A real-time monitoring and optimization system for energy consumption and carbon emissions in steel enterprises

Abstract

This invention relates to the field of process monitoring and optimization technology in the steel industry, and discloses a real-time monitoring and optimization system for energy consumption and carbon emissions in steel enterprises. The system includes a data acquisition layer, a data processing and storage layer, an energy and carbon analysis and optimization layer, and an application interaction layer connected in sequence. The data acquisition layer is deployed at each stage of the entire process to collect process parameters related to energy consumption and carbon emissions in real time. The data processing and storage layer cleans and integrates the data, and calculates process-level and enterprise-level carbon emission indicators in real time based on the principle of material and energy balance. The energy and carbon analysis and optimization layer performs panoramic monitoring, trend analysis, dynamic optimization, and intelligent early warning based on real-time data and integrated models. The application interaction layer provides a human-machine interface and supports the issuance of optimization commands.

Description

Technical Field

[0001] This invention relates to the field of process monitoring and optimization in the steel industry, and in particular to a real-time monitoring and optimization system for energy consumption and carbon emissions in steel enterprises. Background Technology

[0002] Steel production involves enormous energy consumption and high carbon emission intensity, making the need for refined energy efficiency and emission management increasingly urgent for enterprises. Currently, the industry's commonly used methods for energy consumption and carbon emission statistics still rely primarily on manual recording and periodic aggregation. This results in significant data delays, and data is scattered across departments, making integration difficult. This extensive management model fails to capture the instantaneous impact of production fluctuations on energy consumption and carbon emissions, leading to delayed energy conservation and carbon reduction decisions and hindering effective guidance for production operations.

[0003] While the development of technologies such as the Internet of Things and big data has made real-time monitoring possible, existing solutions are mostly limited to data display of single processes or independent systems, lacking the ability to conduct collaborative analysis of energy and carbon flows covering the entire process of ironmaking, steelmaking, and rolling. Information silos are severe, and system early warning and optimization functions are insufficient. Summary of the Invention

[0004] This invention specifically relates to a real-time monitoring and optimization system for energy consumption and carbon emissions in steel enterprises. It aims to achieve synchronous real-time acquisition of energy consumption and key process parameters by deploying dedicated sensors in each process step, thus solving the data lag problem caused by traditional manual data recording. To achieve the above objectives, the specific technical solution of this invention's real-time monitoring and optimization system for energy consumption and carbon emissions in steel enterprises is as follows: A real-time monitoring and optimization system for energy consumption and carbon emissions in steel enterprises includes: a data acquisition layer configured to be deployed in the ironmaking, steelmaking, rolling and auxiliary processes of the steel enterprise; the data acquisition layer includes an energy consumption sensing unit for real-time acquisition of energy consumption data of each process and a carbon emission sensing unit for real-time acquisition of carbon emission-related process parameters of each process; the energy consumption data includes the instantaneous flow rate and cumulative amount of electricity, natural gas, coal and steam; the carbon emission-related process parameters include the input amount of raw materials and fuels, online analysis data of material composition, and exhaust gas temperature, flow rate and carbon dioxide concentration of key exhaust ports. The data processing and storage layer, which communicates with the data acquisition layer to obtain real-time data, includes a real-time data access module, a data cleaning and fusion module, an energy and carbon calculation engine, and a time-series database. The real-time data access module receives and parses real-time monitoring data from the data acquisition layer. The data cleaning and fusion module performs outlier identification, missing data processing, and timestamp alignment on the real-time monitoring data to generate standardized process-level energy and carbon data units. The energy and carbon calculation engine, based on the principles of material and energy balance, incorporates a real-time carbon emission calculation model to calculate the instantaneous carbon emission rate, cumulative carbon emission, and carbon emission intensity per ton of product for each process based on the process-level energy and carbon data units. The time-series database stores all historical and real-time monitoring and calculation results. The energy and carbon analysis and optimization layer, connected to the data processing and storage layer, acquires processed data and calculation results. It includes an energy and carbon panoramic monitoring module, an energy efficiency and carbon emission analysis module, a dynamic optimization model library, and an intelligent early warning module. The energy and carbon panoramic monitoring module is configured to display a full-process dynamic energy flow diagram, carbon flow diagram, and key performance indicator dashboard. The energy efficiency and carbon emission analysis module is configured to perform trend analysis, correlation analysis, and benchmarking analysis on energy consumption and carbon emission data. The dynamic optimization model library integrates a machine learning-based energy consumption prediction model and a multivariate optimization algorithm with energy efficiency improvement and minimum carbon emissions as optimization objectives, used to generate real-time optimization operation suggestions. The intelligent early warning module is configured to trigger tiered early warning signals based on dynamic thresholds and trend predictions. The application interaction layer, connected to the energy and carbon analysis and optimization layer, provides a human-computer interaction interface for different user roles, enabling visual configuration, data drilling, report generation, and the issuance of optimization commands. It can also issue optimization commands to the production control system.

