An industrial internet system for full-process decision support of molten steel production

Abstract

The application provides an industrial internet system for smelting steel production whole-process decision support, comprising: a device twin module, which constructs a device digital twin; a production twin module, which constructs a whole-process production digital twin platform; a unified data module, which hierarchically stores unified industrial data; a production organization module, which constructs an industrial knowledge graph and a scheduling task set, and obtains an optimal production organization scheme; a risk prediction and maintenance module, which predicts a future failure probability distribution of equipment and a remaining service life of equipment, constructs an equipment risk portrait, and generates a predictive maintenance execution plan; a quality prediction and control module, which performs quality control and obtains a quality control scheme; and a decision support module, which performs whole-process decision support. The application realizes intelligent twin collaboration from equipment to decision in steel production by constructing a whole-process industrial internet platform, and significantly improves production efficiency, quality consistency, equipment reliability and green low-carbon level.

Description

Technical Field

[0001] This invention relates to the field of steel production management technology, and in particular to an industrial internet system for decision support throughout the entire process of steel melting production. Background Technology

[0002] As a crucial pillar of the national economy, the steel industry is facing core challenges in cost reduction and efficiency improvement, green and low-carbon transformation, and intelligent transformation. Long-process steel enterprises involve numerous production stages, dense equipment, and large amounts of data, making it difficult for traditional production management models to achieve transparent and refined control throughout the entire process. Industrial Internet platforms, as the core carrier of intelligent manufacturing, can break down data barriers between equipment, production, management, and the supply chain, providing technical support for the digital transformation of steel enterprises, helping them overcome development bottlenecks, and achieve high-quality development.

[0003] Steelmaking enterprises, particularly those focused on steel smelting and rolling, are characterized by long processes, high energy consumption, dense equipment, and strong process continuity. Existing production management models for these enterprises typically include plan-driven production management, process-by-process and workshop-by-workshop management, experience-based on-site scheduling, and information-based management. However, these models suffer from several drawbacks: 1) Data silos: Equipment, production, and operational data are scattered and cannot be efficiently integrated. 2) Lack of a unified twin model: Each process is modeled independently, lacking end-to-end status coordination. 3) Lagging quality control: Quality inspection relies on finished product sampling, making real-time prediction and intervention impossible. 4) Lack of holistic decision-making: Each business module makes independent decisions, lacking a multi-objective collaborative mechanism. 5) Weak low-carbon and environmental protection capabilities: Energy and carbon management is extensive, lacking refined scheduling capabilities. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide an industrial internet system for decision support throughout the entire steel production process. By constructing a full-process industrial internet platform, intelligent twin collaboration between equipment and decision-making in steel production can be achieved, significantly improving production efficiency, quality consistency, equipment reliability, and green and low-carbon levels.

[0005] To achieve the above objectives, the present invention provides the following solution: an industrial internet system for decision support throughout the entire steelmaking process, comprising: The device twin module is used to acquire fully connected device data and construct a digital twin of the device based on the fully connected device data; The production twin module is used to integrate and construct a full-process production data chain, process-level digital twin model, full-process state space model, and process constraint model based on the steel production process, to obtain a full-process production digital twin platform. A unified data module is used to combine equipment digital twins, the end-to-end production digital twin platform, and acquired operational data to generate a standardized dataset, and then store the standardized dataset hierarchically to obtain unified industrial data. The production organization module is used to construct an industrial knowledge graph and a set of scheduled tasks based on the unified industrial data, and to optimize the set of scheduled tasks by combining a dynamic scheduling mechanism to obtain the optimal production organization scheme. The risk prediction and maintenance module is used to predict the future failure probability distribution and remaining usable life of the equipment based on the equipment digital twin, the unified industrial data and the optimal production organization scheme, and to construct an equipment risk profile to generate a predictive maintenance execution plan. The quality prediction and control module is used to predict quality risks and make quality forecasts based on the equipment digital twin, the unified industrial data, and the optimal production organization scheme, so as to complete quality control and obtain a quality control scheme. The decision support module is used to combine the optimal production organization plan, the predictive maintenance execution plan, the quality control plan, the low-energy and low-carbon production plan, and the risk prevention and control plan to make comprehensive decisions and restructure business processes throughout the entire process, thereby completing full-process decision support. The equipment twin module, the production twin module, the unified data module, the production organization module, the risk prediction and maintenance module, the quality prediction and control module, and the decision support module are interconnected.

[0006] Optionally, the device twin module includes: The equipment data unit is used to construct an equipment access list. Based on the equipment access list, it collects the operating status data, process data, equipment health data, control system data, maintenance and repair data, and alarm event data of each type of equipment to obtain the full equipment connection data. The equipment modeling unit is used to construct the plant area number, production line number, equipment type, equipment serial number and component number based on the fully connected data of the equipment to obtain a unified equipment coding system. According to the four-level structure of system, equipment, component and measuring point, hierarchical modeling is performed on each piece of equipment to obtain the equipment hierarchical model. The temperature health score, vibration health score, current health score, pressure health score and maintenance status score are weighted and integrated to obtain the equipment health index. A digital twin unit, combining the device's static profile, the device's fully connected data, the unified coding system, the device's hierarchical model, and the device's health index, yields a digital twin of the device.

