation is 1 + 6.67% × 0.3 ≈ 1.02. At the same time, if the deviation direction is consistent for three consecutive times, the adjustment coefficient of the compensation factor will be increased from 0.3 to 0.5 to enhance the correction strength. The energy consumption node and energy type corresponding to the deviation are then located (e.g., the deviation mainly originates from coal consumption in the boiler room). The emission factor corresponding to the energy type of the node is extracted from the dynamic emission factor matrix (e.g., the current coal emission factor is 2.6 kg CO2 / kg). The original emission factor is then multiplied by the feedback compensation factor to obtain the updated value (e.g., 2.6 × 1.02 ≈ 2.652 kg CO2 / kg). At the same time, an iteration cycle is set (e.g., updated once per hour). After each update, the change in emission factor and the corresponding deviation data are recorded. The compensation effect is statistically analyzed through a sliding window (e.g., the past 24 hours). If the deviation of a node still exceeds the threshold after multiple compensations, the baseline value of the underlying basic emission factor layer is re-verified to ensure that the dynamic emission factor matrix continuously matches the actual emission situation. Finally, the optimized dynamic emission factor matrix is ​​output.

[0120] In this embodiment of the invention, real carbon emission concentration data collected by high-precision monitoring equipment within the park is introduced through a real-time feedback layer. This breaks the limitation of relying solely on final energy structure data and theoretical models to calculate predicted concentrations, providing an objective and realistic benchmark for optimizing emission factors. This avoids the problem of discrepancies between predictions and reality caused by model assumptions (such as fixed diffusion coefficients and deviations in theoretical energy conversion efficiency values). Secondly, by calculating the relative deviation between real-time and predicted concentrations and constructing differentiated feedback compensation factors (such as dynamically adjusting compensation coefficients based on the magnitude of the deviation and enhancing correction strength by combining the continuous trend of the deviation), refined and targeted adjustments to emission factors are achieved. This avoids factor fluctuations caused by over-correction under small deviations and enables rapid response under significant deviations, ensuring that the correction direction and magnitude closely match actual emission changes. Furthermore, The online iterative optimization mechanism based on feedback compensation factors enables the dynamic emission factor matrix to be updated in real time according to the park's operating conditions (such as changes in equipment load, fluctuations in weather conditions, and adjustments in energy supply structure). This eliminates the lag of traditional static emission factors or fixed-period updated factors, ensuring that emission factors are always adapted to the current node's operating status and the overall park environment. Ultimately, the dynamic emission factor matrix optimized by real-time feedback has significantly improved data accuracy, directly providing more realistic core calculation parameters for subsequent carbon emission simulations. This reduces simulation errors in carbon emission results, making the park's carbon emission analysis results more accurately reflect actual emissions. It provides more reliable data support for real-time decision-making such as short-term emission reduction and control, and equipment operation and maintenance optimization. It also enhances the adaptability of the hierarchical adaptive calculation model to complex and dynamic park scenarios.

[0121] Optionally, the step of combining the dynamic emission factor matrix, the multi-source real-time data, and the energy consumption data to simulate carbon emissions at the energy consumption node using a hybrid model to obtain the carbon emission results for the energy consumption node includes:

[0122] The energy consumption data is broken down according to energy type to obtain the consumption amount of each energy type of the energy consumption node;

[0123] The basic carbon emissions of the energy consumption node are obtained by multiplying the individual consumption amounts element by element with the emission factors of the corresponding energy type in the dynamic emission factor matrix.

[0124] Based on the multi-source real-time data, a bidirectional long short-term memory network model is used for fitting to obtain the nonlinear correction amount of the energy consumption node;

[0125] The preliminary carbon emission results for the energy consumption node are obtained by weighted fusion of the baseline carbon emissions and the nonlinear correction.

[0126] The preliminary carbon emission results are subjected to outlier detection and removal, and the data is labeled in conjunction with the collection frequency of the energy consumption data to generate the carbon emission results including timestamps, energy type-specific carbon emission values, and total carbon emission values.

