Non-ferrous metal industry chain coordination flexible scheduling method based on knowledge graph

Abstract

The non-ferrous metal industry chain collaborative flexible scheduling method based on a knowledge graph provided by the application belongs to the technical field of non-ferrous metal smelting intelligent manufacturing, and the application firstly deeply fuses the quantitative evaluation of furnace wall residue pollution and scheduling decision by constructing a co-linear smelting residue inheritance knowledge graph, fundamentally overcomes the technical defects of high quality risk and large capacity waste caused by the experience judgment of the traditional method, and significantly improves the product qualification rate when high-impurity and low-impurity orders are mixed-line produced.

Description

Technical Field

[0001] This invention belongs to the field of intelligent manufacturing technology for non-ferrous metal smelting, specifically involving a collaborative flexible scheduling method for the non-ferrous metal industry chain based on knowledge graphs. Background Technology

[0002] Non-ferrous metal smelting is a crucial foundational industry of the national economy, characterized by high temperatures, continuous production, and frequent switching between multiple product types. With the continuous growth in market demand for diverse, small-batch non-ferrous metal products, co-line smelting systems—that is, sequentially producing different grades and purity requirements of non-ferrous metals on the same smelting line—have become a common production organization model in the industry. During co-line smelting, after each heat, a certain thickness of residue inevitably remains on the inner surface of the furnace wall, containing various impurities from the previous heat. These residues are gradually released and enter the melt under the high-temperature smelting conditions of subsequent heats, posing a risk of cross-contamination to the chemical composition and performance indicators of subsequent products. How to scientifically quantify this inter-heat residual contamination effect and incorporate it into the production scheduling decision-making system is a key issue that non-ferrous metal smelting enterprises urgently need to address to achieve high-quality, flexible production.

[0003] Extensive research has been conducted by scholars and enterprises both domestically and internationally regarding the optimization of production scheduling and operation in the smelting process. At the model building level, existing research primarily employs mathematical programming methods to establish smelting batch planning and furnace scheduling models, and utilizes genetic algorithms and heuristic algorithms for solution. At the dynamic scheduling level, some solutions achieve dynamic adjustment of production plans by using real-time production data to drive digital twin models or rule-based phased scheduling models.

[0004] In recent years, knowledge graph technology has been increasingly introduced into the field of production scheduling due to its advantages in organizing multi-source heterogeneous data and mining semantic relationships between entities in complex industrial scenarios. For example, existing research has achieved dynamic scheduling of steel production lines by constructing integrated multi-model driven dynamic knowledge graphs; some scholars have also proposed methods for constructing large-scale knowledge graphs for the non-ferrous metals sector. However, most of the above-mentioned technical solutions focus on goals such as capacity balancing, improving equipment utilization, or optimizing order delivery cycles, failing to deeply characterize the physicochemical mechanisms of cross-furnace propagation of furnace wall residues in co-line smelting scenarios. Existing production scheduling methods, when assessing the risk of contamination between furnaces, either rely entirely on manual experience or use only simple impurity concentration thresholds as the switching basis, lacking quantitative characterization and modeling of key factors such as the dynamic decay characteristics of furnace wall residue deposits and the sensitivity of interactions between different impurity categories. This leads to production scheduling schemes that are either overly conservative, resulting in wasted capacity due to frequent insertion of cleaning furnaces, or too risky, with low-purity orders being contaminated by high-impurity residues, leading to substandard product quality. Furthermore, in actual production, the dynamic insertion of new orders is common. When faced with temporarily inserted high-impurity orders, the existing scheduling system often lacks the ability to systematically recalculate and optimize the pollution isolation measures in the existing production plan, making it difficult to achieve truly flexible closed-loop production scheduling. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this section, the abstract and title of the invention. Such simplifications or omissions shall not be used to limit the scope of the present invention.

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] To address the aforementioned technical problems, the present invention provides the following technical solution: A knowledge graph-based collaborative flexible scheduling method for the non-ferrous metals industry chain includes: treating each furnace in a co-line smelting system as a graph node, establishing edges between adjacent furnaces based on impurity sensitivity coefficients, and outputting a knowledge graph of co-line smelting residue inheritance; extracting high-impurity furnaces and adjacent furnaces to form a residue inheritance chain based on the knowledge graph, calculating the residue release intensity coefficient, marking the contamination propagation area and affected low-impurity orders, and outputting cross-furnace contamination inheritance results; delaying low-impurity orders by a safe waiting time based on the contamination inheritance results, and inserting slow-release furnaces and isolation cleaning furnaces between contaminated furnace pairs, and outputting flexible scheduling results based on residue inheritance constraints; linking the ore supply time with the flexible scheduling results, predicting furnace wall parameters for new orders and inserting them into the residue inheritance chain, clearing the original isolation furnaces, recalculating the contamination area, and re-inserting slow-release and isolation furnaces, and outputting the industry chain collaborative flexible scheduling results.

[0008] In a preferred embodiment of the present invention, the map node includes: measuring the thickness, impurity type, and temperature of the furnace wall adhesion layer after each furnace switching, which are respectively used as furnace wall retention thickness data, furnace wall impurity type data, and furnace wall hot temperature data; recording the thickness decay curve of each furnace over time, calculating the initial decay rate at the start of the decay curve and the time required for the thickness to decrease from the initial value to a specified proportion of the initial value as the decay time constant; and using the historical furnace switching sequence, furnace wall impurity type data, initial decay rate, decay time constant, and furnace wall hot temperature data as map node attributes.

[0009] As a preferred embodiment of the present invention, the establishment of an edge between adjacent furnaces based on the impurity sensitivity coefficient includes: for adjacent furnaces, establishing a directed edge with residual inheritance weights; The residual inheritance weight is equal to the furnace wall retention thickness data of the previous batch multiplied by the inter-impurity category sensitivity coefficient of the subsequent batch relative to the previous batch; the inter-impurity category sensitivity coefficient is read from the preset inter-impurity category sensitivity coefficient matrix.

