A method, system, medium, and computer program product for cyclic steady state solution of a comminution and screening process flow

Abstract

The application discloses a kind of broken screening process circulation steady-state solving method, system, medium and computer program product, belong to industrial process modeling and numerical solving technical field, comprising: obtaining the process configuration data of broken screening process;Directional graph model is constructed based on process configuration data;Establish equipment operating condition calculation model for each node;Cyclic structure identification is executed to directed graph;Linear pre-computation is executed to linear region, and the external input boundary of each cyclic subgraph is obtained;According to solving level, each cyclic subgraph is executed iteration solving, and the node output data of this round is obtained;In iteration solving process, iteration stabilization processing is executed;After reaching the minimum iteration number of pre-set, based on the error between the adjacent iteration output data of each node in cyclic subgraph, all node convergence determination is executed, when the error of all participating determination nodes meets pre-set convergence condition, determine that cyclic subgraph reaches steady state;Steady-state calculation result is output.

Description

Technical Field

[0001] This application belongs to the field of industrial process modeling and numerical solution technology, specifically involving a cyclic steady-state solution method for crushing and screening process, a cyclic steady-state solution system for crushing and screening process, a computer-readable medium, and a computer program product. Background Technology

[0002] Crushing and screening processes are widely used in mining engineering, sand and gravel aggregate processing, hydropower station construction, municipal engineering, and infrastructure construction. A typical production line usually includes multiple pieces of equipment such as jaw crushers, cone crushers, impact crushers, vertical shaft impact crushers, vibrating screens, distributors, conveying equipment, and stockpiles. The equipment is connected by material flow, forming multiple stages including crushing, screening, grading, return material, and finished product output.

[0003] In practical engineering, crushing and screening processes are often not unidirectional linear processes, but rather involve one or more closed-loop cycles. For example, material oversize by the screening machine returns to the previous crushing stage, is crushed again, and then re-enters the screening machine; in multi-stage screening scenarios, there may also be situations where primary and secondary closed loops coexist. The closed-loop cycle ensures that the equipment input comes not only from fresh material upstream but also from recycled material downstream, thus forming a directed network with feedback throughout the system.

[0004] Among the existing publicly available technologies, several technical directions adjacent to this invention have been disclosed: First, existing closed-circuit crushing and screening processes disclose a process route in which the oversize material is returned to the crushing equipment for further crushing; Second, existing ore crushing and beneficiation systems disclose equipment configuration and operating data for high-recycle-rate closed-circuit screening; Third, existing chemical process simulation technologies disclose representing the flow chart as a graph structure and identifying strongly connected components to schedule simulation calculations; Fourth, existing methods for full-cycle steady-state simulation of fluidized catalytic cracking systems and simulation systems for crushing and screening equipment have been developed.

[0005] While the aforementioned existing technologies have solved problems related to specific process configuration, chemical process simulation scheduling, fluidized system steady-state simulation, or equipment wear simulation, in engineering practice, steady-state calculation of crushing and screening processes mainly suffers from problems such as reliance on manual identification of the cyclic structure, easy contamination of linear calculations by cyclic feedback, difficulty in stable transmission of multi-size material flow, easy oscillation or divergence in high circulation ratio scenarios, difficulty in direct adaptation of traditional process simulation methods, difficulty in directly applying the cyclic solution of general chemical process simulation software to crushing and screening processes with discrete solid material particle size distribution as the core, and long-term reliance on experience and repeated manual trial calculations in industry practice.

[0006] Therefore, there is a lack of a cyclic steady-state solution method that can stably obtain steady-state results under conditions of high cycle ratio, multiple cycle coupling, and nonlinear device models. Summary of the Invention

[0007] The purpose of this invention is to provide a method for solving the cyclic steady-state problem of crushing and screening processes, which can effectively solve the problem in the above-mentioned background technology that complex closed-circuit crushing and screening processes rely on manual identification of the cyclic path, manual estimation of the cyclic ratio, and repeated trial calculations to obtain material balance results.

[0008] Furthermore, this invention also addresses the problems in closed-circuit crushing and screening processes, such as the linear initial conditions being easily contaminated by cyclic feedback, the tendency for iteration to oscillate or diverge under multi-cycle coupling and high return ratio conditions, the difficulty in controlling the multi-size particle size distribution error, and the difficulty in obtaining stable steady-state material balance results throughout the entire process.

[0009] Therefore, to achieve the above objectives, the technical solution adopted in the first aspect of the present invention is as follows: A method for solving the cyclic steady-state problem of a crushing and screening process includes the following steps: Obtain process configuration data for the crushing and screening process, wherein the process configuration data includes equipment information for the crushing and screening process and material connection relationships between equipment; A directed graph model is constructed based on the process configuration data. The directed graph model includes nodes and directed edges connecting the nodes. The nodes correspond to the equipment in the crushing and screening process. The directed edges correspond to the material transfer relationship between the equipment. The directed edges carry the material flow rate and the multi-size particle size distribution vector of the material. A device condition calculation model is established for each node. The device condition calculation model is used to calculate the output material data of the node based on the material data of the input node and the equipment parameters of the node. Cyclic structure identification is performed on the directed graph to determine at least one cyclic subgraph and the linear region in the directed graph other than the cyclic subgraph, and the solution level of each cyclic subgraph is determined. The cyclic subgraph represents a closed-loop structure with material feedback relationship in the crushing and screening process. Perform linear pre-computation on the linear region to obtain the external input boundaries of each cyclic subgraph; According to the solution hierarchy, iterative solution is performed on each cyclic subgraph. For each cyclic subgraph, based on the external input boundary and the feedback data of the previous cycle, the multi-source material flow and its multi-size particle size distribution vector entering the same node are merged, and the equipment condition calculation model of each node is called to perform the current round of calculation to obtain the node output data of the current round. During the iterative solution process, iterative stabilization is performed. After reaching the preset minimum number of iterations, based on the error between adjacent iteration output data of each node in the cyclic subgraph, a full node convergence determination is performed. When the errors of all participating nodes meet the preset convergence condition, the cyclic subgraph is determined to have reached a steady state. Output the steady-state calculation results.

