Intelligent control optimization method and device for silicon steel sheet production of generator

Abstract

This invention discloses an intelligent control optimization method and device for the production of silicon steel sheets for generators, relating to the field of generator technology. The method includes: interactively acquiring design requirement parameters and production process sequences for silicon steel sheet components; performing performance coupling decomposition of the process chain to generate multiple process-level performance optimization targets; matching and aggregating control optimization algorithms to obtain M sets of process-level performance optimization targets; using these as constraints to perform cross-process parameter tuning on the M sets of production processes, outputting M sets of control parameters, and executing corresponding full-process production operations. This invention solves the technical problems of low production efficiency and unstable product quality in traditional silicon steel sheet production due to the lack of effective collaborative optimization between process steps. It achieves the technical effect of improving silicon steel sheet production efficiency and product quality stability through cross-process collaborative parameter tuning optimization, thereby enhancing the performance control accuracy and energy efficiency level of key magnetic components in energy-saving generators.

Description

Technical Field

[0001] This invention relates to the field of generator technology, and more specifically to an intelligent control optimization method and device for the production of silicon steel sheets for generators. Background Technology

[0002] As a crucial component of high-efficiency power systems, the performance of energy-saving generators is highly dependent on the physical quality and magnetic stability of their core magnetic component—silicon steel sheets. Existing silicon steel sheet manufacturing processes often employ static parameter configuration and independent process control, making it difficult to guarantee performance throughout the entire process to meet the precise control requirements of energy-saving generators for core performance aspects such as iron loss and permeability. Furthermore, traditional optimization methods lack performance transmission mechanisms between multiple processes, resulting in a lack of synergy in parameter tuning, ultimately affecting the energy efficiency of energy-saving generators under high-frequency operation or complex load scenarios. Summary of the Invention

[0003] This application provides an intelligent control optimization method and device for the production of silicon steel sheets for generators, which solves the technical problem of low production efficiency and unstable product quality caused by the lack of effective collaborative optimization between process steps in the traditional silicon steel sheet production process.

[0004] The first aspect of this application provides an intelligent control optimization method for the production of silicon steel sheets for generators. The method includes: interactively obtaining design requirement parameters and a production process sequence for the silicon steel sheet component; based on the production process sequence, performing process chain performance coupling decomposition on the design requirement parameters, mapping the design requirement parameters to multiple process-level performance optimization targets corresponding to multiple process-level production processes in the production process sequence; matching control optimization algorithms based on the multiple process-level performance optimization targets and multiple process-level production processes, and aggregating and outputting M sets of process-level performance optimization targets corresponding to M sets of process-level production processes based on the consistency of the optimization algorithms; using the M sets of process-level performance optimization targets as constraints, performing cross-process collaborative parameter tuning optimization on the M sets of process-level production processes, and outputting M sets of collaborative production control parameters; and using the M sets of collaborative production control parameters to execute the entire process of intelligent collaborative production of the silicon steel sheet component for generators.

[0005] A second aspect of this application provides an intelligent control optimization device for the production of silicon steel sheets for generators. The device includes: a basic parameter acquisition module, used to interactively obtain design requirement parameters and component manufacturing process sequences for silicon steel sheet components; a process chain performance decomposition module, used to perform process chain performance coupling decomposition on the design requirement parameters based on the component manufacturing process sequence, mapping the design requirement parameters to multiple process-level performance optimization targets corresponding to multiple process-level manufacturing processes in the component manufacturing process sequence; and a performance optimization target aggregation module, used to... Based on the multiple process-level performance optimization objectives and multiple process-level production processes, after matching control optimization algorithms, and based on the consistency of optimization algorithms, aggregate and output M sets of process-level performance optimization objectives corresponding to M sets of process-level production processes; cross-process collaborative parameter tuning optimization module, which is used to perform cross-process collaborative parameter tuning optimization on the M sets of process-level production processes with the M sets of process-level performance optimization objectives as constraints, and output M sets of collaborative production control parameters; intelligent collaborative production module, which is used to use the M sets of collaborative production control parameters to execute the full-process intelligent collaborative production of the silicon steel sheet components for generators.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The intelligent control optimization method and device for the production of silicon steel sheets for generators provided in this application relate to the field of generator technology. It obtains the design requirements and production process sequence of silicon steel sheets through interaction, performs performance coupling decomposition of the process chain, and forms multi-process-level performance optimization targets. Based on the matching and consistency principles of control optimization algorithms, these targets are aggregated into M sets of optimization targets. Then, cross-process collaborative parameter tuning is implemented to generate M sets of control parameters, which are applied to the entire process of intelligent collaborative production. This solves the technical problem in the traditional silicon steel sheet production process where there is a lack of effective collaborative optimization between process steps, leading to low production efficiency and unstable product quality. It achieves intelligent collaborative production throughout the entire process through cross-process collaborative parameter tuning optimization, improving production efficiency and product quality stability, and ultimately enhancing the performance control accuracy and energy efficiency of key magnetic components in energy-saving generators. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1A schematic diagram of the intelligent control optimization method for producing silicon steel sheets for generators provided in this application embodiment;

