Industrial park enterprise monthly carbon emission non-equipment transformation measuring and calculating method, system and equipment based on block chain and CMAR algorithm, and medium

Abstract

The invention is suitable for the field of carbon emission measurement and calculation, and discloses an industrial park enterprise monthly carbon emission non-equipment transformation measurement and calculation method, system, equipment and medium based on a block chain and a CMAR algorithm, and the method comprises the steps: obtaining enterprise historical energy ledger data, and generating an electricity-carbon conversion coefficient sequence; processing the electricity-carbon conversion coefficient sequence through a centralized moving average regression model, and determining a monthly electricity-carbon conversion coefficient; constructing a three-source data channel, and calculating the monthly total carbon emission of the enterprise in combination with a monthly electricity-carbon conversion coefficient; and calculating a current-month actual electricity-carbon conversion coefficient based on the monthly total carbon emission of the enterprise, monitoring the current-month electricity consumption state of the enterprise, triggering a corresponding correction mechanism in a Bayesian online correction and abnormity self-repairing strategy according to a monitoring result, and writing enterprise carbon emission measurement and calculation data and a correction log into a block chain for evidence storage. According to the method, the calculation-storage-certificate integration of the region blocks is used, so that the high efficiency and low error of enterprise carbon emission measurement and calculation are realized.

Description

Technical Field

[0001] This invention relates to the field of carbon emission measurement, and in particular to a method, system, equipment, and medium for measuring monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithm. Background Technology

[0002] With the upgrading of environmental protection requirements, carbon emission monitoring and accounting for enterprises in industrial parks has become a key link for enterprises to operate in compliance with regulations, participate in carbon trading, and achieve green transformation.

[0003] Currently, there are three main methods for calculating corporate carbon emissions, each with its own characteristics but also significant shortcomings: ① Direct measurement mode by installing sensors at the end of equipment: Its advantage is high measurement accuracy, and the data directly reflects the actual emissions. However, it also has prominent problems: the emission end is mostly a harsh environment with high temperature and high corrosion; sensors are not only expensive but also have high maintenance costs and are easily damaged. For older factories, installing sensors requires modifying existing equipment, which is difficult and often requires production shutdowns, significantly impacting production. ② The inventory method based on annual reports is also a commonly used accounting method for many companies. It mainly relies on data such as energy consumption reports and production ledgers submitted by companies at the end of the year to calculate the annual carbon emissions in one go. The disadvantages of this method are obvious: firstly, it is too lagging and cannot support real-time decision-making in the carbon trading market; secondly, the data reliability is insufficient, relying entirely on self-reporting by companies without third-party verification or real-time data endorsement, which easily leads to data distortion. ③ The averaging calculation method using fixed default factors. This method uses a fixed average coefficient published by the state or industry as the basis for calculation, multiplying it by the company's energy consumption to obtain carbon emissions. This method is simple to operate and requires no additional investment, but it ignores the dynamic changes of companies. Even if carbon emission efficiency has improved, the fixed factor cannot be adjusted in time, resulting in a large deviation between the calculated results and actual emissions, and failing to reflect the company's emission reduction efforts.

[0004] Therefore, there is a need for a carbon emission measurement method that requires no hardware modification to on-site equipment, is low in cost, can output monthly data, and has adaptive correction capabilities and data reliability. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides a method, system, equipment, and medium for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithms. This solves the problem that existing carbon emission monitoring methods require the installation of sensors at the end of production equipment, resulting in high construction difficulty and high downtime costs. At the same time, it solves the problems of data lag, inability to provide high-frequency monthly data, and inaccurate calculations caused by fixed default factors that cannot adapt to sudden changes in enterprise operating conditions, which are caused by the traditional annual billing-based inventory method.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, this invention provides a method for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and the CMAR algorithm, including: Obtain historical energy ledger data of enterprises and construct an annual electricity-carbon raw table. Perform data cleaning on the raw table to generate an electricity-carbon conversion factor sequence. The carbon conversion factor sequence is processed by a centralized moving average regression model to obtain the carbon conversion prediction value range, which is used to determine the monthly carbon conversion factor. Construct a three-source data channel and obtain regional power transaction traceability data verified by blockchain hashing to obtain the comprehensive carbon emission factor of the regional power grid. Based on the three data channels, combined with the monthly electricity carbon conversion factor and the regional power grid comprehensive electricity carbon emission factor, the company's monthly total carbon emissions are obtained. The actual carbon emission conversion factor for the current month is calculated based on the company's total monthly carbon emissions. The company's monthly electricity consumption status is monitored. Based on the monitoring results, the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy is triggered, and the company's carbon emission calculation data and correction log are written to the blockchain for evidence storage.