[0127] Specifically, the energy types included in the selected data (such as electricity, natural gas, coal, biomass energy, etc.) are classified and labeled using energy metering ledgers or sensor data (e.g., electricity data labeled "electricity-10kV", natural gas data labeled "gas-municipal"). This decomposes the total energy consumption data of each node into the consumption of each individual energy type. For example, if a workshop's monthly total energy consumption is 10,000 kWh equivalent, this is broken down into 6,000 kWh of electricity consumption and approximately 4,000 kWh of natural gas consumption (converted by calorific value), forming a breakdown of consumption for each energy type. The breakdown consumption is then multiplied element-by-element by the emission factor corresponding to the energy type in the dynamic emission factor matrix. The real-time emission factor for each energy type is then matched from the dynamic emission factor matrix (e.g., electricity emission factor is 0.5 kg CO2 / kWh, natural gas is 2.0 kg CO2 / m³). 3 Then, for each energy type, calculate the consumption multiplied by the emission factor. For example, the electricity consumption is 6000 kWh, multiplied by 0.5 kg CO2 / kWh to get 3000 kg CO2, and the natural gas consumption is 2000 m³ / kWh. 3 Multiply by 2.0 kg CO2 / m 34000 kg CO2 is obtained. The calculation results of all energy types are summed to obtain the base carbon emission of this node (e.g., 7000 kg CO2). Based on multi-source real-time data, a bidirectional long short-term memory network model is used to fit and obtain the nonlinear correction amount. First, the multi-source real-time data (e.g., real-time equipment load rate, operating temperature, ambient humidity, production cycle, etc.) is standardized (converted to values ​​in the [0,1] interval). The processed data is divided into training samples according to the time series (including historical data and corresponding actual carbon emission deviations). The model is trained to learn the nonlinear correlation in the data (e.g., carbon emissions increase nonlinearly when the load rate exceeds 80%). The real-time data of the current node is input into the trained model, and the model output value is the nonlinear correction amount for the base carbon emission (e.g., when the base value is 7000 kg CO2, the model outputs a correction amount of +350 kg CO2, reflecting the additional emissions caused by high load). The baseline carbon emissions and the nonlinear correction are weighted and fused to obtain the preliminary carbon emission results. The fusion weights of the two are set (e.g., the baseline carbon emissions weight is 0.8 and the nonlinear correction weight is 0.2. The weights can be dynamically adjusted according to historical simulation errors. When the error is small, the weight of the baseline value is increased). The preliminary result is calculated by the formula: Preliminary result = Baseline value × Baseline weight + Correction amount × Correction weight. For example, the baseline value is 7000kgCO2 × 0.8 + 350kgCO2 × 0.2 = 5600 + 70 = 5670kgCO2, which gives the preliminary carbon emission result for this node.

[0128] Outlier detection and removal are performed on the preliminary carbon emission results. Labeling is completed based on the energy consumption data collection frequency. The interquartile range (IQR) method is used to detect outliers, calculating the upper quartile (Q3) and lower quartile (Q1) of the preliminary result sequence. Values ​​exceeding the range [Q1-1.5×IQR, Q3+1.5×IQR] are identified as outliers and removed (e.g., abnormally high values ​​due to momentary equipment failure). Timestamps (e.g., 2025-10-24 08:00:00, 08:15:00) are added to the retained valid data based on the energy consumption data collection frequency (e.g., once every 15 minutes). Simultaneously, the carbon emission values ​​for each energy type (e.g., 3000 kg of electricity, 4000 kg of natural gas) and the total carbon emission value (5670 kg) are summarized, ultimately generating a structured carbon emission result containing timestamps, individual values, and the total value.