[0007] Optionally, the equipment access list includes ironmaking areas, steelmaking areas, steel rolling areas, public energy areas, and logistics and warehousing areas; The ironmaking area includes the blast furnace body, hot blast stove, blower, TRT residual pressure power generation equipment, furnace top charging system, pulverized coal injection system, water cooling system, dust removal equipment, and slag and iron processing equipment. The steelmaking area includes a converter, hot metal pretreatment equipment, refining furnace, continuous casting machine, ladle turret, tundish, oxygen lance system, dust removal system, and hot metal temperature measurement and sampling equipment. The steel rolling area includes a heating furnace, roughing mill, finishing mill, flying shear, cooling bed, coiler, straightener, hydraulic station, and lubrication station; The public energy area includes fans, water pumps, air compressors, oxygen generators, gas holders, substations, boilers, steam pipelines, and circulating water systems; The logistics warehousing area includes belt conveyors, overhead cranes, molten iron tank cars, ladle cars, finished product warehouse hoisting equipment, weighbridges, AGVs, or transport vehicles.

[0008] Optionally, the production twin module includes: The production data unit is used to acquire data from the raw material stage, sintering stage, ironmaking stage, steelmaking stage, continuous casting stage, and rolling stage based on the steel production process, to obtain the main production line data, and based on the main production line data, to design a full-link traceability mechanism for furnace batch numbers from raw materials to customer orders, thus obtaining a full production process data chain. The process twin unit is used to combine the equipment digital twin with the production process data chain to construct digital twins of raw material yard, sintering process, blast furnace ironmaking, converter steelmaking, continuous casting, rolling, energy pipeline network and warehousing and logistics, and obtain process-level digital twin models. A state twin unit is used to construct a full-process state space model based on the steel production process and establish inter-process constraint relationships to obtain a process constraint model; the full-process state space model includes equipment state, production state, quality state, energy state, safety and environmental protection state, logistics state, and cost state; The digital twin platform unit is used to integrate the entire production process data chain, the process-level digital twin model, the entire process state space model, and the process constraint model to obtain a full-process production digital twin platform.

[0009] Optionally, the unified data module includes: The business data unit is used to acquire production management data, quality management data, energy management data, safety and environmental protection data, as well as supply chain and financial data, to obtain operational data; The data governance unit is used to combine the operational data, the equipment digital twin, and the end-to-end production digital twin platform to construct a master data system. Based on the master data system, it unifies field names, units, codes, and definitions, and constructs data quality rules to perform missing value checks, outlier checks, duplicate data checks, timestamp consistency checks, unit conversion checks, and logical relationship checks to obtain a standardized dataset. The master data system includes equipment master data, material master data, product master data, process route master data, quality standard master data, energy medium master data, organizational personnel master data, customer order master data, supplier master data, and financial account master data. A hierarchical storage unit is used to construct a hierarchical data lake to store the standardized dataset, thereby obtaining unified industrial data; wherein, the hierarchical data lake includes a raw data layer, a detailed data layer, a thematic data layer, and an application data layer.

[0010] Optionally, the production organization module includes: The relation generation unit is used to extract entities, construct entity relationships, and generate a causal relationship network based on the unified industrial data to obtain an industrial knowledge graph; the entities include equipment, processes, products, materials, batches, orders, customers, process parameters, quality indicators, energy media, safety risks, environmental indicators, operators, work teams, and suppliers. The scheduling task unit is used to acquire sales order data, inventory data, equipment capacity and health data, process route data, quality constraint data, and energy and environmental constraint data based on the unified industrial data and the industrial knowledge graph, to obtain a scheduling dataset. Based on the scheduling dataset, orders are converted into production tasks, production tasks are broken down into process tasks, and then a mapping of furnace batches, casting batches, and rolling batches is constructed to obtain a set of scheduling tasks. The constraint optimization unit is used to construct order delivery constraints, process sequence constraints, equipment capacity constraints, equipment health constraints, continuous casting continuity constraints, rolling specification switching constraints, quality process window constraints, energy supply constraints, and environmental emission constraints based on the scheduled task set, thereby obtaining a set of scheduling constraints. The scheme output unit is used to construct optimization objectives and comprehensive objective functions based on the scheduling task set and the scheduling constraint set, obtain an initial scheduling model, output an initial plan, introduce a dynamic scheduling mechanism into the initial scheduling model to adjust and optimize the initial plan, obtain a multi-objective intelligent scheduling optimization model, and output the optimal production organization scheme.

[0011] Optionally, the dynamic scheduling mechanism is used to modify the initial plan in real time according to the status of on-site equipment, adjust the production sequence according to quality risks, optimize the production rhythm according to energy load, adjust the load of high-energy-consuming processes according to environmental emission constraints, and output the optimal production organization plan.

[0012] Optionally, the risk prediction and maintenance module includes: The equipment feature unit is used to extract real-time operating data, historical operating data, alarm and fault data, inspection and maintenance data, and operating condition and production data from the equipment digital twin, the unified industrial data, and the optimal production organization scheme to obtain the equipment full life cycle dataset. Based on the equipment full life cycle dataset, time-series features, frequency domain features, trend features, load-related features, and behavioral features are extracted to obtain the equipment feature vector. The model framework unit is used to calculate the equipment health index and classify the health level using the equipment feature vector and scoring function, and then integrate the rule model, statistical model, machine learning model, deep learning model and industrial knowledge graph reasoning to obtain the fusion model framework. The lifespan prediction unit is used to construct a fault prediction model based on the fusion model framework, using the equipment feature vector, the equipment health index and the health level, to predict the fault probability, obtain the future fault probability distribution of the equipment, and then use the future fault probability distribution of the equipment to construct a RUL model to predict the remaining usable lifespan of the equipment. The risk profiling unit is used to conduct risk assessment using the equipment health index, the health level, the probability distribution of future equipment failures, and the remaining usable life of the equipment, to calculate the equipment risk score, output the risk level, the risk location, and the affected process, and obtain the equipment risk profile. The maintenance execution unit is used to combine equipment risk profiles and production schedules to construct equipment maintenance strategy decision schemes, generate predictive maintenance execution plans based on the equipment maintenance strategy decision schemes, execute the predictive maintenance plans, and collect actual fault conditions, maintenance effects, downtime and costs in real time to obtain maintenance execution feedback data. The maintenance execution feedback data is then used for model closed-loop optimization.

