Steel production process optimization system and method based on industrial simulation

Abstract

The application discloses a steel production process optimization system and method based on industrial simulation, and relates to the technical field of steel production process simulation optimization, wherein the system comprises: a configuration module, which is used for configuring actual running lines and simulation running lines for each production link of a steel production line; and a pre-rehearsal module, which is used for collecting real-time working condition data and adding the real-time working condition data to the actual running lines. The application can identify potential fault information in advance before a real fault occurs, can change a traditional post-response mode into a pre-warning mode, and can process the fault in advance, so as to reduce the influence of the fault on production in the steel production process. Through the setting of the associated cut point and the actual cut point, the exchange can be performed at a time resolution of milliseconds, the control accuracy and stability are ensured, and the response speed during the exchange is ensured.

Description

Technical Field

[0001] This invention relates to the field of steel production process simulation and optimization technology, specifically to a steel production process optimization system and method based on industrial simulation. Background Technology

[0002] The steel production process is characterized by its long flow, strong coupling, high temperature and pressure, and large time delay, involving multiple stages such as sintering, blast furnace ironmaking, converter steelmaking, refining, continuous casting, hot rolling, and cold rolling. An abnormality in any stage can lead to serious production accidents such as casting interruption, steel leakage, steel accumulation, and rolling cracking, causing huge economic losses and safety risks. Currently, mainstream production process control systems mainly adopt a threshold alarm + post-event processing mode: sensors collect operating parameters such as temperature, pressure, flow rate, and vibration in real time; when the parameters exceed preset safety thresholds, the system triggers an alarm, and operators or the automatic control system then take corresponding measures. However, this mode has the following shortcomings; 1. Control lag: There is a delay of hundreds of milliseconds or even several seconds from the occurrence of an abnormal operating condition to the response of sensor detection and signal transmission. The failure evolution speed of many metallurgical processes is extremely fast, and traditional lag control often cannot intervene in a timely manner. Second, the simulation and control are disconnected: there is no closed-loop linkage mechanism between the simulation model and the real-time control, the simulation results cannot actively intervene in the actual operation, and the predictive role of simulation is not fully developed.

[0003] In view of this, the present invention proposes a steel production process optimization system and method based on industrial simulation to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide a steel production process optimization system and method based on industrial simulation to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a steel production process optimization system based on industrial simulation, comprising: The configuration module is used to configure the actual operating line and the simulated operating line for each production stage of the steel production line. The pre-run module is used to collect real-time operating condition data and add the real-time operating condition data to the actual operating line. The simulation operating line performs pre-run analysis based on the real-time operating condition data and judges in advance whether there is fault information in the real-time operating condition data based on the pre-run analysis. The exchange module is used to determine the predicted fault point corresponding to the real-time operating condition data based on the fault information when the existence of fault information is determined, to determine the predicted control parameters based on the predicted fault point, and to exchange the predicted control parameters on the simulation line to the actual operating line using a preset exchange assistant. The optimization module is used to set the exchange frequency threshold, optimize the production process that reaches the exchange frequency threshold, and apply the optimization results to the corresponding production process.

[0006] In a preferred embodiment, the configuration module includes: The division unit is used to divide a steel production line into multiple production stages; The first configuration unit is used to configure the corresponding actual operating line for each production stage; The second configuration unit is used to perform industrial simulation for each production stage, obtain the simulation model corresponding to the production stage, and configure the simulation run line for the corresponding simulation model.

[0007] In a preferred embodiment, the pre-rendering module includes: The data acquisition unit is used to collect real-time operating condition data of the corresponding production process, wherein the real-time operating condition data is the comprehensive parameter of the corresponding production process. The collection unit is used to determine multiple historical operating condition data sequences. Each historical operating condition data sequence includes the operating condition data sequence, fault information, and historical control parameters. A fault information database is established based on multiple historical operating condition data sequences. The pre-simulation unit is used to adapt real-time operating condition data to the time points on the actual operating line. The simulated operating line performs a pre-simulation of the actual operating line based on real-time operating condition data and adjacent historical operating condition data. Based on a preset fault information database, the real-time operating condition data and adjacent historical operating condition data are analyzed, and the fault information of the historical operating condition data sequence corresponding to the real-time operating condition data and adjacent historical operating condition data that reach a preset similarity threshold is used as the fault information of the real-time operating condition data. The actual operating line and the simulated operating line are both set with equidistant time points.

