Smeltery multi-process-section collaborative optimization method based on automatic control

Abstract

The invention discloses a smelting plant multi-process-section collaborative optimization method based on automatic control, and relates to the technical field of smelting intelligent collaborative control. The method comprises the following steps: carrying out quantitative calculation on the coupling degree between full-process variables and full-process performance indexes in a whole plant real-time state data set; automatically identifying process coupling control variables and screening redundant variables to form a collaborative optimization variable set; taking the collaborative optimization variable set as a decision variable in the whole-process collaborative optimization twin model, calculating a multi-process set value combination of each collaborative optimization variable through a rolling real-time optimization algorithm, performing dynamic simulation verification on the multi-process set value combination in a three-dimensional virtual simulation environment, and outputting collaborative optimization set values; according to the method, a closed-loop optimization control system based on cooperation of data driving and a process mechanism is established, so that cooperative regulation and control of set values of multiple working sections in the whole process of a smeltery are realized, and the production stability, safety and overall operation performance are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smelting intelligent collaborative control, and in particular to a smelter multi-process section collaborative optimization method based on automation control. BACKGROUND

[0002] With the development of the metallurgical industry towards intelligence and green, the automation and informatization level of smelter production process continues to improve, and sensors, automation instruments and distributed control systems (DCS) and other general control or regulation systems are commonly used in each process section to realize online monitoring and remote control of temperature, pressure, flow, concentration and other parameters, so that the running state of the main equipment can be visualized to meet the basic safety production and process stability requirements. At present, some gold smelting enterprises have built online monitoring systems for roasting, acid making, leaching and water treatment processes, and have basic infrastructures such as remote instruments, industrial networks and video monitoring, which can realize process parameter recording, alarm prompting and remote start-stop operations, and gradually deploy data acquisition, visual display and digital twin platforms to assist production operation supervision and decision analysis. On this basis, the control center presents the real-time data of different process sections in a centralized manner, so that operators can view the running information of the full-process equipment in the general control or regulation system interface, realize centralized supervision and trend analysis of multi-system data, and thus provide support for stable operation of smelting production.

[0003] The existing technical system plays an important role in real-time monitoring, remote operation and local automatic regulation, but as the complexity of smelting production deepens, its requirements for collaboration between multiple processes and overall operation efficiency continue to increase. In the traditional mode, the general control or regulation systems of each process section are often independent of each other, and mainly rely on the experience of post personnel for operation scheduling, and cannot realize quantitative description of the coupling relationship between variables after process-level data fusion, resulting in that the device set value update is still mainly based on single-section optimization, while the overall plant indicators such as energy efficiency, yield and emission level have strong coupling characteristics, and local optimization is not global optimization. When the production fluctuates or the raw material conditions change, the decentralized general control or regulation system cannot quickly link multiple set values for collaborative optimization, which limits the resilience and energy efficiency improvement space of the full-process operation. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a smelter multi-process section collaborative optimization method based on automation control to solve the collaborative control problem of multi-process section set value unable to be linked and optimized due to insufficient full-process data fusion.

[0006] To solve the above technical problems, the present application provides the following technical solutions: The application provides a smelting plant multi-process section collaborative optimization method based on automatic control, which comprises the following steps of collecting smelting plant real-time production data, performing time alignment, data cleaning and mechanism constraint calculation to form a smelting plant real-time state data set; The coupling degree between all-process variables in the smelting plant real-time state data set and all-process performance indexes is quantitatively calculated, process coupling control variables are automatically identified and redundant variables are screened out to form a collaborative optimization variable set; In the all-process collaborative optimization twin model, the collaborative optimization variable set is taken as a decision variable, a multi-process set value combination of each collaborative optimization variable is calculated through a rolling real-time optimization algorithm, the multi-process set value combination is dynamically simulated and verified in a three-dimensional virtual simulation environment, and a collaborative optimization set value is output; The collaborative optimization set value is replaced by an original set value according to the relationship of each process section, all-process related production variables are automatically adjusted, and an optimization execution feedback data set is continuously recorded and generated in the production execution process.

[0007] As a preferred scheme of the smelting plant multi-process section collaborative optimization method based on automatic control, wherein the smelting plant real-time production data comprises process parameters, material composition and flow, energy and environmental indexes and equipment running states of each process section.

[0008] As a preferred scheme of the smelting plant multi-process section collaborative optimization method based on automatic control, wherein the smelting plant real-time production data comprises process parameters, material composition and flow, energy and environmental indexes and equipment running states of each process section. The smelting plant main clock is taken as a global reference time, the time stamp of smelting plant real-time production data with different sampling periods is corrected, and time-aligned time sequence data is obtained; The time-aligned time sequence data is complemented by interpolation, abnormal noise is eliminated by filtering, and physically unreasonable data is corrected to obtain standard smelting plant state data; The standard smelting plant state data is corrected and state variable is calculated to obtain a smelting plant real-time state data set.