[0005] Furthermore, the carbon emission sensing unit in the data acquisition layer specifically includes a non-dispersive infrared online carbon dioxide concentration analyzer deployed on the flue gas ducts of key emission sources such as coke ovens, blast furnaces, converters, and heating furnaces, as well as a nuclear scale or belt scale for monitoring the conveying volume of pulverized coal, coke, and ferroalloy solid materials; the energy consumption sensing unit includes intelligent flow meters and energy meters for monitoring electricity, gas, and steam.

[0006] Furthermore, when performing data cleaning, the data cleaning and fusion module uses a threshold judgment method based on process knowledge rules and a sliding window statistical method to identify abnormal data, and uses forward filling or linear interpolation methods to fill in the data. At the same time, it uses network time protocol to ensure that the data timestamps of the whole system are synchronized.

[0007] Furthermore, the real-time calculation model in the energy carbon calculation engine is specifically configured as follows: for carbon emissions generated by fuel combustion, the actual carbon content based on real-time fuel composition analysis is multiplied by the dynamic oxidation rate calculated based on real-time monitoring of flue gas composition; for carbon emissions generated by process chemical reactions, a model based on the real-time stoichiometric relationship between material input and output is used for calculation, and the carbon emissions of each process are summarized to the enterprise level in real time.

[0008] Furthermore, the multivariate optimization algorithm in the dynamic optimization model library is a hybrid optimization algorithm based on the fusion of process mechanism model and data-driven model. It solves a multi-objective programming problem under the constraints of equipment safety, product quality and production plan, and outputs a set of optimal set values ​​for blast furnace oxygen enrichment rate, hot blast temperature and air-fuel ratio of steel rolling heating furnace.

[0009] Furthermore, the early warning triggering mechanism of the intelligent early warning module includes: comparing the real-time carbon emission intensity with a static threshold set based on historical best levels; comparing it with a dynamic threshold based on dynamic adjustments to production rhythm; and using a prediction model in the energy efficiency and carbon emission analysis module or the dynamic optimization model library to issue an early warning for trends that may exceed the threshold.

[0010] Furthermore, the system also includes an energy and carbon digital twin module, which is connected to the data processing and storage layer and the energy and carbon analysis and optimization layer. The energy and carbon digital twin module integrates the equipment's three-dimensional model, real-time data, and mechanism model to construct a digital twin model of key energy-consuming equipment and carbon emission sources, which is used to simulate and verify the energy and carbon performance under different optimization strategies.

[0011] Furthermore, the optimization instruction issuance function of the application interaction layer is configured to automatically encapsulate the optimization operation suggestions after manual confirmation into standard format instructions, and securely issue them to the production execution system or distributed control system through OPC UA or a specific API interface.

[0012] Furthermore, the system also includes a unified data service and application interface layer, which connects to the data processing and storage layer to obtain data and provides data services to various modules in the energy and carbon analysis and optimization layer. The unified data service and application interface layer is configured to encapsulate the data access functions of the data processing and storage layer and provide standardized data query, indicator calculation and model call API services for the energy and carbon panoramic monitoring module, energy efficiency and carbon emission analysis module and external third-party systems.

[0013] Furthermore, the machine learning models in the dynamic optimization model library have online learning capabilities, enabling them to automatically adjust model parameters based on actual energy and carbon data feedback after the optimization instructions are executed, thereby achieving continuous self-evolution of prediction and optimization capabilities.

[0014] Compared with existing technologies, the advantages of this invention are as follows: By deploying dedicated sensors in each process step, this invention achieves synchronous real-time acquisition of energy consumption and key process parameters, solving the data lag problem caused by traditional manual data recording. The data processing layer can automatically clean, align, and store multi-source data to form standardized data units. Based on the principle of material and energy balance, the carbon emission calculation engine can dynamically calculate the carbon emission intensity of each process and the entire process using real-time acquired fuel composition, flue gas data, etc., improving the calculation accuracy from previous experience-based estimations to real-time calculations based on actual data.

