A cement production process carbon emission calculation method and system based on cloud edge collaboration
Patent Information
- Application Number
- CN202611145402.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-30
- Publication Date
- 2026-10-23
AI Technical Summary
然而，该方案仍仅采用单一来源的监测数据进行碳排放计算，缺乏多源数据融合校验机制，数据可靠性不足；未对碳排放计算结果的测量不确定度进行量化评定，难以满足碳交易市场对数据质量的高要求；边缘端与云端的模型参数未能实现动态闭环更新，随着生产工艺变化，模型计算精度下降；对水泥生产过程多工序的碳排放缺乏精细化的分解与归因能力，无法识别各工序环节的具体减排潜力
[0015]经由上述的技术方案可知，与现有技术相比，本发明提供了一种基于云边协同的水泥生产过程碳排放计算方法及系统，具有以下有益效果：本发明通过同时采用排放因子法、物料平衡法和烟气直测法三种独立方法并行计算碳排放量，并对多路径计算结果进行交叉校验和加权融合，有效克服了单一计算方法存在的系统性偏差问题，显著提高了碳排放计算结果的可靠性和准确性，当各计算结果不一致时自动触发告警，便于人工复核，弥补了现有技术仅采用单一数据源的不足；本发明在云边协同架构中集成了碳排放量测量不确定度的自动化评定功能，通过对各输入量的测量不确定度进行传播计算，以蒙特卡洛模拟获得合成标准不确定度，使碳排放计算结果具有可量化的置信区间，满足全国碳排放权交易市场对数据质量的严格要求，为企业碳配额履约和碳资产管理提供可靠数据支撑；本发明利用云端强大的计算能力对碳排放预测模型进行增量学习训练，并将优化后的模型参数周期性下发至边缘端更新本地计算模型，实现了“边缘采集计算—云端训练优化—边缘参数更新”的完整闭环，确保碳排放计算模型随生产工艺参数的变化和数据的积累而持续优化，避免了模型老化导致的计算偏差；本发明通过对水泥生产各主要工序分别进行碳排放核算，结合云端关键排放工序识别功能，能够精准定位碳排放强度最高的工序环节，为企业制定针对性减排方案提供决策依据，克服了现有技术仅输出企业级总排放量的局限；通过边缘端就近处理数据，降低了数据传输延迟和云端计算压力，实现了水泥生产过程碳排放的近实时计算与上报，满足水泥行业碳排放月度信息化存证和年度报告报送的时效性要求。
Smart Images
Figure CN122964773A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of carbon emission monitoring and calculation technology, and in particular to a method and system for calculating carbon emissions in the cement production process based on cloud-edge collaboration. Background Technology
[0002] The cement industry is a significant source of carbon emissions in the global industrial sector, accounting for as much as 27% of total industrial carbon emissions, second only to the steel industry. Carbon emissions during cement production mainly originate from two sources: the decomposition of raw material carbonates (such as limestone) and the combustion of fuels (such as coal). Carbon emissions from the decomposition of raw material carbonates account for approximately 60% to 70% of total emissions from cement production. As the national carbon emissions trading market gradually incorporates the cement industry into its regulatory scope, cement companies face an urgent need for accurate carbon emission data accounting. The Ministry of Ecology and Environment has issued technical specifications such as the "Guidelines for Enterprise Greenhouse Gas Emission Accounting and Reporting (Cement Industry)," clarifying the accounting boundaries and emission sources at both the clinker production facility and enterprise levels.
[0003] Currently, cement companies mainly rely on the following methods for carbon emission accounting: First, they rely on periodic verification by third-party verification agencies to obtain carbon emission data. However, this method has long data collection intervals, poor real-time performance, and difficulty in guaranteeing data accuracy. Second, they use a continuous emission monitoring system (CEMS) to measure flue gas concentration and flow rate using actual measurement methods to calculate carbon dioxide emissions. However, due to the high uncertainty in flue gas concentration and flow rate measurements, the accuracy of the measurement results is low. Third, they use emission factor methods or material balance methods to calculate emissions based on activity data. However, these methods rely on a large amount of manually entered activity data, which has problems such as data lag and difficulty in identifying abnormal data.