[0008] As a preferred embodiment of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in this invention, the method involves: processing the carbon emission conversion coefficient sequence using a centralized moving average regression model to obtain a carbon emission conversion prediction range, which is used to determine the monthly carbon emission conversion coefficient, including: The sequence of carbon conversion coefficients is received as input, and the golden ratio decay coefficient and Holt-Winters smoothing algorithm are introduced into the centered moving average regression model to calculate the model parameters after residual correction. The system receives annually updated enterprise electricity consumption data, drives the model training window to scroll forward, and obtains the range of predicted carbon conversion values ​​based on the model parameters after residual correction.

[0009] As a preferred embodiment of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in this invention, the step of triggering the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy according to the monitoring results includes: The carbon conversion deviation rate is obtained by calculating the actual carbon conversion factor of the current month and the actual carbon conversion factor of the previous month. If the carbon conversion deviation rate exceeds the coefficient fluctuation threshold, an online correction instruction is obtained to trigger Bayesian online correction. The electricity consumption deviation rate is obtained by calculating the electricity consumption of the enterprise in the current month and the electricity consumption of the same period last year. If the electricity consumption deviation rate exceeds the threshold of sudden change in operating conditions, an abnormal rollback instruction is obtained to trigger the abnormal self-repair strategy.

[0010] As a preferred embodiment of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in this invention, the Bayesian online correction includes: A prior distribution is constructed based on the carbon conversion factor for electricity within a predetermined time period. By combining the prior distribution with the actual carbon conversion factor for the current month, and using Bayes' theorem, the posterior accuracy and the expected value of the carbon conversion factor are obtained. The online correction magnitude is calculated based on the expected value of the carbon conversion factor. If the online correction magnitude reaches the first preset condition, the model parameters are updated using the expected value of the carbon conversion factor; if the online correction magnitude reaches the second preset condition, an anomaly self-repair strategy is implemented.

[0011] As a preferred embodiment of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in this invention, the anomaly self-repair strategy includes: Discarding the current short-cycle training window, we obtain the monthly electricity consumption and carbon emissions of enterprises within a preset time period in the past, and combine them with the annual electricity-carbon raw table to form a new rolling sample sequence; Based on the new rolling sample sequence, the carbon conversion factor is recalculated, and the abnormal correction magnitude is obtained by Bayesian online correction. If the abnormal correction magnitude reaches the third preset condition, retraining is performed; if the abnormal correction magnitude reaches the fourth preset condition, the long short-term memory network is started for retraining.

[0012] As a preferred embodiment of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in this invention, the method includes: constructing a three-source data channel and obtaining regional power transaction traceability data verified by blockchain hashing to obtain the comprehensive carbon emission factor of the regional power grid, including: A three-source data channel is constructed by the marketing system database, the metering automation system database, and the regional power trading database; The power source traceability data of the region's power trading is obtained through the regional power trading database, and the power source traceability data information is written into the blockchain transaction chain and hash verification is performed. Based on the power trading source traceability data that has been hash-verified, and combined with various power source carbon emission factor information, the comprehensive power carbon emission factor of the regional power grid is obtained.

[0013] As a preferred embodiment of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in this invention, the method for obtaining the total monthly carbon emissions of enterprises includes: Based on the marketing system database and the metering automation system database, the company’s total monthly electricity consumption and distributed photovoltaic data are obtained, and the company’s actual monthly electricity consumption is calculated. The fossil fuel emissions are calculated based on the company's actual electricity consumption for the month. By obtaining green electricity consumption data from the marketing system database and combining it with the company's total monthly electricity consumption and the regional power grid's comprehensive carbon emission factor, the company's actual carbon emissions from electricity consumption can be calculated. The company's total monthly carbon emissions are obtained by combining the company's actual electricity consumption carbon emissions with its fossil fuel emissions.

[0014] 8. A system for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithm, using the method as described in any one of claims 1-7, characterized in that it includes: Secondly, this invention provides a system for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting, based on blockchain and the CMAR algorithm, including: The original table construction module is used to obtain historical energy ledger data of enterprises and construct annual electricity-carbon original tables. Data cleaning is performed on the original tables to generate a sequence of electricity-carbon conversion factors. The model prediction module is used to process the carbon conversion factor sequence through a centralized moving average regression model to obtain the carbon conversion prediction value range, which is used to determine the monthly carbon conversion factor. The data channel construction module is used to construct a three-source data channel and obtain regional power transaction traceability data verified by blockchain hash, thereby obtaining the comprehensive carbon emission factor of the regional power grid. The calculation module is used to obtain the company's total monthly carbon emissions based on the three-source data channels, combined with the monthly electricity carbon conversion factor and the regional power grid comprehensive electricity carbon emission factor. The self-correction module is used to calculate the actual electricity carbon conversion factor for the current month based on the company's total monthly carbon emissions, monitor the company's monthly electricity consumption status, trigger the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy according to the monitoring results, and write the company's carbon emission calculation data and correction log to the blockchain for evidence storage.