[0015] The invention's built-in prediction and optimization model can dynamically optimize key process parameters under production constraints and generate specific operational suggestions to assist production personnel in achieving energy efficiency improvements and carbon emission reductions. The digital twin module provides offline verification methods for optimization strategies, reducing the risk of directly adjusting production. Simultaneously, an intelligent early warning mechanism can provide tiered warnings for abnormal situations. Finally, through a unified application interaction layer, optimization commands can be securely issued to the production control system. Attached Figure Description

[0016] Figure 1 This is a system overview and data flow diagram of the present invention; Figure 2 This is the core calculation and optimization decision-making flowchart of the present invention; Figure 3 This is a flowchart of the carbon trading strategy execution sub-process of the present invention; Figure 4 This is a flowchart of the production control strategy execution and feedback sub-process of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the following description, in conjunction with preferred embodiments and appendices, provides further details. Figure 1-4 This invention will be described in detail. This embodiment is only for explaining the invention and is not intended to limit the scope of protection of the invention.

[0018] Example 1 This embodiment details a real-time monitoring and optimization system for energy consumption and carbon emissions in an iron and steel enterprise, comprising: a data acquisition layer configured to be deployed in the ironmaking, steelmaking, rolling, and auxiliary processes of the iron and steel enterprise; the data acquisition layer including energy consumption sensing units for real-time acquisition of energy consumption data for each process and carbon emission sensing units for real-time acquisition of carbon emission-related process parameters for each process; the energy consumption data including instantaneous flow and cumulative amounts of electricity, natural gas, coal, and steam; and the carbon emission-related process parameters including the input of raw materials and fuels, online analysis data of material composition, and exhaust gas temperature, flow rate, and carbon dioxide concentration at key exhaust ports; and a data processing and storage layer communicatively connected to the data acquisition layer to obtain real-time data, comprising a real-time data access module, a data cleaning and fusion module, an energy and carbon calculation engine, and a time-series database; the real-time data access module receiving and parsing real-time monitoring data from the data acquisition layer; the data cleaning and fusion module performing outlier identification, missing data processing, and timestamp alignment on the real-time monitoring data to generate standardized process-level energy and carbon data units; and the energy and carbon calculation engine, based on the principles of material balance and energy balance, incorporating a real-time carbon emission calculation... The calculation model is used to calculate the instantaneous carbon emission rate, cumulative carbon emission, and carbon emission intensity per ton of product in real time based on the process-level energy and carbon data units. The time-series database is used to store all historical and real-time monitoring and calculation results. The energy and carbon analysis and optimization layer is connected to the data processing and storage layer to obtain processed data and calculation results. It includes an energy and carbon panoramic monitoring module, an energy efficiency and carbon emission analysis module, a dynamic optimization model library, and an intelligent early warning module. The energy and carbon panoramic monitoring module is configured to display the dynamic energy flow diagram, carbon flow diagram, and key performance indicator dashboard of the entire process. The energy efficiency and carbon emission analysis module is configured to... The system is configured to perform trend analysis, correlation analysis, and benchmarking analysis on energy consumption and carbon emission data. The dynamic optimization model library integrates a machine learning-based energy consumption prediction model and a multivariate optimization algorithm with energy efficiency improvement and minimum carbon emissions as optimization objectives, used to generate real-time optimization operation suggestions. The intelligent early warning module is configured to trigger tiered early warning signals based on dynamic thresholds and trend predictions. An application interaction layer, connected to the energy and carbon analysis and optimization layer, provides a human-machine interface for different user roles, enabling visual configuration, data drilling, report generation, and optimization command issuance, and can also issue optimization commands to the production control system. In this embodiment, the system is specifically constructed according to its architecture. At the steel plant site, a large number of smart energy meters, gas flow meters, belt scales, nuclear scales, and non-dispersive infrared gas analyzers are installed on the key equipment pipelines of sintering, coking, ironmaking, steelmaking, rolling, and power processes according to the design drawings. These sensors together constitute the data acquisition layer, continuously collecting various types of data at a second-level frequency. This real-time data is transmitted to a server cluster located in the central computer room through an industrial ring network.The server cluster forms the data processing and storage layer. Real-time database software is deployed within this layer to receive and parse data streams. A dedicated data cleaning program automatically identifies and corrects outliers based on preset process rules (such as the normal range of blast furnace blast temperature). The processed, well-organized data is then stored in a time-series database. Simultaneously, an energy and carbon calculation service deployed in the same cluster dynamically calculates the instantaneous energy consumption and carbon emission indicators for each process at minute-by-minute intervals, based on real-time fuel composition (such as received carbon content) and precise material metering. In the energy and carbon analysis and optimization layer, a separate analysis server is deployed. The software platform running on this server retrieves data from the time-series database and dynamically displays the plant's energy and carbon flow graphs on a large monitoring screen. The background analysis model continuously performs trend calculations and optimization solutions. The application interaction layer is represented by a web client interface on the operator's workstation. Engineers configure parameters, view analysis reports, and, after confirmation, send optimization settings to the production control system.