[0009] As a preferred scheme of the smelting plant multi-process section collaborative optimization method based on automatic control, wherein the smelting plant real-time production data comprises process parameters, material composition and flow, energy and environmental indexes and equipment running states of each process section. The smelting plant real-time state data set is quantitatively calculated, and the coupling degree between all-process variables in the smelting plant real-time state data set and all-process performance indexes is quantitatively calculated, process coupling control variables are automatically identified and redundant variables are screened out to form a collaborative optimization variable set; The coupling degree between all-process variables in the smelting plant real-time state data set and all-process performance indexes is quantitatively calculated, process coupling control variables are automatically identified and redundant variables are screened out to form a collaborative optimization variable set; According to the coupling degree index, the full section variables are sorted by the coupling degree from high to low, and the full section variables whose coupling degrees meet the preset coupling degree threshold condition are screened out as the process coupling control variable set according to the sorting result.

[0010] As a preferred scheme of the multi-process section collaborative optimization method of the smelting plant based on automation control, wherein: the forming of the collaborative optimization variable set refers to the redundancy analysis on the mutual influence relationship between the full section variables in the process coupling control variable set, the full section variables whose correlation meets the preset redundancy condition are removed, and the process coupling control variable set after removing the redundancy is determined as the collaborative optimization variable set.

[0011] As a preferred scheme of the multi-process section collaborative optimization method of the smelting plant based on automation control, wherein: the full-process collaborative optimization twin model is formed by digitally modeling the equipment structure, process flow and control logic of each process section of the smelting plant, and parameterizing and dynamically updating by using the full-plant real-time state data set.

[0012] As a preferred scheme of the multi-process section collaborative optimization method of the smelting plant based on automation control, wherein: in the full-process collaborative optimization twin model, the collaborative optimization variable set is taken as a decision variable, and the multi-process set value combination of each collaborative optimization variable is calculated by a rolling real-time optimization algorithm, and the specific steps are as follows, According to the digital mapping relationship, each full section variable in the collaborative optimization variable set is one-to-one mapped with the process node of each process section of the smelting plant, the collaborative optimization variable set is set as a decision variable in the full-process collaborative optimization twin model, the comprehensive performance index is taken as an optimization objective, and a rolling real-time optimization problem is constructed; By using a rolling real-time optimization algorithm, the rolling real-time optimization problem is solved from the current working condition state as a starting point, the running track of the multi-process section of the smelting plant in a future period of time is predicted, and a collaborative optimization variable change track meeting the process constraint is searched within a preset optimization time domain; The multi-process set value combination of the current collaborative optimization variable is intercepted from the collaborative optimization variable change track.

[0013] As a preferred scheme of the multi-process section collaborative optimization method of the smelting plant based on automation control, wherein: in the three-dimensional virtual simulation environment, the multi-process set value combination is dynamically simulated and verified, and the collaborative optimization set value is output, and the specific steps are as follows, The numerical value of each collaborative optimization variable in the multi-process set value combination is mapped to the full-process virtual object in the three-dimensional virtual simulation environment, respectively, the full-process virtual object is dynamically simulated according to the multi-process set value combination with the current working condition state as a starting point, and a full-process performance simulation result is obtained; According to the preset process performance index and process constraint, the full-process performance simulation results of the multi-process setting value combination are judged, if any full-process performance simulation result is not up to standard, the multi-process setting value combination is rejected and returned to re-optimization, and if all are up to standard, the multi-process setting value combination is confirmed as the collaborative optimization setting value.

[0014] As a preferred scheme of the multi-process section collaborative optimization method of the smelting plant based on automatic control, wherein: the three-dimensional virtual simulation environment is obtained by collecting the process flow chart, equipment layout chart, equipment geometric structure parameters and field sensing point information of the smelting plant, reconstructing the spatial structure of each process section and equipment by using three-dimensional construction software, and hanging the real-time production data to the corresponding process equipment node according to the digital mapping relationship.

[0015] As a preferred scheme of the multi-process section collaborative optimization method of the smelting plant based on automatic control, wherein: the collaborative optimization setting value is replaced by the original setting value according to the relationship of each process section, and the full-section related production variables are automatically adjusted, and the optimization execution feedback data set is continuously recorded during the production execution process. The collaborative optimization setting value is sequentially issued to the smelting plant automatic control link according to the relationship of each process section, replaces the original running setting value to drive the corresponding execution mechanism, and automatically adjusts the full-section related production variables to the process setting range indicated by the collaborative optimization setting value. The real-time change data of the full-section related production variables is continuously recorded during the production execution process, and the optimization execution feedback data set is generated by organizing and archiving.

[0016] The beneficial effects of the present application are: accurate and reliable full-plant state data is constructed by real-time data cleaning and fusion, the quantitative analysis and redundancy screening of the coupling relationship between the full-section variables and the full-process performance are realized, and the key variable set for collaborative optimization is formed; and relying on the dynamically updated full-process collaborative optimization twin model, the multi-process setting value combination is calculated by using the rolling real-time optimization algorithm, and then automatically issued to the general control or regulation system for execution after safety and process feasibility verification by the three-dimensional virtual simulation, so as to establish a closed-loop optimization control system driven by data and coordinated by process mechanism, so that the general control or regulation system can realize the collaborative regulation of the full-process multi-section setting value, and improve the stability, safety and overall operation performance of the smelting production. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creating laborious work.

[0018] Figure 1 This is a flowchart of a multi-process collaborative optimization method for a smelter based on automated control.

[0019] Figure 2 A flowchart for obtaining the real-time status dataset of the entire plant.

[0020] Figure 3 This is a flowchart for extracting combinations of multiple process settings.