[0008] In a preferred embodiment, the pre-playing unit further includes: The analysis unit compares real-time operating condition data and adjacent historical operating condition data with multiple historical operating condition data sequences. During the comparison process, it gradually eliminates historical operating condition data sequences that do not meet the preset similarity threshold with the real-time operating condition data and adjacent historical operating condition data until a unique historical operating condition data sequence that meets the preset similarity threshold with the real-time operating condition data and adjacent historical operating condition data is found. If so, it is determined that there is fault information in the real-time operating condition data and adjacent historical operating condition data, and the historical operating condition data sequence containing the fault information is taken as the fault operating condition data sequence. If neither the real-time operating condition data nor the adjacent historical operating condition data meets the preset similarity threshold with any historical operating condition data sequence, it is determined that there is no fault information in the real-time operating condition data and adjacent historical operating condition data.

[0009] In a preferred embodiment, the switching module includes: The first setting unit is used to determine the predicted fault point corresponding to the real-time operating condition data based on the fault information when it is determined that there is fault information, and to bind the predicted control parameters corresponding to the predicted fault point to the predicted fault point. The second setting unit is used to pre-configure execution points on the actual operating line. The execution points correspond to real-time operating data, and the execution points correspond to accompanying points on the simulation operating line. The execution points and accompanying points slide synchronously on the actual operating line and the simulation operating line, respectively. The switching unit is used to adapt the fault condition data sequence forward onto the simulation operation line through the predicted fault point. The starting point of the fault condition data sequence on the simulation operation line is used as the pre-configured execution point on the actual operation line. The predicted fault start point and the predicted fault point are used as the associated space. The positions of the predicted fault start point and the predicted fault point on the actual operation line are used as the fault start point and subsequent fault point, respectively. The space between the fault start point and the subsequent fault point is used as the native space. An switching assistant is set between the associated space and the native space. The switching assistant is used to send the predicted control parameters in the associated space on the simulation operation line to the actual operation line.

[0010] In a preferred embodiment, the switching unit further includes: Determine the docking duration, and based on the docking duration, set multiple corresponding associated tangent points and actual tangent points on the simulation operation line and the actual operation line respectively. The associated tangent points and actual tangent points correspond to each other to form a tangent point pair. For multiple associated tangent points in the associated space, the corresponding predicted operating condition data and predicted control parameters are adapted respectively, and the actual operating condition data is filled into the corresponding actual tangent points. The exchange assistant connects to the execution point and the accompanying point. The exchange assistant carries the execution point and the accompanying point and slides on multiple accompanying tangent points and multiple actual tangent points until the accompanying tangent point and the actual tangent point reach the preset conditions. Then, the exchange assistant is triggered to send the predictive control parameters of the next accompanying tangent point to the corresponding actual tangent point, so as to realize the exchange of predictive control parameters on the simulation line to the actual line.

[0011] In a preferred embodiment, the optimization module includes: The aggregation unit is used to obtain the exchange frequency between the actual running line and the simulated running line corresponding to each production link, mark the original space that is greater than the preset exchange frequency threshold, and take the production link corresponding to the marked original space as the point to be optimized. The optimization unit is used to perform multi-factor analysis on the production process corresponding to the point to be optimized, optimize the corresponding production process based on the results of the multi-factor analysis, and apply the optimized production process.

[0012] This invention also provides a method for optimizing steel production processes based on industrial simulation, comprising the following steps: For each stage of the steel production line, both an actual operating line and a simulated operating line are configured. Collect real-time operating condition data and add it to the actual operating line. The simulation operating line performs advance analysis based on the real-time operating condition data and determines in advance whether there is fault information in the real-time operating condition data based on the advance analysis. When fault information is confirmed, the predicted fault point corresponding to the real-time operating data is determined based on the fault information, the predicted control parameters are determined based on the predicted fault point, and the predicted control parameters on the simulation line are exchanged to the actual operating line using a preset exchange assistant. Set an exchange frequency threshold, optimize the production process that reaches the exchange frequency threshold, and apply the optimization results to the corresponding production process.

[0013] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention configures an independent simulation operation line for each production link and uses real-time operating data for advance simulation analysis. The system can identify potential fault information in advance before the actual fault occurs. This can transform the traditional post-event response mode into an early warning mode, and can handle faults in advance, thereby reducing the impact of faults on steel production.