[0019] In this embodiment, the equipment selection and installation for the data acquisition layer are as follows: At key emission points such as coke oven flues, blast furnace hot blast stove chimneys, converter primary dust removal flues, and rolling mill heating furnace flues, ABB or Siemens brand non-dispersive infrared (NDIR) online carbon dioxide analyzers are installed to directly and continuously monitor the percentage of CO2 concentration in the flue gas. Schenck nuclear scales are installed on belt conveyors for pulverized coal injection, coke conveying, and ferroalloy loading to continuously measure solid materials online. For electricity consumption, multi-functional energy meters are installed in the high-voltage power distribution rooms of each process; for natural gas, oxygen, and steam, Emerson series vortex flow meters are installed on the main and branch pipes. All instrument signals are connected to a nearby PLC or data acquisition station and then uploaded to the system network. Direct measurement of CO2 concentration combined with flue gas flow provides crucial direct data support for real-time carbon emission calculation, while accurate measurement of solid materials provides the foundation for material balance calculation.

[0020] In this embodiment, the specific workflow of the data cleaning and fusion module is as follows: The real-time data access module receives tens of thousands of measurement point data points per second. The cleaning process first performs "threshold judgment." For example, if the blast furnace hot blast temperature data instantaneously exceeds 1350°C or falls below 800°C, it is marked as abnormal because this is impossible in the process. Secondly, "sliding window statistics" are performed. The program continuously analyzes the data from the same measurement point over the past 5 minutes. If a data point deviates from the moving average by more than 3 standard deviations, it is also marked. For missing data identified as abnormal or due to communication interruption, the system performs "forward filling" (using the previous valid value) or "linear interpolation" according to a strategy. Simultaneously, all data acquisition stations and servers synchronize with the central clock server via the NTP protocol to ensure the uniformity of data timestamps across the entire system. This embodiment effectively handles common problems in industrial settings such as signal jumps, noise interference, and communication interruptions through an automated data cleaning and synchronization mechanism.

[0021] In this embodiment, the core calculation model of the carbon calculation engine operates as follows: For carbon emissions from the combustion of fuels such as coal and gas, the system obtains the carbon content of the coal fed into the furnace in real time (updated daily or in real time) from the data platform, and obtains the oxygen content of the flue gas at the boiler or heating furnace outlet in real time. It then dynamically calculates the carbon oxidation rate under the current operating conditions using a formula, thereby calculating the CO2 generation from the emission source in real time. For carbon emissions from chemical reactions in processes such as steelmaking, the system reads the input quantities of materials such as molten iron, scrap steel, and lime, as well as the output quantities of molten steel and slag in real time. Based on the material balance of major elements such as iron, and combined with empirical output coefficients, it calculates the amount of CO2 generated by the chemical reaction in this process in real time. The carbon emissions from all processes are accumulated every second, and the total carbon emissions and intensity at the plant level are displayed in real time on the interface. This embodiment changes the traditional method of using fixed emission factors or monthly inventory checks for rough estimation, and achieves dynamic and accurate calculation based on real-time operating data and process mechanisms. By using dynamic carbon oxidation rate and real-time material balance, the accounting results can more accurately reflect the instantaneous impact of production fluctuations on carbon emissions.