[0129] In this embodiment of the invention, firstly, energy consumption data is split according to energy type and matched with dynamic emission factors to calculate basic carbon emissions. This breaks away from the traditional extensive model of multiplying total energy consumption by a fixed average factor—the emission factors of different energy sources (such as electricity and natural gas) differ significantly. Calculating each category after splitting the data accurately matches the actual emission contribution of each energy source, avoiding the basic error caused by the average calculation of total energy consumption masking the differences in individual components, thus laying a precise underlying data foundation for subsequent simulation results. Secondly, a bidirectional long short-term memory network (Bi-LSTM) is used to fit multi-source real-time data to obtain nonlinear correction values. This specifically addresses the nonlinear correlation between carbon emissions and factors such as equipment load and ambient temperature in real-world scenarios (e.g., emissions will nonlinearly surge when the load rate exceeds 80% due to decreased equipment efficiency). Compared to traditional linear correction models, Bi-LSTM can more accurately capture time series data. The complex dynamic relationships within the columns make the correction amounts more closely match actual working conditions, significantly reducing the deviation caused by the disconnect between linear assumptions and real-world scenarios. Furthermore, by dynamically weighting and integrating the baseline emissions and nonlinear corrections, rather than simply adding them together, the contribution ratio of the two can be adjusted in real time based on historical simulation errors (e.g., increasing the weight when the baseline error is small). This preserves the stability of the baseline calculations while leveraging the dynamic adaptability of the nonlinear corrections, avoiding imbalances caused by a single factor dominating the results. Finally, outlier detection and removal (e.g., abnormally high values ​​caused by equipment failure) and timestamp annotation not only eliminate the interference of noisy data on the results but also give the carbon emission results a time dimension and detailed breakdown (e.g., the carbon emission ratio of each energy type). This ensures that the results accurately reflect the actual emission levels at each node and support subsequent refined management needs such as periodic emission trend analysis and source tracing of key energy emissions. In summary, the entire process forms a precise transformation link from raw data to high-quality carbon emission results. Compared with traditional methods, the simulation error is significantly reduced, and the results are more in line with the actual energy consumption and emission patterns of the park. It can provide accurate data support for short-term emission reduction and control (such as real-time adjustment of the load of high-emission equipment) and provide detailed basis for long-term energy efficiency optimization (such as identifying high-emission energy types).

[0130] Optionally, the step of dynamically analyzing and interpreting carbon emissions based on the carbon emission results of each energy consumption node through an attention mechanism to obtain the carbon emission analysis results of the park in the current time period includes:

[0131] The carbon emission results of the energy consumption nodes are input into the attention mechanism model;

[0132] The attention mechanism model is used to assign weights to the carbon emission values ​​of the energy type and the total carbon emission value in the carbon emission results, thereby determining the contribution of each energy consumer in the energy consumption node to carbon emissions.

[0133] Based on the contribution level, a dynamic analysis is performed on the energy-consuming individuals to obtain the carbon emission sources and key driving factors of the park;

[0134] By combining the carbon emission sources and key driving factors of the park, the carbon emission analysis results of the park during the current time period are generated.