[0004] In recent years, cloud-edge collaborative technology has begun to be applied in the field of carbon emission monitoring. Among existing technologies, some solutions have proposed intelligent sensing and calculation methods for carbon emissions in cement enterprises based on cloud-edge collaboration. The basic idea is to collect carbon emission data in real time at the edge and send it to the cloud, where the cloud determines carbon emission quotas based on the data. However, this solution still only uses monitoring data from a single source for carbon emission calculation, lacking a multi-source data fusion and verification mechanism, resulting in insufficient data reliability; it does not quantify the measurement uncertainty of the carbon emission calculation results, making it difficult to meet the high data quality requirements of the carbon trading market; the model parameters at the edge and cloud cannot achieve dynamic closed-loop updates, and the model calculation accuracy decreases with changes in production processes; it lacks the ability to finely decompose and attribute carbon emissions from multiple processes in cement production, failing to identify the specific emission reduction potential of each process stage.
[0005] Therefore, proposing a cloud-edge collaborative carbon emission calculation method and system for cement production to address the difficulties in existing technologies is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0006] In view of this, the present invention provides a cloud-edge collaborative carbon emission calculation method and system for cement production process, which can improve the accuracy and reliability of carbon emission calculation through multi-source data fusion verification, uncertainty assessment and dynamic updating of cloud model, and meet the strict data quality requirements of the cement industry carbon trading market.
[0007] To achieve the above objectives, the present invention adopts the following technical solution: A method for calculating carbon emissions in a cement production process based on cloud-edge collaboration includes: S1. At the edge, multiple sensors deployed in various cement production processes collect multi-source initial monitoring data in real time. S2. The edge performs data cleaning and anomaly detection on the initial monitoring data from multiple sources to obtain preprocessed data from multiple sources. S3. The edge end uses multi-source preprocessed data to call the preset carbon emission calculation model library and uses multiple methods to calculate the carbon emissions of each process in parallel to obtain multi-path calculation results. S4. The edge end performs cross-validation on the multi-path calculation results to detect whether the deviation between different calculation methods exceeds the preset validation threshold. If the deviation between the calculation results of each path is less than the preset verification threshold, the edge terminal uses a weighted fusion algorithm to calculate the fused carbon emissions; if any deviation exceeds the preset verification threshold, the edge terminal generates a data anomaly alarm and sends it to the cloud. S5. Based on the propagation law of measurement uncertainty of multi-source preprocessed data at the edge end, the Monte Carlo simulation method is used to calculate the combined standard uncertainty of fused carbon emissions. S6. The edge device packages the carbon emission calculation results, including: the fused carbon emission amount, the composite standard uncertainty of the fused carbon emission amount, and the carbon emission amount of each process, and uploads them to the cloud. S7. The cloud performs global analysis based on the received data, including: predicting historical carbon emission sequence data based on long short-term memory neural networks to generate enterprise carbon emission trend prediction results; and identifying key emission processes in cement production based on the carbon emission amounts of each process. S8. Based on the carbon emission calculation results uploaded from the edge device, the cloud performs incremental learning training on the prediction model deployed in the cloud to generate optimized model parameters; the cloud then sends the optimized model parameters to the edge device, which updates the relevant parameters in the carbon emission calculation model library accordingly.
[0008] Optionally, in the above method, the multi-source initial monitoring data in S1 includes: Flue gas emission monitoring data, energy consumption data, material consumption data, and production process parameter data.
[0009] Optionally, in S2 of the above method, the edge performs data cleaning and anomaly detection on the initial monitoring data from multiple sources, specifically as follows: Calculate the deviation between the real-time data and the historical average of each data source. If the deviation exceeds the preset control limit, it is judged as abnormal data. For data that is determined to be abnormal, the edge is automatically marked and removed, and the abnormal event type and timestamp are recorded at the same time.
[0010] The above methods, optionally, include a carbon emission calculation model library in S3, which includes: emission factor method calculation model, material balance method calculation model and flue gas direct measurement method calculation model; It also includes: an alternative fuel carbon content correction model, which is used to correct carbon emissions based on the biomass carbon content ratio of alternative fuels when cement production lines use alternative fuels, deducting carbon emissions from biomass sources from total emissions.