[0010] As a preferred embodiment of the present invention, the construction of the residual inheritance chain includes: extracting furnace wall impurity category data from the knowledge graph, taking furnaces belonging to a preset high impurity category subset as starting nodes, continuously tracing downstream along directed edges, extracting subsequent furnaces reachable through one or more directed edges, until encountering a low impurity category furnace, and arranging the extracted furnaces in production order to form a residual inheritance chain.

[0011] As a preferred embodiment of the present invention, the calculation of the residual release intensity coefficient includes: calculating the residual release intensity coefficient of each pair of adjacent furnaces in the residual inheritance chain; the residual release intensity coefficient is equal to the residual inheritance weight of the pair of adjacent furnaces multiplied by the ratio of the initial decay rate of the previous furnace to the average initial decay rate of the metal category to which the previous furnace belongs in the same historical furnaces; if historical data is insufficient, the ratio is one.

[0012] As a preferred embodiment of the present invention, the marking of the contamination propagation area and the affected low-impurity orders includes: marking the furnace batches with the residual release intensity coefficient greater than a preset contamination threshold as contamination propagation areas, and recording the furnace batch numbers of subsequent furnace batches whose furnace wall impurity category data belong to a preset low-impurity category subset as affected low-impurity orders.

[0013] In a preferred embodiment of the present invention, the safe waiting period for delaying low-impurity orders based on the decay time constant includes: obtaining the contamination propagation area and the corresponding low-impurity order number, and delaying the switching start time of the low-impurity order by a safe waiting period; the safe waiting period is equal to a preset multiple multiplied by the decay time constant of the previous high-impurity furnace cycle; during the delay operation, the slow-release furnace cycle and the isolation cleaning furnace cycle are executed sequentially after the actual measurement of the decay curve of the high-impurity furnace cycle is completed; the actual switching start time of the low-impurity order is the latest of the actual completion time of the decay curve measurement and the end time of all slow-release and isolation furnace cycles.

[0014] As a preferred embodiment of the present invention, the insertion of a slow-release furnace and an isolation cleaning furnace between contaminated furnace pairs includes: inserting a slow-release furnace between two adjacent furnaces in the contamination propagation area; the impurity tolerance threshold of the slow-release furnace is equal to half of the upper limit of the allowable impurity concentration corresponding to the previous high-impurity furnace; and inserting an isolation cleaning furnace after the slow-release furnace, wherein the melting temperature of the isolation cleaning furnace is equal to the furnace wall hot temperature data of the previous high-impurity furnace plus a preset temperature difference, and the melting duration is equal to a preset multiple multiplied by the decay time constant of the previous high-impurity furnace.

[0015] As a preferred embodiment of the present invention, the step of predicting furnace wall parameters for new orders and inserting them into the residual inheritance chain includes: real-time monitoring of new high-impurity orders, obtaining the metal category and upper limit of metal impurity content of the new orders; obtaining the predicted value of furnace wall impurity category from a preset metal category and typical impurity category comparison table based on the metal category; taking the median of historical furnaces of the same metal category as the predicted value of furnace wall retention thickness and the predicted value of decay time constant, and taking the default value if the number of historical furnaces is less than the preset number; inserting the new order as a virtual furnace into the current residual inheritance chain, traversing each furnace in the chain, and if the upper limit of metal impurity content of the new order is lower than the upper limit of equivalent impurity content of the metal category produced in the current furnace, and there is no untreated contamination propagation area between the current furnace and the next furnace, then inserting the new order immediately after the current furnace; otherwise, appending it to the end of the chain.

[0016] As a preferred embodiment of the present invention, the step of recalculating the contaminated area and re-inserting the slow-release and isolation furnaces after clearing the original isolation furnaces includes: removing all slow-release furnaces and isolation cleaning furnaces inserted in the chain due to the previous contamination propagation area, retaining only the original production furnaces and the virtual furnaces of new orders, to obtain the basic chain; recalculating the residual release intensity coefficient of all adjacent furnace pairs in the basic chain, for new orders as edges of the preceding furnace, the initial decay rate of the new order is taken as the average initial decay rate of the same type of furnace in history for the same metal category, if the historical data is insufficient, the default value is taken; for new orders as edges of the following furnace, the residual inheritance weight is calculated using the predicted value of the furnace wall impurity category and the predicted value of the furnace wall retention thickness of the new order; marking the furnace pairs with the recalculated residual release intensity coefficient greater than the preset contamination threshold as new contamination propagation areas, and re-inserting the slow-release furnaces and isolation cleaning furnaces.

[0017] The beneficial effects of this invention are as follows: Compared with the prior art, the technical effects of this invention are as follows: By constructing a knowledge graph of residual inheritance in co-line smelting, this invention, for the first time, deeply integrates the quantitative assessment of furnace wall residual contamination with production scheduling decisions, fundamentally overcoming the technical defects of traditional methods that rely on experience-based judgment, resulting in high quality risks and large capacity waste, and significantly improving the product qualification rate when high-impurity and low-impurity orders are produced on the same production line; At the same time, through the automatic identification of contamination propagation areas driven by the residual release intensity coefficient and the dynamic waiting and two-stage purification insertion mechanism based on the decay time constant, the contamination path is accurately blocked while minimizing unnecessary capacity loss; Furthermore, through the closed-loop process of automatic removal of old isolated furnaces and systematic recalculation of contamination areas after the insertion of new orders, the production scheduling scheme is rapidly and adaptively adjusted to dynamic orders, and combined with the ore supply linkage mechanism, the optimization scope is extended to the entire industrial chain, effectively improving the overall production efficiency and on-time order delivery capability of enterprises. Attached Figure Description

[0018] Figure 1 This is a flowchart of the collinear melting residual inheritance knowledge graph construction method described in this invention.

[0019] Figure 2 This is a flowchart of the cross-furnace pollution propagation analysis method in this invention.