[0021] Figure 4 A flowchart for determining the set of variables for collaborative optimization.

[0022] Figure 5 This is a comparative analysis chart of energy consumption control effects.

[0023] Figure 6 This is a chart comparing the stability of yield. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-6 This is one embodiment of the present invention, which provides a method for collaborative optimization of multiple process stages in a smelter based on automated control, comprising the following steps: S1. Collect real-time production data from the smelter, and perform time alignment, data cleaning, and mechanism constraint calculations to form a real-time status dataset for the entire plant.

[0028] It should be noted that the real-time production data of the smelter includes process parameters of each process section, material composition and flow rate, energy and environmental indicators, and equipment operating status.

[0029] S1.1. Using the smelter's master clock as the global reference time, timestamp correction is performed on the smelter's real-time production data with different sampling periods to obtain time-aligned timing data.

[0030] Furthermore, a unified global time index table is generated using the smelter's master clock, covering the time window to be analyzed and dividing the time points according to the minimum sampling period. The original real-time production data of each process segment, equipment, and raw data are mapped to the corresponding time points in the global time index table based on their original timestamps through nearest neighbor matching or linear interpolation. For high-frequency transient event data that cannot be reasonably completed through interpolation, their original timestamps are retained and marked as event-type data, forming time-aligned time-series data under a unified global time index.

[0031] S1.2. Time-aligned time-series data are interpolated to fill in missing values, filtered to eliminate abnormal noise, and physically unreasonable data is corrected to obtain standard plant-wide status data.

[0032] Furthermore, for missing values ​​of each variable in the time-aligned time series data, based on its physical context and adjacent valid data points, piecewise linear interpolation is used to fill in the missing time intervals of a certain variable in the time-aligned time series data. The two most recent valid data points before and after the missing interval are selected, and a linear function is constructed using their corresponding timestamps and values. The interpolation result is calculated according to the linear function at the missing time points, and all continuous missing intervals are completed piecewise. For data points with abnormal noise, sliding window mid-range filtering or wavelet denoising methods are applied for smoothing. The outlier value is determined based on the equipment operating status (e.g., temperature rise when the equipment is stopped, air volume is non-zero when the fan is not running). Data that violates the basic physical laws of the metallurgical process (e.g., material imbalance, energy non-conservation, temperature exceeding the material melting point, etc.) is identified and corrected to generate standard plant-wide status data that conforms to the process logic.

[0033] S1.3. Correct the standard plant-wide status data and calculate the status variables to obtain the plant-wide real-time status dataset.

[0034] Furthermore, the standard plant-wide status data is compared item by item with the preset metallurgical process mechanism consistency rules to identify and correct data points that violate the metallurgical process mechanism consistency rules. Based on the corrected measurable variables, and according to the material flow direction, energy transfer path and chemical reaction ratio relationship in the process flow, the unmeasurable key state variables (such as intermediate product composition, reaction thermal efficiency and actual furnace temperature) are calculated through algebraic operations or conservation formulas. By fusing the corrected original variables and the calculated state variables, a real-time plant-wide status dataset that satisfies the process logic and covers all process sections of the plant is generated.

[0035] It should be noted that the consistency rule of metallurgical process mechanism is set based on the basic engineering principles of material conservation, energy conservation, chemical reaction stoichiometry, and physical boundaries of equipment operation (such as reasonable ranges of temperature, pressure, and flow rate) in each process, combined with actual process design parameters and historical stable operation data.

[0036] S2. Quantify the degree of coupling between variables of all sections and performance indicators of the entire process in the real-time status data of the whole plant, automatically identify process coupling control variables and filter out redundant variables to form a set of collaborative optimization variables.

[0037] S2.1. Centralize the real-time status data of the entire plant into variables of all work sections and form sample data pairs with the full-process performance indicators at the corresponding time.

[0038] Furthermore, variable data for all process segments are extracted from the real-time status dataset of the entire plant, including process parameters, material composition, flow rate, energy consumption, and equipment status of each process; performance index data for the entire process at the same timestamp are acquired simultaneously, including metal yield, unit energy consumption, emission intensity, and product quality grade; and the set of variables for each process segment at each time point is paired with the corresponding set of performance indexes for the entire process to form structured sample data pairs.

[0039] It should be noted that the full-process performance indicators are automatically generated by collecting the output, raw material input, energy medium consumption, environmental emission monitoring data, and final product quality test results of each process in the smelter, under a unified time benchmark, and according to statistical caliber and accounting cycle (such as shift, day, and real-time sliding window).

[0040] like Figure 6 As shown, the metal yield of the smelter varies significantly over time under different control methods. The yield curve of the collaborative optimization control group is more stable throughout the entire time period, and the overall distribution of data points is close to the reference mean, indicating that the stability of the process is effectively improved by setting the collaborative optimization of key variables. The enlarged partial view further shows that the collaborative optimization group can still recover quickly and maintain a high level after the disturbance occurs, demonstrating that the response capability and robustness of the method of this invention to yield control under dynamic operating conditions are superior to traditional control methods.

[0041] S2.2 Quantitatively calculate the degree of coupling between variables of the entire process section and performance indicators of each process using sample data, and obtain the coupling degree index.

