Method and device for screening of ore blending schemes

Abstract

The application provides a screening method and device for ore blending schemes, relates to the technical field of steel smelting, and is applied to electronic equipment. The electronic equipment is configured with a screening tool. The architecture of the screening tool comprises a main process and N initial sub-processes. The method comprises the following steps: the main process acquires M candidate ore blending schemes and ore blending constraints; the main process preheats the environment of the N initial sub-processes according to the ore blending constraints, and obtains N target sub-processes; the main process distributes the M candidate ore blending schemes to the corresponding target sub-processes, so that the N target sub-processes process the corresponding candidate ore blending schemes in parallel; after the N target sub-processes are processed in parallel, the main process receives M processing results corresponding to the M candidate ore blending schemes from the N target sub-processes; and the main process screens a target ore blending scheme from the M candidate ore blending schemes based on the M processing results and a preset screening standard.

Description

Technical Field

[0001] This invention relates to the field of iron and steel smelting technology, and in particular to a method and apparatus for screening ore blending schemes. Background Technology

[0002] In the process of optimizing batching in ironmaking, complex material ratios, process parameter constraints, and multi-stage processes are typically involved. Due to the complexity of the input data combinations and the strictness of the algorithm constraints, direct optimization calculations often face the risk of being "unsolvable," wasting computational resources and affecting system response efficiency. In actual ironmaking production, raw material yards typically stock more than 30 types of iron ore powder. Considering the combination of different silos and the testing of substitute powders, the theoretically feasible ratio combinations often increase exponentially. Even after initial screening by manual experience, the number of candidate solutions for subsequent refined calculations usually still reaches hundreds or even thousands. Traditional single-threaded calculation of a single solution takes about 0.3-0.5 seconds, and calculating thousands of solutions takes several minutes, which is slow and cannot meet the 'instant response' requirements of batching adjustments on the production site. Summary of the Invention

[0003] This invention provides a method and apparatus for screening ore blending schemes, which solves the technical problem in related technologies where the processing speed is slow due to single-threaded calculation of ore blending schemes, thus failing to meet the needs of rapid processing.

[0004] In a first aspect, embodiments of the present invention provide a method for screening ore blending schemes, applied to an electronic device, wherein the electronic device is configured with a screening tool, and the architecture of the screening tool includes a main process and N initial sub-processes; The method includes: The main process acquires M candidate ore blending schemes and ore blending constraints. Each of the M candidate ore blending schemes includes X types of materials and a range of materials that correspond one-to-one with the X types of materials. The ore blending constraints include process constraints. The main process preheats the environment of the N initial subprocesses according to the ore allocation constraints to obtain N target subprocesses; The main process distributes the M candidate ore blending schemes to the corresponding target subprocesses, so that the N target subprocesses process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes. After the N target subprocesses have processed in parallel, the main process receives M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. The main process selects the target ore blending scheme from the M candidate ore blending schemes based on the M processing results and preset screening criteria, and uses the processing result corresponding to the target ore blending scheme as the target ore blending ratio.

[0005] Optionally, the main process preheats the environment of the N initial subprocesses according to the ore allocation constraints to obtain N target subprocesses, including: for each of the N initial subprocesses, the main process loads the task processing function, the ore allocation algorithm model and the ore allocation constraints into the independent memory space of the initial subprocess through an initialization function to obtain the target subprocess.

[0006] Optionally, the ore blending constraints also include the basic component matrix of the X materials and the linear constraint boundary of the X materials. The ore blending algorithm model is a model constructed based on sintering, pelletizing and blast furnace processes. The ore blending algorithm model takes minimizing the cost of molten iron as the objective function and the ore blending algorithm model is constrained by the ore blending constraints.

[0007] Optionally, the task processing function is used to call the ore blending algorithm model and the ore blending constraints to solve the candidate ore blending scheme after the target subprocess receives the candidate ore blending scheme, and to monitor the calculation time during the calculation of the ore blending algorithm model. If the calculation time exceeds a preset circuit breaker threshold, the calculation process of the ore blending algorithm model is terminated and a timeout result is returned. If the calculation time does not exceed the circuit breaker threshold, the solution result of the ore blending algorithm model is returned. The processing result is either the timeout result or the solution result.

[0008] Optionally, the main process distributes the M candidate mineral allocation schemes to the corresponding target subprocesses, including: the main process obtaining a preset adjustment factor; the main process calculating the target capacity of the task block based on the number of candidate mineral allocation schemes M, the number of target subprocesses N, and the adjustment factor; the main process dividing the M candidate mineral allocation schemes into several task blocks according to the target capacity of the task blocks; and the main process dynamically distributing the several task blocks to the target subprocesses that are currently idle through an asynchronous scheduling mechanism.