[0010] Figure 2 A schematic diagram of the intelligent control optimization device for producing silicon steel sheets for generators provided in this application embodiment.

[0011] Figure labeling: 11 Basic parameter acquisition module, 12 Process chain performance decomposition module, 13 Performance optimization target aggregation module, 14 Cross-process collaborative parameter tuning and optimization module, 15 Intelligent collaborative production module. Detailed Implementation

[0012] This application provides an intelligent control optimization method and device for the production of silicon steel sheets for generators, which solves the technical problem of low production efficiency and unstable product quality caused by the lack of effective collaborative optimization between process steps in the traditional silicon steel sheet production process.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, this application provides an intelligent control optimization method for the production of silicon steel sheets for generators, the method comprising:

[0016] P10: Interact to obtain the design requirements parameters and component manufacturing process sequence for silicon steel sheet components.

[0017] Specifically, through data interaction interfaces with the design platform and production management system, the system automatically acquires the design requirement parameters and corresponding component manufacturing process sequences for silicon steel sheet components. These design requirement parameters include, but are not limited to, the target component's shape and dimensions, tolerance range, magnetic properties (such as permeability and loss value), mechanical performance requirements (such as tensile strength and yield strength), and surface condition standards. These parameters are generally stored in the design database and transmitted to the production optimization system via structured data formats (such as XML or JSON). During data acquisition, the system calls the integrated CAD / PDM system interface to extract key design indicators such as the design file number, material grade, and usage conditions that match the component to be produced. These indicators are then parametrically parsed and standardized to ensure the consistency and identifiability of subsequent optimization model inputs.

[0018] Simultaneously, by accessing historical production configurations in the MES (Manufacturing Execution System), the system retrieves the corresponding production process sequence for this type of silicon steel sheet component within existing process templates. This process sequence is a set of chronologically ordered steps, typically including standard processes such as hot rolling, cold rolling, annealing, coating, and shearing. Each step records the corresponding processing equipment number, adjustable control parameter range, process control elements (such as tension, temperature, and linear speed), and preset quality control thresholds. To ensure that this process sequence matches design requirements, the system introduces a product family process template matching mechanism. This mechanism compares and correlates the key performance requirements in the design specifications with the performance capabilities in the process template library, selecting target process sequences that meet the constraints and using them as input for process optimization.

[0019] This interactive acquisition process is completed through an automated interface, requiring no manual intervention and offering real-time performance and traceability. It supports production change management and customized adjustments. During execution, the system performs integrity checks on the received design parameters and process sequences, and generates standardized data structures as input for process chain performance coupling analysis. This provides operational basic data support for subsequent process-level performance target extrapolation and optimization algorithm matching.

[0020] P20: Based on the component manufacturing process sequence, the design requirement parameters are decomposed into process chain performance coupling, and the design requirement parameters are mapped to multiple process-level performance optimization targets corresponding to multiple process-level manufacturing processes in the component manufacturing process sequence.

[0021] Furthermore, step P20 in this embodiment of the application also includes:

[0022] P21: Based on the performance regulation effects of multiple silicon steel sheets in the multiple process-level production processes, construct an inter-process performance transfer topology; P22: Quantify the coupling strength between process nodes in the inter-process performance transfer topology, and output the inter-process performance coupling topology; P23: Perform gradient sensitivity analysis between adjacent downstream nodes on the inter-process performance coupling topology to locate the dominant decomposition path of the process chain; P24: Back-map the design requirement parameters along the chain gradient of the dominant decomposition path of the process chain to output multiple process-level performance boundaries as the performance optimization targets of the multiple process-level processes.