[0013] Optionally, the quality prediction and control module includes: The quality feature unit is used to acquire full-process quality data in raw materials, production process and finished products based on the equipment digital twin, the unified industrial data and the optimal production organization scheme, construct a furnace batch quality data chain using the full-process quality data, and extract composition deviation, temperature fluctuation, pulling speed fluctuation, cooling rate and rolling force change in the furnace batch quality data chain to obtain quality feature vector; The quality risk prediction unit is used to construct a quality prediction model using the quality feature vector, obtain the probability of various quality risks, and perform causal reasoning using the industrial knowledge graph based on the probability of various quality risks to obtain the causal chain of quality problems. The quality control unit is used to calculate a quality risk score and classify the quality risk level based on the causal chain of the quality problem, so as to predict the quality of molten iron in the raw material stage, the quality of molten steel in the steelmaking stage, the defects in the continuous casting stage, and the performance of finished products in the rolling stage, thereby obtaining a quality prediction. Then, quality control is carried out based on the quality prediction to obtain a quality control plan.

[0014] Optionally, the decision support module includes: The low-energy unit is used to construct an energy medium balance model and a unit product energy consumption model based on the unified industrial data and the full-process production digital twin platform, so as to carry out energy and carbon coordinated scheduling and obtain a low-energy and low-carbon production scheme. The risk prevention and control unit is used to construct a safety risk model and an environmental exceedance prediction model based on the unified industrial data and the industrial knowledge graph, so as to carry out safety risk early warning and environmental emission prediction and control, and obtain a risk prevention and control plan. The integrated decision-making unit is used to combine the optimal production organization scheme, the predictive maintenance execution plan, the quality control scheme, the low-energy and low-carbon production scheme, and the risk prevention and control scheme to identify decision conflicts and make multi-objective collaborative decisions, thereby obtaining a comprehensive decision-making scheme for the entire process. The comprehensive decision-making scheme for the entire process is then used to restructure the business process and complete the full-process decision support.

[0015] This invention discloses the following technical effects by providing an industrial internet system for decision support throughout the entire steelmaking process: 1. By constructing a fully connected data acquisition mechanism for equipment, a unified equipment coding system, a hierarchical model, and a health index fusion algorithm, we can achieve real-time monitoring of equipment status, health assessment, and construction of digital twins, thereby improving the accuracy of equipment management.

[0016] 2. By constructing a data chain for the entire production process, a process-level digital twin model, a full-process state-space model, and a process constraint model, a digital twin of the entire process from raw materials to finished products is achieved, supporting process-level simulation and state tracking. This enables the connection between equipment and the production process, building an integrated digital twin platform encompassing equipment, processes, and production lines.

[0017] 3. By constructing a master data system, data quality rules, and hierarchical data lake, and unifying fields, codes, units, and definitions, the problem of multi-source heterogeneous data fusion can be solved, data consistency and availability can be improved, and high-quality, standardized, and unified industrial data can be provided for upper-level intelligent decision-making.

[0018] 4. By constructing an industrial knowledge graph, scheduling task sets, and a dynamic scheduling mechanism, multi-objective constraint optimization is achieved; optimal production organization schemes are generated, improving scheduling efficiency. This serves as a crucial bridge from data to decision-making, directly impacting production efficiency and resource utilization.

[0019] 5. By integrating multi-dimensional features and multi-model frameworks to predict equipment lifespan and construct risk profiles, predictive maintenance of equipment can be achieved, reducing unplanned downtime and maintenance costs, improving equipment availability and production continuity, and ensuring system reliability.

[0020] 6. By constructing a batch quality data chain, quality feature vector, and quality risk causal reasoning model, it is possible to achieve full-process quality prediction and closed-loop control, improve product consistency, and directly improve the product qualification rate and customer satisfaction.

[0021] 7. By constructing energy and carbon coordinated scheduling models and safety and environmental risk models, multi-objective integrated decision-making and process reengineering are achieved. This supports efficient production decisions that are low-carbon, low-energy-consumption, and low-risk, achieving globally optimal decision support and driving enterprise digital transformation.

[0022] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. 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.

[0024] Figure 1 This is a schematic diagram of the system architecture provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the optimal production organization scheme provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the full-process decision support process provided in an embodiment of the present invention. Detailed Implementation

[0025] 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.

[0026] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] like Figure 1As shown, this invention provides an industrial internet system for decision support throughout the entire steelmaking process, comprising: 1. For example Figure 2 As shown, the device twin module is used to acquire fully connected device data and construct a digital twin of the device based on the fully connected device data; the device twin module includes: 1.1 Equipment Data Unit, used to construct an equipment access list, and based on the equipment access list, to collect operating status data, process data, equipment health data, control system data, maintenance and repair data and alarm event data for each type of equipment to obtain full equipment connection data.

[0028] The equipment access list includes ironmaking area, steelmaking area, steel rolling area, public energy area and logistics and warehousing area.

[0029] The ironmaking area includes the blast furnace body, hot blast stove, blower, TRT residual pressure power generation equipment, furnace top charging system, pulverized coal injection system, water cooling system, dust removal equipment, and slag and iron processing equipment.