[0009] Optionally, before the main process acquires M candidate ore blending schemes and ore blending constraints, the method further includes: acquiring material information corresponding to multiple materials, the material information including the initial proportion range of the materials; for each of the multiple materials, discretizing the initial proportion range corresponding to the material according to a preset step size to obtain a set of discrete proportion ranges; combining the discrete proportion ranges in the set of discrete proportion ranges of the multiple materials to generate Y initial ore blending schemes; performing pre-screening verification on the Y initial ore blending schemes in sequence, the pre-screening verification including at least parameter boundary logic verification, component conservation verification, and process mutual exclusion and equipment capability verification; if the initial ore blending scheme passes the pre-screening verification, the initial ore blending scheme is used as a candidate ore blending scheme and input into the main process; if the initial ore blending scheme fails the pre-screening verification, the initial ore blending scheme is rejected.

[0010] Secondly, embodiments of the present invention provide a screening device for a ore blending scheme. The device is configured with a screening tool, the architecture of which includes a main process and N initial sub-processes. The device includes: The acquisition module is used to acquire M candidate ore blending schemes and ore blending constraints through the main process. Each of the M candidate ore blending schemes includes X kinds of materials and a material range that corresponds one-to-one with the X kinds of materials. The ore blending constraints include process constraints. The preheating module is used to preheat the environment of the N initial sub-processes according to the ore mixing constraints through the main process, so as to obtain N target sub-processes; The distribution module is used to distribute the M candidate ore blending schemes to the corresponding target subprocesses through the main process, so that the N target subprocesses can process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes. After the N target subprocesses process in parallel, the main process receives the M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. The filtering module is used to select a target ore blending scheme from the M candidate ore blending schemes through the main process based on the M processing results and preset filtering criteria, and to use the processing result corresponding to the target ore blending scheme as the target ore blending ratio.

[0011] Thirdly, embodiments of the present invention provide an electronic device, including: processor; A memory for storing instructions to be executed by the processor, wherein the processor is configured to execute the instructions to implement the method as described in the first aspect.

[0012] Fourthly, embodiments of the present invention provide a storage medium that, when instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to perform the method described in the first aspect.

[0013] Fifthly, embodiments of the present invention provide a computer program product, the program product comprising a computer program, the computer program being executed by a processor as described in the first aspect.

[0014] According to an embodiment of the present invention, a method for screening ore blending schemes is provided, applied to an electronic device. The electronic device is configured with a screening tool, the architecture of which includes a main process and N initial sub-processes. The method includes: the main process acquiring M candidate ore blending schemes and ore blending constraints, each of the M candidate ore blending schemes including X materials and material ratios corresponding to the X materials, the ore blending constraints including process constraints; the main process preheating the N initial sub-processes according to the ore blending constraints to obtain N target sub-processes; the main process selecting the M candidate ore blending schemes... The plan is distributed to the corresponding target subprocesses so that the N target subprocesses process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes. After the N target subprocesses process in parallel, the main process receives the M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. Based on the M processing results and preset screening criteria, the main process selects the target ore blending scheme from the M candidate ore blending schemes and uses the processing results corresponding to the target ore blending scheme as the target ore blending ratio. By constructing N target subprocesses based on the hardware resources of electronic devices by the main process, and pre-initializing the resident solution environment of each target subprocess according to the ore blending constraints, it is ensured that each target subprocess has a computing framework consistent with the process. Thus, when processing M candidate ore blending schemes in parallel, it can not only make full use of the multi-process concurrency capability to greatly improve the computing speed, but also fundamentally guarantee the accuracy and reliability of the evaluation of each candidate ore blending scheme. It achieves high throughput and fast response while accurately selecting the optimal target ore blending scheme, effectively solving the technical problem of difficulty in balancing processing efficiency and result accuracy in the existing technology. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 The flowchart of a screening method for ore blending schemes provided in one embodiment of the present invention is shown; Figure 2 The structure of a screening device for a ore blending scheme provided in one embodiment of the present invention is shown; Figure 3 A structural block diagram of an electronic device provided in one embodiment of this application is shown. Detailed Implementation

[0017] As described in the background section, the optimization calculation process for ironmaking batching typically involves complex material ratios, process parameter constraints, and multi-stage processes. Due to the complexity of the input data combinations and the strictness of the algorithm constraints, direct optimization calculations often face the risk of being "unsolvable," wasting computational resources and affecting system response efficiency. In actual ironmaking production, raw material yards typically stock more than 30 types of iron ore powder. Considering the combination of different silos and the testing of substitute ore powders, the theoretically feasible ratio combinations often increase exponentially. Even after initial screening by manual experience, the number of candidate solutions for subsequent refined calculations usually still reaches hundreds or even thousands. Traditional single-threaded calculation of a single solution takes about 0.3-0.5 seconds, and calculating thousands of solutions takes several minutes, resulting in slow processing speed that cannot meet the 'instant response' requirements of batching adjustments on the production site.