[0022] In this embodiment, the hybrid optimization algorithm in the dynamic optimization model library is specifically applied as follows: Taking the blast furnace process as an example, the optimization model combines the mechanistic model of blast furnace smelting (such as heat balance and material balance) and the data-driven model (such as a fuel ratio prediction neural network trained based on historical data). In each calculation cycle, the model receives hundreds of real-time parameters such as the current blast temperature, blast pressure, top pressure, pulverized coal injection rate, and top gas composition. Under set constraints (such as the upper limit of blast temperature, the theoretical combustion temperature range, and pig iron quality requirements), it runs a multi-objective optimization algorithm (such as NSGA-II) to optimize for "lowest fuel consumption per ton of iron" and "lowest carbon emissions per ton of iron". Finally, the algorithm outputs a set of suggested settings, such as slightly adjusting the oxygen enrichment rate from the current 2.5% to 2.7% and increasing the hot blast temperature from 1180°C to 1185°C.

[0023] In this embodiment, the triggering logic of the intelligent early warning module is as follows: The system sets three early warning lines for the carbon emission intensity per ton of steel for each process. The first is the "static red line," set at 110% of the average of the best three months historically; exceeding this value triggers a Level 1 (yellow) warning. The second is the "dynamic yellow line," a reasonable upper limit dynamically calculated based on the current shift's production plan (e.g., steel grade) and current output; exceeding this line triggers a Level 2 (orange) warning. The third is the "trend warning," where the module uses a simplified time series model to predict the carbon emission intensity for the next hour; if the predicted value exceeds the "static red line," a Level 3 (blue) pre-alarm is triggered in advance. All warnings are communicated to relevant engineers via interface flashing, sound, and push notifications.

[0024] In this embodiment, the energy and carbon digital twin module is deployed on a dedicated graphics server. Based on the steel plant's 3D design model, it integrates real-time data from the data processing and storage layer, and incorporates simplified mechanistic models (such as heat transfer and combustion reaction kinetic equations) of key equipment like blast furnaces and heating furnaces. On the engineer's client, a virtual "blast furnace twin" can be accessed, whose appearance, internal temperature field, and airflow distribution are synchronized with the physical blast furnace or simulated through model calculations. Engineers can adjust parameters such as "oxygen enrichment rate" and "coal injection rate" on this twin. The system quickly simulates the changing trends of key indicators such as fuel ratio and carbon emissions over a future period after the adjustment and displays the comparison results in chart form.

[0025] In this embodiment, the instruction issuance function of the application interaction layer is implemented as follows: After the optimization model calculates a set of optimal air-fuel ratio setpoints for the rolling mill heating furnace, these recommended values ​​are displayed as prominent cards on the monitoring screen of the rolling mill operator. After the operator confirms that the process conditions allow, they click the "Issue" button on the card. The system then automatically encapsulates these setpoints and the associated rolling plan number into a standardized instruction package conforming to the OPC UA specification, and securely sends it to the distributed control system (DCS) server of the rolling mill through a dedicated industrial firewall. After receiving the package, the DCS server issues it as a setpoint to the heating furnace combustion controller, automatically completing the parameter adjustment.

[0026] In this embodiment, the unified data service and application interface layer is deployed as a set of independent microservices. This layer encapsulates all direct read and write operations to the time-series database below (the data processing and storage layer) and provides a unified RESTful API to the outside world. For example, the energy and carbon monitoring module needs to call the "GetRealtimeEnergyFlow" API to obtain real-time energy flow data, the energy efficiency and carbon emission analysis module calls the "CalculateCarbonIntensity" API to calculate the carbon intensity of the past 24 hours, and the company's ERP system can synchronously obtain the latest warning information through the "GetWarningList" API. All data access and calculation logic is centrally managed and maintained in this layer.

[0027] In this embodiment, the machine learning models (such as neural networks for predicting coke ratio) in the dynamic optimization model library have online learning capabilities. Each time an optimization suggestion (such as adjusting the ore-coke ratio in a blast furnace) is issued and executed, the system continuously collects actual production data (such as actual coke ratio, molten iron temperature, etc.) for the following hours. These new input-output data pairs are automatically added to the model's training dataset. The model service automatically initiates a retraining process once a week during off-peak business hours, updating the neural network weights using all historical data containing new data, thus enabling the model's predictive ability to adaptively adjust to the slow changes in production conditions and equipment status. Traditional optimization models, once deployed, may experience performance degradation over time and with changes in production conditions. The online learning mechanism allows the model to continuously optimize itself using the latest production feedback data, ensuring that prediction and optimization results always maintain high accuracy.