[0020] Figure 3 A schematic diagram of the structure of an electronic device for implementing the knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain in this embodiment of the invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0022] like Figures 1-2 As shown, the knowledge graph-based collaborative flexible production scheduling method for the non-ferrous metal industry chain of this invention includes: S1: Treat each furnace of the collinear melting system as a graph node, establish edges between adjacent furnaces according to the impurity sensitivity coefficient, and output the knowledge graph of collinear melting residue inheritance.

[0023] S2: Based on the knowledge graph, extract the high-impurity furnace batches and adjacent furnace batches to form a residual inheritance chain, calculate the residual release intensity coefficient, mark the pollution propagation area and the affected low-impurity orders, and output the cross-furnace pollution inheritance results.

[0024] S3: Based on the contamination inheritance result, delay the low-impurity order by a safe waiting time based on the decay time constant, and insert the slow-release furnace and the isolation cleaning furnace between the contamination furnace pairs, and output the flexible scheduling result based on the residual inheritance constraint.

[0025] S4: Based on the flexible production scheduling results and the linkage of ore supply time, predict furnace wall parameters for new orders and insert residual inheritance chains, clear the original isolated furnaces, recalculate the contaminated area and re-insert the slow-release and isolated furnaces, and output the industry chain collaborative flexible production scheduling results.

[0026] It should be noted that this embodiment first performs knowledge graph construction processing of the residual inheritance state of the collinear melting system, aiming to transform the physical state of the furnace wall of each production heat and the residual transfer relationship between heats into computable graph structure data. The specific operations are as follows: S1.1: For each heat executed sequentially in a collinear melting system, the heat number is used as the starting point. The labels were used to measure the thickness of the furnace wall coating, the type of impurities in the coating, and the temperature of the coating after the furnace cycle was completed.

[0027] The thickness of the furnace wall adhesion layer is denoted as The unit is millimeters. This data is obtained by laser ranging or ultrasonic thickness measurement at preset measurement points on the furnace wall. Furnace wall impurity category data is denoted as... From the preset impurity category set The set is pre-divided into a high-impurity category subset H and a low-impurity category subset L, which have no overlap. The furnace wall hot-state temperature data is denoted as... The temperature is measured in degrees Celsius and is collected in real time using embedded thermocouples or infrared thermal imaging devices.

[0028] After obtaining the above static parameters, the dynamic decay process of the furnace wall adhesion layer thickness over time was further monitored. The time of furnace switchover completion was taken as the zero point of timing. Record the thickness change curve over time. The monitoring continues until the thickness drops to a preset safe thickness threshold. The preset safe thickness threshold is pre-set based on the metal type and refractory material type of furnace i. Specifically, it is the statistical median of the minimum safe thickness of the residual adhesion layer on the furnace wall during normal production of that metal type (with a statistical sample of no less than 20 groups). If there is no historical data for that metal type, it is set to 0.15 mm by default. This value is set based on the safety engineering experience value of commonly used non-ferrous metal smelting furnace walls and can be calibrated by the user during system initialization.

[0029] To ensure the accuracy of curve fitting, at least three measured thickness values ​​at different times were collected. A continuous thickness decay curve function was obtained by fitting the curve using the least squares method or interpolation method.

[0030] Ideally, based on the fitted curve, the initial decay rate and decay time constant are calculated, where the initial decay rate is the value of the thickness curve at... The absolute value of the slope of the tangent line at point A is expressed as: ; The unit is millimeters per second. This parameter quantitatively characterizes the rate at which furnace wall residues peel off or volatilize in the initial stage after furnace switching.

[0031] The decay time constant is the thickness from the initial value decay to The required time is expressed as: ; The unit is seconds, and it is applicable to various practical working conditions such as exponential decay and power-law decay.

[0032] The historical furnace switching order and the furnace wall impurity category data for each furnace were used to... Initial decay rate decay time constant Furnace wall hot temperature data The equivalent impurity content upper limit of the metal produced in that furnace and the corresponding upper limit of the impurity content are stored together as the node attributes of the knowledge graph. Each furnace is abstracted as a node in the knowledge graph, and the node attributes fully reflect the residual contamination capacity and thermal evolution characteristics of that furnace. Among them, the calibration of the upper limit of the equivalent impurity content is performed on the actual production collinear smelting system; if online calibration is not possible due to production continuity limitations, it is allowed to be performed on a dedicated calibration test furnace with the same furnace inner diameter, refractory material grade, and heating power density as the production furnace (deviation not exceeding ±5%). The calibration result is multiplied by a correction factor (background concentration of production furnace / background concentration of calibration furnace) and then used in the production system. Taking electrolytic copper as the reference metal product, the specific calibration process is as follows: After the production furnace switch is completed, the surface state of the preset measurement area of ​​the furnace wall is standardized (including removing floating dust, controlling the temperature of the area to be consistent with the historical calibration conditions, with a temperature difference not exceeding ±5℃). A standard coating agent containing a specified type of impurity is manually applied to the standardized area, and the coating concentration increases in a geometric or arithmetic gradient from the preset starting concentration. At each concentration gradient, one batch of electrolytic copper is smelted according to normal production processes, and the key performance indicators of the produced copper ingots are tested (for electrolytic copper, the key performance indicator is conductivity). When this key performance indicator first falls below the preset lower limit of the factory quality standard, the coating concentration at that concentration gradient is recorded, and divided by the effective area of ​​the furnace wall to obtain the upper limit of the equivalent impurity content (unit: g / cm²) for that impurity category in the metal product. For different metal products (such as aluminum and zinc), the key performance indicators and judgment criteria are determined according to the national or industry standards for that product (for example, for aluminum alloys, the key performance indicator may be, but is not limited to, elongation or tensile strength). The calibration results are stored in a preset mapping table as paired data (impurity category, upper limit of equivalent impurity content).

[0033] Each set of calibration data in the mapping table includes information on the furnace wall temperature, furnace metal type, and refractory material batch at the time of calibration, to ensure that the calibration value closest to the conditions is selected during subsequent queries.