[0018] According to an embodiment of the present invention, a method for screening ore blending schemes is provided. The main process acquires M candidate ore blending schemes and ore blending constraints. Each of the M candidate ore blending schemes includes X types of materials and a material range corresponding to each of the X types of materials. The ore blending constraints include process constraints. The main process preheats the environment of N initial subprocesses according to the ore blending constraints to obtain N target subprocesses. The main process distributes the M candidate ore blending schemes to the corresponding target subprocesses, so that the N target subprocesses process the corresponding candidate ore blending schemes in parallel, obtaining processing results corresponding to the candidate ore blending schemes. The processing results include solutions corresponding to the candidate ore blending schemes. After the N target subprocesses process in parallel, the main process receives M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. Based on the M processing results and a preset screening criterion, the main process selects a target ore blending scheme from the M candidate ore blending schemes and uses the processing results corresponding to the target ore blending scheme as the target ore blending ratio.

[0019] The technical solution of the present invention and how the technical solution of the present invention solves the above-mentioned technical problems will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0020] Figure 1 The flowchart illustrates a screening method for a mineral blending scheme provided by an embodiment of the present invention. It should be understood that, as... Figure 1 The screening method for the ore blending scheme shown can be executed by electronic equipment. This electronic equipment can be a physical server, a server cluster consisting of multiple servers, or a cloud server capable of cloud computing. The electronic equipment is configured with a screening tool, the architecture of which includes a main process and N initial child processes. The main process can perform global management and resource scheduling of the electronic equipment, thus enabling the implementation of... Figure 1 Before proceeding with the method shown, the main process can determine the number of initial parallel child processes N based on the hardware resources of the electronic device, such as the number of CPU cores and memory capacity. N is a positive integer greater than 1. Then, a process pool containing N initial child processes can be created, for example, using Python's `multiprocessing.Pool` to create the process pool. Basic configurations can then be performed on the N initial child processes, allowing the main process to manage these N initial child processes. Figure 1 As shown, the screening method for ore blending schemes provided in this embodiment of the invention includes steps 110 to 150.

[0021] Step 110: The main process obtains M candidate ore blending schemes and ore blending constraints. Each of the M candidate ore blending schemes includes X types of materials and a range of materials corresponding to the X types of materials. The ore blending constraints include process constraints.

[0022] In this embodiment of the invention, each candidate ore blending scheme includes X types of materials and a material range corresponding to each of the X types of materials, where X is a positive integer greater than 1. The material range includes the upper and lower proportional boundaries of the materials in the candidate ore blending scheme, that is, the solution corresponding to each material in the candidate ore blending scheme conforms to the material range in the candidate ore blending scheme. Since the production of molten iron involves multiple processes, such as sintering, pelletizing, or blast furnace processes, the types and ranges of materials in each candidate ore blending scheme can be different from other candidate ore blending schemes. Ore blending constraints describe multiple constraints in the process of generating molten iron, where process processing constraints include process physical constants such as element recovery rate, baseline yield, and processing cost that do not change with the material ratio in the processing of each process. For example, process processing constraints may include the solvent recovery rate rrF or the sintering metal recovery rate rrM in the sintering process.

[0023] Step 120: The main process preheats the environment of the N initial subprocesses according to the ore allocation constraints to obtain N target subprocesses.

[0024] In this embodiment of the invention, N is a positive integer greater than 1. Since the ore blending constraint is a universal and fixed parameter in the process of producing molten iron, the main process can preheat the environment of the N initial subprocesses according to the ore blending constraint. For example, the ore blending constraint can be loaded into the independent memory space corresponding to the N subprocesses to obtain N target subprocesses. This makes it convenient for the N target subprocesses to obtain the ore blending constraint from the main process again during subsequent calculations.

[0025] Step 130: The main process divides and distributes the M candidate ore blending schemes to the corresponding target subprocesses, so that the N target subprocesses process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes.

[0026] In this embodiment of the invention, the main process can employ an asynchronous preemptive distribution mechanism. The main process can obtain the working status of N target child processes, i.e., whether the target child processes are in an idle state. Then, it preferentially distributes candidate mineral allocation schemes to the target child processes in the idle state through an asynchronous task queue, so that the N target child processes are always in a saturated working state. In the process of distributing candidate mineral allocation schemes to the target child processes in the idle state, this embodiment of the invention does not limit the number of candidate mineral allocation schemes included in a single distribution operation.

[0027] In this embodiment of the invention, the target subprocess can call the ore blending algorithm model described below to solve the candidate ore blending scheme, and obtain the ore blending ratio that satisfies the material range included in the candidate ore blending scheme and meets the ore blending constraints as the solution corresponding to the candidate ore blending scheme.

[0028] Step 140: After the N target subprocesses have processed in parallel, the main process receives the M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses.

[0029] In this embodiment of the invention, during the parallel processing of N target subprocesses, to improve processing efficiency, the main process establishes a non-blocking response mechanism to achieve streaming reception of results. Specifically, the main process does not need to wait for all task blocks to finish processing; instead, it monitors the communication channel in real time. Whenever any target subprocess completes its currently assigned candidate ore allocation scheme and obtains a set of processing results, the target subprocess immediately sends the set of results asynchronously to the main process. Through streaming monitoring, the main process gradually receives and temporarily stores M processing results from the N target subprocesses while the subprocesses continue to compute, thereby eliminating the waiting time during the aggregation phase.