[0011] Optionally, in S4 of the above method, a weighted fusion algorithm is used to calculate the fused carbon emissions, specifically as follows: The weight coefficients of the weighted fusion algorithm are dynamically determined based on the historical confidence level of the corresponding computation path; The historical confidence level of the calculation path is obtained by statistically analyzing the historical deviations between the calculation results of each calculation path and the results of third-party verification. The weighting coefficients are determined by: calculating the average relative deviation of each calculation path from the third-party verification results in the past accounting period. The smaller the average deviation, the higher the confidence level. The weighting coefficients are proportional to the confidence level, and the sum of the weighting coefficients of all paths is 1.
[0012] Optionally, in S5, the Monte Carlo simulation method is used to calculate the combined standard uncertainty of the aggregated carbon emissions, specifically as follows: Determine the probability distribution type and distribution parameters of each input quantity. Perform N random samplings on each input quantity, where N≥10000; The carbon emissions corresponding to each sampling are calculated based on the carbon emission calculation model. The standard deviation of the results from N calculations is used as the combined standard uncertainty.
[0013] Optionally, in S7, when generating enterprise carbon emission trend prediction results based on historical carbon emission sequence data using a long short-term memory neural network, carbon quota profit and loss prediction and carbon trading strategy suggestions are generated by combining the carbon emission benchmark of the cement industry and the national carbon emission trading market price.
[0014] A cloud-edge collaborative carbon emission calculation system for cement production processes, executing any one of the cloud-edge collaborative carbon emission calculation methods for cement production processes described above, characterized in that it includes: Edge devices, or industrial edge computing gateways, are deployed in the central control room or field cabinet room of a cement production line. The cloud platform is deployed on a remote server and communicates with edge devices. Edge devices include: The multi-source data acquisition module is used to collect multi-source initial monitoring data in real time through various sensors deployed in various processes of cement production; The data preprocessing module, connected to the multi-source data acquisition module, is used to perform data cleaning and anomaly detection on the initial monitoring data from multiple sources. The carbon emission calculation engine is connected to the data preprocessing module. The carbon emission calculation engine has a built-in carbon emission calculation model library, which is used to calculate the carbon emissions of each process in parallel using multiple methods. The data verification and fusion module is connected to the carbon emission calculation engine and is used to cross-verify the multi-path calculation results and generate the fused carbon emission using a weighted fusion algorithm. The uncertainty assessment module, connected to the data verification and fusion module, is used to calculate the combined standard uncertainty of the fused carbon emissions based on the propagation law of measurement uncertainty. The edge storage and communication module is connected to the data verification and fusion module and the uncertainty assessment module, respectively, and is used to cache calculation results locally and communicate data with the cloud platform. The cloud platform includes: The data receiving and storage module is connected to the edge storage and communication module to receive and store the carbon emission calculation results uploaded from the edge. The global analysis module, connected to the data receiving and storage module, is used to predict historical carbon emission sequence data based on long short-term memory neural networks and generate enterprise carbon emission trend prediction results. The process emission analysis module, connected to the data receiving and storage module, is used to identify key emission processes in cement production based on process-level emission decomposition data. The decision support module is connected to the global analysis module and the process emission analysis module respectively. It is used to compare the prediction results with the preset emission quotas and generate carbon emission management decision suggestions. The model training module, connected to the data receiving and storage module, is used to incrementally learn and train the cloud prediction model based on multiple batches of carbon emission calculation results uploaded from the edge device, generate optimized model parameters, and send them to the edge device.