[0010] Preferably, the multi-size particle size distribution vector includes mass flow data or mass ratio data corresponding to at least three particle size levels, and the particle size levels correspond to the project sieve aperture definition, product specifications, raw material particle size range, or preset standard particle size definition.

[0011] Preferably, the step of performing cycle structure identification on the directed graph, determining at least one cycle subgraph and the linear regions in the directed graph excluding the cycle subgraphs, and determining the solution level of each cycle subgraph, wherein the cycle subgraph represents a closed-loop structure with material feedback relationships in the crushing and screening process, including: Perform strongly connected component decomposition on the directed graph to obtain the strongly connected components of the directed graph; A cyclic subgraph is determined from the strongly connected components based on at least one of the following: the device type corresponding to the node contained in each strongly connected component, the connection port type between nodes, the material attribute identifier carried by the directed edge, or the return edge identifier.

[0012] Preferably, the strongly connected component decomposition employs any one of the Tarjan algorithm, Kosaraju algorithm, or Gabow algorithm.

[0013] Preferably, determining the solution level of each cyclic subgraph includes: Construct a directed acyclic dependency graph between each cyclic subgraph. The output data of one of the cyclic subgraphs after convergence is used as the external input boundary data of the downstream cyclic subgraphs that depend on this cyclic subgraph. Determine the solution hierarchy order of each cyclic subgraph based on the directed acyclic dependency graph.

[0014] Preferably, the step of performing linear pre-computation on the linear region to obtain the external input boundaries of each cyclic subgraph includes: The linear pre-computation stage does not take the feedback value of the previous round for the cyclic feedback edge, but only calculates the output of linear nodes that do not belong to the cyclic subgraph and the external input boundary that enters the cyclic subgraph.

[0015] Preferably, the merging of multi-source material flows entering the same node and their multi-size particle size distribution vectors includes: Obtain the material flow data and corresponding particle size ratio data carried by each input directed edge entering the node; The mass ratio data is weighted and merged according to the material flow rate of each input directed edge to obtain the synthetic input granularity distribution of the node.

[0016] Preferably, the iterative stabilization process includes performing at least one of the following on node output data, circulating flow data, or node state vectors: damped update, state vector mixing, outlier suppression, negative value truncation, or local rollback.

[0017] Preferably, the damping update is to perform a weighted mixing of the current round candidate state vector and the previous round state vector, and the mixed state vector is used to replace the current round node output data. The current round candidate state vector is the node state data directly output by the equipment operating condition calculation model. The weight of the current round candidate state vector is β, and the weight of the previous round state vector is 1-β. The value range of β is 0 < β ≤ 1.

[0018] Preferably, the state vector mixing includes Anderson mixing processing, which constructs accelerated state vectors based on historical multi-round node output data.

[0019] Preferably, the full node convergence determination includes: For each participating decision node in the cyclic subgraph, calculate the absolute error and relative error of the output flow data of the node in the two adjacent rounds; When the absolute error is not greater than the first preset threshold or the relative error is not greater than the second preset threshold, the node is determined to meet the flow convergence condition. When all participating nodes in the cyclic subgraph satisfy the flow convergence condition, the cyclic subgraph is determined to satisfy the flow steady-state condition.

[0020] Preferably, after the cyclic subgraph has initially converged, one or more final correction iterations are performed, and the final steady-state result is submitted when the maximum node output change in the final correction iteration is lower than a third preset threshold.

[0021] Preferably, before or after outputting the steady-state calculation results, a material balance verification is performed. The material balance verification includes verifying the balance relationship between the total fresh input, the total final product output, the internal circulation volume, and the loss bypass volume, and verifying the quality balance deviation of each particle size range to generate a material balance report.

[0022] Preferably, the steady-state calculation results include at least one of the following: input flow data, output flow data, circulating flow data, circulation ratio, equipment load rate data, product particle size distribution data, convergence status data, material balance report, or calculation result identifier that can be written back to CAD drawing elements.

[0023] Secondly, the present invention provides a cyclic steady-state solution system for a crushing and screening process, comprising: The acquisition module is used to acquire process configuration data of the crushing and screening process flow, the process configuration data including equipment information of the crushing and screening process flow and material connection relationships between equipment; A directed graph model construction module is used to construct a directed graph model based on the process configuration data. The directed graph model includes nodes and directed edges connecting the nodes. The nodes correspond to the equipment in the crushing and screening process, and the directed edges correspond to the material transfer relationship between the equipment. The directed edges carry the material flow rate and the multi-size particle size distribution vector of the material. The equipment condition calculation module is used to establish an equipment condition calculation model for each node. The equipment condition calculation model is used to calculate the output material data of the node based on the material data of the input node and the equipment parameters of the node. The cycle structure identification module is used to perform cycle structure identification on the directed graph, identify at least one cycle subgraph and the linear region in the directed graph other than the cycle subgraph, and determine the solution level of each cycle subgraph. The cycle subgraph represents a closed-loop structure with material feedback relationship in the crushing and screening process. The linear pre-computation module is used to perform linear pre-computation on the linear region to obtain the external input boundaries of each cyclic subgraph; The iterative solution module is used to perform iterative solution on each loop subgraph according to the solution hierarchy. For each loop subgraph, based on the external input boundary and the feedback data of the previous loop, the multi-source material flow and its multi-size particle size distribution vector entering the same node are merged, and the equipment condition calculation model of each node is called to perform the current round of calculation to obtain the node output data of the current round. A stabilization module is used to perform iterative stabilization processing during the iterative solution process; The convergence determination module is used to perform full node convergence determination based on the error between adjacent iteration output data of each node in the cyclic subgraph after reaching the preset minimum number of iterations. When the errors of all participating nodes meet the preset convergence conditions, the cyclic subgraph is determined to have reached a steady state. The results output module is used to output the steady-state calculation results.