[0014] 2. This invention, through the setting of the exchange assistant, the accompanying space, and the original space, triggers the exchange of predictive control parameters when the predicted operating condition data of the system simulation and the actual operating condition data reach consistency. This ensures the synchronization and accuracy of intervention actions with the actual production state, reduces the side effects caused by premature or late control, and enables early regulation of faults before they occur. In particular, by setting the accompanying cut-off point and the actual cut-off point, the exchange can be carried out at a millisecond-level time resolution, ensuring the precision and stability of control, while also ensuring the response speed during the exchange. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a system block diagram of the present invention.

[0017] Figure 2 This is a schematic diagram of the connection of the exchange assistant of the present invention.

[0018] Figure 3 This is a flowchart of the method of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0020] Example 1, please refer to Figure 1 and Figure 2 As shown in this embodiment, the steel production process optimization system based on industrial simulation includes: The configuration module is used to configure the actual operating line and the simulated operating line for each production stage of the steel production line. The pre-run module is used to collect real-time operating condition data and add the real-time operating condition data to the actual operating line. The simulation operating line performs pre-run analysis based on the real-time operating condition data and judges in advance whether there is fault information in the real-time operating condition data based on the pre-run analysis. The exchange module is used to determine the predicted fault point corresponding to the real-time operating condition data based on the fault information when the existence of fault information is determined, to determine the predicted control parameters based on the predicted fault point, and to exchange the predicted control parameters on the simulation line to the actual operating line using a preset exchange assistant. The optimization module is used to set the exchange frequency threshold, optimize the production process that reaches the exchange frequency threshold, and apply the optimization results to the corresponding production process.

[0021] In one embodiment, the configuration module includes: The division unit is used to divide a steel production line into multiple production stages; The first configuration unit is used to configure the corresponding actual operating line for each production stage; The second configuration unit is used to perform industrial simulation for each production stage, obtain the simulation model corresponding to the production stage, and configure the simulation run line for the corresponding simulation model.

[0022] It should be noted that the steel production line is decoupled into multiple independent and manageable production links according to process logic and equipment boundaries, such as: sintering, pelletizing, coking, blast furnace, converter, refining (LF / RH), continuous casting, heating furnace, rough rolling, finish rolling, coiling, shearing, pickling, cold rolling, annealing, galvanizing, etc. For each segmented production stage, an actual operating line is established on the industrial internet platform. This actual operating line is essentially a real-time data stream from all sensors and actuators in that production stage, stored in a time-series format, accurately reflecting the state of the physical equipment in that stage. For each production stage, a high-precision simulation model (i.e., industrial simulation) is built using industrial simulation software. This simulation model can be a mechanistic model (based on physicochemical equations), a data model (based on machine learning, such as neural networks or LSTM), or a hybrid model. The simulation model is deployed on edge computing nodes or in the cloud and configured with a simulation operating line. This simulation operating line receives input from real-time operating data from the actual operating line and performs a pre-simulation of the real-time operating data based on the real-time operating data and adjacent operating data, thus predicting the future state of the real-time operating data.

[0023] In one embodiment, the pre-rendering module includes: The data acquisition unit is used to collect real-time operating condition data of the corresponding production process, wherein the real-time operating condition data is the comprehensive parameter of the corresponding production process. The collection unit is used to determine multiple historical operating condition data sequences. Each historical operating condition data sequence includes the operating condition data sequence, fault information, and historical control parameters. A fault information database is established based on multiple historical operating condition data sequences. The pre-simulation unit is used to adapt real-time operating condition data to the time points on the actual operating line. The simulated operating line performs a pre-simulation of the actual operating line based on real-time operating condition data and adjacent historical operating condition data. Based on a preset fault information database, the real-time operating condition data and adjacent historical operating condition data are analyzed, and the fault information of the historical operating condition data sequence corresponding to the real-time operating condition data and adjacent historical operating condition data that reach a preset similarity threshold is used as the fault information of the real-time operating condition data. The actual operating line and the simulated operating line are both set with equidistant time points.