[0030] The steelmaking area includes a converter, hot metal pretreatment equipment, refining furnace, continuous casting machine, ladle turret, tundish, oxygen lance system, dust removal system, and hot metal temperature measurement and sampling equipment.

[0031] The steel rolling area includes a heating furnace, roughing mill, finishing mill, flying shear, cooling bed, coiler, straightener, hydraulic station, and lubrication station.

[0032] The public energy area includes fans, water pumps, air compressors, oxygen generators, gas holders, substations, boilers, steam pipelines, and circulating water systems.

[0033] The logistics warehousing area includes belt conveyors, overhead cranes, molten iron tank cars, ladle cars, finished product warehouse hoisting equipment, weighbridges, AGVs, or transport vehicles.

[0034] Operating status data: power on, power off, standby, fault, maintenance, alarm, etc.

[0035] Process data: parameters such as temperature, pressure, flow rate, speed, load, current, voltage, vibration, rotational speed, and liquid level.

[0036] Equipment health data: bearing temperature, vibration spectrum, motor current, lubrication status, hydraulic pressure, cooling water flow rate, etc.

[0037] Control system data: control commands, interlock status, valve opening, setpoints, and feedback values ​​from PLC, DCS, and SCADA.

[0038] Maintenance and repair data: inspection records, fault records, repair records, spare parts replacement records, lubrication records.

[0039] Alarm event data: alarm time, alarm level, alarm object, alarm reason, handling measures, and recovery time.

[0040] For existing PLC and DCS systems, data access is achieved through the OPC UA protocol; for older equipment, smart sensors and edge acquisition terminals are added; for mobile devices, such as overhead cranes, vehicles, and ladle cars, 5G, RFID, UWB, and BeiDou positioning are used for access; for energy equipment, access is achieved through smart meters, flow meters, pressure gauges, thermometers, etc.; for high-frequency vibration equipment, edge computing gateways are used for local preprocessing to reduce the pressure on the central platform.

[0041] 1.2 Equipment modeling unit, used to construct plant area number, production line number, equipment type, equipment serial number and component number based on the fully connected data of the equipment, to obtain a unified equipment coding system. According to the four-level structure of system, equipment, component and measuring point, hierarchical modeling is performed on each piece of equipment to obtain the equipment hierarchical model. The temperature health score, vibration health score, current health score, pressure health score and maintenance status score are weighted and integrated to obtain the equipment health index.

[0042] The structure consists of four levels: system, equipment, components, and measuring points. For example, a rolling mill can be divided into: rolling mill system; main drive equipment; rolls; bearings; hydraulic system; lubrication system; cooling system; electrical control system; and various temperature, pressure, vibration, and current measuring points.

[0043] 1.3 Digital twin unit, combining the device static archive, the device fully connected data, the unified coding system, the device hierarchical model and the device health index, to obtain a device digital twin.

[0044] The digital twin of the equipment includes: equipment static profile, equipment real-time status, equipment health index, equipment fault history, equipment maintenance knowledge, equipment operation trend, and equipment risk level.

[0045] 2. For example Figure 2 As shown, the production twin module is used to integrate and construct a full-process production data chain, a process-level digital twin model, a full-process state-space model, and a process constraint model based on the steel production process, to obtain a full-process production digital twin platform; the production twin module includes: 2.1 Production Data Unit, used to acquire data from the raw material stage, sintering stage, ironmaking stage, steelmaking stage, continuous casting stage and rolling stage based on the steel production process, to obtain the main production line data, and based on the main production line data, to design a full-link traceability mechanism from raw materials to customer orders for furnace batch numbers, to obtain the full production process data chain.

[0046] Raw material stage: iron ore powder batches, coke batches, limestone batches, alloy material batches, raw material chemical composition, raw material inventory, raw material incoming inspection data, and raw material consumption data.

[0047] Sintering or pelletizing process: ingredient ratio, moisture content of the mixture, sintering machine speed, ignition temperature, material layer thickness, sinter strength, sinter basicity, and yield.

[0048] Ironmaking process: blast furnace air volume, blast temperature, blast pressure, oxygen enrichment rate, pulverized coal injection rate, furnace top pressure, furnace temperature, molten iron temperature, molten iron silicon content, molten iron sulfur content, tapping time, and slag-to-iron ratio.

[0049] Steelmaking process: molten iron composition, scrap steel addition amount, oxygen flow rate, blowing time, final temperature, final carbon content, alloy addition amount, molten steel composition, and refining time.

[0050] Continuous casting process: molten steel temperature, casting speed, liquid level in the crystallizer, cooling water volume, water volume in the secondary cooling zone, billet surface temperature, billet defects, and billet specifications.

[0051] Steel rolling process: heating furnace temperature, billet loading time, billet unloading temperature, rolling speed, reduction, rolling force, final rolling temperature, cooling rate, finished product dimensions, and surface quality.

[0052] 2.2 Process Twin Unit, used to combine the equipment digital twin with the production process data chain to construct digital twins of raw material yard, sintering process, blast furnace ironmaking, converter steelmaking, continuous casting, rolling, energy pipeline network and warehousing and logistics, thus obtaining process-level digital twin models.

[0053] 2.3 State twin unit, used to construct a full-process state space model based on the steel production process, and establish inter-process constraint relationships to obtain a process constraint model; the full-process state space model includes equipment state, production state, quality state, energy state, safety and environmental protection state, logistics state, and cost state.

[0054] The entire process status space includes: equipment status, production status, quality status, energy status, safety and environmental protection status, logistics status, and cost status.