[0024] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0025] Fourthly, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the method described in the first aspect.

[0026] This invention abstracts the crushing and screening process into a directed graph model carrying material flow rate and multi-size particle size distribution vectors. Nodes represent process objects such as crushing, screening, distribution, conveying, buffering, or stockpiling, while edges represent material flow directions. The system first identifies the cyclic structure in the directed graph, dividing the entire process into linear regions and cyclic subgraphs. Then, it performs linear pre-calculation on the linear regions to obtain the external input boundaries entering the cyclic subgraphs. Next, it iteratively solves each cyclic subgraph according to the dependency hierarchy, employing stabilization techniques such as damped updates, state vector mixing, outlier suppression, negative flow rate or negative particle size mass truncation, and full node convergence determination during the iteration process. Finally, it outputs the steady-state material balance result for the entire process.

[0027] The loop structure identification does not only output strongly connected components in the general graph theory sense, but further combines equipment type, connection port, oversize material attributes, undersize material attributes, or return edge identifiers to determine the loop subgraph corresponding to the actual material flow closed loop. During the loop subgraph solving process, the system merges the external fresh input with the feedback of the previous loop, and performs weighted merging of the multi-particle size distribution vectors of multi-source material flows to obtain the node synthesis input. After reaching the preset minimum number of iterations, the system performs flow error and particle size vector error checks on all nodes participating in the judgment within the loop subgraph, and outputs the steady-state material balance result after meeting the convergence condition.

[0028] Thus, compared with the prior art, the present invention has the following beneficial effects: (1) It can automatically identify the closed-loop structure in the crushing and screening process, reducing the reliance on manual experience to judge the loop path; (2) It obtains the pure external input boundary of the loop subgraph through linear pre-calculation, reducing the pollution of the initial conditions by loop feedback; (3) It supports layered solution of multiple loop subgraphs, which is suitable for multi-closed-loop, multi-stage screening and multi-equipment coupling scenarios; (4) It incorporates multi-size particle size distribution vectors into iterative calculation and convergence judgment, which is suitable for the calculation needs of the sand and gravel aggregate industry with particle size as the core; (5) It improves the solution stability in high loop ratio scenarios through iterative stabilization, anomaly detection and final correction iteration; (6) The output results include flow rate, particle size distribution, loop ratio, equipment load rate and material balance verification, which can directly serve equipment selection, process verification and working condition optimization. Attached Figure Description

[0029] Figure 1 This is a flowchart of the steps of a cyclic steady-state solution method for a crushing and screening process according to the present invention; Figure 2 This is a process flow diagram of the present invention; Figure 3 This is another step flow diagram of the cyclic steady-state solution method for a crushing and screening process according to the present invention; Figure 4This is a schematic diagram of a cyclic steady-state solution system for a crushing and screening process according to the present invention; Figure 5 This is another schematic diagram of a cyclic steady-state solution system for a crushing and screening process according to the present invention. Detailed Implementation

[0030] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following detailed description of specific embodiments based on the present invention is provided in conjunction with the accompanying drawings and preferred embodiments.

[0031] In existing technologies, designers typically rely on experience to determine which return path belongs to a loop, making it difficult to automatically identify multiple loops, nested loops, or cross-equipment coupled loops. In linear calculations, directly using the loop feedback flow in the initial calculation may lead to unstable upstream boundary conditions, affecting subsequent loop iterations. Sand and gravel materials are usually represented by multiple particle size ranges; the crushing and screening processes alter the mass distribution of each particle size, resulting in errors not only in the total flow rate but also in the particle size vector. When the amount of circulating material approaches or exceeds the amount of fresh feed, simple iterative methods are sensitive to initial values, easily leading to numerical oscillations, divergence, or false convergence. Furthermore, general chemical process simulation software primarily targets continuous phases, thermodynamic states, and chemical unit operations; its loop solutions typically rely on tearing flow or global equations, making it difficult to directly apply to crushing and screening processes centered on discrete solid material particle size distributions. Designers often rely on empirical estimations, manual adjustments to the loop ratio, or multiple trial calculations to find "seemingly reasonable" operating conditions, a process that is time-consuming and lacks stability. Especially under conditions of multiple cycles and high return ratio, manual iteration often results in large oscillations or divergence, making it impossible to obtain a convergent solution.

[0032] Based on this, existing technologies have not yet developed a general solution method for solid multi-size material flow, closed-loop over-screen return, high circulation ratio feedback, and CAD process diagram calculation results output in crushing and screening processes. This method is capable of automatically identifying cyclic subgraphs, isolating linear external input boundaries, performing layered cyclic iterations, stabilizing the process, and submitting steady-state results with full node error and material balance reports.

[0033] This application provides a cyclic steady-state solution method that can stably obtain steady-state results under conditions of high cycle ratio, multiple cycle coupling, and nonlinear equipment model for the directed graph structure, multi-size material flow, and closed-loop feedback relationship of crushing and screening processes, aiming to solve the above-mentioned problems.