[0024] It should be noted that the collected real-time operating condition data are comprehensive parameters, including multiple physical quantities such as temperature, pressure, flow rate, speed, current, vibration, and component analysis values, rather than single parameters. The real-time operating condition data represents the comprehensive parameters for the corresponding production stage, and the comprehensive parameters for each production stage are not entirely the same. The system collects a large number of fault cases (each case is called a historical operating condition data sequence). Each case includes: a continuous operating condition data sequence (time series segment) collected before the fault occurred, the fault information corresponding to that segment of operating condition data (such as interrupted casting, steel leakage, etc.), and the effective control parameters taken to resolve the fault. Each fault case is stored as an independent record in the fault information database. The historical operating condition data sequence is not a single sequence of operating condition values, but a triplet structure: {operating condition data segment (time series, fault information (fault type)}. In subsequent similarity comparisons and associated cut-off point adaptations, only the operating condition data sequence portion of the control parameters (operation records) is used for matching and adaptation. The pre-simulation unit places the collected real-time operating condition data onto the actual operating line's time point. The simulated operating line retrieves the current time point and several adjacent historical operating condition data points from the actual operating line in real time, forming a short-term operating condition window (composed of real-time operating condition data and adjacent historical operating condition data). The short-term operating condition window is compared with the historical operating condition data sequence in the fault information database. The comparison method commonly used is Dynamic Time Warping (DTW) or Euclidean distance to calculate the overall similarity. If there is a historical operating condition data sequence whose similarity exceeds a preset similarity threshold (e.g., 90%), the system determines that there is fault information in the real-time operating condition data and adjacent historical operating condition data, and uses the corresponding historical operating condition data sequence as the fault operating condition data sequence of the real-time operating condition data.

[0025] In one embodiment, the pre-playing unit further includes: The analysis unit compares real-time operating condition data and adjacent historical operating condition data with multiple historical operating condition data sequences. During the comparison process, it gradually eliminates historical operating condition data sequences that do not meet the preset similarity threshold with the real-time operating condition data and adjacent historical operating condition data until a unique historical operating condition data sequence that meets the preset similarity threshold with the real-time operating condition data and adjacent historical operating condition data is found. If so, it is determined that there is fault information in the real-time operating condition data and adjacent historical operating condition data, and the historical operating condition data sequence containing the fault information is taken as the fault operating condition data sequence. If neither the real-time operating condition data nor the adjacent historical operating condition data meets the preset similarity threshold with any historical operating condition data sequence, it is determined that there is no fault information in the real-time operating condition data and adjacent historical operating condition data.

[0026] It should be noted that the analysis unit performs an iterative and progressive matching process, rather than a one-time full database comparison. The gradual elimination process begins by using a broad similarity threshold to divide the historical operating condition data sequence sample set into a candidate set and a rejection set. In the next round of comparison, more dimensions of operating condition data are added or the time window is shortened to more rigorously filter the candidate set. This process is repeated, eliminating more inconsistent historical operating condition data sequences at each step. The ultimate goal is to select a unique set of historical operating condition data sequences that reaches the preset similarity threshold with the current short-term operating condition window (real-time operating condition data and adjacent historical operating condition data). The historical operating condition data sequence containing the fault information is then considered the fault operating condition data sequence, thereby reducing multiple or incorrect matches caused by noise. If, during the elimination process, the candidate set becomes empty at a certain step, or after completing all steps, no historical operating condition data sequence reaches the preset similarity threshold, the system considers the real-time operating condition data and adjacent historical operating condition data to have no fault information.

[0027] In one embodiment, the switching module includes: The first setting unit is used to determine the predicted fault point corresponding to the real-time operating condition data based on the fault information when it is determined that there is fault information, and to bind the predicted control parameters corresponding to the predicted fault point to the predicted fault point. The second setting unit is used to pre-configure execution points on the actual operating line. The execution points correspond to real-time operating data, and the execution points correspond to accompanying points on the simulation operating line. The execution points and accompanying points slide synchronously on the actual operating line and the simulation operating line, respectively. The switching unit is used to adapt the fault condition data sequence forward (towards the elapsed time) from the predicted fault point onto the simulation line. The starting point of the fault condition data sequence on the simulation line is used as the pre-configured execution point on the actual line. The space between the predicted fault start point and the predicted fault point is used as the associated space. The positions of the predicted fault start point and the corresponding positions of the predicted fault point on the actual line are used as the fault start point and the subsequent fault point, respectively. The space between the fault start point and the subsequent fault point is used as the native space. An switching assistant is set between the associated space and the native space. The switching assistant is used to send the predicted control parameters in the associated space on the simulation line to the actual line.

[0028] In one embodiment, the switching unit further includes: Determine the docking duration, and based on the docking duration, set multiple corresponding associated tangent points and actual tangent points on the simulation operation line and the actual operation line respectively. The associated tangent points and actual tangent points correspond to each other to form a tangent point pair. For multiple associated tangent points in the associated space, the corresponding predicted operating condition data and predicted control parameters are adapted respectively, and the actual operating condition data is filled into the corresponding actual tangent points. The exchange assistant connects to the execution point and the accompanying point. The exchange assistant carries the execution point and the accompanying point and slides on multiple accompanying tangent points and multiple actual tangent points until the accompanying tangent point and the actual tangent point reach the preset conditions. Then, the exchange assistant is triggered to send the predictive control parameters of the next accompanying tangent point to the corresponding actual tangent point, so as to realize the exchange of predictive control parameters on the simulation line to the actual line.