[0015] As can be seen from the above technical solution, compared with the prior art, the present invention provides a cloud-edge collaborative carbon emission calculation method and system for cement production processes, which has the following beneficial effects: The present invention calculates carbon emissions in parallel by simultaneously employing three independent methods—emission factor method, material balance method, and direct flue gas measurement method—and cross-validates and weights the multi-path calculation results, effectively overcoming the systematic bias problem of single calculation methods, significantly improving the reliability and accuracy of carbon emission calculation results. When the calculation results are inconsistent, an alarm is automatically triggered, facilitating manual review and compensating for the shortcomings of existing technologies that only use a single data source. The present invention integrates an automated assessment function for carbon emission measurement uncertainty in the cloud-edge collaborative architecture. By propagating the measurement uncertainty of each input quantity and obtaining the combined standard uncertainty through Monte Carlo simulation, the carbon emission calculation results have a quantifiable confidence interval, meeting the stringent data quality requirements of the national carbon emission trading market and facilitating enterprises' carbon quota compliance. This invention provides reliable data support for carbon asset management. It utilizes the powerful computing capabilities of the cloud to incrementally train the carbon emission prediction model and periodically distributes the optimized model parameters to the edge to update the local calculation model. This achieves a complete closed loop of "edge data acquisition and calculation—cloud training and optimization—edge parameter update," ensuring that the carbon emission calculation model continuously optimizes with changes in production process parameters and data accumulation, avoiding calculation biases caused by model aging. By calculating carbon emissions for each major process in cement production and combining this with the cloud's key emission process identification function, this invention can accurately locate the process with the highest carbon emission intensity, providing a basis for enterprises to formulate targeted emission reduction plans and overcoming the limitation of existing technologies that only output total enterprise-level emissions. By processing data locally at the edge, it reduces data transmission latency and cloud computing pressure, achieving near real-time calculation and reporting of carbon emissions in the cement production process, meeting the timeliness requirements for monthly information storage and annual report submission of carbon emissions in the cement industry. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0017] Figure 1A flowchart of a cloud-edge collaborative carbon emission calculation method for cement production process provided by the present invention; Figure 2 A structural diagram of an edge device for a cloud-edge collaborative carbon emission calculation system for cement production processes, provided by this invention; Figure 3 This invention provides a cloud platform structure diagram for a cloud-edge collaborative carbon emission calculation system for cement production processes. Detailed Implementation
[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0019] In this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0020] Reference Figure 1 As shown, this invention discloses a method for calculating carbon emissions in a cement production process based on cloud-edge collaboration, including: S1. At the edge, multiple sensors deployed in various cement production processes collect multi-source initial monitoring data in real time. S2. The edge performs data cleaning and anomaly detection on the initial monitoring data from multiple sources, removing abnormal and duplicate data to obtain preprocessed data from multiple sources. S3. The edge end uses multi-source preprocessed data to call the preset carbon emission calculation model library and uses multiple methods to calculate the carbon emissions of each process in parallel to obtain multi-path calculation results. S4. The edge end performs cross-validation on the multi-path calculation results to detect whether the deviation between different calculation methods exceeds the preset validation threshold. If the deviation between the calculation results of each path is less than the preset verification threshold, the edge terminal uses a weighted fusion algorithm to calculate the fused carbon emissions; if any deviation exceeds the preset verification threshold, the edge terminal generates a data anomaly alarm and sends it to the cloud. S5. Based on the propagation law of measurement uncertainty of multi-source preprocessed data at the edge end, the Monte Carlo simulation method is used to calculate the combined standard uncertainty of fused carbon emissions. S6. The edge device packages the carbon emission calculation results, including: the fused carbon emission amount, the composite standard uncertainty of the fused carbon emission amount, and the carbon emission amount of each process, and uploads them to the cloud. S7. The cloud performs global analysis based on the received data, including: predicting historical carbon emission sequence data based on long short-term memory neural networks to generate enterprise carbon emission trend prediction results; and identifying key emission processes in cement production based on the carbon emission amounts of each process. S8. Based on the carbon emission calculation results uploaded from the edge device, the cloud performs incremental learning training on the prediction model deployed in the cloud to generate optimized model parameters; the cloud then sends the optimized model parameters to the edge device, which updates the relevant parameters in the carbon emission calculation model library accordingly.
[0021] Furthermore, in S1, the multi-source initial monitoring data includes: Flue gas emission monitoring data, energy consumption data, material consumption data, and production process parameter data.
[0022] Furthermore, in S2, the edge performs data cleaning and anomaly detection on the initial monitoring data from multiple sources, specifically as follows: Calculate the deviation between the real-time data and the historical average of each data source. If the deviation exceeds the preset control limit, it is judged as abnormal data. For data that is determined to be abnormal, the edge is automatically marked and removed, and the abnormal event type and timestamp are recorded at the same time.