[0042] Furthermore, for each pair of full-process variables and full-process performance indicators, based on the sample data pairs formed under the corresponding timestamps, the statistical correlation between the two is quantitatively analyzed. According to the variable type and correlation characteristics, methods such as mutual information or Pearson correlation coefficient are used to calculate the coupling degree between each full-process variable and all full-process performance indicators one by one. All calculation results are stored in a structured manner to generate a coupling degree index containing variable name, corresponding performance indicator and coupling degree value.

[0043] It should be noted that the degree of coupling is calculated one by one to form a multidimensional numerical result of the coupling degree, expressed as follows: ; In the formula, It is a variable for the entire work section. With overall process performance indicators The Pearson correlation coefficient between the two is a dimensionless index used to measure the degree of linear correlation between them. It is a variable for the entire work section. With overall process performance indicators The covariance is a statistic that reflects the joint trend of change of the two samples, obtained by calculating the sample mean of the product of their respective deviations from their sample means at the same time stamp. It is the mathematical expectation of the product of deviations, i.e., the definition of covariance, representing the variables across the entire work section. With overall process performance indicators The average of the products of deviations from their respective means; It is a variable for the entire work section. The sample standard deviation measures the variables across the entire work section. The degree of dispersion around its mean; It is a full-process performance indicator The sample standard deviation is used to measure the overall performance of the process. The degree of dispersion around its mean; It is a variable for the entire work section. The sample mean, which is the average value of the variables across the entire work section within the observation time window; It is a full-process performance indicator The sample mean, that is, the average value of the overall process performance index within the same time window; It is time series data of variables for a certain entire process section (such as oxygen flow rate, temperature, or equipment operating frequency in a certain process). It is time series data of a certain whole process performance index (such as metal yield, unit energy consumption or emission intensity, etc.); the variables of the whole process section and the whole process performance index have different physical dimensions, but when calculating the coupling degree index, through mean removal and standardization, they are transformed into dimensionless statistics to characterize the synchronous change relationship between the two.

[0044] like Figure 5 As shown, comparing the unit energy consumption trends of the traditional control group and the collaborative optimization control group, the collaborative optimization group exhibits lower overall energy consumption and smaller fluctuations throughout the entire time period. Most data points are distributed within the ideal reference range, indicating that the variable coupling screening mechanism proposed in this invention can effectively reduce the overall energy consumption level. A magnified view further reveals that within a typical disturbance window, the energy consumption response curve of the collaborative optimization group changes more gradually and has a shorter recovery time, reflecting its stronger energy-saving control accuracy and adjustment capability.

[0045] S2.3. Sort all process variables from high to low according to the coupling degree index, and select all process variables whose coupling degree meets the preset coupling degree threshold condition based on the sorting results, as the set of process coupling control variables.

[0046] Furthermore, the coupling degree indices calculated for each full-process variable and its corresponding full-process performance indicators are summarized, and the maximum value is taken as the representative coupling degree value of that full-process variable. All full-process variables are sorted in descending order of their representative coupling degree values ​​to form a full-process variable coupling sorting list. The full-process variable coupling sorting list is traversed according to the coupling degree threshold, and full-process variables with representative coupling degree values ​​greater than or equal to the coupling degree threshold are selected. For example, when the coupling degree threshold is set to 0.3, full-process variables with representative coupling degree values ​​not less than 0.3 are determined to have a coupling effect on the full-process performance indicators and are selected to enter the process coupling control variable set. The selected full-process variables form the process coupling control variable set.

[0047] It should be noted that the coupling degree threshold is set by statistically analyzing the distribution characteristics of the coupling degree indicators between variables of the entire process section and the performance indicators of the entire process recorded under historical stable operating conditions of the smelter; based on the statistical distribution of the coupling degree indicators in the historical stable operating condition data, the value range of the coupling degree threshold is set to 0.2 to 0.5.

[0048] S2.4. Perform redundancy analysis on the mutual influence relationships among variables of all processes in the process coupling control variable set, remove variables of all processes whose correlation meets the preset redundancy conditions, and determine the process coupling control variable set after redundancy removal as the collaborative optimization variable set.

[0049] Furthermore, all observed values ​​of each process segment variable under a unified global time index are obtained from the set of process coupling control variables. Each process segment variable is then paired with the remaining process segment variables in the set of process coupling control variables to form multiple pairs. Based on the time series values ​​corresponding to the process segment variables in each pair, the expected value of the product of the two variables after deviating from the sample mean is calculated in the sample space. The expected value is then normalized using the sample standard deviation as a scaling factor, resulting in a correlation value representing the strength of the linear association between the process segment variables for each pair. After obtaining all correlation values ​​representing the strength of the association between the process segment variables, each correlation value is compared with a preset redundancy condition. The comparison operation identifies process segment variable pairs whose correlation values ​​satisfy the preset redundancy condition. For each pair that satisfies the preset redundancy condition... For all pairs of variables with remaining conditions, the representative coupling degree values ​​between the variables and the overall process performance indicators, the dispersion of the variables under historical stable operating conditions, and the effectiveness of the adjustment methods of the variables in the metallurgical process control are compared. Based on the comparison results, variables with low contribution to the improvement of overall process performance indicators or unclear adjustment effects are identified as redundant variables and eliminated. Only variables with control value and more critical impact on performance improvement goals (such as material ratio parameters of key processes, energy medium supply, reaction temperature or pressure setpoints, and equipment operation setpoints directly involved in closed-loop control) are retained. After all pairs have undergone comparison and elimination operations, only the remaining variables in the process coupling control variable set are the remaining variables. This remaining set is directly determined as the collaborative optimization variable set.