[0055] For example, the constraints between processes include: blast furnace molten iron production affects converter production rhythm; converter smelting rhythm affects continuous casting times; continuous casting billet temperature affects hot delivery and charging; rolling rhythm affects finished product delivery; energy supply capacity affects production load; and environmental emission indicators constrain production organization methods.

[0056] 2.4 The digital twin platform unit is used to integrate the production process data chain, the process-level digital twin model, the full-process state space model, and the process constraint model to obtain a full-process production digital twin platform.

[0057] 3. For example Figure 2 As shown, a unified data module is used to combine equipment digital twins, a full-process production digital twin platform, and acquired operational data to generate a standardized dataset, which is then stored hierarchically to obtain unified industrial data. The unified data module includes: 3.1 Business data unit, used to acquire production management data, quality management data, energy management data, safety and environmental protection data, as well as supply chain and financial data, to obtain operational data.

[0058] Production management data includes: production plan, shift plan, daily work plan, process output, plan completion rate, downtime, changeover time, and production anomaly records.

[0059] Quality management data includes: raw material inspection, molten iron composition, molten steel composition, billet defects, finished product dimensions, mechanical properties, surface defects, and customer quality objections.

[0060] Energy management data includes: electricity consumption, water consumption, oxygen consumption, nitrogen consumption, gas generation, gas consumption, steam consumption, compressed air consumption, and waste heat and energy recovery.

[0061] Safety and environmental protection data include: safety hazards, hazardous work permits, personnel location, gas alarms, dust emissions, flue gas emissions, wastewater indicators, noise monitoring, and online environmental monitoring data.

[0062] Supply chain and financial data include: purchase orders, raw material prices, supplier deliveries, inventory amounts, sales orders, customer delivery requirements, production costs, energy costs, equipment maintenance costs, and product gross profit.

[0063] 3.2 The data governance unit is used to combine the operational data, the equipment digital twin, and the full-process production digital twin platform to construct a master data system. Based on the master data system, the field names, units, codes, and definitions are unified, and data quality rules are constructed to perform missing value checks, outlier checks, duplicate data checks, timestamp consistency checks, unit conversion checks, and logical relationship checks to obtain a standardized dataset.

[0064] The master data system includes equipment master data, material master data, product master data, process route master data, quality standard master data, energy medium master data, organizational personnel master data, customer order master data, supplier master data, and financial account master data. 3.3 Hierarchical storage units are used to construct a hierarchical data lake for storing the standardized dataset, thereby obtaining unified industrial data; wherein, the hierarchical data lake includes: Raw data layer: Stores raw data from devices, systems, and business processes; Detailed data layer: Cleaned standard detailed data; Thematic data layer: organized by themes such as production, quality, energy, and equipment; Application Data Layer: Data services geared towards intelligent decision-making and business applications.

[0065] 4. For example Figure 2 As shown, the production organization module is used to construct an industrial knowledge graph and a set of scheduled tasks based on the unified industrial data, and to perform constraint optimization on the set of scheduled tasks using a dynamic scheduling mechanism to obtain the optimal production organization scheme; the production organization module includes: 4.1 Relationship generation unit, used to extract entities based on the unified industrial data, construct entity relationships, and generate a causal relationship network to obtain an industrial knowledge graph.

[0066] Entities include equipment, processes, products, materials, batches, orders, customers, process parameters, quality indicators, energy media, safety risks, environmental indicators, operators, work teams, and suppliers.

[0067] Entity relationships include: equipment participating in production batches, batches originating from raw material batches, products corresponding to customer orders, process parameters affecting quality indicators, energy media supplying production processes, equipment malfunctions causing quality fluctuations, shift operations affecting production efficiency, and supplier raw materials affecting molten iron composition.

[0068] Industrial knowledge graphs can support: fault cause reasoning, quality problem tracing, energy consumption anomaly analysis, production bottleneck identification, cost anomaly explanation, intelligent question answering, and decision support.

[0069] 4.2 The scheduling task unit is used to acquire sales order data, inventory data, equipment capacity and health data, process route data, quality constraint data, and energy and environmental constraint data based on the unified industrial data and the industrial knowledge graph, to obtain a scheduling dataset. Based on the scheduling dataset, orders are converted into production tasks, production tasks are broken down into process tasks, and then a mapping of furnace batches, casting batches, and rolling batches is constructed to obtain a set of scheduling tasks.

[0070] 4.3 Constraint optimization unit, used to construct order delivery constraints, process sequence constraints, equipment capacity constraints, equipment health constraints, continuous casting continuity constraints, rolling specification switching constraints, quality process window constraints, energy supply constraints, and environmental emission constraints based on the scheduling task set, to obtain a scheduling constraint set.

[0071] 4.4 Scheme output unit, used to construct optimization objectives and comprehensive objective functions based on the scheduling task set and the scheduling constraint set, obtain an initial scheduling model, output an initial plan, introduce a dynamic scheduling mechanism into the initial scheduling model to adjust and optimize the initial plan, obtain a multi-objective intelligent scheduling optimization model, and output the optimal production organization scheme.

[0072] The dynamic scheduling mechanism is used to modify the initial plan in real time based on the status of on-site equipment, adjust the production sequence based on quality risks, optimize the production rhythm based on energy load, adjust the load of high-energy-consuming processes based on environmental emission constraints, and output the optimal production organization plan.