[0030] Step 150: The main process selects a target ore blending scheme from the M candidate ore blending schemes based on the M processing results and preset screening criteria, and uses the processing result corresponding to the target ore blending scheme as the target ore blending ratio.

[0031] In this embodiment of the invention, after the main process receives M processing results, it can score the M processing results according to preset screening criteria, and sort the M candidate ore blending schemes according to the scores of the M processing results, selecting the target ore blending scheme from the M candidate ore blending schemes. The preset screening criteria can include one or more criteria, such as the cost of molten iron corresponding to the candidate ore blending scheme, or the corresponding energy consumption. If the preset screening criteria include multiple criteria, the evaluation weights corresponding to multiple criteria can be set according to production needs, achieving flexible adjustment of multiple objectives. The target ore blending scheme can include one or more candidate ore blending schemes, for example, the target ore blending scheme can be the 5 candidate ore blending schemes with the lowest cost among the M candidate ore blending schemes. The target ore blending ratio can be used for subsequent ore blending production, and the target ore blending ratio includes the material ratios corresponding to X physical substances in the candidate ore blending schemes.

[0032] By constructing N target subprocesses based on the hardware resources of electronic devices by the main process, and pre-initializing the resident solution environment of each target subprocess according to the ore blending constraints, it is ensured that each target subprocess has a computing framework consistent with the process. Thus, when processing M candidate ore blending schemes in parallel, it can not only make full use of the multi-process concurrency capability to greatly improve the computing speed, but also fundamentally guarantee the accuracy and reliability of the evaluation of each candidate ore blending scheme. It achieves high throughput and fast response while accurately selecting the optimal target ore blending scheme, effectively solving the technical problem of difficulty in balancing processing efficiency and result accuracy in the existing technology.

[0033] In this embodiment of the invention, to enable the target subprocess to have a faster processing speed and reduce communication overhead during the initialization process, the construction of the target subprocess can be carried out using solver environment warm-up technology. Specifically, step 120, in which the main process performs environment warm-up on the N initial subprocesses according to the ore allocation constraints to obtain N target subprocesses, may include the following steps: For each of the N initial subprocesses, the main process loads the task processing function, the ore allocation algorithm model, and the ore allocation constraints into the independent memory space of the initial subprocess through an initialization function to obtain the target subprocess.

[0034] In this embodiment of the invention, for each of the N initial subprocesses, when each initial subprocess starts, the main process loads the general and fixed ore blending constraints, task processing functions, and ore blending algorithm model into the independent memory space corresponding to the initial subprocess all at once through an initialization function (e.g., init_worker), constructs an instance of the ore blending algorithm model, and obtains the target subprocess. This means that for subsequent calculations of different candidate ore blending schemes, the target subprocess does not need to repeatedly construct the ore blending algorithm model architecture; it only needs to update the decision variables in the candidate ore blending schemes, i.e., the material range of each material.

[0035] In this embodiment of the invention, to address the time-consuming initialization of the ore blending algorithm model (such as loading a large raw material database and constructing the ore blending constraint matrix), a "process preheating and resident" mechanism is adopted. When creating the process pool, an initializer mechanism is used to preload fixed ore blending constraints, such as process parameters (sintering / pelletizing / blast furnace process parameters, quality indicators) and the ore blending algorithm model, into the independent private memory corresponding to the initial subprocess at the beginning of each initial subprocess startup. This avoids the repeated model construction steps for each candidate ore blending scheme, and only requires inputting the changing decision variables (ratio combinations), significantly reducing the I / O overhead of a single calculation.

[0036] In this embodiment of the invention, the ore blending constraint also includes the basic component matrix of the X materials and the linear constraint boundary of the X materials. The ore blending algorithm model is a multi-process linkage optimal solution model constructed based on sintering, pelletizing and blast furnace processes. The ore blending algorithm model takes minimizing the cost of molten iron as the objective function and the ore blending algorithm model is constrained by the ore blending constraint.

[0037] In this embodiment of the invention, X is a positive integer greater than 1. The basic composition matrix of X materials contains a static database of the chemical composition (e.g., TFe, SiO2, CaO, MgO, Al2O3, S, P, etc.), physical properties (moisture, loss on ignition), price, and material classification labels (hematite / magnetite / flux, etc.) of all stock materials. The linear constraint boundary of X materials includes the "red line" indicators set by the operator for the proportion of X materials, including the upper and lower limits of target components (e.g., basicity [1.8, 2.0], SiO2 [0, 6]), the total limit of the proportion of various ores (e.g., the lower limit of hematite proportion (Lhes), the lower limit of magnetite proportion (Lmas), the lower limit of limonite proportion (Lgos)), and the extreme value of single material proportion [0, 10], etc.