[0023] It should be understood that the design requirement parameters are decomposed into process chain performance coupling based on the component manufacturing process sequence. First, design requirement parameters typically consist of multiple dimensions of physical, chemical, and mechanical performance indicators, which are core elements ensuring the final product meets design requirements. The goal of process chain performance coupling decomposition is to analyze the interactions between different process steps, determine how each step affects various performance indicators of the final product, and refine these performance indicators into performance optimization goals at each process level. For example, based on the obtained component manufacturing process sequence, the design requirement parameters are initially decomposed. Design requirement parameters (such as dimensional accuracy, magnetic properties, and surface quality) are performance indicators of the final product, and achieving these indicators requires multiple process-level manufacturing processes. For instance, achieving magnetic property indicators (such as iron loss and magnetic flux density) depends not only on temperature control in the annealing process but also on factors such as rolling force in the cold rolling process and coating quality in the coating process. Therefore, it is necessary to decompose the design requirement parameters into multiple process-level performance optimization goals to ensure that the optimization goals of each process are consistent with the overall design requirements.

[0024] Furthermore, based on the performance regulation effects of multiple production processes on silicon steel sheets, a performance transfer topology between processes is constructed. Specifically, the impact of each process on the performance of silicon steel sheets is analyzed, and the performance transfer relationships between processes are determined. For example, the rolling force and rolling speed in the cold rolling process affect the thickness uniformity and internal stress distribution of the silicon steel sheet, and these factors directly affect the grain growth in the annealing process, thus affecting the final magnetic properties. By analyzing this causal relationship between processes, a topological structure for performance transfer between processes is constructed, clarifying the role and mutual influence of each process in the process chain.

[0025] Next, the coupling strength between process nodes is quantified in the constructed inter-process performance transfer topology. The purpose of coupling strength quantification is to calculate the degree of mutual influence between each process node, and thus assess their combined impact on design requirements parameters. Coupling strength reflects the tightness and interdependence of performance transfer between processes. For example, there may be a strong coupling relationship between temperature control in the annealing process and rolling force in the cold rolling process, because the stress distribution after cold rolling affects the grain growth direction and size during annealing. By quantifying this coupling strength, it is possible to more clearly identify which processes have the most critical performance transfer. This step, through numerical methods, transforms the coupling relationship between each process node into operable numerical results, thereby providing a clear technical basis for subsequent optimization calculations.

[0026] Next, gradient sensitivity analysis is performed on the inter-process performance coupling topology between adjacent downstream nodes. Gradient sensitivity analysis is a method used to identify key influencing factors. By calculating the gradient change of each process parameter with respect to the performance of downstream processes, the dominant decomposition path in the process chain is located. For example, analyzing the gradient effect of the rolling force change in the cold rolling process on the magnetic properties after annealing, if the gradient is large, it indicates that the rolling force in the cold rolling process is one of the key factors affecting the magnetic properties. Through this analysis, it is possible to determine which process parameters have the most significant impact on the performance of the final product, thus providing a basis for setting optimization objectives.

[0027] Finally, along the dominant decomposition path of the process chain, a chain-like gradient back-mapping is used to map the design requirement parameters, outputting multiple process-level performance boundaries as performance optimization targets for each process. Specifically, starting from the design requirement parameters of the final product, the performance optimization target for each process is derived backward along the identified dominant decomposition path. For example, if the final product requires low iron loss and high magnetic induction intensity, back-mapping can determine the temperature range that needs to be controlled in the annealing process, the rolling force range that needs to be controlled in the cold rolling process, and so on. These process-level performance boundaries are the targets that subsequent optimization algorithms need to achieve, ensuring that the optimization results of each process can collaboratively realize the design requirements of the final product.

[0028] Through the above process, the system not only maps design requirement parameters to performance optimization targets at each process level, but also ensures that every process in the production process can operate efficiently while guaranteeing product quality through process chain performance coupling decomposition and optimization.

[0029] Furthermore, step P21 in the embodiments of this application also includes:

[0030] P21-1: By backtracking local production test data, obtain multiple sets of performance regulation quantification values ​​for the multiple process-level production processes; P21-2: Aggregate the multiple sets of performance regulation quantification values ​​by performance regulation type to obtain P types of performance regulation; P21-3: Using the component production process sequence as a directed connection constraint, treat the multiple process-level production processes as multiple topology nodes, and use the P types of performance regulation to traverse the multiple sets of performance regulation quantification values ​​to fit the performance transmission path, outputting P one-dimensional performance transmission topologies; P21-4: Spatially overlap the P one-dimensional performance transmission topologies to obtain the inter-process performance transfer topology.