[0034] S1.2: For any two adjacent furnace batches Establish a line from furnace to furnace Pointing to furnace A directed edge. This directed edge carries a residual inherited weight. The expression is: ; Among them, for adjacent furnace pairs, This refers to the thickness of the furnace wall remaining from the previous furnace cycle. This represents the sensitivity coefficient between impurity categories. The sensitivity coefficient between impurity categories is read from a pre-defined sensitivity coefficient matrix. This matrix is ​​an m×m square matrix, where row numbers correspond to the impurity category of the pollution source, and column numbers correspond to the impurity category of the contaminated metal. Each element in the matrix... The numerical range is from 0 to 1, and it is experimentally calibrated in the following way: Under the same furnace wall conditions (including furnace wall refractory material grade, furnace temperature, and furnace atmosphere), known categories The residue was coated at a standard concentration onto a pre-defined measurement area on the furnace wall, and then the smelting category was... The metal. Detection of the produced metal from... The concentration of a specific impurity element introduced by the residue is calculated, and the ratio of this concentration to the standard coating concentration is defined as the sensitivity coefficient S between impurity categories. , To ensure the representativeness of the calibration, for each pair of impurity combinations ( , The calibration experiment needs to be repeated at different furnace wall hot temperatures, for example, in the low-temperature, medium-temperature, and high-temperature zones within the normal smelting temperature range. The weighted average of the coefficients measured at each temperature is taken as the final stored value. If it is not possible to calibrate all combinations, a conservative estimate between 0 and 1 can be assigned to the uncalibrated combinations based on the similarity of the chemical properties of the elements from a pre-set periodic table association database of impurity elements. This estimate can then be corrected using monitoring data in subsequent actual production. The matrix is ​​an m×m dimensional square matrix, with row numbers corresponding to the impurity category of the pollution source. The column number corresponds to the impurity category of the contaminated metal. .

[0035] Finally, all the nodes and their attributes, as well as all directed edges and their weights, are combined to output a collinear smelting residue inheritance knowledge graph. This graph records the residue propagation paths and intensities between furnace runs in a structured form. It should be noted that this embodiment is the first to realize a computable graph model of the smelting residue inheritance effect. Compared to the traditional production scheduling method that relies on manual experience to judge the contamination risk between furnace runs, this technical solution can accurately and quantitatively characterize the residue transfer intensity of each pair of adjacent furnace runs, overcoming the technical difficulties of invisible and difficult-to-quantify furnace wall conditions.

[0036] Furthermore, based on the knowledge graph of residue inheritance from collinear melting, this embodiment performs cross-furnace residue propagation analysis to quantitatively identify contamination risk areas where high-impurity furnace residues propagate to low-impurity furnaces. The specific operations are as follows: S2.1: From the constructed knowledge graph, first extract all furnaces whose furnace wall impurity category data belongs to the preset high impurity category subset H, as the starting node; for each starting high impurity furnace, trace downstream along the directed edge direction to extract all subsequent furnaces (not limited to direct adjacency) reachable by one or more directed edges, until a low impurity category furnace is encountered; arrange the extracted furnaces in sequence according to the actual production switching order to form a sequence, which is defined as the residual inheritance chain. Let the furnace sequence in the residual inheritance chain be... ,in, This is the first high-impurity furnace batch. This refers to the last furnace before the chain terminates. `n` is a positive integer representing the total number of furnaces included in the residual inheritance chain. Adjacent furnace pairs in the chain... Each edge corresponds one-to-one with a directed edge in the knowledge graph, and each edge carries a residual inheritance weight. .

[0037] This chain fully preserves the path information of high-impurity furnace residues being transferred downstream, avoiding redundant calculations that would require traversing all furnaces in the entire graph.

[0038] S2.2: For each pair of adjacent furnaces in the residual inheritance chain Calculate the residual release intensity coefficient: ; in, For furnace The average initial decay rate of the metal category in historical similar furnaces.

[0039] Among them, the historical average initial decay rate The calculation method is as follows: Select the batch number from the historical production database. For all furnaces of the same metal type, the initial decay rate of these furnaces is statistically analyzed, and the arithmetic mean is calculated. To ensure the reliability of the statistics, the number of furnaces of the same type must be at least a preset number, such as 5. When the historical data is less than the preset number, [the following is omitted]. At this point, the ratio is 1, indicating that the attenuation characteristics of the current furnace are consistent with the historical average level, and no additional amplification or reduction effect is produced.

[0040] Residual release intensity coefficient This means that, when considering the number of furnaces... After the actual attenuation characteristics deviated from the historical average level, from the furnace batch Transfer to furnace +1 represents the relative intensity of residual pollution. If this coefficient is greater than 1, it indicates that the actual pollution release intensity is higher than the historical average; if it is less than 1, it indicates that it is lower than the historical average.

[0041] It should be noted that in this invention It is only used for relative comparison with a preset pollution threshold; its absolute value is not involved in any direct calculations of physical quantities. To maintain consistency with the threshold in terms of dimensions, weights are applied during implementation. Divide by a baseline thickness for normalization to transform the residual inherited weights into dimensionless ratios.

[0042] S2.3: Traverse all adjacent furnace pairs in the residual inheritance chain and mark the furnace pairs with residual release intensity coefficients greater than the preset pollution threshold as pollution propagation areas.

[0043] The calibration of the preset pollution threshold includes: selecting all furnace pairs corresponding to quality non-conformities caused by residual pollution on the furnace wall that have occurred and been confirmed by quality inspection from the historical production database. The selection criteria are: the non-conformity item of the furnace is that the impurity element exceeds the standard, and the high impurity category of the previous furnace is consistent with the excessive impurity element. The residual release intensity coefficient of these furnace pairs is calculated, and the minimum value of all residual release intensity coefficient values ​​is taken as the pollution threshold under that metal category. If there are less than 3 sets of historical quality accident data under the same metal category, the pollution threshold is taken as 1 + (the upper limit of the national standard for total impurities - the average total impurities in historical normal production) / the average total impurities in historical normal production. If there is no historical normal production data, the threshold can be taken as 1.5.