[0015] Thirdly, the present invention provides an electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the method for calculating the monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm.

[0017] Compared with existing technologies, the beneficial effects of this invention are as follows: By integrating historical energy consumption data of enterprises with existing three-source data from the power grid, and combining it with a centralized moving average regression model, carbon emission calculation parameters, logs, and results data are stored; a multi-source data fusion carbon emission calculation model is constructed to accurately measure the individual carbon emissions of enterprises, achieving accurate monthly carbon emission calculation for enterprises; and an innovative Bayesian online correction and anomaly self-repair strategy is introduced, which can not only correct the natural drift of the model over time, but also sensitively capture sudden changes in operating conditions and automatically switch between rollback or retraining mechanisms, ensuring the accuracy of calculation throughout the entire life cycle. The method of this invention is easy to operate, relies on existing data collected by the power grid, does not require the deployment and installation of sensors on the enterprise side, and does not rely on statistical reports for calculation, significantly reducing the cost of carbon emission calculation and possessing broad applicability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the overall process of a method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting, based on blockchain and CMAR algorithm, according to an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram illustrating the self-healing mechanism of an anomaly in the method for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithm, according to an embodiment of the present invention. Detailed Implementation

[0021] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0022] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithm, is provided, including: S100: Obtain historical energy ledger data of the enterprise and construct an annual electricity-carbon raw table. Perform data cleaning on the raw table to generate an electricity-carbon conversion factor sequence. S200: The carbon conversion coefficient sequence is processed by a centralized moving average regression model to obtain the carbon conversion prediction value range, which is used to determine the monthly carbon conversion coefficient; S300: Construct three-source data channels and obtain regional power transaction traceability data verified by blockchain hashing to obtain the comprehensive carbon emission factor of the regional power grid; S400: Based on three data channels, combined with the monthly electricity carbon conversion factor and the regional power grid comprehensive electricity carbon emission factor, the company’s monthly total carbon emissions are obtained. S500: Calculates the actual electricity carbon conversion factor for the current month based on the enterprise's total monthly carbon emissions, monitors the enterprise's monthly electricity consumption status, triggers the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy based on the monitoring results, and writes the enterprise's carbon emission calculation data and correction log to the blockchain for evidence storage.

[0023] It should be noted that industrial park enterprises have complex production processes and operate continuously. If the direct measurement method with added sensors is adopted, it is necessary to modify the old exhaust pipes by drilling holes. This not only involves high equipment investment but also incurs economic losses from production stoppages. Traditional inventory methods mainly rely on annual reports, which suffer from serious "data lag" and cannot reflect subtle changes in carbon emissions caused by order fluctuations, green electricity access, or equipment aging in a company's monthly operations. In addition, static calculation factors are difficult to cope with non-linear changes in enterprise operating conditions, resulting in large deviations between the calculated results and the actual situation, making it difficult to meet the dual requirements of data timeliness and accuracy in the carbon trading market.

[0024] Therefore, to address the aforementioned issues of high calculation costs and time lag, the following steps (S100-S500) are used to obtain historical energy ledger data from enterprises and construct annual electricity-carbon raw tables, generating a series of electricity-carbon conversion coefficients. A centralized moving average regression model is then used to process the data to obtain monthly electricity-carbon conversion coefficients. Through three data channels, the enterprise's monthly total carbon emissions and other relevant information are calculated. Simultaneously, online monitoring is conducted to ensure that enterprises implement Bayesian online correction or anomaly self-repair strategies to correct their electricity-carbon emissions. Finally, the raw carbon emission calculation data, carbon emission results data, and related logs are stored on the blockchain for verification by third parties.

[0025] Example 2, refer to Figures 1-2 As an embodiment of the present invention, based on the above embodiment, a method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting is provided based on blockchain and CMAR algorithm.

[0026] In this embodiment of the application, step S100 involves acquiring historical energy ledger data of the enterprise and constructing an annual electricity-carbon raw table, performing data cleaning on the raw table, and generating an electricity-carbon conversion factor sequence, including: Specifically, we obtain annual energy ledger data from enterprises for 2020–2024, including the physical consumption of coal, natural gas, gasoline and diesel, purchased steam, and purchased electricity, as well as the corresponding carbon emissions, to form an annual electricity-carbon raw table. At the same time, we obtain the rated parameters and functional energy categories of the enterprise's main energy-consuming equipment to form a standardized archive table.

[0027] Furthermore, the annual electricity-carbon raw data table was cleaned by removing outliers, standardizing units, and interpolating for missing values, based on electricity consumption. Total carbon emissions are the independent variable. Define the carbon conversion factor as the dependent variable. (Unit: kgCO2e / kWh), generating the initial sequence of electrocarbon conversion factors.