[0135] Specifically, the carbon emission results of energy consumption nodes are input into the attention mechanism model. The structured data in the carbon emission results are organized, including the carbon emission values ​​of energy type at each time stamp (such as the emissions corresponding to electricity and natural gas), the total carbon emission value, and the associated node attributes (such as equipment number and functional area type). This data is converted into a vector form that the model can recognize (such as concatenating the hourly sub-values ​​with the total emission value to form a feature vector), and input into the model in time series order to provide structured input for subsequent weight calculation. The attention mechanism model assigns weights to the component values ​​and total emissions in the carbon emission results to determine the contribution of each energy-consuming entity. The model calculates the correlation strength between the emission data of each energy-consuming entity (such as the electricity consumption of a gas boiler or a production line) and the total carbon emissions. For example, if the cosine similarity between the emission value of natural gas and the total emission value is 0.92 in a certain period, which is significantly higher than that of other energy types, the model assigns it an attention weight of 0.45. Then, combining the weight with the actual emissions of the entity, the contribution percentage of the entity to carbon emissions is calculated as: contribution = weight × component emission value / total emission value × 100% (e.g., 38%), thus clarifying the degree of emission impact of different entities. Dynamic analysis of individual energy consumers based on their contribution levels reveals the sources and key drivers of carbon emissions in the industrial park. Individual energy consumers are ranked in descending order of contribution, and the top 30% are identified as primary sources of carbon emissions (e.g., coal consumption in boiler rooms and electricity consumption in refrigeration plants). For these primary sources, driving factors are analyzed in conjunction with their operational data (e.g., daily operating hours and load rates of coal-consuming individuals). For example, if a significant increase in boiler room contribution is observed between 8:00 and 12:00 on weekdays, with a corresponding load rate of 90%, then high-load operation is identified as the key driver for this source. Simultaneously, the temporal changes in contribution levels are tracked to identify the time points of sudden increases / decreases in emissions and the corresponding individuals, supplementing the driving factors (e.g., equipment maintenance causing a sudden drop in emissions from a production line). Finally, combining the carbon emission sources and key driving factors of the park, the carbon emission analysis results for the current period are generated, integrating core indicators, including the total carbon emissions of the park, the contribution ratio of each major source (e.g., coal accounts for 45%, electricity accounts for 30%), peak emission periods and values; interpretive content is written to clarify the specific impact of key driving factors (e.g., coal consumption is the primary source, mainly due to the boiler room operating at full capacity during peak production periods, contributing 42% of the emissions); and visualization charts (e.g., contribution pie charts, emission time-series curves, and driving factor annotations) are attached, ultimately forming an analysis result that can reflect the overall emission level of the park and accurately locate key emission sources and their causes, providing clear guidance for emission reduction decisions.

[0136] In this embodiment of the invention, an attention mechanism model is used to weight the component values ​​and total emissions in the carbon emission results. This breaks away from the traditional, crude approach of focusing solely on total emissions while ignoring individual differences. By quantifying the correlation strength between each type of energy consumption (e.g., a single device, a single energy type) and total carbon emissions and calculating its contribution, high-contribution emission entities (e.g., natural gas consumption devices contributing 38%) can be accurately identified. This avoids masking key emission sources due to averaging, allowing the park to clearly understand which entities are the core targets for emission reduction and providing clear targets for subsequent regulation. Secondly, dynamic analysis based on contribution overcomes the limitations of static analysis. By tracking the temporal changes in contribution (e.g., a sudden increase in boiler room contribution between 8-12 am on weekdays) and combining operational data to mine key driving factors (e.g., high load operation), it can capture both normal emission patterns (e.g., high emissions during fixed periods) and identify temporary impacts (e.g., a sudden drop in emissions due to equipment maintenance). Compared to traditional post-event summary static reports, this approach is more in line with the dynamic changes in park carbon emissions in real time, helping managers respond promptly to sudden emission fluctuations. The resulting analysis combines data indicators with causal interpretations, solving the interpretability problem of the black-box output of traditional models. The report not only includes hard indicators such as total emissions and contribution percentages, but also clearly marks the specific impact of key driving factors (e.g., coal consumption contributes 42% due to full-load operation during peak hours). This allows managers to not only know how much is emitted, but also why it is emitted in this way, avoiding the problem of blindly formulating emission reduction measures based solely on data values. Finally, from the perspective of actual management value in the park, this technology upgrades carbon emission analysis from general statistics to precise guidance. The park can formulate targeted emission reduction plans (such as adjusting production shifts and optimizing equipment load) for high-contribution equipment (such as boiler rooms) and key driving factors (such as peak load), rather than indiscriminately investing resources. This significantly improves emission reduction efficiency and resource utilization. At the same time, the dynamic tracking capability allows the park to evaluate the effectiveness of emission reduction measures in real time (e.g., whether the contribution of a certain piece of equipment has decreased after adjustment), forming a closed-loop management of analysis-decision-evaluation. This effectively transforms carbon emission data into implementable management strategies, promoting the transformation of park carbon management from passive statistics to proactive regulation.