[0029] It should be noted that the execution point is connected to the control terminal (such as a PLC controller) of the corresponding production process equipment, enabling the transmission of control parameters to the equipment. The execution point corresponds to the latest real-time operating condition data. After determining the fault information, the corresponding fault operating condition data sequence is found. Based on the time length of this fault operating condition data sequence, the location of the predicted fault point on the simulation line is determined. The execution point is configured on the actual operating line, and a companion point is set on the simulation line for each execution point (the execution point and the companion point are like two cursors, sliding synchronously on the actual operating line and the simulation line respectively). The predicted fault starting point is then moved to... The simulated running line segment between predicted fault points is used as the associated space; on the actual running line, the segment between the fault initiation point and subsequent fault points is used as the native space; the system adapts and fills the fault condition data sequence (from symptom to fault) that leads to fault information onto each associated tangent point in the associated space (only when the predicted condition data and real-time condition data are consistent; if only the trend is consistent, the normalized change of the fault condition data sequence is superimposed on the baseline value of the current actual condition to generate a predicted condition sequence suitable for the current condition); at the same time, the current and subsequent actual condition data are filled into the native space. At each actual cut point, an exchange assistant is set between the associated space and the original space. During the time progression, the exchange assistant triggers an exchange based on preset conditions. The exchange assistant continuously compares the cut point pairs where the execution point is located: when sliding on a cut point pair, the exchange assistant determines whether a preset condition is met. The preset condition is that the predicted operating condition data corresponding to the associated cut points of multiple consecutive cut point pairs is consistent with the real-time operating condition data at the actual cut points, or the trends of the associated space and the original space are consistent (the consistency criterion is that the predicted operating condition data and the real-time operating condition data have completely identical values ​​or trends). If the preset conditions are met, the exchange assistant will send the predictive control parameters of the next associated cut point to its corresponding actual cut point (i.e., the next actual cut point). When the execution point slides to the next actual cut point, it will directly use the predictive control parameters it receives (i.e., the predictive control parameters sent by the next associated cut point to the next actual cut point) to realize the exchange of predictive control parameters on the simulation line to the actual line (i.e., send the predictive control parameters associated with the predictive operating condition data corresponding to the next associated cut point to the next actual cut point, and the execution point will execute the predictive control parameters when it slides to the next actual cut point). Specifically, by combining the real-time operating condition data corresponding to the predicted fault initiation point with the fault operating condition data sequence, the corresponding predicted operating condition data and corresponding predictive control parameters (i.e., real-time operating condition data of the fault information existing under the current trend and predictive control parameters for resolving the fault information (degree) corresponding to the real-time operating condition data) are adapted to multiple associated cutting points in the associated space. The predicted operating condition data and predictive control parameters corresponding to each associated cutting point are different. The method for obtaining the predicted operating condition data and corresponding predictive control parameters adapted to multiple associated cutting points is as follows: if the actual operating condition data and the predicted operating condition data are completely consistent, the historical operating condition data and historical control data corresponding to the fault operating condition data sequence are adapted to multiple associated cutting points according to the time sequence to serve as the predicted operating condition data and predictive control parameters; if the actual operating condition data and the predicted operating condition data have the same trend but different values, the normalized change of the fault operating condition data sequence is superimposed on the current actual operating condition baseline value to generate a predicted operating condition sequence suitable for the current operating condition, thus obtaining multiple predicted operating condition data and corresponding predictive control data. For example, in the scenario of continuous casting crystallizer leakage prediction: the fault condition data sequence is (liquid level 98mm, temperature 1520℃, casting speed 1.2m / min), (96mm, 1523℃, 1.3), (93mm, 1527℃, 1.5), (89mm, 1532℃, 1.8), corresponding to control parameters: [0, 0, 0, 15% speed reduction]; real-time condition data: liquid level 100mm, temperature 1518℃, casting speed 1.1m / min, the trend is consistent with the historical data; the normalized changes of each dimension are superimposed: liquid level change [0, -2, -5, -11] - prediction [100, 98, 95, 89]mm; temperature change [0, +3, +7, +12]. - Predicted [1518, 1521, 1525, 1530]℃; pulling speed change [0, +0.1, +0.3, +0.7] - Predicted [1.1, 1.2, 1.4, 1.8]m / min; For each predicted operating condition data, the corresponding predicted control parameters are recalculated using the same control strategy (such as mapping rules in the fault information database or machine learning model (i.e., control system)); q1′(100mm, 1518℃, 1.1) - control parameter 0, q2′ predicted operating condition data (98mm, 1521℃, 1.2) - control parameter 2% speed reduction, q3′ predicted operating condition data (95mm, 1525℃, 1.4) - control parameter 5% speed reduction, q4′ predicted operating condition data (89mm, 1530℃, 1.8) - Control parameter 15% speed reduction; when the actual cut point q2 (98mm, 1521℃, 1.2) is consistent with the accompanying cut point q2′, the exchange is triggered, and the 5% speed reduction of the next accompanying cut point q3′ is sent to q3 in advance to achieve early intervention; (DTW algorithm can also be used: extract the feature vector of the fault condition data sequence (feature vector includes slope, peak value, second-order difference, etc.), take several real-time condition data of the execution point passing through the original space as the benchmark, and reconstruct the complete predicted condition sequence according to the benchmark and feature vector, that is, obtain multiple predicted condition data and corresponding predicted control data); the control parameter corresponding to the actual cut point (i.e., real-time control parameter) is obtained through the same control strategy (such as the mapping rule in the fault information database or the machine learning model (trained through multiple historical condition data and historical control parameters)), and the predicted control parameter and control parameter include the current output value of the PLC and the actuator feedback value of the corresponding production link; For example, given existing actual operating lines t1, t2…t17 (t1 and t2 are time points, and t1…t10 are partial sections of the actual operating lines), and simulated operating lines t1`, t2`…t17` (t1, t2…t17 correspond one-to-one with t1`, t2`…t17`); when the simulated operating lines are used to preview the actual operating lines (based on real-time operating data corresponding to t5` and previous historical operating data (t1`…t4`), the preview reveals fault information in the real-time operating data corresponding to t5`. The fault condition data sequence corresponding to the fault information determines the predicted fault point t12' (i.e., based on the trend of the fault condition data sequence or the matching of data on the fault condition data sequence, the fault condition data sequence controls the fault information at t12'); the space between the predicted fault starting point t5' and the predicted fault point t12' on the simulation line is used as the associated space, and the space between t5 and t12 on the actual line is used as the original space. Multiple associated tangent points (q1', q2'...q7') are assigned between t5' and t12', and between t5 and t12. The actual cut-off points (q1, q2…q7) are paired with q1'. The interval between adjacent companion cut-off points and adjacent actual cut-off points is equal to the docking time (the docking time is the time required from the exchange assistant triggering the exchange to the actual operating line receiving the predictive control parameters (i.e., the docking time T includes the exchange assistant triggering the exchange time + the data transmission time). For example, if the docking time is 200ms, the interval between adjacent companion cut-off points and adjacent actual cut-off points is also 200ms). As the execution point slides continuously, when the execution point slides to the actual cut-off point q2… When the real-time operating data of q2 is consistent with the predicted operating data of q2', the exchange assistant sends the predictive control parameters corresponding to the next associated cut-off point q3' of q2' to the actual cut-off point q3. When the execution point slides to the actual cut-off point q3, the predictive control parameters corresponding to q3 (here, the predictive control parameters refer to the predictive control parameters that can repair the faults existing in the real-time operating data of q3) are executed directly. Among them, the predicted operating data and predictive control parameters corresponding to each associated cut-off point q1', q2'...q7' are different. Furthermore, by configuring an independent simulation line for each production stage and using real-time operating data for advance analysis, the system can identify potential fault information before actual faults occur. This transforms the traditional reactive response mode into an advance warning mode, enabling proactive fault handling and reducing production disruptions caused by faults in steel production. Through the settings of the exchange assistant, associated space, and native space, when the predicted operating data from the system simulation matches the actual operating data, the exchange of predictive control parameters is triggered. This ensures the synchronization and accuracy of intervention actions with the actual production state, achieves explicit compensation for system execution delays, reduces the side effects of premature or late control, and allows for proactive fault control before faults occur. In particular, the setting of associated and actual tangent points allows the exchange to be performed at millisecond-level time resolution, ensuring the precision and stability of control while maintaining a fast response speed during the exchange.