[0044] For each marked area of ​​contamination spread Further determine the furnace batch Does the impurity category data belong to the preset low impurity category subset L? If so, then the furnace... For affected low-impurity orders, record their furnace number. If the furnace number... If the residue belongs to the high impurity category, the search continues downwards along the directed edges in the original production sequence or knowledge graph until the first low impurity category furnace is encountered (this furnace may not be in the pre-extraction sequence of the current residue inheritance chain and needs to be added). This low impurity furnace is then marked as an affected order. This search rule ensures that even if high impurity residues pass through multiple high impurity furnaces consecutively, the risk of contamination to low impurity orders can still be traced.

[0045] S2.4: Organize the above analysis results into cross-furnace contamination inheritance results, which include at least the start and end furnace pairs of all marked contamination propagation areas; and the furnace number of the affected low-impurity order corresponding to each contamination propagation area.

[0046] Optionally, the residual release intensity coefficient values ​​for each furnace batch can be included for reference in subsequent production scheduling adjustments.

[0047] It should be noted that this embodiment constructs a residual inheritance chain to limit the contamination propagation path of high-impurity furnaces to a local sequence, avoiding a full-map traversal search and significantly reducing computational complexity.

[0048] Furthermore, after obtaining the cross-furnace contamination inheritance results, this embodiment performs flexible scheduling correction processing based on residue inheritance constraints. By delaying the switchover time of low-impurity orders and inserting transition furnaces between contaminated furnace pairs, the propagation of high-impurity residues to low-impurity orders is blocked or weakened. The specific operation is as follows: S3.1: Read each contamination propagation region and its corresponding affected low-impurity order furnace number from the cross-furnace contamination inheritance results. Assume the contamination propagation region consists of high-impurity furnaces. and its adjacent furnace Composition, including furnace number For high-impurity furnace batches, furnace batches For affected low-impurity orders or intermediate furnaces pointing to low-impurity orders.

[0049] First, calculate the theoretical time delay. ,in, This refers to the original start time of the low-impurity order in the original production schedule. , For a safety waiting factor greater than 1, for example, it can be divided into 3 levels based on the residual release intensity coefficient of the pollution transmission area, with different levels corresponding to different values. The values ​​can be selected, and the gears can be divided according to actual needs. Values ​​of 1.5, 2, and 2.5 are acceptable. The above value selection rule is based on the exponential decay model of furnace wall residue concentration over time. ; go through After a certain period of time, the theoretical remaining proportion of the residual concentration is: . When the values ​​are 1.5, 2, and 2.5, the corresponding theoretical residual ratios are approximately 22.3%, 13.5%, and 8.2%, respectively. These correspond to acceptable safety margin levels for different pollution risks, allowing for recalibration in the system initialization interface based on actual production data, such as historical pass rate statistics. The above theoretical decay model is based on the inherent mathematical properties of an exponential function, rather than experimentally determined values ​​for a specific metal.

[0050] Considering the completion of the attenuation curve measurement and the physical time required for the new furnace insertion, the final actual switching start time was determined to be the maximum of the three factors: ; in, This is the actual completion time of the decay curve measurement for this high-impurity furnace. This is the moment when all the inserted slow-release furnace cycles and isolation cleaning furnace cycles have ended.

[0051] Among them, the safe waiting time Internally, the production scheduling system coordinates the preparation of metal raw materials required for slow-release furnaces and the preparation work for furnace switchover; waiting for high-impurity furnaces... The attenuation curve has been measured and is now complete. Then, immediately execute the inserted slow-release furnace and isolation cleaning furnace in sequence. Among them, the actual switchover start time of the low-impurity order is the latest of the measured completion time of the decay curve and the end time of all slow-release and isolation furnaces.

[0052] S3.2: For two adjacent furnace runs in each contamination propagation zone, in the furnace run... After that, furnace batch Previously, a slow-release furnace was inserted, denoted as furnace s. The purpose of this slow-release furnace was to use medium-purity metal raw materials to flush the furnace wall, gradually diluting the concentration of residues.

[0053] Specifically, the furnace number is read from the preset impurity category-allowed impurity concentration upper limit mapping table. Furnace wall impurity category data The corresponding upper limit of permissible impurity concentration, in parts per million (ppm), is taken as half of this upper limit as the impurity tolerance threshold for the slow-release furnace. This setup ensures that the slow-release furnace can tolerate a certain amount of residue, while preventing high-concentration contaminants from being directly transferred to low-impurity orders.

[0054] The impurity categories of the metals used in the slow-release furnace are automatically matched by the production scheduling system based on the currently available metal raw materials. The specific matching rule is: all impurities with a tolerance threshold of not less than a certain value are selected from the current stockpile. Furthermore, the metal raw material concentration must not exceed the upper limit of the allowable impurity concentration corresponding to the previous high-impurity furnace batch. If multiple candidate metals exist simultaneously, the one with the closest impurity tolerance threshold shall be selected first. The metal, and select the main element from it. Metals belonging to the same group or period in the periodic table, and whose binary phase diagram shows a minimum eutectic temperature of not less than 800°C, are selected as feed for the slow-release furnace. If multiple candidate metals exist, the metal with the highest minimum eutectic temperature is prioritized. If no metal meets the criteria, a insufficient feed alarm is issued, requiring manual selection.

[0055] The melting time of the slow-release furnace is not strictly limited, but at least the furnace wall must undergo one complete melting cycle to fully dilute the residue.

[0056] S3.3: After the slow-release furnace s, furnace number Let's insert another isolation cleaning furnace session, denoted as furnace session c. The purpose of this isolation cleaning furnace session is to thoroughly remove residues from the furnace walls using a high-temperature, long-duration melting process.