[0050] It should be noted that the redundancy condition is based on the statistical distribution characteristics of the correlation values ​​between variables in all sections during the historical stable operation of the smelter. It is set by statistically organizing the correlation values ​​between all variables in all sections within the time window of the historical stable operation and observing the distribution pattern of the correlation values ​​in the high, medium and low correlation intervals.

[0051] S3. In the full-process collaborative optimization twin model, the collaborative optimization variable set is used as the decision variable. The multi-process set value combination of each collaborative optimization variable is calculated by the rolling real-time optimization algorithm. The multi-process set value combination is dynamically simulated and verified in a three-dimensional virtual simulation environment, and the collaborative optimization set value is output.

[0052] It should be noted that the full-process collaborative optimization twin model is formed by digitally modeling the equipment structure, process flow and control logic of each process section of the smelter, and by using the real-time status dataset of the entire plant for parameter calibration and dynamic updates.

[0053] Training a full-process collaborative optimization twin model involves inputting process parameters, material composition and flow rate, energy and environmental indicators, equipment operating status, and corresponding full-process performance indicators from a historical stable operating period into the full-process collaborative optimization twin model according to a unified time base. The full-process collaborative optimization twin model then predicts the state variable vector sequence for the entire process segment and compares the results with the real-time status dataset of the entire plant and the actual observed sequences of the full-process performance indicators. Based on the error comparison results, iterative adjustments are made to the metallurgical process parameters, equipment characteristic parameters, and dynamic response parameters in the full-process collaborative optimization twin model, gradually bringing the predicted results closer to the observed results. Once the error converges to a preset range, the parameters are recorded as the current optimal parameter set. During subsequent production, new real-time status datasets from the entire plant are continuously received, and the error comparison and parameter adjustment operations are repeated. This allows the full-process collaborative optimization twin model to dynamically update with changes in production status and maintain the accuracy of its future operational trajectory predictions.

[0054] S3.1 Based on the digital mapping relationship, each full-process variable in the collaborative optimization variable set is mapped one-to-one with the process nodes of each process section of the smelter. In the full-process collaborative optimization twin model, the collaborative optimization variable set is set as the decision variable, and the comprehensive performance index is used as the optimization target to construct a rolling real-time optimization problem.

[0055] Furthermore, based on the digital mapping relationship, the equipment number, measuring point number, and process segment identifier recorded in the digital mapping relationship for each full-process variable in the collaborative optimization variable set are mapped one-to-one with the process nodes of each process segment in the smelter, completing the mapping operation from the collaborative optimization variable set to specific process nodes. In the full-process collaborative optimization twin model, each full-process variable in the collaborative optimization variable set is set as a decision variable, and the allowable range of change, the rate of change limit, and the time scale of action are limited for each decision variable. Taking the comprehensive performance index as the optimization objective, weight coefficients are configured by selecting full-process performance indicators such as metal yield, unit energy consumption, emission intensity, and product quality grade, and the full-process performance indicators and weight coefficients are combined into a single comprehensive evaluation value to clarify the direction of improvement of the comprehensive performance index in the optimization process. Based on the above settings, the collaborative optimization variable set, comprehensive performance index, and smelter process constraints, equipment operation constraints, and safety constraints are written into the same optimization problem description to form a rolling real-time optimization problem.

[0056] It should be noted that the comprehensive performance index is a single indicator used to evaluate the overall operating effect of a smelter, which is formed by weighting the performance indicators of the entire process, such as metal yield, unit energy consumption, emission intensity, and product quality grade, according to the weight coefficients. Smelter process constraints are limitations on the material flow, energy conversion, and chemical reaction stoichiometry. They are obtained by determining the reasonable balance between materials, energy, and chemical reactions through the consistency rules of metallurgical process mechanisms and historical stable operating data. Equipment operation constraints are limitations on the equipment's operating capacity, startup status, and allowable working range. These constraints are obtained by identifying the equipment's allowable operating range, maximum load capacity, and operating limitations through equipment design parameters, equipment maintenance records, and real-time equipment operating status. Safety constraints are safety boundary limits for operating conditions such as temperature, pressure, and flow rate to avoid hazards. They are obtained by identifying the boundary ranges of unsafe operating conditions such as temperature, pressure, and flow rate through long-term safety production monitoring data of the smelter.

[0057] It should also be noted that, based on the full-process collaborative optimization twin model and combined with on-site video data, real-time data backend services and visualization interaction capabilities, it is possible to realize functions such as global display of process flow, operation monitoring linkage, three-dimensional visualization early warning, and dynamic inspection. Relevant process parameters and equipment status are linked to the full-process collaborative optimization twin model in real time, which can realize a one-to-one correspondence between data and equipment spatial structure. Digital mapping includes the mapping of monitoring parameters of various equipment in the process flow, the mapping of online monitoring data such as liquid level, material level, flow rate, temperature and pressure in the process flow, the mapping of monitoring video data, the mapping of weather forecasts, and production management. It can be set to display various parameters of key equipment in real time, and can also click on the equipment to pop up the interface of various parameters of the equipment when needed. At the same time, in important sections equipped with video monitoring, you can click on the virtual camera at the location to pop up the monitoring screen captured by the real camera and see the real situation on site.