[0034] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings. Figures 1 to 5 The diagram illustrates a method for solving the cyclic steady-state problem of a crushing and screening process according to the present invention. One embodiment of the present invention provides a method for solving the cyclic steady-state problem of a crushing and screening process, such as... Figure 1 As shown, it includes the following steps: Step S1: Obtain process configuration data for the crushing and screening process flow, wherein the process configuration data includes equipment information for the crushing and screening process flow and material connection relationships between equipment; Step S2: Construct a directed graph model based on the process configuration data. The directed graph model includes nodes and directed edges connecting the nodes. The nodes correspond to the equipment in the crushing and screening process. The directed edges correspond to the material transfer relationship between the equipment. The directed edges carry the material flow rate and the multi-size particle size distribution vector of the material. Step S3: Establish an equipment condition calculation model for each node. The equipment condition calculation model is used to calculate the output material data of the node based on the material data of the input node and the equipment parameters of the node. Step S4: Perform cycle structure identification on the directed graph, determine at least one cycle subgraph and the linear region in the directed graph other than the cycle subgraph, and determine the solution level of each cycle subgraph. The cycle subgraph represents a closed-loop structure with material feedback relationship in the crushing and screening process. Step S5: Perform linear pre-computation on the linear region to obtain the external input boundary of each cyclic subgraph; Step S6: Perform iterative solution for each cyclic subgraph according to the solution hierarchy. For each cyclic subgraph, based on the external input boundary and the feedback data of the previous cycle, merge the multi-source material flow and its multi-size particle size distribution vector entering the same node, call the equipment condition calculation model of each node to perform the current calculation, and obtain the node output data of the current cycle. Step S7: During the iterative solution process, an iterative stabilization process is performed. Step S8: After reaching the preset minimum number of iterations, based on the error between adjacent iteration output data of each node in the cyclic subgraph, perform full node convergence determination. When the errors of all participating nodes meet the preset convergence condition, determine that the cyclic subgraph has reached a steady state. Step S9: Output the steady-state calculation results.

[0035] Specifically, in step S2, the directed graph of the crushing and screening process is G=(V,E); where V represents the set of equipment nodes and E represents the set of directed edges between equipment. Any node v∈V represents a process object such as a crusher, screening machine, feeder, distributor, conveying equipment, or stockpile; any edge e=(u,v)∈E represents the flow of material from node u to node v.

[0036] The multi-level material vector carried by edge e can be represented as: M_e=[m_1,m_2,...,m_n]; where m_i represents the material mass or flow rate of the i-th level, n is the number of levels, and n≥3. The total flow rate of edge e is: Q_e=Σm_i, i=1...n.

[0037] When node v has multiple input edges, its total input flow is: Q_in(v) = Σ_{e∈In(v)}Q_e; its synthesized input granularity vector can be weighted and merged according to the flow: M_in(v,k) = [Σ_{e∈In(v)}Q_e × P_e(k)] / [Σ_{e∈In(v)}Q_e]; where In(v) represents the set of input edges entering node v, and P_e(k) represents the quality proportion of the k-th granularity level in edge e. If Σ_{e∈In(v)}Q_e is zero, the synthesized input granularity vector of this node can be set to a zero vector or a preset empty input vector can be used.

[0038] Specifically, in step S3, the node output can be uniformly represented as: M_out(v)=F_v(M_in(v),P_v) Where F_v represents the equipment operating condition calculation model of node v, and P_v represents the equipment parameters. For crushing equipment, F_v may include the crushed product particle size curve or particle size conversion matrix; for screening equipment, F_v may include the screen aperture size, screening efficiency, and throughput matrix; for material distribution equipment, F_v may include the diversion ratio of each outlet.

[0039] For a crushing equipment node, the material entering the node can be represented as the total flow rate Q_in(v) and the particle size vector M_in(v) = [M_in(v,1),M_in(v,2),...,M_in(v,n)]. The crushing equipment node calculates and outputs the particle size vector of the material based on the equipment operating parameters θ_v, feed rate, discharge port settings, crushed product particle size curve, particle size transformation matrix, or equivalent empirical model.

[0040] M_out(v,k)=Σ_{i=1}^{n}A_v(i,k;θ_v,Q_in(v))×M_in(v,i) Wherein, A_v(i,k;θ_v,Q_in(v)) represents the crushing transfer coefficient of node v of the crushing equipment, which converts the input particle size i into the output particle size k under the operating parameters θ_v and the input flow rate Q_in(v). The crushing transfer coefficient can be derived from the particle size curve of the equipment manufacturer's product, experimental data, empirical models, project configuration parameters, or combinations thereof. This invention does not limit the specific form of the crushing function, but only requires that the equipment operating condition calculation model can convert the total input flow rate and the multi-particle size distribution vector into the total output flow rate and the multi-particle size distribution vector for use in cyclic subgraph iteration, stabilization processing, and convergence determination.

[0041] Specifically, in step S4, for any cyclic subgraph C, its cycle ratio can be expressed as: R_C = Q_recycle(C) / Q_fresh(C) Where Q_recycle(C) is the internal feedback material flow rate of the cycle subgraph, and Q_fresh(C) is the external fresh input flow rate entering the cycle subgraph.

[0042] It is necessary to explain the iterative stabilization of the cyclic subgraph. The state vector of the cyclic subgraph in the t-th iteration can be represented as X_t, and the candidate state obtained by calculating the equipment condition is F(X_t). The stabilization process can be expressed as: X_(t+1)=S(F(X_t),X_t,H_t) Where S represents the stabilization operator, and H_t represents the set of historical iteration states. The stabilization operator S can take different forms depending on the operating conditions; when damped updates are used, H_t may not participate in the calculation. In this case, damped updates are a special case of the above stabilization operator, which can be written as: X_(t+1)=β×F(X_t)+(1-β)×X_t Where β is the stabilization coefficient, and its value ranges from 0 to 1.