[0031] Optionally, the construction process of the performance transfer topology between processes can be further refined. By backtracking production test data, aggregating performance control types, fitting performance transmission paths, and spatial overlap of topologies, a precise description of the performance transfer relationship between processes can be achieved.

[0032] First, by retrospectively analyzing local production test data, multiple sets of quantitative values ​​for performance regulation at various process levels were obtained. These data reflect the specific impact of parameter variations at each process stage on the performance of silicon steel sheets during actual production. For example, by analyzing data on the thickness uniformity and internal stress distribution of silicon steel sheets under different rolling forces and speeds in the cold rolling process, and grain size and magnetic properties under different temperatures and holding times in the annealing process, specific quantitative values ​​for performance regulation can be obtained. This data forms the basis for subsequent analysis and provides empirical support for determining the types of performance regulation.

[0033] After data acquisition, the aforementioned multiple sets of performance regulation quantification values ​​are further aggregated by type. Based on the mechanism, direction, and magnitude of the parameter regulation's effect on the target performance, these quantification values ​​are classified into P performance regulation types. For example, some regulation types mainly affect magnetic properties, while others mainly affect mechanical properties. The system uses clustering analysis or principal component decomposition techniques to perform multidimensional mapping on the quantification values ​​and extract features, grouping highly similar performance regulation behaviors into one category. In this way, dimensional complexity can be significantly reduced while maintaining the integrity of the expression of process performance regulation patterns, providing standardized input for the subsequent topology fitting process.

[0034] Subsequently, based on the acquired component manufacturing process sequence, multiple process-level processes are set as directed nodes in the topology graph. The connection order between nodes is completely consistent with the process execution order, forming an initial directed graph structure. Under this graph structure, the system sequentially uses each performance control type as a constraint, traverses the quantified performance control values ​​between processes, and uses modeling methods such as linear regression, Bayesian networks, or graph neural networks to fit the path of the transmission behavior of control values ​​between nodes, quantifying the performance influence direction and transmission strength of each path, thereby outputting P one-dimensional performance transmission topologies. Each topology graph corresponds to the flow characteristics of a performance control type in the process chain. For example, for the "magnetic performance control type," from the rolling force of the cold rolling process affecting the grain size of the annealing process, to the final realization of magnetic properties, a complete magnetic performance transmission path is fitted. These one-dimensional topology graphs collectively characterize the propagation structure of different control types in the multi-process chain, which is a prerequisite for constructing a comprehensive performance transfer model.

[0035] After fitting P single-dimensional performance transfer topologies, these graphs are spatially overlapped. Specifically, all topology graphs are superimposed on the same directed graph structure, and the influence of different performance control types under the same connection path is weighted and fused to generate an inter-process performance transfer topology graph containing composite control relationships. This topology graph not only describes whether performance transfer relationships exist between process nodes, but also records the influence intensity and direction of different control types in the form of edge weights, thus comprehensively reflecting the degree of synergy and coupling characteristics between process steps in achieving the final design goal. For example, the cold rolling process node manifests as an influence on grain size in the magnetic performance transfer path and an influence on thickness uniformity in the dimensional accuracy transfer path. Through this spatial overlap, the final inter-process performance transfer topology can comprehensively and accurately describe the performance transfer relationships between processes, providing a solid foundation for subsequent process chain performance coupling decomposition and optimization.

[0036] Furthermore, step P22 in the embodiments of this application also includes:

[0037] P22-1: Calculate performance control weights for the multiple sets of performance control quantization values ​​using the P performance control types to locate multiple sets of performance control intensity coefficients; P22-2: Based on the P performance control types, perform cross-node serialization processing on the multiple sets of performance control intensity coefficients to obtain P performance control intensity sequences; P22-3: Preset a cross-stage control intensity threshold; P22-4: Use the cross-stage control intensity threshold to traverse the P performance control intensity sequences to locate P sets of collaborative control nodes; P22-5: Sort the multiple sets of performance control intensity coefficients within their groups to locate multiple independent control nodes; P22-6: Reconstruct the inter-process performance transfer topology based on the P sets of collaborative control nodes and multiple independent control nodes, and output the inter-process performance coupling topology, wherein the coupling nodes have performance control intensity weights.

[0038] Specifically, based on the existing inter-process performance transfer topology, a quantitative calculation and structural reconstruction mechanism for performance regulation intensity can be further introduced to generate an inter-process performance coupling topology containing intensity weights, thereby providing a foundation for accurately identifying key regulation nodes and performing collaborative optimization.