[0030] In one embodiment, the optimization module includes: The aggregation unit is used to obtain the exchange frequency between the actual running line and the simulated running line corresponding to each production link, mark the original space that is greater than the preset exchange frequency threshold, and take the production link corresponding to the marked original space as the point to be optimized. The optimization unit is used to perform multi-factor analysis on the production process corresponding to the point to be optimized, optimize the corresponding production process based on the results of the multi-factor analysis, and apply the optimized production process.

[0031] It should be noted that the system continuously monitors each native space in each production stage for actual use (i.e., an exchange occurs) due to the intervention of the exchange assistant. For each exchange, the system increments a counter for the corresponding production stage process. The exchange frequency = (number of times the process segment is exchanged) / (total number of times the process segment is executed) * 100%; The system compares this frequency with a preset exchange frequency threshold (e.g., 10%); Production links corresponding to the original space exceeding this exchange frequency threshold are marked as points to be optimized, indicating that the robustness of this production link itself is poor and it relies too much on dynamic intervention; Exchange case data of all production links corresponding to points to be optimized are collected and multi-factor analysis is performed using analytical methods; Analytical methods include: association rule mining (e.g., Apriori algorithm), principal component analysis (PCA), or using machine learning models (e.g., random forest) to find the key input factors (key input factors may include: fluctuations in incoming material properties, changes in ambient temperature, equipment wear and tear, operating habits, etc.) and their influence weights, and obtain the analysis results; Based on the analysis results, specific process parameter modification suggestions are proposed; For example: modifying the baseline value of the setting table, adjusting the gain of the PID controller, adding a feedforward compensation link, or changing the equipment maintenance cycle; The optimized production links are deployed to the control system of the actual production line; At the same time, the simulation model of the corresponding production links is also updated, so that the next simulation and exchange are based on a better foundation, thereby achieving the optimization of the corresponding production process of the steel production links.

[0032] Example 2, please refer to Figure 3 As shown in this embodiment, the steel production process optimization method based on industrial simulation includes the following steps: For each stage of the steel production line, both an actual operating line and a simulated operating line are configured. Collect real-time operating condition data and add it to the actual operating line. The simulation operating line performs advance analysis based on the real-time operating condition data and determines in advance whether there is fault information in the real-time operating condition data based on the advance analysis. When fault information is confirmed, the predicted fault point corresponding to the real-time operating data is determined based on the fault information, the predicted control parameters are determined based on the predicted fault point, and the predicted control parameters on the simulation line are exchanged to the actual operating line using a preset exchange assistant. Set an exchange frequency threshold, optimize the production process that reaches the exchange frequency threshold, and apply the optimization results to the corresponding production process.

[0033] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A steel production process optimization system based on industrial simulation, characterized in that, include: The configuration module is used to configure the actual operating line and the simulated operating line for each production stage of the steel production line. The pre-run module is used to collect real-time operating condition data and add the real-time operating condition data to the actual operating line. The simulation operating line performs pre-run analysis based on the real-time operating condition data and judges in advance whether there is fault information in the real-time operating condition data based on the pre-run analysis. The exchange module is used to determine the predicted fault point corresponding to the real-time operating condition data based on the fault information when the existence of fault information is determined, to determine the predicted control parameters based on the predicted fault point, and to exchange the predicted control parameters on the simulation line to the actual operating line using a preset exchange assistant. The optimization module is used to set the exchange frequency threshold, optimize the production process that reaches the exchange frequency threshold, and apply the optimization results to the corresponding production process.

2. The steel production process optimization system based on industrial simulation according to claim 1, characterized in that, The configuration module includes: The division unit is used to divide a steel production line into multiple production stages; The first configuration unit is used to configure the corresponding actual operating line for each production stage; The second configuration unit is used to perform industrial simulation for each production stage, obtain the simulation model corresponding to the production stage, and configure the simulation run line for the corresponding simulation model.

3. The steel production process optimization system based on industrial simulation according to claim 1, characterized in that, The pre-simulation module includes: The data acquisition unit is used to collect real-time operating condition data of the corresponding production process, wherein the real-time operating condition data is the comprehensive parameter of the corresponding production process. The collection unit is used to determine multiple historical operating condition data sequences. Each historical operating condition data sequence includes the operating condition data sequence, fault information, and historical control parameters. A fault information database is established based on multiple historical operating condition data sequences. The pre-simulation unit is used to adapt real-time operating condition data to the time points on the actual operating line. The simulated operating line performs a pre-simulation of the actual operating line based on real-time operating condition data and adjacent historical operating condition data. Based on a preset fault information database, the real-time operating condition data and adjacent historical operating condition data are analyzed, and the fault information of the historical operating condition data sequence corresponding to the real-time operating condition data and adjacent historical operating condition data that reach a preset similarity threshold is used as the fault information of the real-time operating condition data. The actual operating line and the simulated operating line are both set with equidistant time points.