[0043] When using a vector blending approach based on historical states, H_t can be represented as the set of historical states from the most recent p rounds: H_t={X_(tp),X_(t-p+1),...,X_t} In one Anderson hybrid implementation, the residual r_i can be set as F(X_i) - X_i. Under the condition that Σγ_i = 1, the set of coefficients γ_i that minimizes ||Σγ_ir_i|| is obtained, and the next round state is generated as follows: X_(t+1)=Σγ_i×F(X_i),i=tp,...,t Therefore, damped updates are used to suppress abrupt changes in the state of a single round of candidate results relative to the previous round, while Anderson mixing is used to improve iterative stability or convergence speed by utilizing historical residual information from multiple rounds. After the above stabilization processing, the system can also truncate, backtrack, or diagnose anomalies for negative flow rates, negative granularity masses, infinite results, or non-numerical results.

[0044] It is necessary to explain the convergence determination in step S8. For a node v in the cyclic subgraph, the output flows in round t and round (t-1) are Q_t(v) and Q_(t-1)(v) respectively. Then the absolute error and relative error can be expressed as: E_abs(v)=|Q_t(v)-Q_(t-1)(v)| E_rel(v)=|Q_t(v)-Q_(t-1)(v)| / max(Q_(t-1)(v),δ) Here, δ represents the minimum flow threshold to prevent division by zero. When all participating nodes in the cyclic subgraph satisfy E_abs(v)≤ε_abs or E_rel(v)≤ε_rel, and the preset minimum number of iterations is reached, the cyclic subgraph is considered to meet the flow convergence condition. Here, "absolute error or relative error" is used to accommodate both small and large flow nodes: for small flow nodes close to zero, absolute error is used to prevent the relative error from being amplified; for larger flow nodes, relative error is used to reflect the engineering proportion error. To avoid small flow nodes prematurely meeting the absolute error condition, the system can further require multiple consecutive rounds to meet the convergence condition, or combine granular vector error, the minimum flow threshold, and the final correction iteration for joint determination.

[0045] In particular, if it is necessary to control the particle size distribution error simultaneously, the particle size vector error can be further calculated: E_psd(v)=max_k|P_t(v,k)-P_(t-1)(v,k)| When E_psd(v) is not greater than the preset granularity error threshold, the node is considered to meet the granularity convergence condition.

[0046] Table 1 below shows the key technical parameters of this invention. Table 1

[0047] The technical solution of this application will be described in detail below with reference to specific embodiments.

[0048] In directed graph construction, after reading the process equipment and connection relationships, the system maps each equipment to a node and the material connection relationships to directed edges. Each directed edge carries the total flow and granularity vector. For CAD integration scenarios, nodes and edges can also carry CAD primitive IDs, equipment IDs, port numbers, and extended data identifiers; for independent computing scenarios, nodes and edges can come from project snapshots, JSON files, or databases.

[0049] During the cycle structure identification process, cycle structure identification is performed on the directed graph, preferably using the strongly connected component decomposition method. If a strongly connected component contains two or more nodes, the system further combines the node device type, connection port, oversize material attributes, undersize material attributes, material category, or return edge identifier to determine whether the strongly connected component corresponds to an actual material flow closed loop. When there are oversize return, oversize material return, closed-loop crushing, or multi-level feedback relationships between nodes, it is treated as a cycle subgraph. For multiple cycle subgraphs, the system further constructs the dependencies between them; if the output of one cycle subgraph enters another cycle subgraph, the former is the upstream cycle subgraph of the latter.

[0050] In one specific determination method, the system first counts the node set V_C and the internal directed edge set E_C for each strongly connected component C, and reads the source node type, target node type, source port, target port, material category, oversize attribute, undersize attribute, and return edge identifier carried by each directed edge. The following determination rule can be adopted: if SCC(C) satisfies the number of nodes |V_C|≥2, and there exists e∈E_C such that e.return identifier∈{oversize return, super-particle return, closed-loop return}, and e.target node type∈{crushing equipment, pre-crushing feeding equipment, pre-crushing buffer equipment}, then C is determined to be a cyclic subgraph.

[0051] When the screening equipment and the crushing equipment are not directly connected, but are connected by a conveying device, a buffer device or a feeding device, the directed path judgment method is adopted: if there is a directed path in the strongly connected component C that leads from the screen outlet of the screening equipment through the conveying device, the buffer device or the feeding device to the inlet of the crushing equipment, and the material category on this path is screen oversize, oversized material or material to be further crushed, then the strongly connected component is determined as a closed-loop crushing cycle subgraph.

[0052] As an exclusion rule, if a strongly connected component C consists only of connections between conveying equipment, buffer equipment, or stockpiles, and there are no oversize return, oversize return, closed-loop return, or crushing loop identifiers, and there is no directed path from the oversize outlet of the screening equipment back to the inlet of the crushing equipment, then it will not be treated as a cyclic subgraph that needs to be solved iteratively, but can be treated as a linear region or a normal connected structure.

[0053] In linear pre-computation, topological sorting and sequential computation are performed on linear nodes that do not belong to the cyclic subgraph. For external edges entering the cyclic subgraph, the system uses the outputs of the already computed linear nodes as fresh inputs to the cyclic subgraph. For feedback edges inside the cyclic subgraph, they are not used as fresh inputs during the linear pre-computation stage to avoid contaminating the external boundary conditions with feedback flow.