[0023] Furthermore, in S3, the carbon emission calculation model library includes: emission factor method calculation model, material balance method calculation model, and flue gas direct measurement method calculation model; Among them, the emission factor method calculation model is used to calculate carbon emissions from fuel combustion based on fossil fuel consumption and emission factors; The material balance calculation model is used to calculate process carbon emissions based on the carbon content of raw materials and clinker; The direct flue gas measurement method calculation model is used to calculate the measured carbon emissions based on the CO2 concentration and flue gas flow rate in the flue gas. The carbon emission calculation model library also includes: an alternative fuel carbon content correction model, which is used to correct carbon emissions based on the biomass carbon content ratio of alternative fuels when cement production lines use alternative fuels, deducting carbon emissions from biomass sources from the total emissions.
[0024] Furthermore, in S4, a weighted fusion algorithm is used to calculate the fused carbon emissions, specifically as follows: The weight coefficients of the weighted fusion algorithm are dynamically determined based on the historical confidence level of the corresponding computation path; The historical confidence level of the calculation path is obtained by statistically analyzing the historical deviations between the calculation results of each calculation path and the results of third-party verification. The weighting coefficients are determined by: calculating the average relative deviation of each calculation path from the third-party verification results in the past accounting period. The smaller the average deviation, the higher the confidence level. The weighting coefficients are proportional to the confidence level, and the sum of the weighting coefficients of all paths is 1.
[0025] Furthermore, in S5, the combined standard uncertainty of the merged carbon emissions is calculated using the Monte Carlo simulation method, specifically as follows: Determine the probability distribution type and distribution parameters of each input quantity. Perform N random samplings on each input quantity, where N≥10000; The carbon emissions corresponding to each sampling are calculated based on the carbon emission calculation model. The standard deviation of the N calculation results is used as the combined standard uncertainty. Furthermore, the composite standard uncertainty of carbon emissions is integrated, which at least covers the propagation components of the measurement uncertainty of flue gas flow rate, CO2 concentration, fuel consumption, and material carbon content.
[0026] Furthermore, in S7, when generating corporate carbon emission trend prediction results based on historical carbon emission sequence data using long short-term memory neural networks, carbon quota profit and loss predictions and carbon trading strategy suggestions are generated by combining the carbon emission benchmark of the cement industry and the national carbon emission trading market price.
[0027] A cloud-edge collaborative carbon emission calculation system for cement production processes, executing any one of the cloud-edge collaborative carbon emission calculation methods for cement production processes described above, characterized in that it includes: Edge devices, or industrial edge computing gateways, are deployed in the central control room or field cabinet room of a cement production line. The cloud platform is deployed on a remote server and communicates with edge devices. Reference Figure 2 As shown, the edge device includes: The multi-source data acquisition module is used to collect multi-source initial monitoring data in real time through various sensors deployed in various processes of cement production; The data preprocessing module, connected to the multi-source data acquisition module, is used to perform data cleaning and anomaly detection on the initial monitoring data from multiple sources. The carbon emission calculation engine is connected to the data preprocessing module. The carbon emission calculation engine has a built-in carbon emission calculation model library, which is used to calculate the carbon emissions of each process in parallel using multiple methods. The data verification and fusion module is connected to the carbon emission calculation engine and is used to cross-verify the multi-path calculation results and generate the fused carbon emission using a weighted fusion algorithm. The uncertainty assessment module, connected to the data verification and fusion module, is used to calculate the combined standard uncertainty of the fused carbon emissions based on the propagation law of measurement uncertainty. The edge storage and communication module is connected to the data verification and fusion module and the uncertainty assessment module, respectively, and is used to cache calculation results locally and communicate data with the cloud platform. Reference Figure 3 As shown, the cloud platform includes: The data receiving and storage module is connected to the edge storage and communication module to receive and store the carbon emission calculation results uploaded from the edge. The global analysis module, connected to the data receiving and storage module, is used to predict historical carbon emission sequence data based on long short-term memory neural networks and generate enterprise carbon emission trend prediction results. The process emission analysis module, connected to the data receiving and storage module, is used to identify key emission processes in cement production based on process-level emission decomposition data. The decision support module is connected to the global analysis module and the process emission analysis module respectively. It is used to compare the prediction results with the preset emission quotas and generate carbon emission management decision suggestions. The model training module, connected to the data receiving and storage module, is used to incrementally learn and train the cloud prediction model based on multiple batches of carbon emission calculation results uploaded from the edge device, generate optimized model parameters, and send them to the edge device.