[0028] Example 2 This embodiment provides a specific implementation of the system in a long-process integrated steel enterprise (covering the entire process of coking, sintering, ironmaking, steelmaking and rolling).

[0029] At the data acquisition level, in addition to the basic sensors described in Implementation Method 1, this embodiment particularly strengthens the tracking of carbon-containing material flows. More weighing and composition detection devices (such as online neutron activation analyzers) are deployed at the material inlets (e.g., coke oven coal towers, blast furnace ore bins, converter high-level silos) and outlets (e.g., coke bins, molten iron ladles, steel ladles) of each process to ensure accurate measurement of key material flows such as iron and carbon. The energy-carbon calculation engine in the data processing layer constructs a material and carbon balance model covering the entire plant. This model not only calculates carbon emissions at each process but also tracks the migration and transformation of carbon elements in real time throughout the entire process. It displays in real time the specific destination and distribution of raw materials such as purchased coal, coke, and scrap steel, to intermediate products such as coke, molten iron, and molten steel, ultimately emitting carbon dioxide or entering finished / by-product processes.

[0030] At the optimization level, the algorithms in the dynamic optimization model library are designed specifically for the characteristics of long processes. For example, it can perform coupled optimization across processes: when peak electricity prices are predicted for a period of time, the optimization algorithm may suggest appropriately increasing coke load in the ironmaking process to increase blast furnace output, while simultaneously suggesting using more scrap steel in the steelmaking process to reduce molten iron consumption, and dynamically adjusting the ratio of self-generated electricity to grid-purchased electricity. Its goal is to minimize the total cost of the entire process (not just process energy consumption) and the net carbon emissions of the entire process while ensuring the overall plant production plan. The application interaction layer provides a "Full Process Optimization" view, which centrally displays the optimization suggestions across processes and their expected benefits.

[0031] Example 3 This embodiment provides a specific implementation of the system in a short-process electric arc furnace steel plant.

[0032] To address the intermittent nature and high impact loads of electric arc furnace steelmaking, the data acquisition layer is equipped with a higher-frequency power quality analyzer and a continuous emission monitoring system (CEMS). The CEMS directly, in real-time, and continuously measures the concentrations and flow rates of CO2, CO, and O2 in the electric arc furnace flue gas, providing the most direct first-hand data for carbon emission accounting and cross-checking the results with the material balance method. Simultaneously, the system accesses power generation data from distributed energy sources such as photovoltaic and wind power within the plant, as well as the real-time carbon intensity factor from the power grid.

[0033] The carbon calculation engine employs a hybrid accounting model of "primarily direct measurement, supplemented by material balance verification" to calculate the carbon emissions of each steel furnace in real time. The model combines grid carbon intensity prediction, time-of-use electricity pricing, and on-site renewable energy generation prediction to minimize "electricity costs + carbon costs." It optimizes operations such as the electric arc furnace's smelting power curve, transformer tap settings, and the activation of backup natural gas burners, generating a "green steelmaking" scheduling scheme. A digital twin module establishes electrical and smelting models of the electric arc furnace to simulate and verify the impact of different power curves on power consumption per ton of steel, furnace life, and carbon emissions.