[0028] In this embodiment of the application, step S200 processes the carbon conversion factor sequence using a centralized moving average regression model to obtain a carbon conversion prediction range, which is used to determine the monthly carbon conversion factor, including: Specifically, the centered moving average regression (CMAR) model is used to calculate the carbon conversion factor for electricity. The calculation formula is as follows:

[0029] in (Golden section decay coefficient) It is the moving average of the previous n periods. With goodness of fit Adaptive adjustment; For the residuals, Holt-Winters smoothing is used for correction; This is the carbon conversion factor for the previous year.

[0030] The training window is set up so that after acquiring the monthly electricity consumption data of enterprises in January of each year, it is automatically updated once to output the data for the next 12 months of the current year. Predicted values ​​and standard errors The range of predicted values ​​for carbon dioxide conversion was obtained. (Z=1.96 corresponds to the 95% confidence interval).

[0031] In this embodiment of the application, step S300 involves constructing a three-source data channel and obtaining regional power transaction traceability data verified by blockchain hashing to obtain the comprehensive carbon emission factor of the regional power grid, including the following steps A1-A3: A1: Construct a three-source data channel through the marketing system database, the metering automation system database, and the regional power trading database; Specifically, the marketing system data API: By using the user ID information corresponding to enterprises in the industrial park, it obtains the enterprise's monthly electricity bill information and extracts the enterprise's total monthly electricity consumption. ), green electricity ( ).

[0032] Metering Automation System Data API: Based on the meter numbers of the power generation side and the grid connection side of the distributed photovoltaic systems built by industrial users in the park, the power generation of the distributed photovoltaic systems can be obtained respectively. Internet power consumption .

[0033] Regional Power Transaction Data API: Obtain power source traceability data for the industrial park where the power transaction institution is located, including coal-fired power generation, existing hydropower generation, wind power generation, photovoltaic power generation, nuclear power generation, and new hydropower generation.

[0034] A2: Obtain power source traceability data for the region through the regional power trading database, write the original message, timestamp, and unique ID of the power source traceability data into the blockchain transaction chain, and perform hash verification.

[0035] It should be noted that by using existing user electricity consumption and generation settlement data in the current power system, as well as regional power transaction data verified by blockchain, and combining enterprise electricity consumption with carbon emission conversion factors, a high-precision monthly carbon emission calculation model for enterprises is constructed to accurately determine user emissions. This solves the shortcomings of traditional direct measurement methods, which require enterprises to shut down production and deploy a large number of high-temperature, corrosion-resistant, and explosion-proof sensors at key boilers, steam pipes, and other nodes, while simultaneously building a new data acquisition network to collect carbon emission data from each node. No new primary instruments are required, achieving "zero hardware modification," without interfering with normal enterprise production, with no maintenance costs, and the marginal cost of expansion approaching zero.

[0036] A3: Based on the power trading source traceability data that has been hash-verified, and combined with various power source carbon emission factor information, the comprehensive power carbon emission factor of the regional power grid is obtained. Specifically, a hash check is performed, and if the hashes do not match, an exception alarm is triggered and subsequent calculations are rejected.

[0037] Based on hash-verified power trading source traceability data, combined with various power source carbon emission factor information, the comprehensive power carbon emission factor of the regional power grid is calculated:

[0038] In the formula, For the first blockchain-verified Electricity trading volume for similar power sources For the corresponding number Carbon emission factors of similar power sources.

[0039] Among them, the carbon emission factor information of various power sources mainly includes coal-fired power plants, hydropower, solar photovoltaic power and wind power.

[0040] It should be noted that, after blockchain hash verification, it is updated monthly. Synchronized with monthly grid electricity consumption, this method effectively addresses the shortcomings of the traditional "provincial annual average emission factor" method, which only publishes data once a year and fails to reflect seasonal variations in the regional grid power supply structure (e.g., the actual coal / hydro / wind / solar ratio can differ by 15% to 45% between the "high-water season" and the "low-water season"), leading to inaccurate monthly carbon emission measurements. By upgrading the regional grid carbon emission factor from "unchanged for a year" to "updated monthly," synchronizing the carbon emission factor with electricity consumption, the system can systematically reduce the error of the carbon emission factor by more than 15%.

[0041] In this embodiment of the application, step S400, based on three data channels and combining the monthly electricity carbon conversion factor with the regional power grid comprehensive electricity carbon emission factor, yields the company's total monthly carbon emissions, including the following steps B1-B4: B1: Based on the marketing system database and the metering automation system database, obtain the company's total monthly electricity consumption and distributed photovoltaic data, and calculate the company's actual monthly electricity consumption. Specifically, the company's actual monthly electricity consumption is calculated by combining the company's total monthly electricity consumption with the power generation and grid connection of distributed photovoltaic power.