[0057] Among them, the melting temperature of the isolation cleaning furnace. Set as: ; ; in, The preset temperature difference aims to maximize the cleaning temperature without damaging the furnace wall. This represents the upper limit of the temperature that the furnace wall refractory material can withstand. The constant 50 (unit: °C) represents the upper limit of the safe temperature rise allowed during a single cleaning operation, using a typical magnesia-chromium refractory material as an example in this embodiment. Exceeding this value may cause thermal shock cracking of the furnace wall. Adjustments are made based on thermal shock stability data provided by the material supplier; this invention does not impose a uniform limitation on this.

[0058] This temperature setting is higher than the normal melting temperature, designed to accelerate the volatilization, decomposition, or peeling off of residues. The melting duration of the isolated cleaning furnace is also specified. Set as: ; ; in, For preset multiples, The reference decay time constant is 600 seconds. This value is the order of magnitude of the decay time constant of residues of most non-ferrous metals (with copper as a typical example) on the walls of commonly used magnesia-chromium furnaces in industry, and is a well-known engineering experience reference value in the field.

[0059] This duration ensures that the furnace walls undergo sufficient high-temperature treatment to reduce the residue concentration below a safe threshold. The metal raw materials used in the isolation cleaning furnace can be low-cost industrial pure metals or recycled materials, and the final product is typically treated as waste or a secondary product, not entering the high-end product supply chain.

[0060] It should be noted that the two-stage purification mechanism of the present invention, which inserts a slow-release furnace and an isolation cleaning furnace, first uses a medium-tolerance furnace to adsorb most of the residue, and then uses a high-temperature long-time furnace for thorough cleaning. Compared with using an isolation furnace alone, this can greatly reduce cleaning costs.

[0061] S3.4: Integrate the above-mentioned delayed treatment, slow-release furnace insertion, and isolation cleaning furnace insertion operations into the original production schedule to generate a new production sequence. In this sequence, the original contamination spread area... Replaced with And furnace number The start time will be postponed accordingly. Furthermore, for furnace runs that are delayed multiple times due to multiple contamination zones, the maximum delay will be taken.

[0062] The final output is a flexible scheduling result based on residual inheritance constraints, which includes an updated furnace sequence list; the planned start and end times for each furnace; the identifiers and process parameters of newly added slow-release furnaces and isolation cleaning furnaces; and the adjusted switchover start times for each low-impurity order.

[0063] It should be noted that the embodiments of the present invention dynamically correlate the delay time with the actual residual decay characteristics of the furnace by using a safe waiting time based on the decay time constant, thereby avoiding the blindness of fixed-duration delays, i.e., excessively long delays waste production capacity, while excessively short delays cause pollution residues.

[0064] Furthermore, after obtaining the flexible production scheduling results based on residual inheritance constraints, this embodiment further performs supply chain linkage correction processing, coordinating the production scheduling plan with the upstream ore supply plan, and responding to the dynamic insertion demand of new high-impurity orders, thereby achieving closed-loop adaptive adjustment of the production scheduling scheme. The specific operations are as follows:

[0065] S4.1: Adjust the arrival time of the ore supply plan based on the time position of the slow-release furnace and the isolation cleaning furnace in the flexible production scheduling results.

[0066] Specifically, for each slow-release furnace cycle, the required ore arrival time is synchronously adjusted to a preset buffer period, such as 8 hours, before the planned start time of that slow-release furnace cycle. This buffer period is used to reserve time windows for ore arrival inspection, storage and transfer, and furnace preparation. Its specific value can be configured according to the supply chain responsiveness and in-plant logistics efficiency.

[0067] For isolated cleaning furnaces, since lower-cost industrial pure metals or recycled materials are typically used, and the quality requirements for the ore are relatively low, no special adjustment to the material arrival time is required; materials can be directly drawn from the plant's stockpile. If the plant's stockpile is insufficient, the procurement and arrival of materials should be arranged in advance according to the same rules as for slow-release furnaces.

[0068] These adjustments ensure that the ore required for slow-release furnaces and isolation cleaning furnaces arrives on time, preventing production schedules from being disrupted due to material shortages. This linkage mechanism integrates the production scheduling system with the supply chain management system, achieving coordinated optimization of production planning and material supply.

[0069] S4.2: Real-time monitoring of new high-impurity orders arising during the production process. New orders refer to temporary production demands for high-impurity metals inserted after the original production schedule has been formulated and partially executed. For each new order, the metal category is read. and the upper limit of metal impurity content These two parameters are obtained directly from the order information system.

[0070] Since the new orders have not yet been put into actual production, the measured parameters of their furnace wall adhesion layer are unavailable. Therefore, a prediction method based on historical data is adopted: (1) According to metal category The reference table is obtained from a pre-defined metal category-typical impurity category lookup table. This lookup table is pre-defined by those skilled in the art based on the types of residual impurities commonly found in collinear smelting of each metal category; for example, copper metals correspond to copper impurity categories, and aluminum metals correspond to alumina inclusion categories.

[0071] (2) The predicted values ​​of furnace wall retention thickness and decay time constant need to be selected from the historical production database. All furnaces with the same metal category as the new order and whose furnace wall impurity category data belong to the high impurity category subset H should be selected. The median of the corresponding parameters of these furnaces should be taken as the predicted value. If the number of historical furnaces with the same high impurity category is insufficient, the operator can set the value according to actual needs. This invention does not impose a unique limitation. After sufficient historical furnace data (20 sets) is accumulated in subsequent actual production, the system will automatically switch to using the actual statistical median as the default value to achieve adaptive parameter updates without the need for continuous manual maintenance.

[0072] (3) Using the same method as the prediction of furnace wall retention thickness, select all furnaces with the same metal category as the new order from the historical production database, and take the median of the initial decay rate of these furnaces as the predicted value; if the number of similar furnaces in the past is insufficient, take the default value of the initial decay rate calibrated at the system initialization as 0.001 mm / s.