[0054] During the iterative solution of the cyclic subgraph, for each cyclic subgraph, the system uses the external input boundary and the feedback from the previous cycle as input, and performs equipment node calculations round by round. The crushing equipment substitutes the total synthetic input flow rate and multi-size particle size distribution vector of the current cycle into the crushed product particle size curve, particle size transformation matrix, or equivalent empirical model to obtain the candidate total output flow rate and mass flow rate of each particle size in the current cycle; the screening equipment allocates the input particle size vector to the oversize, undersize, or intermediate product outlet according to the throughput matrix; the material distribution equipment allocates the input flow rate and particle size vector to multiple downstream paths according to the distribution ratio; the conveying and buffering equipment transfers or buffers the material.

[0055] When handling iterative stabilization and anomaly processing, the system performs iterative stabilization when the cycle ratio is high, node output changes drastically, or cyclic feedback causes fluctuations in candidate outputs. Stabilization can process the total output flow, granular vector, or state vector as a whole. If the mixed result contains negative quality or anomalous values, the system truncates negative values ​​to zero and triggers anomaly processing for non-numerical or infinite results. If the cyclic flow increases rapidly during continuous iterations, a divergence risk can be identified, and the system can reduce the stabilization coefficient, revert to the previous stable state, or diagnose non-convergence.

[0056] Finally, regarding the convergence determination and final correction, it should be noted that after reaching the preset minimum number of iterations, the system performs a full node convergence check on all participating nodes within the cyclic subgraph. Only when all nodes meet the preset error conditions is the cyclic subgraph considered to have reached a steady state. To further eliminate iteration lag errors, several final correction iterations can be performed after initial convergence until the change in the output of the largest node is lower than the correction threshold.

[0057] Example 1: As Figure 2 The process flow diagram shown illustrates the connections between the jaw crusher, primary screen, cone crusher, secondary screen, vertical shaft impact crusher, and finished product stockpile, as well as two closed-loop circulation structures. This embodiment uses a three-stage crushing and two-stage screening sand and gravel aggregate production line as an example. The process structure includes a jaw crusher, primary screen, cone crusher, secondary screen, vertical shaft impact crusher, and finished product stockpile. This process includes two closed-loop circulations: the oversize material from the primary screen is returned to the cone crusher to form SCC1, and the oversize material from the secondary screen is returned to the vertical shaft impact crusher to form SCC2.

[0058] In this embodiment, the solution method for SCC1 is as follows: the fresh input entering the primary screen after primary crushing is 300 t / h. The primary screen and the cone crusher form SCC1. The system first initializes the cyclic feedback of SCC1 to 0, and then iterates according to the feedback relationship between the material on the primary screen and the material returned to the cone crusher. Table 2 below shows the parameters and states related to the number of iterations.

[0059] Table 2

[0060] After SCC1 converges, the steady-state total feed rate is approximately 333.30 t / h, the circulating feed rate is approximately 33.30 t / h, and the circulation ratio is approximately 11.10%. In this embodiment, the undersize material from the primary screen is used as the external input to the secondary screen and SCC2. According to the material balance at the primary screen node, the amount of material entering the secondary screen is 333.30 - 33.30 = 300.00 t / h. If there are bypasses, losses, or metering corrections in actual engineering, the external input to SCC2 can be determined based on the actual material balance result calculated using the linear path.

[0061] The solution method for SCC2 is as follows: SCC2 consists of a secondary screen and a vertical shaft impact crusher. Its external input comes from the undersize material of the primary screen, that is, the material output from SCC1 after convergence and entering the secondary screen, consistent with the aforementioned material balance, taken as 300.00 t / h. The oversize material of the secondary screen returns to the vertical shaft impact crusher to form a closed loop with a high return ratio. To illustrate the effect of the stabilization treatment, this embodiment assumes that the candidate feedback quantity of the oversize material of the secondary screen is approximately 50% of the total feed quantity of the previous round of vertical shaft impact crusher, and damping update is used in this cycle subgraph, with a stabilization coefficient β of 0.95. Table 3 below gives the corresponding parameters and states. In Table 3, "Candidate Total Feed Amount" is F(X_t) directly obtained from the equipment operating condition calculation model, and "Stabilized Total Feed Amount" is the state submitted to the next iteration according to X_(t+1)=β×F(X_t)+(1-β)×X_t; "Relative Change with Previous Round" in Table 3 is the relative error between the stabilized total feed amount of two adjacent rounds, corresponding to E_rel in the convergence judgment.

[0062] Table 3

[0063] Ultimately, the total feed rate of SCC2 is approximately 599.87 t / h, the recycled feed rate is approximately 299.87 t / h, the external input is 300.00 t / h, and the recycling ratio is approximately 99.96%. This example illustrates that, in scenarios with high recycling ratios, this invention, through external input boundary locking, piecewise solution of the cyclic subgraph, and damping stabilization processing, can gradually transform candidate outputs into stable submission states and obtain a self-consistent closed-loop steady-state solution.

[0064] This invention constructs a directed graph model of the crushing and screening process carrying material properties; establishes calculation models of the operating conditions of crushing, screening, and material distribution equipment; identifies the cyclic structure and divides it into linear regions and cyclic subgraphs; performs linear pre-calculation on the linear regions to obtain the external input boundaries of the cyclic subgraphs; performs iterative solution of the cyclic subgraphs according to the solution hierarchy; performs stabilization processing such as damping update, state vector mixing, outlier suppression, negative flow or negative particle size truncation or local backtracking during the iteration process; performs anomaly detection and full node convergence determination; and outputs steady-state material balance results after each cyclic subgraph converges. Figure 3 Another flowchart of the method of the present invention is provided.

[0065] Thus, the beneficial effects corresponding to Example 1 are as follows: (1) The system can automatically identify the primary screen-cone crusher cycle and the secondary screen-vertical shaft crusher cycle; (2) The system can perform segmented solutions for weak and strong cycles respectively, avoiding mutual contamination of multiple cycles; (3) The system can still gradually reduce errors and output stable circulation flow and product flow in high circulation ratio scenarios; (4) The steady-state results can be used for equipment load verification, process scheme comparison, product particle size analysis and subsequent CAD drawing result rewriting.