[0058] S3.2. Using the rolling real-time optimization algorithm, the rolling real-time optimization problem is solved starting from the current operating condition. The operating trajectory of multiple process sections of the smelter in the future is predicted, and the trajectory of the change of collaborative optimization variables that meet the process constraints is searched within the preset optimization time domain.

[0059] Furthermore, the current operating state corresponding to the collaborative optimization variable set is extracted from the real-time status dataset of the entire plant, and the current operating state is set as the starting point of the rolling real-time optimization problem in the full-process collaborative optimization twin model. Multiple prediction times are divided according to the time step within the preset optimization time domain, generating various candidate collaborative optimization variable change sequences for the collaborative optimization variable set. Each candidate collaborative optimization variable change sequence is input into the full-process collaborative optimization twin model to extrapolate the operating trajectory of multiple process sections of the smelter in the future time range time step by step. During the extrapolation of each operating trajectory, the smelter's process constraints, equipment operation constraints, and safety constraints are simultaneously set as constraints. Candidate collaborative optimization variable change sequences that violate the smelter's process constraints, equipment operation constraints, and safety constraints are eliminated or adjusted to the allowable range during the search process. Through traversal search or intelligent search methods, the collaborative optimization variable change trajectory with the best comprehensive performance index within the preset optimization time domain and which satisfies all smelter process constraints, equipment operation constraints, and safety constraints is selected from the operating trajectories corresponding to all candidate collaborative optimization variable change sequences. This trajectory is used as the collaborative optimization variable change trajectory obtained by the rolling real-time optimization algorithm at that optimization time.

[0060] It should be noted that the optimization time domain is the time range covered by the rolling real-time optimization algorithm in future operation prediction and collaborative optimization variable change sequence search, used to limit the length of the collaborative optimization variable change trajectory; the optimization time domain is set based on the characteristics of the smelter's production adjustment response cycle, equipment dynamic inertia and process performance index feedback delay, and after combining the statistical results of the time window required for each process section of the smelter to complete one effective adjustment under historical stable operating conditions, the time length that can balance the effectiveness of prediction and the real-time performance of optimization is selected.

[0061] S3.3 Extract the combination of multi-process setpoints of the current collaborative optimization variable from the trajectory of the collaborative optimization variable change.

[0062] Furthermore, by comparing the discrete time series used in the collaborative optimization variable change trajectory with the unified global time index corresponding to the current control moment, the time index position corresponding to the current control moment is determined at the position where the two time values ​​are equal. At this time index position, the target value of each full-section variable in the collaborative optimization variable set in the collaborative optimization variable change trajectory is read sequentially, and organized according to the correspondence between the process section and equipment to which the full-section variable belongs. The target values ​​of all full-section variables in the collaborative optimization variable set at the current control moment are combined in a unified order to form the multi-process set value combination of the current collaborative optimization variable.

[0063] S3.4 Map the values ​​of each collaborative optimization variable in the multi-process setting value combination to the full-process virtual object in the three-dimensional virtual simulation environment. Starting from the current working condition, perform dynamic simulation on the full-process virtual object according to the multi-process setting value combination to obtain the full-process performance simulation results.

[0064] Furthermore, the values ​​of each co-optimization variable in the multi-process setpoint combination are written into the attribute fields of the corresponding process equipment nodes in the 3D virtual simulation environment according to the digital mapping relationship. This ensures that the virtual objects of the entire process in the 3D virtual simulation environment are consistent with the current multi-process setpoint combination of the co-optimization variables at the parameter level. The state variables of the virtual objects of the entire process in the 3D virtual simulation environment are initialized using the current operating status recorded in the real-time status data of the entire plant. After the virtual objects of the entire process have completed the state initialization, the dynamic calculation function inside the 3D virtual simulation environment is enabled. When the operation process of the virtual objects of the entire process is advanced in a time-step manner, the time step is moved forward one by one according to the unified global time index in the 3D virtual simulation environment. At each time step, the state variable update result of the virtual objects of the entire process is recalculated according to the setpoint values ​​of the state variables and co-optimization variables of the virtual objects of the entire process at the previous moment. This continuously simulates the operation and change process of the smelter over time. After the dynamic simulation is completed, the process parameters, material composition and flow rate, energy and environmental indicators and product quality simulation results output from the 3D virtual simulation environment are summarized according to the statistical caliber of the entire process performance indicators to obtain the entire process performance simulation results.

[0065] It should be noted that the 3D virtual simulation environment is obtained by collecting the process flow diagram, equipment layout diagram, equipment geometric parameters and on-site sensor point information of the smelter, using 3D construction software to reconstruct the spatial structure of each process section and equipment, and attaching real-time production data to the corresponding process equipment nodes according to digital mapping relationships.

[0066] Additional notes: The 3D virtual simulation environment can simulate the equipment operation process and provides color coding, animation effects and trend display functions, making the comparison of the performance results of the whole process more intuitive. The 3D virtual simulation supports color display of equipment operation status, alarm location and highlight display of fault areas.