[0073] It is important to clarify that for the same new virtual furnace order, all predicted furnace wall physical parameters (including but not limited to predicted furnace wall retention thickness, predicted decay time constant, and predicted initial decay rate) are calculated based on the same set of historical data. Therefore, these parameters should remain consistent in all subsequent calculations involving this virtual furnace order (whether as a preceding or succeeding node). For example, when calculating the residual inheritance weight of the edge (j→new) pointing from the previous furnace j to the new order, the predicted furnace wall impurity category and furnace wall retention thickness of the new order should be used; similarly, when calculating the residual inheritance weight of the edge (new→j+1) pointing from the new order to the next furnace j+1, the same set of predicted furnace wall retention thickness and furnace wall impurity category values ​​should be used to ensure the consistency and rationality of the calculation.

[0074] S4.3: When a new high-impurity order arrives, first determine the arrival time of the order: (1) If a new order arrives before the flexible scheduling result based on residual inheritance constraints is issued to the production execution system, i.e. before actual smelting begins, all slow-release furnaces and isolation cleaning furnaces inserted in the current residual inheritance chain due to the previous contamination spread area are directly removed, and only the original production furnaces are retained, resulting in a basic chain without any virtual transition furnaces. (2) If the previously inserted partial release or isolation furnace has been actually produced when the new order arrives, the executed furnace is retained as an immutable node, and only the remaining isolation furnace that has not yet been executed is cleared to obtain a partial fixed base chain.

[0075] S4.4: Treat new orders as virtual furnaces and insert them into the base chain or a portion of the fixed base chain obtained in S4.3. Specifically, the insertion rules are as follows, which aim to comprehensively assess the impact of the insertion position on the overall chain's contamination propagation pattern, and not just limited to local adjacent edges: Specifically, following the order from the head to the tail of the chain, each adjacent furnace pair (j, j+1) in the current chain is traversed, where j and j+1 are the original production furnaces or fixed, immutable nodes; for each furnace pair (j, j+1), the simulation virtually inserts a new order between them, forming a new local sequence.<j,new,j+1> Calculate the residual release intensity coefficients of the two newly formed directed edges j→new and new→j+1. When a new order is used as a subsequent node, its furnace wall retention thickness, decay time constant, and other parameters are based on the predicted values ​​in S4.2. To determine whether the insertion position is a candidate safe position, the following two conditions must be met simultaneously: (a) The upper limit of the metal impurity content of new orders is lower than the upper limit of the equivalent impurity content of the metal category produced in furnace j, to ensure that the pollution source will not contaminate the new orders in reverse; (b) The residual release intensity coefficients of the two new edges calculated do not exceed the preset pollution threshold, ensuring that the insertion of new orders will not generate new pollution transmission risks in the local area; If multiple candidate safe locations exist, the location that minimizes the sum of the residual release intensity coefficients of the two newly formed edges is selected as the final insertion location. This selection strategy aims to minimize the additional contamination risk brought about by insertion and ensure the overall safety margin of the entire chain. If no candidate safe location is found after traversing all furnace pairs, the new order is appended to the end of the chain, i.e., as the last furnace produced, to avoid contaminating subsequent low-impurity orders.

[0076] Based on the new basic chain after inserting new orders, the residual release intensity coefficient of all adjacent furnace pairs is recalculated according to step S2. Furnace pairs exceeding the threshold are marked as new contamination propagation areas. Slow-release furnaces and isolation cleaning furnaces are then re-inserted according to step S3. Finally, a flexible production scheduling result for the industrial chain is generated, including: the updated complete furnace sequence, including the original furnaces, virtual furnaces for new orders, newly inserted slow-release furnaces, and isolation cleaning furnaces; the planned start and end times for each furnace, with the ore supply time adjusted accordingly; all marked contamination propagation areas and their corresponding treatment measures; and structured data that can be directly accessed by downstream systems, such as supply chain management and production execution systems.

[0077] The initial decay rate of the new order virtual furnace is determined using the same method as the predicted decay time constant in S4.2, which is to take the median of historical furnaces of the same metal category that belong to the high impurity category subset H; if historical data is insufficient, the default value calibrated during system initialization is used.

[0078] It should be noted that this embodiment achieves full-chain coordination from production planning to material supply by linking the production scheduling results with the ore supply time, avoiding production scheduling failures caused by material shortages. Furthermore, for new high-impurity orders, a furnace wall parameter prediction method based on historical medians is adopted, which can reasonably estimate the contamination risk without the need for actual trial production, thus solving the technical problem of data loss when dynamic orders are inserted.

[0079] The knowledge graph-based collaborative flexible scheduling device for the non-ferrous metal industry chain provided in this embodiment of the invention can execute the knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0080] Figure 3 This is a schematic diagram of the structure of an electronic device for implementing the knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain, as described in this embodiment of the invention. The electronic device is intended to represent various forms of digital computers, such as laptops, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0081] like Figure 3As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0082] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0083] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 executes the various methods and processes described above, such as a knowledge graph-based collaborative flexible scheduling method for the non-ferrous metals industry chain.

[0084] In some embodiments, the knowledge graph-based collaborative flexible scheduling method for the non-ferrous metals industry chain can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the knowledge graph-based collaborative flexible scheduling method for the non-ferrous metals industry chain described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to execute the knowledge graph-based collaborative flexible scheduling method for the non-ferrous metals industry chain by any other suitable means (e.g., by means of firmware).