[0066] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects.

[0067] Secondly, the present invention provides a cyclic steady-state solution system for a crushing and screening process, such as... Figure 4 As shown, the cyclic steady-state solution system 200 for the crushing and screening process includes: The acquisition module 201 is used to acquire process configuration data of the crushing and screening process flow, including equipment information and material connection relationships between equipment in the crushing and screening process flow; the directed graph model construction module 202 is used to construct a directed graph model based on the process configuration data, the directed graph model including nodes and directed edges connecting the nodes, the nodes corresponding to equipment in the crushing and screening process flow, the directed edges corresponding to material transfer relationships between equipment, and the directed edges carrying material flow rate and multi-size particle size distribution vector of the material; the equipment condition calculation module 203 is used to establish an equipment condition calculation model for each node, the equipment condition calculation model being used to calculate the output material data of the node based on the material data of the input node and the equipment parameters of the node; the cycle structure identification module 204 is used to perform cycle structure identification on the directed graph, identify at least one cycle subgraph and the linear regions in the directed graph other than the cycle subgraph, and determine the solution level of each cycle subgraph, the cycle subgraph representing the... The crushing and screening process includes a closed-loop structure with material feedback relationships; a linear pre-calculation module 205, used to perform linear pre-calculation on the linear region to obtain the external input boundaries of each cyclic subgraph; an iterative solution module 206, used to perform iterative solution on each cyclic subgraph according to the solution hierarchy, for each cyclic subgraph, based on the external input boundaries and the feedback data of the previous cycle, merging the multi-source material flow and its multi-size particle size distribution vector entering the same node, calling the equipment condition calculation model of each node to perform the current round of calculation, and obtaining the node output data of the current round; a stabilization processing module 207, used to perform iterative stabilization processing during the iterative solution process; a convergence judgment module 208, used to perform full node convergence judgment based on the error between adjacent iterative output data of each node in the cyclic subgraph after reaching the preset minimum number of iterations, and when the errors of all participating judgment nodes meet the preset convergence conditions, it is determined that the cyclic subgraph has reached a steady state; and a result output module 209, used to output the steady-state calculation results. Figure 5This is another schematic diagram of a system for solving the cyclic steady-state problem in a crushing and screening process. The system includes a directed graph modeling module, an equipment condition calculation module, a cycle structure identification module, a linear pre-calculation module, a cycle subgraph solving module, an iterative stabilization module, a full node convergence determination module, a material balance verification module, and a steady-state result output module. The directed graph modeling module abstracts crushing equipment, screening equipment, distribution equipment, conveying equipment, buffer equipment, or stockpiles in the crushing and screening process into nodes, and abstracts the material transfer relationships between equipment into directed edges carrying flow rate and multi-size particle size distribution vectors. The equipment condition calculation module calculates node outputs based on the equipment parameters of the crushing, screening, and distribution equipment. The cycle structure identification module identifies cycle subgraphs formed by oversize return material, oversized material return material, closed-circuit crushing, or multi-level feedback relationships. The linear pre-calculation module obtains the external input boundaries entering the cycle subgraph. The cycle subgraph solving module, iterative stabilization module, and full node convergence determination module complete the iteration, stabilization, and steady-state determination of the cycle subgraph. The material balance verification module and the result output module are used to generate steady-state material balance results and calculation result identifiers that can be written back to CAD elements.

[0068] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment includes only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for solving the cyclic steady-state problem of a crushing and screening process, characterized in that, Includes the following steps: Obtain process configuration data for the crushing and screening process, wherein the process configuration data includes equipment information for the crushing and screening process and material connection relationships between equipment; A directed graph model is constructed based on the process configuration data. The directed graph model includes nodes and directed edges connecting the nodes. The nodes correspond to the equipment in the crushing and screening process. The directed edges correspond to the material transfer relationship between the equipment. The directed edges carry the material flow rate and the multi-size particle size distribution vector of the material. A device condition calculation model is established for each node. The device condition calculation model is used to calculate the output material data of the node based on the material data of the input node and the equipment parameters of the node. Cyclic structure identification is performed on the directed graph to determine at least one cyclic subgraph and the linear region in the directed graph other than the cyclic subgraph, and the solution level of each cyclic subgraph is determined. The cyclic subgraph represents a closed-loop structure with material feedback relationship in the crushing and screening process. Perform linear pre-computation on the linear region to obtain the external input boundaries of each cyclic subgraph; According to the solution hierarchy, iterative solution is performed on each cyclic subgraph. For each cyclic subgraph, based on the external input boundary and the feedback data of the previous cycle, the multi-source material flow and its multi-size particle size distribution vector entering the same node are merged, and the equipment condition calculation model of each node is called to perform the current round of calculation to obtain the node output data of the current round. During the iterative solution process, iterative stabilization is performed. After reaching the preset minimum number of iterations, based on the error between adjacent iteration output data of each node in the cyclic subgraph, a full node convergence determination is performed. When the errors of all participating nodes meet the preset convergence condition, the cyclic subgraph is determined to have reached a steady state. Output the steady-state calculation results.

2. The method according to claim 1, characterized in that, The multi-size particle size distribution vector includes mass flow data or mass ratio data corresponding to at least three particle size ranges, and the particle size ranges correspond to the project sieve aperture definition, product specifications, raw material particle size range, or preset standard particle size definition.