[0042] B2: Based on the company's actual electricity consumption in the current month, the fossil fuel emissions are calculated; Specifically, the carbon emission calculation for users is based on the company's actual monthly electricity consumption:

[0043] Calculate the carbon emission value of electricity consumption using the company's actual monthly electricity consumption:

[0044] In the formula, Take the regional power grid's comprehensive carbon emission factor from the previous year.

[0045] Furthermore, if Then the carbon emissions from fossil fuels All of the user's carbon emissions come from electricity consumption; like Then calculate the carbon emissions from fossil fuels:

[0046] B3: Obtain green electricity consumption data from the marketing system database, and combine this data with the company's monthly electricity bill issuance and the regional power grid's comprehensive carbon emission factor to calculate the company's actual carbon emissions from electricity consumption.

[0047] B4: The company's total monthly carbon emissions are calculated by combining the company's actual electricity consumption carbon emissions with its fossil fuel emissions.

[0048] In this embodiment of the application, step S500 calculates the actual monthly carbon emission conversion factor based on the enterprise's total monthly carbon emissions, monitors the enterprise's monthly electricity consumption status, triggers the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy based on the monitoring results, and writes the enterprise's carbon emission calculation data and correction log into the blockchain for evidence storage, including the following steps C1-C3: C1: Triggers the Bayesian online correction; Specifically, calculate the actual carbon conversion factor for the current month:

[0049] Will Compared with the actual carbon conversion factor of last month When comparing, At that time, Bayesian online correction is triggered.

[0050] Furthermore, triggering Bayesian online correction includes the following sub-steps: ① Based on the carbon conversion factor of the past 6 months, hyperparameter estimation is performed to construct the prior distribution:

[0051] In the formula, The prior mean; The prior standard deviation; For observation accuracy; To account for errors in fossil fuel carbon emissions, taking into account measurement and data entry errors, the value is ±3%; The measurement error of the electricity meter is ±0.2% based on the current accuracy of the electricity meters used by enterprise users.

[0052] ② Perform Bayesian online correction and calculate posterior accuracy. Expected value of carbon conversion factor :

[0053] ③ The online correction range is calculated based on the expected value of the carbon conversion factor. :

[0054] in, for.

[0055] If 3 times in a row Then expand the model parameter scroll window from the initial 6 scroll windows to 12, and update the model parameters. ;like Then it will perform abnormal self-repair.

[0056] It should be noted that Bayesian online correction updates the data to the blockchain, forming a closed loop of "prediction-measurement-correction-re-prediction". This effectively solves the problem of the traditional annual carbon inventory method of "static factors + manual ledger" which only calibrates once a year and cannot reflect changes in production output and regional power grid structure. It reduces the carbon emission accounting error from the industry average of 15%~25% to ≤5%, and has the ability to adapt to sudden operating conditions, forming a replicable accuracy improvement paradigm.

[0057] C2: Write the enterprise's carbon emission calculation data and correction logs to the blockchain for evidence storage; Specifically, will , , , , , The transaction hash is packaged into JSON and the structured data on-chain method (ac.cn.iie.bc.method.Upload.uploadStruct(String structText,String accessId,String desc,ActionParams params)) is called to write the company's monthly carbon emission information into the blockchain. Other users can then query the original hash and parsing report on the blockchain using the company ID and month, enabling one-click traceability.

[0058] C3: Triggers the abnormal self-repair strategy; Specifically, when When this occurs, a Bayesian rollback is triggered, which is an example of self-correction of the anomaly. This represents the total monthly electricity consumption of enterprises in the same period of the previous year.

[0059] For example, such as Figure 2 The diagram shown is a Bayesian rollback flowchart.

[0060] Furthermore, triggering a Bayesian rollback also includes the following sub-steps: ① Discard the current short-cycle training window and construct a sliding window of n=7 by combining the electricity consumption of the current month with the electricity consumption, carbon emissions and carbon conversion factor of the past 6 months; Based on the updated sliding window, hyperparameter estimation is performed again:

[0061] In the formula, 'With the center point; 'This represents the range of uncertainty.'

[0062] ②Based on the rated power, functional energy category, and seasonal characteristics of the enterprise's main equipment, calculate the monthly carbon conversion factor for electricity. and error Quick estimation:

[0063]

[0064] In the formula, 'To quickly calculate carbon emissions from fossil fuels; OT represents the operating time for the current month; LF represents the seasonal characteristic factor for the current month, reflecting the typical distribution of enterprise electricity consumption as it changes from month to month;' Rated power of the equipment; These are the lower heating value and carbon emission factor of the fossil energy used, respectively.

[0065] ③ Perform Bayesian online correction to calculate the corresponding posterior accuracy and expected value of the carbon conversion factor. Then, based on the expected value of the carbon conversion factor, obtain the anomaly correction magnitude. (The processing method is the same as that of sub-steps ②-③ in C1).