[0137] Combination Figure 2 As shown, a carbon emission analysis device of the present invention includes:

[0138] The data acquisition unit is used to acquire multi-source real-time data and energy consumption data of each energy consumption node in the park within the current time period.

[0139] The feature extraction unit is used to extract multi-dimensional key features of the energy consumption node from the multi-source real-time data of each energy consumption node according to the carbon emission analysis requirements.

[0140] An energy structure analysis unit is used to determine the initial energy structure data of each energy consumption node based on the multi-dimensional key features of each energy consumption node.

[0141] The correction unit is used to perform correlation correction on the initial energy structure data based on the correlation between the energy consumption nodes and other energy consumption nodes in the park, so as to obtain the final energy structure data.

[0142] The emission factor matrix construction unit is used to obtain the dynamic emission factor matrix of the energy consumption node based on the final energy structure data through a hierarchical adaptive calculation model.

[0143] The simulation unit is used to combine the dynamic emission factor matrix, the multi-source real-time data, and the energy consumption data to simulate the carbon emissions of the energy consumption node through a hybrid model, and obtain the carbon emission results of the energy consumption node.

[0144] The analysis unit is used to perform dynamic analysis and interpretation of carbon emissions based on the carbon emission results of each energy consumption node through an attention mechanism, so as to obtain the carbon emission analysis results of the park in the current time period.

[0145] The carbon emission analysis device of the present invention has the same advantages over the prior art as the carbon emission analysis method described above, and will not be repeated here.

[0146] Combination Figure 3 As shown, an electronic device of the present invention includes: a processor and a memory, wherein the memory is used to store computer programs;

[0147] When the computer program is loaded by the processor, it causes the processor to execute the carbon emission analysis method as described above.

[0148] The electronic device of the present invention has the same advantages over the prior art as the carbon emission analysis method described above, and will not be repeated here.

[0149] The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the carbon emission analysis method as described above.

[0150] The computer-readable storage medium of the present invention has the same advantages over the prior art as the carbon emission analysis method described above, and will not be repeated here.

[0151] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A carbon emission analysis method, characterized in that, include: Acquire multi-source real-time data and energy consumption data of each energy consumption node in the park within the current time period; Based on the carbon emission analysis requirements, multi-dimensional key features of the energy consumption nodes are extracted from the multi-source real-time data of each energy consumption node. Based on the multi-dimensional key features of each energy consumption node, the initial energy structure data of the energy consumption node is determined; specifically, this includes: analyzing and classifying the multi-dimensional key features using an energy structure identification model to determine multiple energy structure types and scenario-related features of the energy consumption node; determining the energy type proportion, energy conversion efficiency, and energy consumption fluctuation data of the energy consumption node based on the energy structure types; determining scenario-related impact data strongly correlated with energy consumption based on the scenario-related features; and using the energy type proportion, energy conversion efficiency, energy consumption fluctuation data, and scenario-related impact data of the energy consumption node as the initial energy structure data. Based on the correlation between the energy consumption nodes in the park and other energy consumption nodes, the initial energy structure data is correlated and corrected to obtain the final energy structure data; specifically, this includes: quantifying the impact coefficient of each energy consumption node based on the correlation between each energy consumption node in the park; and correcting the energy type proportion, energy conversion efficiency, energy consumption fluctuation data, and scenario-based impact data based on the impact coefficient to obtain the final energy structure data; A hierarchical adaptive calculation model is used to obtain the dynamic emission factor matrix of the energy consumption nodes based on the final energy structure data. The hierarchical adaptive calculation model includes a basic factor matching layer, an efficiency correction layer, a scenario dynamic adjustment layer, and a deviation calibration layer. Specifically, obtaining the dynamic emission factor matrix of the energy consumption nodes based on the final energy structure data using the hierarchical adaptive calculation model includes: obtaining an initial emission factor vector based on the energy type proportion and a preset industry benchmark emission factor library using the basic factor matching layer; correcting the initial emission factor vector based on the energy conversion efficiency and energy consumption fluctuation data using the efficiency correction layer; dynamically adjusting the efficiency correction factor vector in real time based on the scenario-specific impact data using the scenario dynamic adjustment layer; and determining the deviation between the scenario-adjusted factor matrix and historical actual monitoring data using the deviation calibration layer, and performing self-calibration optimization on the scenario-adjusted factor matrix based on the deviation value to obtain the dynamic emission factor matrix. By combining the dynamic emission factor matrix, the multi-source real-time data, and the energy consumption data, a hybrid model is used to simulate carbon emissions at the energy consumption node, and the carbon emission results of the energy consumption node are obtained. Based on the carbon emission results of each energy consumption node, the carbon emissions are dynamically analyzed and interpreted through an attention mechanism to obtain the carbon emission analysis results of the park in the current time period.