3. The method according to claim 1, characterized in that, The process involves performing cycle structure identification on the directed graph to determine at least one cyclic subgraph and the linear regions in the directed graph excluding the cyclic subgraphs, and determining the solution level of each cyclic subgraph. The cyclic subgraph represents a closed-loop structure with material feedback relationships in the crushing and screening process, including: Perform strongly connected component decomposition on the directed graph to obtain the strongly connected components of the directed graph; A cyclic subgraph is determined from the strongly connected components based on at least one of the following: the device type corresponding to the node contained in each strongly connected component, the connection port type between nodes, the material attribute identifier carried by the directed edge, or the return edge identifier.

4. The method according to claim 3, characterized in that, The strongly connected component decomposition employs any one of the Tarjan algorithm, Kosaraju algorithm, or Gabow algorithm.

5. The method according to claim 3, characterized in that, The determination of the solution level for each cyclic subgraph includes: Construct a directed acyclic dependency graph between each cyclic subgraph. The output data of one of the cyclic subgraphs after convergence is used as the external input boundary data of the downstream cyclic subgraphs that depend on this cyclic subgraph. Determine the solution hierarchy order of each cyclic subgraph based on the directed acyclic dependency graph.

6. The method according to claim 1, characterized in that, The step of performing linear pre-computation on the linear region to obtain the external input boundaries of each cyclic subgraph includes: The linear pre-computation stage does not take the feedback value of the previous round for the cyclic feedback edge, but only calculates the output of linear nodes that do not belong to the cyclic subgraph and the external input boundary that enters the cyclic subgraph.

7. The method according to claim 1, characterized in that, The merging of multi-source material flows entering the same node and their multi-size particle size distribution vectors includes: Obtain the material flow data and corresponding particle size ratio data carried by each input directed edge entering the node; The mass ratio data is weighted and merged according to the material flow rate of each input directed edge to obtain the synthetic input granularity distribution of the node.

8. The method according to claim 1, characterized in that, The iterative stabilization process includes performing at least one of the following on node output data, circulating flow data, or node state vectors: damped update, state vector mixing, outlier suppression, negative value truncation, or local rollback.

9. The method according to claim 8, characterized in that, The damping update involves weighted mixing of the current round candidate state vector and the previous round state vector. The mixed state vector is used to replace the current round node output data. The current round candidate state vector is the node state data directly output by the equipment operating condition calculation model. The weight of the current round candidate state vector is β, and the weight of the previous round state vector is 1-β. The value of β is 0 < β ≤ 1.

10. The method according to claim 8, characterized in that, The state vector mixing includes Anderson mixing processing, which constructs accelerated state vectors based on historical multi-round node output data.

11. The method according to claim 1, characterized in that, The full node convergence determination includes: For each participating decision node in the cyclic subgraph, calculate the absolute error and relative error of the output flow data of the node in the two adjacent rounds; When the absolute error is not greater than the first preset threshold or the relative error is not greater than the second preset threshold, the node is determined to meet the flow convergence condition. When all participating nodes in the cyclic subgraph satisfy the flow convergence condition, the cyclic subgraph is determined to satisfy the flow steady-state condition.

12. The method according to claim 1, characterized in that, After the cyclic subgraph has initially converged, one or more final correction iterations are performed, and the final steady-state result is submitted when the change in the output of the largest node in the final correction iteration is lower than the third preset threshold.

13. The method according to claim 1, characterized in that, Before or after outputting the steady-state calculation results, a material balance verification is performed. The material balance verification includes verifying the balance relationship between the total fresh input, the total final product output, the internal circulation volume, and the loss bypass volume, and verifying the quality balance deviation of each particle size range, and generating a material balance report.

14. The method according to claim 1, characterized in that, The steady-state calculation results include at least one of the following: input flow data, output flow data, circulating flow data, circulation ratio, equipment load rate data, product particle size distribution data, convergence status data, material balance report, or calculation result identifier that can be written back to CAD drawing elements.

15. A cyclic steady-state solution system for a crushing and screening process, characterized in that, include: The acquisition module is used to acquire process configuration data of the crushing and screening process flow, the process configuration data including equipment information of the crushing and screening process flow and material connection relationships between equipment; A directed graph model construction module is used to construct a directed graph model based on the process configuration data. The directed graph model includes nodes and directed edges connecting the nodes. The nodes correspond to the equipment in the crushing and screening process, and the directed edges correspond to the material transfer relationship between the equipment. The directed edges carry the material flow rate and the multi-size particle size distribution vector of the material. The equipment condition calculation module is used to establish an equipment condition calculation model for each node. The equipment condition calculation model is used to calculate the output material data of the node based on the material data of the input node and the equipment parameters of the node. The cycle structure identification module is used to perform cycle structure identification on the directed graph, identify at least one cycle subgraph and the linear region in the directed graph other than the cycle subgraph, and determine the solution level of each cycle subgraph. The cycle subgraph represents a closed-loop structure with material feedback relationship in the crushing and screening process. The linear pre-computation module is used to perform linear pre-computation on the linear region to obtain the external input boundaries of each cyclic subgraph; The iterative solution module is used to perform iterative solution on each loop subgraph according to the solution hierarchy. For each loop subgraph, based on the external input boundary and the feedback data of the previous loop, the multi-source material flow and its multi-size particle size distribution vector entering the same node are merged, and the equipment condition calculation model of each node is called to perform the current round of calculation to obtain the node output data of the current round. A stabilization module is used to perform iterative stabilization processing during the iterative solution process; The convergence determination module is used to perform full node convergence determination based on the error between adjacent iteration output data of each node in the cyclic subgraph after reaching the preset minimum number of iterations. When the errors of all participating nodes meet the preset convergence conditions, the cyclic subgraph is determined to have reached a steady state. The results output module is used to output the steady-state calculation results.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 14.

17. A computer program product comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the method described in any one of claims 1 to 14.

Images