[0038] In this embodiment of the invention, the ore blending algorithm model is an optimization solution model covering sintering, pelletizing, and blast furnace processes, considering the quality and energy consumption linkages between each process. The mathematical expression in the ore blending algorithm model can be described using mixed-integer linear programming (MILP) or quadratic programming, with variables including continuous variables (ore quantity, proportion, etc.) and necessary discrete process selection variables. The objective function in the ore blending algorithm model can be to minimize the cost of molten iron, minimize energy consumption, or maximize output. For example, if minimizing the cost of molten iron is chosen as the objective function in the ore blending algorithm model, then the objective function expression can be:

[0039] In the formula, The unit is the cost of molten iron, in yuan; M is the number of material types, in units. Cost information for each material, in yuan / ton; This represents the quantity of each material in the candidate ore blending scheme, in tons. Constraints on the objective function include, but are not limited to, ore blending constraints such as raw material supply limitations, product quality indicators, and process parameter ranges.

[0040] In this embodiment of the invention, the task processing function is used to call the ore blending algorithm model and the ore blending constraints to solve the candidate ore blending scheme after the target subprocess receives the candidate ore blending scheme. Iteratively, it obtains the material ratio corresponding to minimizing the cost of molten iron within the material range of the candidate ore blending scheme. During the calculation process of the ore blending algorithm model, the calculation time is monitored. If the calculation time exceeds a preset circuit breaker threshold, the calculation process of the ore blending algorithm model is terminated, and a timeout result is returned. If the calculation time does not exceed the circuit breaker threshold, the solution result of the ore blending algorithm model is returned. The processing result is either the timeout result or the solution result.

[0041] In this invention and its embodiments, a solver time monitoring function is added inside the encapsulated task processing function `process_combination`. Since ore blending algorithm models typically belong to mixed-integer linear programming (MILP) or nonlinear programming (NLP) problems, exhibiting NP-hard characteristics, specific parameter combinations may lead to an excessively large branch-and-bound tree search space, preventing rapid convergence. Therefore, a circuit breaker mechanism is a necessary safeguard for industrial applications. If any candidate ore blending scheme fails to converge within a specified time (e.g., 200ms), and the computation time exceeds a preset circuit breaker threshold, the circuit breaker mechanism is triggered, suspending the computation process of the ore blending algorithm model and returning a "computation timeout" status as the timeout result, ensuring that a single abnormal scheme does not block the entire parallel queue. If the computation time does not exceed the circuit breaker threshold, the solution result of the ore blending algorithm model is returned. The solution result is the material ratio within the material range corresponding to the candidate ore blending scheme that minimizes the cost of molten iron and the ore blending constraints. The processing result is either the timeout result or the solution result.

[0042] In this embodiment of the invention, a timeout circuit breaker and memory protection mechanism are introduced to address the risks of "computational non-convergence" or "infinite loop" that may occur in the ore-matching algorithm. A timeout circuit breaker mechanism is embedded in parallel computing. A maximum solution time threshold is set for each target subprocess. If a candidate ore-matching scheme causes the solver to fail to converge for an extended period, the target subprocess automatically terminates its current computation and marks it as "computation timed out," releasing resources to process the next candidate ore-matching scheme and preventing the entire target subprocess from blocking.

[0043] In this embodiment of the invention, considering that the computation of a single candidate mineral blending scheme involves iterations and a large amount of data, extensive inter-process communication (IPC) can become a bottleneck. The main process adopts an adaptive chunking strategy to dynamically divide tasks into chunks, calculating the optimal chunk size based on the total number of candidate mineral blending schemes and the number of subprocesses N. Candidate mineral blending schemes are packaged into multiple task blocks and dynamically distributed to idle target subprocesses through an asynchronous scheduling mechanism. Specifically, the main process sends task blocks to a process pool containing N target subprocesses, where each target subprocess dynamically obtains a task block to be processed based on its own load status. If a target subprocess finishes processing first, the main process distributes subsequent task blocks to that subprocess in real time. This mechanism ensures that the "long tail effect" caused by differences in computation time of individual schemes is eliminated during the processing of M candidate mineral blending schemes.

[0044] In this embodiment of the invention, to implement the task block mechanism, the process of the main process distributing the M candidate mineral allocation schemes to the corresponding target subprocesses in step 130 may include the following steps: the main process obtains a preset adjustment factor, which is used to balance the communication frequency and the load balancing of the subprocesses (for example, a value of 4); the main process calculates the target capacity of the task block based on the number of candidate mineral allocation schemes M, the number of target subprocesses N, and the adjustment factor, whereby the target capacity of the task block represents the number of candidate mineral allocation schemes included in the task block; the main process divides the M candidate mineral allocation schemes into several task blocks based on the target capacity of the task block; and the main process dynamically distributes the several task blocks to the target subprocesses that are currently in an idle state through an asynchronous scheduling mechanism.