2. The carbon emission analysis method according to claim 1, characterized in that, The step of extracting multi-dimensional key features of each energy consumption node from the multi-source real-time data of each energy consumption node according to carbon emission analysis requirements includes: Based on the management scenario of the park, the types of carbon emission analysis needs are determined, including short-term emission reduction monitoring needs, long-term policy formulation needs, and equipment energy efficiency optimization needs. For each type of carbon emission analysis requirement, match the target data type corresponding to the multi-source real-time data of each energy consumption node; Extract the original features of the corresponding dimensions of the target data type; specifically, for the short-term emission reduction monitoring needs, extract the energy consumption peak and valley fluctuation features of the time dimension and the real-time load features of the equipment dimension; for the long-term policy formulation needs, extract the functional area energy consumption difference features of the spatial dimension and the clean energy proportion change features of the policy dimension; for the equipment energy efficiency optimization needs, extract the aging and loss features of the equipment dimension and the capacity-energy consumption correlation features of the operating condition dimension. Data preprocessing is performed on the original features of each extracted dimension to obtain multi-dimensional key features for each energy consumption node.

3. The carbon emission analysis method according to claim 1, characterized in that, The hierarchical adaptive computing model also includes a real-time feedback layer; The step of obtaining the dynamic emission factor matrix of the energy consumption node through a hierarchical adaptive calculation model based on the final energy structure data also includes: The real-time feedback layer obtains real-time carbon emission concentration data collected by the real-time carbon emission monitoring equipment set up in the park. Based on the deviation between the real-time carbon emission concentration data and the estimated carbon emission concentration data generated by the dynamic emission factor matrix, a feedback compensation factor is constructed. The dynamic emission factor matrix is ​​iteratively optimized online using the feedback compensation factor to update the emission factor values ​​in the dynamic emission factor matrix.

4. The carbon emission analysis method according to claim 1, characterized in that, The process of combining the dynamic emission factor matrix, the multi-source real-time data, and the energy consumption data to simulate carbon emissions at the energy consumption node using a hybrid model, and obtaining the carbon emission results for the energy consumption node, includes: The energy consumption data is broken down according to energy type to obtain the consumption amount of each energy type of the energy consumption node; The basic carbon emissions of the energy consumption node are obtained by multiplying the individual consumption amounts element by element with the emission factors of the corresponding energy type in the dynamic emission factor matrix. Based on the multi-source real-time data, a bidirectional long short-term memory network model is used for fitting to obtain the nonlinear correction amount of the energy consumption node; The preliminary carbon emission results for the energy consumption node are obtained by weighted fusion of the baseline carbon emissions and the nonlinear correction. The preliminary carbon emission results are subjected to outlier detection and removal, and the data is labeled in conjunction with the collection frequency of the energy consumption data to generate the carbon emission results including timestamps, energy type-specific carbon emission values, and total carbon emission values.