[0028] In one specific embodiment, a cloud-edge collaborative carbon emission calculation system for cement production process is provided. The system includes one or more edge devices deployed at the production site of a cement production enterprise, and a cloud platform deployed in a remote data center. The edge devices utilize industrial-grade edge computing gateways, deployed in the central control room of the cement production line. They are connected to various types of sensors on the production site via industrial Ethernet, including: CO2 concentration sensors and flue gas flow meters for monitoring emission parameters of kiln tail flue gas and preheater exhaust gas; fuel consumption metering devices for calculating the consumption of raw coal, alternative fuels, etc.; material metering devices for weighing and calculating raw materials, clinker, and blended materials; and sensors for process parameters such as temperature, pressure, and humidity. The cloud platform is deployed on public cloud or enterprise private cloud servers, and communicates securely with the edge devices of various cement production enterprises via the Internet or dedicated lines; The edge device polls and collects data from each sensor at intervals not exceeding 1 minute. The collected raw data includes: CO2 volume concentration C (%) in kiln tail flue gas, flue gas flow rate Q (Nm³ / h), raw coal consumption FC_coal (t), raw coal receiving basis lower heating value NCV (GJ / t), raw coal unit calorific value carbon content CC (tC / GJ), raw meal consumption G_raw (t), raw meal average carbon content C_raw (tC / t), clinker production G_clinker (t), clinker average carbon content C_clinker (tC / t), etc. At the edge, the above data is cleaned and anomaly detected: a moving average method based on a sliding window is used to smooth the continuous data and remove spike data caused by instantaneous sensor failures; for missing data, linear interpolation or neighbor-value filling is used to fill in the missing data; a statistical process control method is used to calculate the deviation between the real-time data and the historical data mean of each data source. If the deviation exceeds 3 times the standard deviation control limit, it is judged as abnormal data, automatically marked and removed, and the abnormal event type and timestamp are recorded for traceability. The edge calls the carbon emission calculation model library and uses three methods to calculate the carbon emissions in parallel within the same accounting period: emission factor method, material balance method, and direct flue gas measurement method. The edge end compares the carbon emissions E calculated by the above three methods pairwise and calculates the relative deviation; If all deviation values are less than the preset verification threshold (usually set to 10%), the data consistency is considered good. The weighted fusion algorithm is used to calculate the fused carbon emissions. The weight coefficients are dynamically determined based on the historical confidence of each method. Specifically, the average relative deviation of each calculation path from the third-party verification results in the past accounting period (such as a quarter) is statistically analyzed. The smaller the average deviation, the higher the confidence. The weight coefficients are proportional to the confidence, and the sum of the weight coefficients is 1. The historical confidence is dynamically updated as new data is accumulated. If any deviation value exceeds the preset verification threshold, the edge terminal will immediately generate a data anomaly alarm message, including the calculation method of the deviation exceeding the standard and the deviation value, and push it to the cloud and the on-site operation and maintenance terminal through the communication module to prompt manual intervention for verification; This embodiment uses the Monte Carlo simulation method to assess the uncertainty of the combined carbon emissions. The specific steps are as follows: Determine the probability distribution of each input quantity: Based on the calibration certificate of the measuring equipment and the on-site usage conditions, determine the probability distribution type and distribution parameters of each input quantity; for example, the measured value of flue gas CO2 concentration follows a normal distribution, and the standard deviation is taken from the repeatability index of the CEMS equipment; the measured value of fuel consumption follows a rectangular distribution, and the half-width is taken from the maximum permissible error of the electronic belt scale. Random Sampling and Calculation: Perform N independent random samplings (N=10000) on each input quantity, and call the carbon emission calculation model to calculate the corresponding carbon emission after each sampling. Statistical analysis: The standard deviation of the results of N calculations is used as the combined standard uncertainty; The edge device packages the fused carbon emissions, composite standard uncertainty, and emission decomposition data of each process level (such as emissions from raw material preparation, clinker calcination, and cement grinding processes, with the clinker calcination system being the core accounting unit) into a data package, adds a timestamp and enterprise identifier, and then uploads it to the cloud platform via an encrypted communication protocol. After receiving the data, the cloud platform performs the following global analysis functions: (1) Carbon emission trend prediction: The cloud-based Long Short-Term Memory Neural Network (LSTM) model uses the monthly carbon emission data of enterprises in the past 12 months as the input