[0085] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0086] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0087] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0088] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0089] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0090] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0091] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0092] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A knowledge graph-based collaborative flexible production scheduling method for the non-ferrous metals industry chain, characterized in that, include: Each heat of the collinear melting system is used as a graph node, and edges are built between adjacent heats according to the impurity sensitivity coefficient. The knowledge graph of collinear melting residue inheritance is then output. Based on the knowledge graph, the high-impurity furnace batches and adjacent furnace batches are extracted to form a residual inheritance chain. The residual release intensity coefficient is calculated, the pollution propagation area and the affected low-impurity orders are marked, and the cross-furnace pollution inheritance results are output. Based on the contamination inheritance result, the low-impurity orders are delayed by a safe waiting time based on the decay time constant, and slow-release furnaces and isolation cleaning furnaces are inserted between contaminated furnace pairs, outputting a flexible scheduling result based on residual inheritance constraints. Based on the flexible production scheduling results linked to the ore supply time, the furnace wall parameters of new orders are predicted and the residual inheritance chain is inserted. After clearing the original isolated furnaces, the contaminated area is recalculated and the slow-release and isolated furnaces are re-inserted, and the industrial chain collaborative flexible production scheduling results are output. The map nodes include: measuring the thickness, impurity type, and temperature of the furnace wall adhesion layer after each furnace switching, which are respectively used as furnace wall retention thickness data, furnace wall impurity type data, and furnace wall hot temperature data; recording the thickness decay curve of each furnace over time, calculating the initial decay rate at the start of the decay curve and the time required for the thickness to decrease from the initial value to a specified proportion of the initial value as the decay time constant; and using the historical furnace switching sequence, furnace wall impurity type data, initial decay rate, decay time constant, and furnace wall hot temperature data as map node attributes. The establishment of an edge between adjacent furnaces based on the impurity sensitivity coefficient includes: for adjacent furnaces, establishing a directed edge with a residual inheritance weight; the residual inheritance weight is equal to the furnace wall retention thickness data of the previous furnace multiplied by the impurity category sensitivity coefficient between the next furnace and the previous furnace. The sensitivity coefficients between impurity categories are read from a preset sensitivity coefficient matrix between impurity categories; The calculation of the residual release intensity coefficient includes: calculating the residual release intensity coefficient of each pair of adjacent furnaces in the residual inheritance chain; the residual release intensity coefficient is equal to the residual inheritance weight of the pair of adjacent furnaces multiplied by the ratio of the initial decay rate of the previous furnace to the average initial decay rate of the metal category to which the previous furnace belongs in the same historical furnaces; if historical data is insufficient, the ratio is one. The method of delaying the low-impurity order based on the decay time constant for a safe waiting period includes: obtaining the contamination propagation area and the corresponding low-impurity order number, and delaying the switching start time of the low-impurity order by a safe waiting period; the safe waiting period is equal to a preset multiple multiplied by the decay time constant of the previous high-impurity furnace cycle; during the delay operation, the slow-release furnace cycle and the isolation cleaning furnace cycle are executed sequentially after the actual measurement of the decay curve of the high-impurity furnace cycle is completed; the actual switching start time of the low-impurity order is the latest of the actual completion time of the decay curve measurement and the end time of all slow-release and isolation furnace cycles; The insertion of slow-release furnaces and isolation cleaning furnaces between contaminated furnace pairs includes: inserting a slow-release furnace between two adjacent furnaces in the contamination propagation area; the impurity tolerance threshold of the slow-release furnace is equal to half of the upper limit of the allowable impurity concentration corresponding to the previous high-impurity furnace; and inserting an isolation cleaning furnace after the slow-release furnace, wherein the melting temperature of the isolation cleaning furnace is equal to the furnace wall hot temperature data of the previous high-impurity furnace plus a preset temperature difference, and the melting duration is equal to a preset multiple multiplied by the decay time constant of the previous high-impurity furnace.

2. The knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain according to claim 1, characterized in that, The construction of the residual inheritance chain includes: The furnaces that belong to the preset high impurity category subset are extracted from the knowledge graph as starting nodes. The furnaces are then traced downstream along directed edges to extract subsequent furnaces that can be reached through one or more directed edges until a low impurity category furnace is encountered. The extracted furnaces are then arranged in production order to form a residual inheritance chain.

3. The knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain according to claim 2, characterized in that, The marked contamination spread areas and affected low-impurity orders include: Furnaces with residual release intensity coefficients greater than a preset pollution threshold are marked as pollution propagation areas, and furnace wall impurity category data of subsequent furnaces are recorded. Furnaces belonging to a preset low impurity category subset are recorded as affected low impurity orders.

4. The knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain according to claim 1, characterized in that, The process of predicting furnace wall parameters for new orders and inserting residual inheritance chains includes: Real-time monitoring of new high-impurity orders to obtain the metal category and upper limit of metal impurity content for new orders; Based on the metal category, the predicted value of the furnace wall impurity category is obtained from the preset metal category and typical impurity category comparison table; The predicted values ​​of furnace wall retention thickness and decay time constant are taken from the median of historical furnaces of the same metal category. If the number of historical furnaces is less than the preset number, the default value is taken. The new order is inserted as a virtual furnace into the current residual inheritance chain. Each furnace in the chain is traversed. If the upper limit of the metal impurity content of the new order is lower than the upper limit of the equivalent impurity content of the metal type produced by the current furnace, and there is no untreated contamination propagation area between the current furnace and the next furnace, then the new order is inserted immediately after the current furnace. Otherwise, it is appended to the end of the chain.

5. The knowledge graph-based collaborative flexible scheduling method for the non-ferrous metal industry chain according to claim 4, characterized in that, The process of recalculating the contaminated area and re-inserting the slow-release and isolation furnaces after removing the original isolation furnaces includes: Remove all slow-release furnaces and isolation cleaning furnaces inserted into the chain due to previous contamination spread areas, and retain only the original production furnaces and new order virtual furnaces to obtain the basic chain; Recalculate the residual release intensity coefficient of all adjacent furnace pairs in the base chain. For new orders as edges of the preceding furnace, the initial decay rate of new orders is taken as the average initial decay rate of the same type of furnace in history for the same metal category. If historical data is insufficient, the default value is taken. For new orders as edges of subsequent furnace cycles, the residual inheritance weight is calculated using the predicted values ​​of furnace wall impurity category and furnace wall retention thickness of the new orders. Furnaces with recalculated residual release intensity coefficients greater than the preset pollution threshold are marked as new pollution propagation areas, and slow-release furnaces and isolation cleaning furnaces are re-inserted.

Images