[0066] like Then retraining will be triggered; if it happens 3 times in a row. The window was extended to 12 months, and the network was retrained using a Long Short-Term Memory (LSTM) network.

[0067] In this embodiment of the application, the retraining using LSTM in step S500 after completing step C2 further includes the following steps S1-S5: S1: Obtain five-dimensional data for the most recent 12 months, including electricity consumption, fossil fuel emissions, carbon emissions from electricity, carbon conversion factor for electricity, and comprehensive carbon emission factor of the regional power grid. When the data missing rate is <5%, linear interpolation is used. After Z-score normalization, the sequence center point is calculated. Uncertainty range For reuse in reasoning:

[0068] The training set is used for the first 10 months, and the validation set is used for the last 2 months. A sliding window is used to ensure online rolling updates.

[0069] S2: Construct a lightweight LSTM model; Specifically, the process involves 5-dimensional data input → LSTM (64 units) → Dropout (0.2) → Dense (1), outputting the estimated carbon conversion factor for the current month, with the prediction error measured by MAE.

[0070] In the formula, , , , For trainable weights, The hidden state of the LSTM is used as the input for the next time step to achieve "memory rolling". This represents the current 5-dimensional income vector; This is a hidden state for perturbation, used for uncertainty sampling; This is the estimated value of the carbon conversion factor for the current month.

[0071] S3: Training loop; Specifically, S1-S2 are executed in a loop, with a batch size of 32, and the Adam optimizer is initialized. In each round, the loss is calculated first through forward propagation, and then the weights are updated through backward propagation. If the MAE does not decrease by 0.01 for five consecutive rounds, the process is stopped early to prevent overfitting. |

[0072] In the formula, For all trainable parameters; This is the carbon conversion factor for the current month; The learning rate is used in the k-th round. loss function For model parameters The gradient vector.

[0073] S4: Uncertainty estimation; Specifically, with Dropout enabled, repeat the forward pass 100 times, take the mean as the most reliable estimate, and use the standard deviation as the error range:

[0074] In the formula, Let be the predicted value obtained from the i-th sampling; The most reliable carbon conversion factor; The prediction error range for the carbon conversion factor.

[0075] S5: If the MAE is verified to be ≤2.1%, save the model weights and normalization parameters; otherwise, roll back to the previous CMAR (Centered Moving Average Regression) model and no longer use the LSTM that was just retrained.

[0076] S6: Output the expected value of the current carbon conversion factor. .

[0077] C4: Carbon emission correction and on-chain data update; Specifically, the expected value using the carbon conversion factor. The system will make corrections, recalculate fossil fuel carbon emissions and total monthly carbon emissions, and simultaneously generate audit logs showing the reasons for the corrections, the magnitude of the corrections, and the confidence probabilities.

[0078] Furthermore, the corrected monthly carbon emission parameters of the enterprise and the audit logs are packaged into JSON and re-uploaded to the blockchain using the unstructured file uploading method: ac.cn.iie.bc.method.Upload.uploadFile(FilefileData,String desc,String accessId,ActionParams params), to ensure that the correction process is traceable and auditable, and to ensure that the final error is still ≤5%.

[0079] It should be noted that by using a two-level triggering mechanism of Bayesian online correction and Bayesian rollback, the shortcomings of traditional static factors or annual models that cannot respond in time when monthly electricity consumption surges or drops due to events such as enterprise maintenance, fuel switching, and production line switching are resolved, which leads to an instantaneous amplification of carbon emission calculation errors. This transforms "sudden changes lead to loss of control" into "sudden changes lead to escalation." Even under extreme operating conditions, the carbon emission calculation error can still be guaranteed to be within 5%, and the entire process is traceable on the chain.

[0080] Example 3 is an embodiment of the present invention. Based on the above embodiments, a practical example analysis of monthly carbon emission calculation for enterprises in industrial parks without equipment retrofitting based on blockchain and CMAR algorithm is provided to verify its feasibility and effectiveness. Using a small chemical enterprise in a western province as the verification object, we obtained its historical energy consumption from 2020 to 2024 as the basis for modeling.

[0081] Select key parameters (see Table 1), model parameters (see Table 2), energy usage data for each category from 2020 to 2024 (see Table 3), and monthly electricity consumption data for the past two years (see Table 4).

[0082] Table 1

[0083] Table 2

[0084] Table 3

[0085] Table 4

[0086] The carbon conversion factor was calculated. The sequence was then input into the CMAR prediction model along with the company's monthly electricity consumption data for the past two years to obtain the monthly data for 2025. See Table 5 for reference.

[0087] Table 5

[0088] The average carbon emission factor of a certain regional power grid in 2024 was 0.710 kgCO2e / kWh.