4. The steel production process optimization system based on industrial simulation according to claim 3, characterized in that, The pre-rehearsal unit also includes: The analysis unit compares real-time operating condition data and adjacent historical operating condition data with multiple historical operating condition data sequences. During the comparison process, it gradually eliminates historical operating condition data sequences that do not meet the preset similarity threshold with the real-time operating condition data and adjacent historical operating condition data until a unique historical operating condition data sequence that meets the preset similarity threshold with the real-time operating condition data and adjacent historical operating condition data is found. If so, it is determined that there is fault information in the real-time operating condition data and adjacent historical operating condition data, and the historical operating condition data sequence containing the fault information is taken as the fault operating condition data sequence. If neither the real-time operating condition data nor the adjacent historical operating condition data meets the preset similarity threshold with any historical operating condition data sequence, it is determined that there is no fault information in the real-time operating condition data and adjacent historical operating condition data.

5. The steel production process optimization system based on industrial simulation according to claim 4, characterized in that, The switching module includes: The first setting unit is used to determine the predicted fault point corresponding to the real-time operating condition data based on the fault information when it is determined that there is fault information, and to bind the predicted control parameters corresponding to the predicted fault point to the predicted fault point. The second setting unit is used to pre-configure execution points on the actual operating line. The execution points correspond to real-time operating data, and the execution points correspond to accompanying points on the simulation operating line. The execution points and accompanying points slide synchronously on the actual operating line and the simulation operating line, respectively. The switching unit is used to adapt the fault condition data sequence forward onto the simulation operation line through the predicted fault point. The starting point of the fault condition data sequence on the simulation operation line is used as the pre-configured execution point on the actual operation line. The predicted fault start point and the predicted fault point are used as the associated space. The positions of the predicted fault start point and the predicted fault point on the actual operation line are used as the fault start point and subsequent fault point, respectively. The space between the fault start point and the subsequent fault point is used as the native space. An switching assistant is set between the associated space and the native space. The switching assistant is used to send the predicted control parameters in the associated space on the simulation operation line to the actual operation line.

6. The steel production process optimization system based on industrial simulation according to claim 5, characterized in that, The switching unit further includes: Determine the docking duration, and based on the docking duration, set multiple corresponding associated tangent points and actual tangent points on the simulation operation line and the actual operation line respectively. The associated tangent points and actual tangent points correspond to each other to form a tangent point pair. For multiple associated tangent points in the associated space, the corresponding predicted operating condition data and predicted control parameters are adapted respectively, and the actual operating condition data is filled into the corresponding actual tangent points. The exchange assistant connects to the execution point and the accompanying point. The exchange assistant carries the execution point and the accompanying point and slides on multiple accompanying tangent points and multiple actual tangent points until the accompanying tangent point and the actual tangent point reach the preset conditions. Then, the exchange assistant is triggered to send the predictive control parameters of the next accompanying tangent point to the corresponding actual tangent point, so as to realize the exchange of predictive control parameters on the simulation line to the actual line.

7. The steel production process optimization system based on industrial simulation according to claim 6, characterized in that, The optimization module includes: The aggregation unit is used to obtain the exchange frequency between the actual running line and the simulated running line corresponding to each production link, mark the original space that is greater than the preset exchange frequency threshold, and take the production link corresponding to the marked original space as the point to be optimized. The optimization unit is used to perform multi-factor analysis on the production process corresponding to the point to be optimized, optimize the corresponding production process based on the results of the multi-factor analysis, and apply the optimized production process.

8. A method for optimizing steel production processes based on industrial simulation, used to implement the steel production process optimization system based on industrial simulation as described in any one of claims 1-7, characterized in that, Includes the following steps: For each stage of the steel production line, both an actual operating line and a simulated operating line are configured. Collect real-time operating condition data and add it to the actual operating line. The simulation operating line performs advance analysis based on the real-time operating condition data and determines in advance whether there is fault information in the real-time operating condition data based on the advance analysis. When fault information is confirmed, the predicted fault point corresponding to the real-time operating data is determined based on the fault information, the predicted control parameters are determined based on the predicted fault point, and the predicted control parameters on the simulation line are exchanged to the actual operating line using a preset exchange assistant. Set an exchange frequency threshold, optimize the production process that reaches the exchange frequency threshold, and apply the optimization results to the corresponding production process.

Images