[0045] In this embodiment of the invention, the adjustment factor is determined based on the communication overhead between the main process and the target child process. The value of the adjustment factor reflects the core trade-off between load balancing and communication overhead in parallel computing. By reasonably setting this factor, a balance can be found between reducing the number of communication operations and maintaining load balancing. To avoid excessive communication overhead caused by frequent interactions between the main process and the child process, the capacity of the task block can be calculated using the following formula, and based on the capacity of each task block, the M candidate mining schemes are packaged into several task blocks: chunk_size = M / (N * Factor); Where `chunk_size` is the capacity of the task block, `M` is the number of candidate mineral blending schemes, `N` is the number of target subprocesses, and `Factor` is an adjustment factor. After determining the capacity and number of task blocks, the main process packages the M candidate mineral blending schemes into multiple task blocks and uses the `pool.imap_unordered` interface to put these task blocks into an asynchronous distribution queue. The N target subprocesses dynamically extract task blocks from the asynchronous distribution queue based on their idle status. After each target subprocess completes the calculation of all candidate mineral blending schemes contained in its current task block locally, it returns the processing result set corresponding to that task block to the main process all at once, thereby reducing the communication frequency between the master and slave processes.

[0046] In this embodiment of the invention, the calculation of the ore blending scheme typically involves complex material ratios, process parameter constraints, and multi-stage processes. Due to the complexity of the input data combinations and the strictness of the algorithm constraints, direct optimization calculations often face the risk of being "unsolvable," wasting computational resources and affecting system response efficiency. Therefore, before inputting candidate ore blending schemes into the target subprocess, the material ratios and configuration parameter data included in the candidate schemes can be pre-screened to determine whether they meet simple constraint requirements, thus eliminating invalid or unreasonable data combinations in advance. This improves computational efficiency and ensures the stability and practicality of the algorithm's solution. The pre-screening criteria include incorrect upper and lower limit settings for various parameters (lower limit higher than upper limit, upper limit 0), abnormal ratios and composition parameters (e.g., total not equal to 100%), and unreasonable ratios and settings of various process materials in sintering pellets or blast furnaces. If any of these errors are detected, the current candidate scheme is eliminated.

[0047] Specifically, before the main process obtains M candidate ore blending schemes and ore blending constraints in step 110, the method further includes: obtaining material information corresponding to multiple materials, the material information including the initial proportion range of the materials; for each of the multiple materials, discretizing the initial proportion range corresponding to the material according to a preset step size to obtain a set of discrete proportion ranges; combining the discrete proportion ranges in the set of discrete proportions of the multiple materials to generate Y initial ore blending schemes; performing pre-screening verification on the Y initial ore blending schemes in sequence, the pre-screening verification including at least parameter boundary logic verification, component conservation verification, and process mutual exclusion and equipment capability verification; if the initial ore blending scheme passes the pre-screening verification, the initial ore blending scheme is used as a candidate ore blending scheme and input into the main process; if the initial ore blending scheme fails the pre-screening verification, the initial ore blending scheme is rejected.

[0048] In this embodiment of the invention, the electronic device acquires material information corresponding to various materials. This material information includes quality inspection composition information, cost data, and process parameters, quality indicators, and proportioning ranges for sintering, pelletizing, and blast furnace processes. The final information representation for each material can be: {"Material Number":S001,"Material Name":Mineral Powder 1,"Material Category":Hematite,"Price":795,Moisture":8.19,TFe":66.35,SiO2":6.34,CaO":0.340,"MgO":0.44,Al2O3":0.42,"S":0.2,"P":0.005,"Loss on Ignition":-1.8,"Initial Proportioning Range":[0%,10%]}. The quality inspection composition information and cost data for each material can be directly retrieved from the pre-iron MES system. The process parameters, quality indicators, and upper and lower limits of the proportioning for sintering, pelletizing, and blast furnace processes are provided by the batching operator based on actual production conditions.

[0049] In this embodiment of the invention, after obtaining the material information corresponding to multiple materials, for each of the multiple materials, the proportion range corresponding to the material is discretized according to a preset step size to obtain a discrete proportion set. The preset step size can be 1%. Discretizing the proportion range corresponding to the material according to the preset step size yields a discrete proportion range set. For example, the hematite ore mentioned above has an initial proportion range of [0%, 10%]. After discretization with a preset step size of 1%, discrete proportion range sets including [0%, 1%], (1%, 2%), etc. are obtained. The discrete proportion range set includes the discrete proportion ranges within the proportion range of the material, that is, [0%, 1%] is a discrete proportion range in the discrete proportion range set. Finally, the discrete proportion ranges corresponding to multiple materials are combined by enumeration to obtain Y initial ore blending schemes.

[0050] In this embodiment of the invention, Y initial ore blending schemes are pre-screened and verified sequentially to eliminate invalid or unreasonable data combinations in advance. The pre-screening and verification includes at least parameter boundary logic verification, component conservation verification, and process mutual exclusion and equipment capability verification. If an initial ore blending scheme passes the pre-screening and verification, it is selected as a candidate ore blending scheme and input into the main process; if an initial ore blending scheme fails the pre-screening and verification, it is eliminated.