sequence to predict the carbon emission trend in the next 3 months. The LSTM model adopts a single hidden layer structure (the number of hidden layer neurons is 64). The input layer receives the monthly emission data of 12 time steps, and the output layer outputs the predicted value for the next 3 months. The model uses mean squared error as the loss function and is trained with Adam optimizer. The LSTM model effectively captures the long-term temporal dependence of carbon emission data through the forget gate, input gate and output gate structure. (2) Identification of key emission processes: Based on process-level emission decomposition data, the cloud calculates the proportion of carbon emissions of each process to the total emissions, sorts them in descending order, and identifies processes with a proportion exceeding 20% as key emission processes. At the same time, the cloud compares the carbon emission intensity of each process of the enterprise with the industry benchmark value and marks the processes that exceed the benchmark value. (3) Carbon trading strategy recommendations: The cloud platform combines the carbon emission trend forecast results, the company's existing carbon emission quotas and the carbon quota price trend in the national carbon market to predict the company's quota profit and loss during the compliance period and provide strategy recommendations for buying or selling carbon quotas. The cloud platform uses multiple batches of carbon emission calculation results uploaded from the edge devices every quarter to incrementally train the LSTM prediction model. Incremental learning avoids the high computational cost of retraining the entire dataset by fine-tuning the model with new data while retaining the existing model parameters. The cloud then packages and distributes the trained and optimized model parameters to the edge devices. After receiving new model parameters, the edge device automatically updates the relevant parameters in the local carbon emission calculation model library, including emission factor values and weighting coefficients, so that the next carbon emission calculation can be performed based on the latest model parameters.
[0029] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.
[0030] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for calculating carbon emissions in a cement production process based on cloud-edge collaboration, characterized in that, include: S1. At the edge, multiple sensors deployed in various cement production processes collect multi-source initial monitoring data in real time. S2. The edge performs data cleaning and anomaly detection on the initial monitoring data from multiple sources to obtain preprocessed data from multiple sources. S3. The edge end uses multi-source preprocessed data to call the preset carbon emission calculation model library and uses multiple methods to calculate the carbon emissions of each process in parallel to obtain multi-path calculation results. S4. The edge end performs cross-validation on the multi-path calculation results to detect whether the deviation between different calculation methods exceeds the preset validation threshold. If the deviation between the calculation results of each path is less than the preset verification threshold, the weighted fusion algorithm is used at the edge to calculate the fused carbon emissions. If any deviation exceeds the preset verification threshold, the edge device will generate a data anomaly alarm and send it to the cloud. S5. Based on the propagation law of measurement uncertainty of multi-source preprocessed data at the edge end, the Monte Carlo simulation method is used to calculate the combined standard uncertainty of fused carbon emissions. S6. The edge device packages the carbon emission calculation results, including: the fused carbon emission amount, the composite standard uncertainty of the fused carbon emission amount, and the carbon emission amount of each process, and uploads them to the cloud. S7. The cloud performs global analysis based on the received data, including: predicting historical carbon emission sequence data based on long short-term memory neural networks to generate enterprise carbon emission trend prediction results; and identifying key emission processes in cement production based on the carbon emission amounts of each process. S8. Based on the carbon emission calculation results uploaded from the edge device, the cloud performs incremental learning training on the prediction model deployed in the cloud to generate optimized model parameters; the cloud then sends the optimized model parameters to the edge device, which updates the relevant parameters in the carbon emission calculation model library accordingly.
2. The method for calculating carbon emissions in a cement production process based on cloud-edge collaboration as described in claim 1, characterized in that, In S1, the initial monitoring data from multiple sources includes: Flue gas emission monitoring data, energy consumption data, material consumption data, and production process parameter data.
3. The method for calculating carbon emissions in a cement production process based on cloud-edge collaboration according to claim 2, characterized in that, In S2, the edge performs data cleaning and anomaly detection on the initial monitoring data from multiple sources, specifically as follows: Calculate the deviation between the real-time data and the historical average of each data source. If the deviation exceeds the preset control limit, it is judged as abnormal data. For data that is determined to be abnormal, the edge is automatically marked and removed, and the abnormal event type and timestamp are recorded at the same time.