[0089] Table 6, Reference Table for Electricity Transaction Statistics by Power Source Category in March 2025 (Unit: kWh): Table 6

[0090] Calculated = 0.3347 tCO2e / MWh Obtain the company's electricity bill for March 2025 and information on its self-invested distributed photovoltaic power generation to calculate the company's actual electricity consumption, refer to Table 7.

[0091] Table 7

[0092] Carbon emissions were calculated based on the company's electricity consumption records in March 2025, as per Table 8, after a third-party investigation.

[0093] Table 8

[0094] The calculation yielded: = 1.69346 million kWh; = 484.043 tCO2e; =12430.0 tCO2e; = 12159.526 tCO2e.

[0095] The error is 3.17% compared to the third-party verification report (11785.16 tCO2e).

[0096] In summary, this invention integrates historical energy consumption data of enterprises with existing three-source data from the power grid, and combines this with a centralized moving average regression model to store carbon emission calculation parameters, logs, and results data. It constructs a multi-source data fusion carbon emission calculation model to accurately measure individual enterprise carbon emissions, achieving precise monthly carbon emission calculations. It innovatively introduces Bayesian online correction and anomaly self-repair strategies, which not only correct the natural drift of the model over time but also sensitively capture sudden changes in operating conditions, automatically switching between rollback and retraining mechanisms to ensure calculation accuracy throughout the entire lifecycle. This method demonstrates high predictive accuracy and practicality at the scale of monthly carbon emission calculation for enterprises. By predicting the monthly electricity carbon conversion factor sequence for the current year using CMAR, it effectively compensates for the shortcomings of high cost, long cycle, and large error in monthly carbon emission calculation for enterprises in industrial parks, providing a reliable technical tool for enterprise monthly carbon emission management.

[0097] Example 4 illustrates a schematic scheme for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithms. It should be noted that the technical solution of this system for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithms belongs to the same concept as the aforementioned method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithms. Details not described in detail in the technical solution of the system for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithms in this example can be found in the description of the aforementioned method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithms.

[0098] This embodiment also provides a system for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm, including: The original table construction module is used to obtain historical energy ledger data of enterprises and construct annual electricity-carbon original tables. Data cleaning is performed on the original tables to generate a sequence of electricity-carbon conversion factors. The model prediction module is used to process the carbon conversion factor sequence through a centralized moving average regression model to obtain the carbon conversion prediction value range, which is used to determine the monthly carbon conversion factor. The data channel construction module is used to construct a three-source data channel and obtain regional power transaction traceability data verified by blockchain hash, thereby obtaining the comprehensive carbon emission factor of the regional power grid. The calculation module is used to obtain the company's total monthly carbon emissions based on the three-source data channels, combined with the monthly electricity carbon conversion factor and the regional power grid comprehensive electricity carbon emission factor. The self-correction module is used to calculate the actual electricity carbon conversion factor for the current month based on the company's total monthly carbon emissions, monitor the company's monthly electricity consumption status, trigger the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy according to the monitoring results, and write the company's carbon emission calculation data and correction log to the blockchain for evidence storage.

[0099] This embodiment also provides an electronic device applicable to the calculation of monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as proposed in the above embodiment.

[0100] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting, as proposed in the above embodiment, based on blockchain and CMAR algorithm.

[0101] The storage medium proposed in this embodiment and the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0102] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0103] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithm, characterized in that... include: Obtain historical energy ledger data of enterprises and construct an annual electricity-carbon raw table. Perform data cleaning on the raw table to generate an electricity-carbon conversion factor sequence. The carbon conversion factor sequence is processed by a centralized moving average regression model to obtain the carbon conversion prediction value range, which is used to determine the monthly carbon conversion factor. Construct a three-source data channel and obtain regional power transaction traceability data verified by blockchain hashing to obtain the comprehensive carbon emission factor of the regional power grid. Based on the three data channels, combined with the monthly electricity carbon conversion factor and the regional power grid comprehensive electricity carbon emission factor, the company's monthly total carbon emissions are obtained. The actual carbon emission conversion factor for the current month is calculated based on the company's total monthly carbon emissions. The company's monthly electricity consumption status is monitored. Based on the monitoring results, the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy is triggered, and the company's carbon emission calculation data and correction log are written to the blockchain for evidence storage.

2. The method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in claim 1, characterized in that, The carbon conversion factor sequence is processed using a centralized moving average regression model to obtain a range of predicted carbon conversion values, which are used to determine the monthly carbon conversion factor, including: The sequence of carbon conversion coefficients is received as input, and the golden ratio decay coefficient and Holt-Winters smoothing algorithm are introduced into the centered moving average regression model to calculate the model parameters after residual correction. The system receives annually updated enterprise electricity consumption data, drives the model training window to scroll forward, and obtains the range of predicted carbon conversion values ​​based on the model parameters after residual correction.