5. The carbon emission analysis method according to claim 4, characterized in that, The step of dynamically analyzing and interpreting carbon emissions based on the carbon emission results of each energy consumption node, using an attention mechanism, to obtain the carbon emission analysis results of the park within the current time period, includes: The carbon emission results of the energy consumption nodes are input into the attention mechanism model; The attention mechanism model is used to assign weights to the carbon emission values ​​of the energy type and the total carbon emission value in the carbon emission results, thereby determining the contribution of each energy consumer in the energy consumption node to carbon emissions. Based on the contribution level, a dynamic analysis is performed on the energy-consuming individuals to obtain the carbon emission sources and key driving factors of the park; By combining the carbon emission sources and key driving factors of the park, the carbon emission analysis results of the park during the current time period are generated.

6. A carbon emission analysis device, characterized in that, include: The data acquisition unit is used to acquire multi-source real-time data and energy consumption data of each energy consumption node in the park within the current time period. The feature extraction unit is used to extract multi-dimensional key features of the energy consumption node from the multi-source real-time data of each energy consumption node according to the carbon emission analysis requirements. An energy structure analysis unit is used to determine the initial energy structure data of each energy consumption node based on its multi-dimensional key features. Specifically, this includes: analyzing and classifying the multi-dimensional key features using an energy structure identification model to determine multiple energy structure types and scenario-related features of the energy consumption node; determining the energy type proportion, energy conversion efficiency, and energy consumption fluctuation data of the energy consumption node based on the energy structure types; determining scenario-related impact data strongly correlated with energy consumption based on the scenario-related features; and using the energy type proportion, energy conversion efficiency, energy consumption fluctuation data, and scenario-related impact data of the energy consumption node as the initial energy structure data. The correction unit is used to perform correlation correction on the initial energy structure data according to the correlation between the energy consumption nodes and other energy consumption nodes in the park, so as to obtain the final energy structure data; specifically, it includes: quantifying the impact coefficient of each energy consumption node according to the correlation between each energy consumption node in the park; and correcting the energy type proportion, energy conversion efficiency, energy consumption fluctuation data and scenario-based impact data according to the impact coefficient to obtain the final energy structure data. An emission factor matrix construction unit is used to obtain the dynamic emission factor matrix of the energy consumption node based on the final energy structure data using a hierarchical adaptive calculation model. The hierarchical adaptive calculation model includes a basic factor matching layer, an efficiency correction layer, a scenario dynamic adjustment layer, and a deviation calibration layer. The process of obtaining the dynamic emission factor matrix of the energy consumption node based on the final energy structure data using the hierarchical adaptive calculation model includes: obtaining an initial emission factor vector based on the energy type proportion and a preset industry benchmark emission factor library using the basic factor matching layer; correcting the initial emission factor vector based on the energy conversion efficiency and energy consumption fluctuation data using the efficiency correction layer to obtain an efficiency-corrected factor vector; dynamically adjusting the efficiency-corrected factor vector in real time based on the scenario-based impact data using the scenario dynamic adjustment layer to obtain a scenario-adjusted factor matrix; and determining the deviation value between the scenario-adjusted factor matrix and historical actual monitoring data using the deviation calibration layer, and performing self-calibration optimization on the scenario-adjusted factor matrix based on the deviation value to obtain the dynamic emission factor matrix. The simulation unit is used to combine the dynamic emission factor matrix, the multi-source real-time data, and the energy consumption data to simulate the carbon emissions of the energy consumption node through a hybrid model, and obtain the carbon emission results of the energy consumption node. The analysis unit is used to perform dynamic analysis and interpretation of carbon emissions based on the carbon emission results of each energy consumption node through an attention mechanism, so as to obtain the carbon emission analysis results of the park in the current time period.

7. An electronic device, characterized in that, include: Processor and memory, the memory being used to store computer programs; When the computer program is loaded by the processor, it causes the processor to execute the carbon emission analysis method as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the carbon emission analysis method as described in any one of claims 1-5.

Images