[0067] S3.5. Based on the preset process performance indicators and process constraints, the full-process performance simulation results of the combination of multiple process setting values ​​are judged. If any full-process performance simulation result fails to meet the standard, it is rejected and returned for re-optimization. If all meet the standard, the combination of multiple process setting values ​​is confirmed as the collaborative optimization setting value.

[0068] Furthermore, the process constraints include smelter process constraints, equipment operation constraints, and safety constraints. The simulation results of each full-process performance, such as metal yield, unit energy consumption, emission intensity, and product quality grade, are compared with the preset process performance indicators within their allowable ranges. Simultaneously, it is verified whether the process variables corresponding to the full-process performance simulation results meet the smelter process constraints, equipment operation constraints, and safety constraints. During the comparison process, if any full-process performance simulation result is found to exceed the preset process performance indicators' allowable range or violate the smelter process constraints, equipment operation constraints, or safety constraints, the current multi-process setpoint combination is judged as a non-compliant multi-process setpoint combination, and the rolling real-time optimization problem is returned to the collaborative optimization variable change trajectory search step for re-execution of optimization. When all full-process performance simulation results are within the preset process performance indicators' allowable range and fully meet the smelter process constraints, equipment operation constraints, and safety constraints, the current multi-process setpoint combination is directly confirmed and recorded as the collaborative optimization setpoint.

[0069] It should be noted that the allowable range of process performance indicators is determined by statistically analyzing the normal fluctuation range of the performance indicators of the entire process under the historical stable operating conditions of the smelter, and combining the range with the metallurgical process design requirements and product quality standards.

[0070] S4. Replace the original settings with collaborative optimization settings according to the relationship between each process segment, automatically adjust the relevant production variables of the entire process segment, and continuously record and generate optimization execution feedback datasets during the production execution process.

[0071] S4.1 The collaborative optimization setpoints are sequentially distributed to the smelter automation control system according to the relationship of each process segment, replacing the original operating setpoints to drive the corresponding actuators, and automatically adjusting the relevant production variables of the entire process segment to the process setting range indicated by the collaborative optimization setpoints.

[0072] Furthermore, the collaborative optimization setpoints are updated to the corresponding setpoint fields in the smelter automation control system according to the process segment affiliation recorded in the collaborative optimization setpoints. This allows the smelter automation control system to replace the original operating setpoints with the collaborative optimization setpoints, driving the actuators to adjust relevant production variables across the entire process segment. The actuators then change the actual values ​​of relevant production variables such as flow rate, temperature, and pressure, bringing these variables closer to the target range indicated by the collaborative optimization setpoints. The smelter automation control system then continuously verifies the deviation between the actual values ​​and the collaborative optimization setpoints using the plant's real-time status dataset and continuously corrects the control output, ultimately ensuring that all relevant production variables remain stably within the process setting range corresponding to the collaborative optimization setpoints.

[0073] S4.2 Continuously record the real-time changes of relevant production variables throughout the entire production process, and generate an optimized execution feedback dataset by organizing and archiving the data.

[0074] Furthermore, using a unified global time index as the time base, real-time measurements of relevant production variables across the entire process section, such as flow rate, temperature, pressure, and equipment start-up / shutdown status, are continuously collected. The collected relevant production variables across the entire process section are then organized in chronological order along with the current collaborative optimization setpoints and equipment response status, and archived according to process section number and variable name. At the end of the recording period, the relevant production variables, collaborative optimization setpoints, and equipment response status across the entire time interval are aggregated into a structured record file, named the Optimization Execution Feedback Dataset, for subsequent collaborative optimization effect analysis and rolling real-time optimization algorithm parameter updates.

[0075] In summary, this invention achieves quantitative analysis and redundancy screening of the coupling relationship between variables in all work sections and the performance of the entire process by constructing accurate and reliable plant-wide status data through real-time data cleaning and fusion, forming a set of key variables for collaborative optimization. Furthermore, relying on a dynamically updated twin model of the entire process collaborative optimization, it calculates combinations of multi-process setpoints using a rolling real-time optimization algorithm. After verifying safety and process feasibility through 3D virtual simulation, the results are automatically distributed to general control or regulation systems for execution. This establishes a closed-loop optimization control system that combines data-driven and process mechanism collaboration, enabling general control or regulation systems to achieve collaborative regulation of setpoints across multiple work sections throughout the entire process, thereby improving the stability, safety, and overall operational performance of smelting production.

[0076] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for multi-process section coordinated optimization of a smelter based on automation control, characterized in that: The application relates to a method for realizing full-process collaborative optimization of a smelting plant. Real-time production data of a smelting plant is collected, time alignment, data cleaning and mechanism constraint calculation are carried out, and a full-plant real-time state data set is formed; The coupling degree between full-section variables and full-process performance indexes in the full-plant real-time state data set is quantitatively calculated, process coupling control variables are automatically identified and redundant variables are screened out, and a collaborative optimization variable set is formed; In a full-process collaborative optimization twin model, the collaborative optimization variable set is taken as a decision variable, a multi-process set value combination of each collaborative optimization variable is calculated through a rolling real-time optimization algorithm, dynamic simulation verification of the multi-process set value combination is carried out in a three-dimensional virtual simulation environment, and collaborative optimization set values are outputted; The collaborative optimization set values are replaced by original set values according to the relationship of each process section, automatic adjustment is carried out on full-section related production variables, and an optimization execution feedback data set is continuously recorded and generated in the production execution process.