3. The method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in claim 2, characterized in that, The corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy triggered based on monitoring results includes: The carbon conversion deviation rate is obtained by calculating the actual carbon conversion factor of the current month and the actual carbon conversion factor of the previous month. If the carbon conversion deviation rate exceeds the coefficient fluctuation threshold, an online correction instruction is obtained to trigger Bayesian online correction. The electricity consumption deviation rate is obtained by calculating the electricity consumption of the enterprise in the current month and the electricity consumption of the same period last year. If the electricity consumption deviation rate exceeds the threshold of sudden change in operating conditions, an abnormal rollback instruction is obtained to trigger the abnormal self-repair strategy.

4. The method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in claim 3, characterized in that, The Bayesian online correction includes: A prior distribution is constructed based on the carbon conversion factor for electricity within a predetermined time period. By combining the prior distribution with the actual carbon conversion factor for the current month, and using Bayes' theorem, the posterior accuracy and the expected value of the carbon conversion factor are obtained. The online correction magnitude is calculated based on the expected value of the carbon conversion factor. If the online correction magnitude reaches the first preset condition, the model parameters are updated using the expected value of the carbon conversion factor; if the online correction magnitude reaches the second preset condition, an anomaly self-repair strategy is implemented.

5. The method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in claim 4, characterized in that, The anomaly self-repair strategy includes: Discarding the current short-cycle training window, we obtain the monthly electricity consumption and carbon emissions of enterprises within a preset time period in the past, and combine them with the annual electricity-carbon raw table to form a new rolling sample sequence; Based on the new rolling sample sequence, the carbon conversion factor is recalculated, and the abnormal correction magnitude is obtained by Bayesian online correction. If the abnormal correction magnitude reaches the third preset condition, retraining is performed; if the abnormal correction magnitude reaches the fourth preset condition, the long short-term memory network is started for retraining.

6. The method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in claim 5, characterized in that, By constructing a three-source data channel and acquiring regional power transaction traceability data verified by blockchain hashing, the comprehensive carbon emission factor of the regional power grid is obtained, including: A three-source data channel is constructed by the marketing system database, the metering automation system database, and the regional power trading database; The power source traceability data of the region's power trading is obtained through the regional power trading database, and the power source traceability data information is written into the blockchain transaction chain and hash verification is performed. Based on the power trading source traceability data that has been hash-verified, and combined with various power source carbon emission factor information, the comprehensive power carbon emission factor of the regional power grid is obtained.

7. The method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in claim 6, characterized in that, The obtained monthly total carbon emissions of the enterprise include: Based on the marketing system database and the metering automation system database, the company’s total monthly electricity consumption and distributed photovoltaic data are obtained, and the company’s actual monthly electricity consumption is calculated. The fossil fuel emissions are calculated based on the company's actual electricity consumption for the month. By obtaining green electricity consumption data from the marketing system database and combining it with the company's total monthly electricity consumption and the regional power grid's comprehensive carbon emission factor, the company's actual carbon emissions from electricity consumption can be calculated. The company's total monthly carbon emissions are obtained by combining the company's actual electricity consumption carbon emissions with its fossil fuel emissions.

8. A system for calculating monthly carbon emissions of enterprises in industrial parks without equipment retrofitting, based on blockchain and CMAR algorithm, using the method described in any one of claims 1-7, characterized in that... include: The original table construction module is used to obtain historical energy ledger data of enterprises and construct annual electricity-carbon original tables. Data cleaning is performed on the original tables to generate a sequence of electricity-carbon conversion factors. The model prediction module is used to process the carbon conversion factor sequence through a centralized moving average regression model to obtain the carbon conversion prediction value range, which is used to determine the monthly carbon conversion factor. The data channel construction module is used to construct a three-source data channel and obtain regional power transaction traceability data verified by blockchain hash, thereby obtaining the comprehensive carbon emission factor of the regional power grid. The calculation module is used to obtain the company's total monthly carbon emissions based on the three-source data channels, combined with the monthly electricity carbon conversion factor and the regional power grid comprehensive electricity carbon emission factor. The self-correction module is used to calculate the actual electricity carbon conversion factor for the current month based on the company's total monthly carbon emissions, monitor the company's monthly electricity consumption status, trigger the corresponding correction mechanism in the Bayesian online correction and anomaly self-repair strategy according to the monitoring results, and write the company's carbon emission calculation data and correction log to the blockchain for evidence storage.

9. An electronic device, characterized in that, include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting based on blockchain and CMAR algorithm as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It includes storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for calculating monthly carbon emissions of industrial park enterprises without equipment retrofitting, as described in any one of claims 1 to 7, based on blockchain and CMAR algorithm.

Images