[0073] 5. For example Figure 3 As shown, the risk prediction and maintenance module is used to predict the future failure probability distribution and remaining usable life of the equipment based on the equipment digital twin, the unified industrial data, and the optimal production organization scheme, and to construct an equipment risk profile to generate a predictive maintenance execution plan; the risk prediction and maintenance module includes: 5.1 Equipment feature unit, used to extract real-time operation data, historical operation data, alarm and fault data, inspection and maintenance data, and operating condition and production data from the equipment digital twin, the unified industrial data, and the optimal production organization scheme to obtain the equipment full life cycle dataset; and based on the equipment full life cycle dataset, extract time-series features, frequency domain features, trend features, load-related features, and behavioral features to obtain the equipment feature vector. 5.2 Model framework unit, used to calculate the equipment health index and classify the health level using the equipment feature vector and scoring function, and then integrate rule model, statistical model, machine learning model, deep learning model and industrial knowledge graph reasoning to obtain the fusion model framework.

[0074] Rule-based model: based on expert experience and judgment; Statistical model: based on the mean, variance, and control limits; Machine learning model: trained based on historical failure data; Deep learning models: used for vibration spectrum and time-series trend recognition; Industrial knowledge graph reasoning: inferring causes based on fault causal chains.

[0075] 5.3 Lifetime prediction unit, used to construct a fault prediction model based on the fusion model framework using the equipment feature vector, the equipment health index and the health level, to predict the fault probability, obtain the future fault probability distribution of the equipment, and then use the future fault probability distribution of the equipment to construct a RUL model to predict the remaining usable life of the equipment.

[0076] 5.4 Risk profiling unit, used to conduct risk assessment using the equipment health index, the health level, the probability distribution of future equipment failures and the remaining usable life of the equipment, to calculate the equipment risk score, output the risk level, risk location and affected process, and obtain the equipment risk profile.

[0077] 5.5 The maintenance execution unit is used to combine equipment risk profiles and production schedules to construct equipment maintenance strategy decision schemes, generate predictive maintenance execution plans based on the equipment maintenance strategy decision schemes, execute the predictive maintenance plans, and collect actual fault conditions, maintenance effects, downtime and costs in real time to obtain maintenance execution feedback data. The maintenance execution feedback data is then used for model closed-loop optimization.

[0078] 6. For example Figure 3 As shown, the quality prediction and control module is used to predict quality risks and make quality forecasts based on the equipment digital twin, the unified industrial data, and the optimal production organization scheme, so as to complete quality control and obtain a quality control scheme; the quality prediction and control module includes: 6.1 Quality feature unit, used to acquire full-process quality data in raw materials, production process and finished products based on the equipment digital twin, the unified industrial data and the optimal production organization scheme, construct furnace batch quality data chain using the full-process quality data, and extract composition deviation, temperature fluctuation, pulling speed fluctuation, cooling rate and rolling force change in the furnace batch quality data chain to obtain quality feature vector.

[0079] 6.2 Quality risk prediction unit, used to construct a quality prediction model using the quality feature vector, obtain the probability of various quality risks, and perform causal reasoning using the industrial knowledge graph based on the probability of various quality risks to obtain the causal chain of quality problems.

[0080] 6.3 Quality control unit, used to calculate quality risk score and classify quality risk level based on the causal chain of the quality problem, so as to predict the quality of molten iron in the raw material stage, the quality of molten steel in the steelmaking stage, the defects in the continuous casting stage, and the performance of finished products in the rolling stage, to obtain quality prediction, and then to carry out quality control according to the quality prediction to obtain a quality control plan.

[0081] 7. For example Figure 3 As shown, the decision support module is used to combine the optimal production organization plan, the predictive maintenance execution plan, the quality control plan, the low-energy and low-carbon production plan, and the risk prevention and control plan to perform comprehensive decision-making and business process reengineering throughout the entire process, thereby completing full-process decision support; the decision support module includes: 7.1 Low-energy unit, used to construct an energy medium balance model and a unit product energy consumption model based on the unified industrial data and the full-process production digital twin platform, so as to carry out energy and carbon coordinated scheduling and obtain a low-energy and low-carbon production scheme.

[0082] Energy and carbon coordinated scheduling, for example: Schedule high-energy-consuming processes during off-peak electricity prices; When there is a surplus of gas, priority should be given to consuming it in heating furnaces; Adjust the rolling rhythm when there is insufficient gas; Reduce the load on high-emission equipment when environmental emissions are close to the threshold. Prioritize the use of waste heat and energy for power generation and steam supply; Establish process-level carbon emission accounting.

[0083] 7.2 Risk control unit, used to construct safety risk model and environmental exceedance prediction model based on the unified industrial data and the industrial knowledge graph, so as to carry out safety risk early warning and environmental emission prediction and control, and obtain risk control plan.

[0084] 7.3 The integrated decision-making unit is used to combine the optimal production organization scheme, the predictive maintenance execution plan, the quality control scheme, the low-energy and low-carbon production scheme, and the risk prevention and control scheme to identify decision conflicts and make multi-objective collaborative decisions, thereby obtaining a comprehensive decision-making scheme for the entire process. The comprehensive decision-making scheme for the entire process is then used to restructure the business process and complete the full-process decision support.

[0085] Decision conflict identification, for example: Meeting order deadlines may increase production load, but this will also increase equipment risks. To reduce energy consumption, the rolling pace may be slowed down, but this will affect delivery. Improving quality may increase alloy consumption, but this will increase costs. To reduce environmental emissions, the sintering load may be reduced, but this will affect the supply of raw materials for the blast furnace.

[0086] Therefore, this invention provides an industrial internet system for decision support throughout the entire steel production process. By constructing a full-process industrial internet platform, it achieves intelligent twin collaboration between equipment and decision-making in steel production, significantly improving production efficiency, quality consistency, equipment reliability, and green and low-carbon levels.