[0051] Among them, parameter boundary logic verification is used to check whether the physical meaning of each process parameter conflicts, including: upper and lower limit inversion verification, verifying all parameters with set ranges to see if there is a logical error of "set lower limit value > set upper limit value" (e.g., min_ratio > max_ratio). Zero value constraint verification: for key component indicators, verifying whether there is a contradictory situation of "upper limit constraint is 0" and "lower limit constraint > 0". Component conservation verification is used for verification based on the physical mass conservation law, including: normalization verification: verifying whether the sum of the percentages of each material component is within a reasonable error range (e.g., between [80%, 120%]) to prevent non-physical data from being included in the calculation due to data entry errors. Proportioning completeness verification: verifying whether the sum of the lower limit proportions of a certain type of material (e.g., hematite, magnetite, limonite) in each process of sintering, pelletizing, or blast furnace exceeds 100% (i.e., Lhes + Lmas + Lgos > 100), or whether the sum of the upper limit proportions is less than 100%. If the sum of the lower limits exceeds the total amount, or the sum of the upper limits is insufficient for the total amount, then the initial ore blending scheme is physically unsolvable. Process mutual exclusion and equipment capacity verification are used to perform hard constraint verification in conjunction with on-site equipment capacity, including: silo limit verification: verifying whether the maximum feed ratio of a single silo × the number of ore powder silos meets the total ratio requirement, preventing exceeding the belt scale's feeding capacity. Specific indicator zeroing verification: for specific recovery rate indicators (such as sintering solvent recovery rate rrF, sintering metal recovery rate rrM) or production indicators, if an abnormal zero value appears, it is considered invalid data. When the initial ore blending scheme triggers any of the above verification rules, it is directly determined that the scheme is unsolvable, without needing to be included in the ore blending algorithm model in the target subprocess for calculation, and then the validity of the next set of initial ore blending scheme data is determined, until M candidate ore blending schemes are determined from Y initial ore blending schemes, and then the following is executed: Figure 1 The method shown.

[0052] In this embodiment of the invention, the material ratios and configuration parameters of the input scheme are intelligently predicted before calculation, filtering out invalid or unreasonable data combinations in advance, thereby improving computational efficiency and ensuring the stability and practicality of the algorithm solution. Multi-process parallel computing is employed, significantly improving the calculation speed of the ore blending scheme and meeting the real-time requirements of large-scale data processing. The ore blending algorithm model has good scalability and adaptability, and can be adjusted according to the actual situation of different enterprises.

[0053] Figure 2 The structure of a screening device for a ore blending scheme provided by an embodiment of the present invention is shown. For example... Figure 2 As shown, the screening device 200 for a ore blending scheme provided in this embodiment of the invention is configured with a screening tool. The architecture of the screening tool includes a main process and N initial subprocesses. The screening device 200 for the ore blending scheme includes: The acquisition module 210 is used to acquire M candidate ore blending schemes and ore blending constraints through the main process. Each of the M candidate ore blending schemes includes X kinds of materials and a material range that corresponds one-to-one with the X kinds of materials. The ore blending constraints include process constraints. The preheating module 220 is used to preheat the environment of the N initial sub-processes according to the ore mixing constraints through the main process, so as to obtain N target sub-processes; The distribution module 230 is used to distribute the M candidate ore blending schemes to the corresponding target subprocesses through the main process, so that the N target subprocesses can process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes. After the N target subprocesses process in parallel, the main process receives the M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. The screening module 240 is used to select a target ore blending scheme from the M candidate ore blending schemes through the main process based on the M processing results and preset screening criteria, and to use the processing result corresponding to the target ore blending scheme as the target ore blending ratio.

[0054] It should be noted that the embodiments of the screening device for ore blending schemes in this specification and the embodiments of the screening method for ore blending schemes in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can be referred to the corresponding embodiments of the screening method for ore blending schemes mentioned above, and the repeated parts will not be described again.

[0055] Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of this application. Figure 3 As shown, the electronic device provided in this application embodiment includes a processor 310 and a memory 320, wherein the memory is used to store instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the above method.

[0056] In an exemplary embodiment, the electronic device may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0057] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory including instructions that can be executed by a processor of a device to perform the described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc. This non-transitory computer-readable storage medium, when the instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to perform... Figure 1 The method shown.

[0058] This application also provides a computer program product, including a computer program, which, when executed by a processor, performs... Figure 1 The method shown.

[0059] The above description does not provide detailed technical specifications regarding the structure of each layer. However, those skilled in the art should understand that layers and regions of desired shapes can be formed using various technical means. Furthermore, to form the same structure, those skilled in the art can also design methods that are not entirely identical to those described above. Additionally, although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be advantageously combined.

[0060] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other modifications and adjustments to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all modifications and adjustments falling within the scope of the invention.

[0061] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims and their equivalents, this invention is also intended to include these modifications and variations.