2. The automated control based multi-process section coordinated optimization method for a smelter of claim 1, wherein: The real-time production data of the smelting plant comprises process parameters, material composition and flow, energy and environmental indexes and equipment running states of each process section.

3. The automated control based multi-process section coordinated optimization method for a smelter of claim 2, wherein: The time alignment, data cleaning and mechanism constraint calculation for forming the full-plant real-time state data set specifically comprises the following steps, A smelting plant master clock is taken as a global reference time, time stamp correction is carried out on smelting plant real-time production data with different sampling periods, time alignment time sequence data is obtained, The time alignment time sequence data is used to complete missing values through interpolation, eliminate abnormal noise through filtering and correct physically unreasonable data, and standard full-plant state data is obtained; The standard full-plant state data is corrected and state variable calculation is carried out, and a full-plant real-time state data set is obtained.

4. The automated control based multi-process section coordinated optimization method for a smelter of claim 3, wherein: The quantitatively calculating the coupling degree between full-section variables and full-process performance indexes in the full-plant real-time state data set specifically comprises the following steps, Full-section variables in the full-plant real-time state data set and full-process performance indexes at corresponding moments are combined to form sample data pairs; The coupling degree between full-section variables and each full-process performance index is quantitatively calculated by using the sample data pairs, and a coupling degree index is obtained; Full-section variables are sorted according to the coupling degree from high to low according to the coupling degree index, and full-section variables with coupling degrees satisfying a preset coupling degree threshold condition are screened out according to the sorting result, and are taken as a process coupling control variable set.

5. The automated control based multi-process section coordinated optimization method for a smelter of claim 1, wherein: The forming of the collaborative optimization variable set refers to redundant analysis on the mutual influence relationship between full-section variables in the process coupling control variable set, elimination of full-section variables with a correlation satisfying a preset redundancy condition, and determination of the process coupling control variable set after the elimination of redundancy as the collaborative optimization variable set.

6. The automated control based multi-process section coordinated optimization method for a smelter of claim 1, wherein: The full-process collaborative optimization twin model is formed by digital modeling of equipment structures, process flows and control logics of each process section of the smelting plant and parameter calibration and dynamic updating by using the full-plant real-time state data set.

7. The automated control based multi-process section coordinated optimization method for a smelter of claim 1, wherein: In the full-process collaborative optimization twin model, the collaborative optimization variable set is taken as a decision variable, a multi-process set value combination of each collaborative optimization variable is calculated through a rolling real-time optimization algorithm, and specific steps are as follows, According to the digital mapping relationship, each full-section variable in the collaborative optimization variable set is one-to-one mapped with a process node of each process section of the smelting plant, the collaborative optimization variable set is set as a decision variable in the full-process collaborative optimization twin model, a comprehensive performance index is taken as an optimization target, and a rolling real-time optimization problem is constructed; By using a rolling real-time optimization algorithm, the rolling real-time optimization problem is solved from a current working condition state as a starting point, a running track of the smelting plant multi-process section in a future period of time is predicted, and a collaborative optimization variable change track meeting process constraints is searched within a preset optimization time domain; A multi-process setting value combination of the current collaborative optimization variable is intercepted from the collaborative optimization variable change track.

8. The automated control based multi-process section coordinated optimization method for a smelter of claim 1, wherein: The specific steps of dynamically simulating and verifying the multi-process setting value combination in the three-dimensional virtual simulation environment and outputting the collaborative optimization setting value are as follows, The numerical value of each collaborative optimization variable in the multi-process setting value combination is mapped to a full-process virtual object in the three-dimensional virtual simulation environment, respectively, the full-process virtual object is dynamically simulated according to the multi-process setting value combination with the current working condition state as a starting point, and a full-process performance simulation result is obtained; According to the preset process performance index and the process constraint, the full-process performance simulation result of the multi-process setting value combination is judged, if any full-process performance simulation result is not up to standard, the multi-process setting value combination is rejected and returned to re-optimization, and if all the full-process performance simulation results are up to standard, the multi-process setting value combination is confirmed as the collaborative optimization setting value.

9. The automated control based multi-process section coordinated optimization method for a smelter of claim 8, wherein: The three-dimensional virtual simulation environment is obtained by collecting a process flow chart, a device layout chart, device geometric structure parameters and field sensing point information of the smelting plant, reconstructing the spatial structure of each process section and device by using three-dimensional construction software, and obtaining real-time production data according to the digital mapping relationship and connecting the data to corresponding process device nodes.

10. The automated control based multi-process section coordinated optimization method for a smelter of claim 7, wherein: The specific steps of replacing the collaborative optimization setting value with the original setting value according to the relationship between each process section, automatically adjusting the full-section related production variables, and continuously recording and generating an optimization execution feedback data set during production execution are as follows, The collaborative optimization setting value is sequentially issued to the smelting plant automation control link according to the relationship between each process section, replaces the original running setting value to drive the corresponding execution mechanism, and automatically adjusts the full-section related production variables to the process setting range indicated by the collaborative optimization setting value; During the production execution process, the real-time change data of the full-section related production variables is continuously recorded, and the optimization execution feedback data set is generated by organizing and archiving.

Images