[0087] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0088] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An industrial internet system for decision support throughout the entire steelmaking process, characterized in that, include: The device twin module is used to acquire fully connected device data and construct a digital twin of the device based on the fully connected device data; The production twin module is used to integrate and construct a full-process production data chain, process-level digital twin model, full-process state space model, and process constraint model based on the steel production process, to obtain a full-process production digital twin platform. A unified data module is used to combine equipment digital twins, the end-to-end production digital twin platform, and acquired operational data to generate a standardized dataset, and then store the standardized dataset hierarchically to obtain unified industrial data. The production organization module is used to construct an industrial knowledge graph and a set of scheduled tasks based on the unified industrial data, and to optimize the set of scheduled tasks by combining a dynamic scheduling mechanism to obtain the optimal production organization scheme. The risk prediction and maintenance module is used to predict the future failure probability distribution and remaining usable life of the equipment based on the equipment digital twin, the unified industrial data and the optimal production organization scheme, and to construct an equipment risk profile to generate a predictive maintenance execution plan. The quality prediction and control module is used to predict quality risks and make quality forecasts based on the equipment digital twin, the unified industrial data, and the optimal production organization scheme, so as to complete quality control and obtain a quality control scheme. The decision support module is used to combine the optimal production organization plan, the predictive maintenance execution plan, the quality control plan, the low-energy and low-carbon production plan, and the risk prevention and control plan to make comprehensive decisions and restructure business processes throughout the entire process, thereby completing full-process decision support. The equipment twin module, the production twin module, the unified data module, the production organization module, the risk prediction and maintenance module, the quality prediction and control module, and the decision support module are interconnected.

2. The industrial internet system for decision support throughout the entire steelmaking process according to claim 1, characterized in that, The device twin module includes: The equipment data unit is used to construct an equipment access list. Based on the equipment access list, it collects the operating status data, process data, equipment health data, control system data, maintenance and repair data, and alarm event data of each type of equipment to obtain the full equipment connection data. The equipment modeling unit is used to construct the plant area number, production line number, equipment type, equipment serial number and component number based on the fully connected data of the equipment to obtain a unified equipment coding system. According to the four-level structure of system, equipment, component and measuring point, hierarchical modeling is performed on each piece of equipment to obtain the equipment hierarchical model. The temperature health score, vibration health score, current health score, pressure health score and maintenance status score are weighted and integrated to obtain the equipment health index. A digital twin unit, combining the device's static profile, the device's fully connected data, the unified coding system, the device's hierarchical model, and the device's health index, yields a digital twin of the device.

3. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 2, is characterized in that... The equipment access list includes ironmaking area, steelmaking area, steel rolling area, public energy area and logistics and warehousing area; The ironmaking area includes the blast furnace body, hot blast stove, blower, TRT residual pressure power generation equipment, furnace top charging system, pulverized coal injection system, water cooling system, dust removal equipment, and slag and iron processing equipment. The steelmaking area includes a converter, hot metal pretreatment equipment, refining furnace, continuous casting machine, ladle turret, tundish, oxygen lance system, dust removal system, and hot metal temperature measurement and sampling equipment. The steel rolling area includes a heating furnace, roughing mill, finishing mill, flying shear, cooling bed, coiler, straightener, hydraulic station, and lubrication station; The public energy area includes fans, water pumps, air compressors, oxygen generators, gas holders, substations, boilers, steam pipelines, and circulating water systems; The logistics warehousing area includes belt conveyors, overhead cranes, molten iron tank cars, ladle cars, finished product warehouse hoisting equipment, weighbridges, AGVs, or transport vehicles.

4. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 3, is characterized in that... The production twin module includes: The production data unit is used to acquire data from the raw material stage, sintering stage, ironmaking stage, steelmaking stage, continuous casting stage, and rolling stage based on the steel production process, to obtain the main production line data, and based on the main production line data, to design a full-link traceability mechanism for furnace batch numbers from raw materials to customer orders, thus obtaining a full production process data chain. The process twin unit is used to combine the equipment digital twin with the production process data chain to construct digital twins of raw material yard, sintering process, blast furnace ironmaking, converter steelmaking, continuous casting, rolling, energy pipeline network and warehousing and logistics, and obtain process-level digital twin models. A state twin unit is used to construct a full-process state space model based on the steel production process and establish inter-process constraint relationships to obtain a process constraint model; the full-process state space model includes equipment state, production state, quality state, energy state, safety and environmental protection state, logistics state, and cost state; The digital twin platform unit is used to integrate the entire production process data chain, the process-level digital twin model, the entire process state space model, and the process constraint model to obtain a full-process production digital twin platform.

5. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 4, characterized in that: The unified data module includes: The business data unit is used to acquire production management data, quality management data, energy management data, safety and environmental protection data, as well as supply chain and financial data, to obtain operational data; The data governance unit is used to combine the operational data, the equipment digital twin, and the end-to-end production digital twin platform to construct a master data system. Based on the master data system, it unifies field names, units, codes, and definitions, and constructs data quality rules to perform missing value checks, outlier checks, duplicate data checks, timestamp consistency checks, unit conversion checks, and logical relationship checks to obtain a standardized dataset. The master data system includes equipment master data, material master data, product master data, process route master data, quality standard master data, energy medium master data, organizational personnel master data, customer order master data, supplier master data, and financial account master data. A hierarchical storage unit is used to construct a hierarchical data lake to store the standardized dataset, thereby obtaining unified industrial data; wherein, the hierarchical data lake includes a raw data layer, a detailed data layer, a thematic data layer, and an application data layer.

6. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 5, is characterized in that... The production organization module includes: The relation generation unit is used to extract entities, construct entity relationships, and generate a causal relationship network based on the unified industrial data to obtain an industrial knowledge graph; the entities include equipment, processes, products, materials, batches, orders, customers, process parameters, quality indicators, energy media, safety risks, environmental indicators, operators, work teams, and suppliers. The scheduling task unit is used to acquire sales order data, inventory data, equipment capacity and health data, process route data, quality constraint data, and energy and environmental constraint data based on the unified industrial data and the industrial knowledge graph, to obtain a scheduling dataset. Based on the scheduling dataset, orders are converted into production tasks, production tasks are broken down into process tasks, and then a mapping of furnace batches, casting batches, and rolling batches is constructed to obtain a set of scheduling tasks. The constraint optimization unit is used to construct order delivery constraints, process sequence constraints, equipment capacity constraints, equipment health constraints, continuous casting continuity constraints, rolling specification switching constraints, quality process window constraints, energy supply constraints, and environmental emission constraints based on the scheduled task set, thereby obtaining a set of scheduling constraints. The scheme output unit is used to construct optimization objectives and comprehensive objective functions based on the scheduling task set and the scheduling constraint set, obtain an initial scheduling model, output an initial plan, introduce a dynamic scheduling mechanism into the initial scheduling model to adjust and optimize the initial plan, obtain a multi-objective intelligent scheduling optimization model, and output the optimal production organization scheme.

7. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 6, is characterized in that... The dynamic scheduling mechanism is used to modify the initial plan in real time based on the status of on-site equipment, adjust the production sequence based on quality risks, optimize the production rhythm based on energy load, adjust the load of high-energy-consuming processes based on environmental emission constraints, and output the optimal production organization plan.

8. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 7, characterized in that: The risk prediction and maintenance module includes: The equipment feature unit is used to extract real-time operating data, historical operating data, alarm and fault data, inspection and maintenance data, and operating condition and production data from the equipment digital twin, the unified industrial data, and the optimal production organization scheme to obtain the equipment full life cycle dataset. Based on the equipment full life cycle dataset, time-series features, frequency domain features, trend features, load-related features, and behavioral features are extracted to obtain the equipment feature vector. The model framework unit is used to calculate the equipment health index and classify the health level using the equipment feature vector and scoring function, and then integrate the rule model, statistical model, machine learning model, deep learning model and industrial knowledge graph reasoning to obtain the fusion model framework. The lifespan prediction unit is used to construct a fault prediction model based on the fusion model framework, using the equipment feature vector, the equipment health index and the health level, to predict the fault probability, obtain the future fault probability distribution of the equipment, and then use the future fault probability distribution of the equipment to construct a RUL model to predict the remaining usable lifespan of the equipment. The risk profiling unit is used to conduct risk assessment using the equipment health index, the health level, the probability distribution of future equipment failures, and the remaining usable life of the equipment, to calculate the equipment risk score, output the risk level, the risk location, and the affected process, and obtain the equipment risk profile. The maintenance execution unit is used to combine equipment risk profiles and production schedules to construct equipment maintenance strategy decision schemes, generate predictive maintenance execution plans based on the equipment maintenance strategy decision schemes, execute the predictive maintenance plans, and collect actual fault conditions, maintenance effects, downtime and costs in real time to obtain maintenance execution feedback data. The maintenance execution feedback data is then used for model closed-loop optimization.

9. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 8, characterized in that: The quality prediction and control module includes: The quality feature unit is used to acquire full-process quality data in raw materials, production process and finished products based on the equipment digital twin, the unified industrial data and the optimal production organization scheme, construct a furnace batch quality data chain using the full-process quality data, and extract composition deviation, temperature fluctuation, pulling speed fluctuation, cooling rate and rolling force change in the furnace batch quality data chain to obtain quality feature vector; The quality risk prediction unit is used to construct a quality prediction model using the quality feature vector, obtain the probability of various quality risks, and perform causal reasoning using the industrial knowledge graph based on the probability of various quality risks to obtain the causal chain of quality problems. The quality control unit is used to calculate a quality risk score and classify the quality risk level based on the causal chain of the quality problem, so as to predict the quality of molten iron in the raw material stage, the quality of molten steel in the steelmaking stage, the defects in the continuous casting stage, and the performance of finished products in the rolling stage, thereby obtaining a quality prediction. Then, quality control is carried out based on the quality prediction to obtain a quality control plan.

10. An industrial internet system for decision support throughout the entire steelmaking process, as described in claim 9, characterized in that, The decision support module includes: The low-energy unit is used to construct an energy medium balance model and a unit product energy consumption model based on the unified industrial data and the full-process production digital twin platform, so as to carry out energy and carbon coordinated scheduling and obtain a low-energy and low-carbon production scheme. The risk prevention and control unit is used to construct a safety risk model and an environmental exceedance prediction model based on the unified industrial data and the industrial knowledge graph, so as to carry out safety risk early warning and environmental emission prediction and control, and obtain a risk prevention and control plan. The integrated decision-making unit is used to combine the optimal production organization scheme, the predictive maintenance execution plan, the quality control scheme, the low-energy and low-carbon production scheme, and the risk prevention and control scheme to identify decision conflicts and make multi-objective collaborative decisions, thereby obtaining a comprehensive decision-making scheme for the entire process. The comprehensive decision-making scheme for the entire process is then used to restructure the business process and complete the full-process decision support.

Images