Claims

1. A method for screening ore blending schemes, characterized in that, Applied to electronic devices, the electronic devices are configured with a filtering tool, the architecture of which includes a main process and N initial child processes; The method includes: The main process acquires M candidate ore blending schemes and ore blending constraints. Each of the M candidate ore blending schemes includes X types of materials and a range of materials that correspond one-to-one with the X types of materials. The ore blending constraints include process constraints. The main process preheats the environment of the N initial subprocesses according to the ore allocation constraints to obtain N target subprocesses; The main process distributes the M candidate ore blending schemes to the corresponding target subprocesses, so that the N target subprocesses process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes. After the N target subprocesses have processed in parallel, the main process receives M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. The main process selects the target ore blending scheme from the M candidate ore blending schemes based on the M processing results and preset screening criteria, and uses the processing result corresponding to the target ore blending scheme as the target ore blending ratio.

2. The method according to claim 1, characterized in that, The main process preheats the environment of the N initial subprocesses according to the ore allocation constraints to obtain N target subprocesses, including: For each of the N initial subprocesses, the main process loads the task processing function, the ore allocation algorithm model, and the ore allocation constraints into the independent memory space of the initial subprocess through the initialization function, thereby obtaining the target subprocess.

3. The method according to claim 2, characterized in that, The ore blending constraints also include the basic component matrix of the X materials and the linear constraint boundary of the X materials. The ore blending algorithm model is a model constructed based on sintering, pelletizing and blast furnace processes. The ore blending algorithm model takes minimizing the cost of molten iron as the objective function and is constrained by the ore blending constraints.

4. The method according to claim 3, characterized in that, The task processing function is used to call the ore blending algorithm model and the ore blending constraints to solve the candidate ore blending scheme after the target subprocess receives the candidate ore blending scheme. During the calculation of the ore blending algorithm model, the calculation time is monitored. If the calculation time exceeds a preset circuit breaker threshold, the calculation process of the ore blending algorithm model is terminated and a timeout result is returned. If the calculation time does not exceed the circuit breaker threshold, the solution result of the ore blending algorithm model is returned. The processing result is either the timeout result or the solution result.

5. The method according to claim 1, characterized in that, The main process distributes the M candidate mineral allocation schemes to the corresponding target subprocesses, including: The main process acquires a preset adjustment factor; The main process calculates the target capacity of the task block based on the number M of the candidate ore blending schemes, the number N of the target subprocesses, and the adjustment factor. The main process divides the M candidate ore allocation schemes into several task blocks according to the target capacity of the task block; The main process uses an asynchronous scheduling mechanism to dynamically distribute the several task blocks to target child processes that are currently idle.

6. The method according to claim 1, characterized in that, Before the main process acquires M candidate ore allocation schemes and ore allocation constraints, the method further includes: Obtain material information for multiple materials, including the initial mixing ratio range of the materials; For each of the various materials, the initial proportion range corresponding to the material is discretized according to a preset step size to obtain a set of discrete proportion ranges; The discrete proportion ranges in the set of discrete proportion ranges of the various materials are combined to generate Y initial ore blending schemes; The Y initial ore blending schemes are pre-screened and verified in sequence. The pre-screening and verification includes at least parameter boundary logic verification, component conservation verification, and process mutual exclusion and equipment capability verification. If the initial ore blending scheme passes the pre-screening verification, the initial ore blending scheme is used as a candidate ore blending scheme and input into the main process; If the initial ore blending scheme fails the pre-screening verification, then the initial ore blending scheme is rejected.

7. A screening device for a ore blending scheme, characterized in that, The device is configured with a filtering tool, the architecture of which includes a main process and N initial subprocesses; The device includes: The acquisition module is used to acquire M candidate ore blending schemes and ore blending constraints through the main process. Each of the M candidate ore blending schemes includes X kinds of materials and a material range that corresponds one-to-one with the X kinds of materials. The ore blending constraints include process constraints. The preheating module is used to preheat the environment of the N initial sub-processes according to the ore mixing constraints through the main process, so as to obtain N target sub-processes; The distribution module is used to distribute the M candidate ore blending schemes to the corresponding target subprocesses through the main process, so that the N target subprocesses can process the corresponding candidate ore blending schemes in parallel to obtain the processing results corresponding to the candidate ore blending schemes. The processing results include the solutions corresponding to the candidate ore blending schemes. After the N target subprocesses process in parallel, the main process receives the M processing results corresponding to the M candidate ore blending schemes from the N target subprocesses. The filtering module is used to select a target ore blending scheme from the M candidate ore blending schemes through the main process based on the M processing results and preset filtering criteria, and to use the processing result corresponding to the target ore blending scheme as the target ore blending ratio.

8. An electronic device, characterized in that, include: processor; A memory for storing instructions to be executed by the processor, wherein the processor is configured to execute the instructions to implement the method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is able to perform the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, The program product includes a computer program that is executed by a processor according to any one of claims 1-6.

Images