[0034] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. A real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise, characterized in that, The application relates to a steel enterprise energy-carbon data acquisition and analysis system. The data acquisition layer is arranged in the ironmaking, steelmaking and rolling processes of a steel enterprise, and comprises energy consumption sensing units for acquiring energy consumption data of each process in real time and carbon emission sensing units for acquiring carbon emission related process parameters of each process in real time; the energy consumption data comprises instantaneous flow and cumulative amount of electric power, natural gas, coal and steam; the carbon emission related process parameters comprise input amount of raw materials and fuels, online analysis data of material composition, and exhaust gas temperature, flow and carbon dioxide concentration of key exhaust ports; The data processing and storage layer is connected with the data acquisition layer to acquire real-time data, and comprises a real-time data access module, a data cleaning and fusion module, an energy-carbon calculation engine and a time series database; the real-time data access module receives and analyzes real-time monitoring data from the data acquisition layer; the data cleaning and fusion module identifies abnormal values, processes missing data and aligns time stamps, and generates standardized process level energy-carbon data units; the energy-carbon calculation engine is internally provided with a carbon emission real-time calculation model according to material balance and energy balance principles, and is used for calculating instantaneous carbon emission rate, cumulative carbon emission amount and carbon emission intensity per ton of product of a process according to the process level energy-carbon data units; and the time series database is used for storing all historical and real-time monitoring and calculation results; The energy-carbon analysis and optimization layer is connected with the data processing and storage layer to acquire processed data and calculation results, and comprises an energy-carbon panoramic monitoring module, an energy efficiency and carbon emission analysis module, a dynamic optimization model library and an intelligent early warning module; the energy-carbon panoramic monitoring module is configured to display dynamic energy flow diagrams, carbon flow diagrams and key performance indicator dashboards; the energy efficiency and carbon emission analysis module is configured to perform trend analysis, correlation analysis and benchmarking analysis on energy consumption and carbon emission data; the dynamic optimization model library is integrated with an energy consumption prediction model based on machine learning and a multi-element optimization algorithm taking energy efficiency improvement and minimum carbon emission as optimization objectives, and is used for generating real-time optimization operation suggestions; The intelligent early warning module is configured to trigger graded early warning signals according to dynamic thresholds and trend prediction; The application interaction layer is connected with the energy-carbon analysis and optimization layer, and provides human-computer interaction interfaces with visual configuration, data drilling, report generation and optimization instruction issuing for different role users, and can issue optimization instructions to a production control system. 2.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The carbon emission sensing units in the data acquisition layer specifically comprise non-dispersive infrared online carbon dioxide concentration analyzers arranged on flues of key emission sources of coke ovens, blast furnaces, converters and heating furnaces, and nuclear scales or belt scales for monitoring conveying amounts of coal powder, coke and iron alloy solid materials; the energy consumption sensing units comprise intelligent flow meters and electric energy meters for monitoring electric power, gas and steam. 3.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, When the data cleaning and fusion module performs data cleaning, threshold judgment method based on process knowledge rules and sliding window statistical method are adopted to identify abnormal data, and forward filling or linear interpolation method is adopted for data filling; meanwhile, network time protocol is used to ensure synchronization of data time stamps of the whole system. 4.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The real-time calculation model in the carbon calculation engine is specifically configured to: for carbon emissions generated by fuel combustion, actual carbon content based on real-time fuel component analysis is multiplied by a dynamic oxidation rate calculated according to real-time monitoring of flue gas components; for carbon emissions generated by process chemical reactions, a model based on real-time material input and output stoichiometric relationship is used for calculation, and carbon emissions of each process are real-time aggregated to the enterprise level. 5.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The multi-element optimization algorithm in the dynamic optimization model library is a hybrid optimization algorithm based on the fusion of process mechanism model and data-driven model, which outputs an optimal set of values such as blast furnace oxygen enrichment rate, hot air temperature, and rolling steel heating furnace air-fuel ratio by solving a multi-objective planning problem under the constraints of equipment safety, product quality, and production plan. 6.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The early warning triggering mechanism of the intelligent early warning module includes: comparing the real-time carbon emission intensity with a static threshold value based on the historical best level setting; comparing it with a dynamic threshold value based on the dynamic adjustment of production rhythm; and through the prediction model in the energy efficiency and carbon emission analysis module or the dynamic optimization model library, a warning is given for the trend that may exceed the threshold value. 7.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The system further comprises an energy-carbon digital twin module connected with the data processing and storage layer and the energy-carbon analysis and optimization layer, which integrates device three-dimensional models, real-time data and mechanism models to build digital twin models of key energy-consuming devices and carbon emission sources for simulating and verifying energy-carbon performance under different optimization strategies. 8.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The optimization instruction issuing function of the application interaction layer is configured to automatically package the optimization operation suggestions confirmed by humans into standard format instructions and issue them to the production execution system or distributed control system through OPC UA or specific API interface. 9.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The system further comprises a unified data service and application interface layer connected with the data processing and storage layer to obtain data and provide data services for each module in the energy-carbon analysis and optimization layer; the unified data service and application interface layer is configured to encapsulate the data access function of the data processing and storage layer, and provide standardized data query, index calculation and model calling API services for the energy-carbon panoramic monitoring module, the energy efficiency and carbon emission analysis module and external third-party systems. 10.The real-time monitoring and optimization system for energy consumption and carbon emission of a steel enterprise according to claim 1, characterized in that, The machine learning model in the dynamic optimization model library has online learning function, which can automatically adjust model parameters according to the actual energy-carbon data feedback after optimization instruction execution, realize continuous self-evolution of prediction and optimization ability.

Images