4. The method for calculating carbon emissions in a cement production process based on cloud-edge collaboration as described in claim 1, characterized in that, In S3, the carbon emission calculation model library includes: emission factor method calculation model, material balance method calculation model and flue gas direct measurement method calculation model; It also includes: an alternative fuel carbon content correction model, which is used to correct carbon emissions based on the biomass carbon content ratio of alternative fuels when cement production lines use alternative fuels, deducting carbon emissions from biomass sources from total emissions.
5. The method for calculating carbon emissions in a cement production process based on cloud-edge collaboration according to claim 1, characterized in that, In S4, a weighted fusion algorithm is used to calculate the fused carbon emissions, specifically: The weight coefficients of the weighted fusion algorithm are dynamically determined based on the historical confidence level of the corresponding computation path; The historical confidence level of the calculation path is obtained by statistically analyzing the historical deviations between the calculation results of each calculation path and the results of third-party verification. The weighting coefficients are determined by: calculating the average relative deviation of each calculation path from the third-party verification results in the past accounting period. The smaller the average deviation, the higher the confidence level. The weighting coefficients are proportional to the confidence level, and the sum of the weighting coefficients of all paths is 1.
6. The method for calculating carbon emissions in a cement production process based on cloud-edge collaboration according to claim 1, characterized in that, In S5, the combined standard uncertainty of the merged carbon emissions is calculated using the Monte Carlo simulation method, specifically as follows: Determine the probability distribution type and distribution parameters of each input quantity. Perform N random samplings on each input quantity, where N≥10000; The carbon emissions corresponding to each sampling are calculated based on the carbon emission calculation model. The standard deviation of the results from N calculations is used as the combined standard uncertainty.
7. The method for calculating carbon emissions in a cement production process based on cloud-edge collaboration as described in claim 1, characterized in that, In S7, when generating corporate carbon emission trend prediction results based on historical carbon emission sequence data using long short-term memory neural networks, carbon quota profit and loss predictions and carbon trading strategy suggestions are generated by combining the carbon emission benchmark of the cement industry and the national carbon emission trading market price.
8. A cloud-edge collaborative carbon emission calculation system for cement production, executing the cloud-edge collaborative carbon emission calculation method for cement production as described in any one of claims 1-7, characterized in that, include: Edge devices, or industrial edge computing gateways, are deployed in the central control room or field cabinet room of a cement production line. The cloud platform is deployed on a remote server and communicates with edge devices. Edge devices include: The multi-source data acquisition module is used to collect multi-source initial monitoring data in real time through various sensors deployed in various processes of cement production; The data preprocessing module, connected to the multi-source data acquisition module, is used to perform data cleaning and anomaly detection on the initial monitoring data from multiple sources. The carbon emission calculation engine is connected to the data preprocessing module. The carbon emission calculation engine has a built-in carbon emission calculation model library, which is used to calculate the carbon emissions of each process in parallel using multiple methods. The data verification and fusion module is connected to the carbon emission calculation engine and is used to cross-verify the multi-path calculation results and generate the fused carbon emission using a weighted fusion algorithm. The uncertainty assessment module, connected to the data verification and fusion module, is used to calculate the combined standard uncertainty of the fused carbon emissions based on the propagation law of measurement uncertainty. The edge storage and communication module is connected to the data verification and fusion module and the uncertainty assessment module, respectively, and is used to cache calculation results locally and communicate data with the cloud platform. The cloud platform includes: The data receiving and storage module is connected to the edge storage and communication module to receive and store the carbon emission calculation results uploaded from the edge. The global analysis module, connected to the data receiving and storage module, is used to predict historical carbon emission sequence data based on long short-term memory neural networks and generate enterprise carbon emission trend prediction results. The process emission analysis module, connected to the data receiving and storage module, is used to identify key emission processes in cement production based on process-level emission decomposition data. The decision support module is connected to the global analysis module and the process emission analysis module respectively. It is used to compare the prediction results with the preset emission quotas and generate carbon emission management decision suggestions. The model training module, connected to the data receiving and storage module, is used to incrementally learn and train the cloud prediction model based on multiple batches of carbon emission calculation results uploaded from the edge device, generate optimized model parameters, and send them to the edge device.