[0039] The first step in this process is to calculate the performance control weights. Using the P performance control types categorized in the previous stage, weights are assigned to each group of quantitative performance control values. This involves introducing an influencing factor calculation model (such as the sensitivity function of control parameters to output performance, normalized weight matrices, etc.) while preserving the performance response change trend. This quantifies the intensity of each control type's effect under specific processes, ultimately yielding a set of performance control intensity coefficients that correspond one-to-one with each process node. These coefficients accurately reflect the dominance and control capability of each process step in achieving the target performance.

[0040] After obtaining the performance regulation intensity coefficients, based on the process execution sequence, cross-node serialization is performed on the intensity coefficients for each regulation type, thereby constructing P performance regulation intensity sequences. Each sequence represents the transmission path and intensity distribution characteristics of that type of regulation factor in the complete process chain, revealing which processes have strong control relationships and which processes are regulation breakpoints or weakly correlated nodes. For example, for magnetic property regulation, a performance regulation intensity sequence is generated from the cold rolling process to the annealing process and then to the coating process. Through this serialization process, the intensity change trend of each process under different performance regulation types can be clearly displayed.

[0041] After the sequence is constructed, a set of cross-stage control intensity thresholds is preset. These thresholds are used to screen high-intensity control paths that span multiple process stages and continuously affect downstream performance output. For example, if the performance control intensity coefficient of a certain process exceeds the preset threshold, the process is considered to have a significant impact on this type of performance control and needs to be optimized in conjunction with other processes. These thresholds are usually set based on historical process statistical analysis and expert rules, and have dynamic adjustment capabilities to ensure adaptability and versatility under different product types and process scenarios.

[0042] Subsequently, using this cross-stage control intensity threshold, each performance control intensity sequence was traversed to screen for node combinations with control intensities continuously exceeding the threshold, identifying P groups of collaborative control nodes. These nodes exhibit significant performance influence relationships, and their coordinated regulation contributes far more to the overall target performance than random node combinations; therefore, they are considered key targets for subsequent cross-process collaborative parameter tuning. Simultaneously, to improve the precision of system control, the control intensity coefficients under each performance control type were ranked within their respective groups to identify process nodes that, while not reaching the collaborative control threshold, possess high intensity within their groups. Although these nodes do not require collaborative control with other processes, they remain important within their own processes and require independent optimization. For example, in the dimensional accuracy control intensity sequence, although the performance control intensity coefficient of the cold rolling process did not exceed the collaborative control threshold, it ranked highly within its group, thus positioning it as an independent control node.

[0043] Finally, based on the P-group of collaborative control nodes and multiple independent control nodes, the inter-process performance transfer topology is reconstructed, outputting the inter-process performance coupling topology. Specifically, the collaborative control nodes and independent control nodes are reintegrated into the inter-process performance transfer topology, forming a new topology. In this new topology, each coupling node has a performance control intensity weight, which reflects the importance and influence of each process under different performance control types. For example, in the reconstructed topology, the annealing process, as a collaborative control node, has a high performance control intensity weight, while the cold rolling process, as an independent control node, has a moderate performance control intensity weight. This topology can be used for subsequent sensitivity analysis, decomposed path backtracking, and optimization constraint generation, providing a highly reliable basic data structure for realizing a full-process, multi-node, asynchronous parameter tuning intelligent control strategy.

[0044] P30: Based on the multiple process-level performance optimization objectives and multiple process-level production processes, after matching the control optimization algorithms, and based on the consistency of the optimization algorithms, aggregate and output the M sets of process-level performance optimization objectives corresponding to the M sets of process-level production processes.

[0045] Furthermore, step P30 in this embodiment of the application also includes:

[0046] P31: Interact to obtain multiple sets of support optimization performance and multiple sets of support optimization algorithms for multiple sample production processes; P32: Based on the knowledge graph, associate and store the multiple sample production processes, multiple sets of support optimization performance, and multiple sets of support optimization algorithms to obtain a control optimization algorithm library; P33: Use the multiple process-level performance optimization objectives and multiple process-level production processes as dual retrieval matching constraints to retrieve and call multiple control optimization algorithms from the control optimization algorithm library; P34: Based on the consistency of optimization algorithms, aggregate the multiple process-level performance optimization objectives and multiple process-level production processes according to the multiple control optimization algorithms to obtain the M sets of process-level production processes and M sets of process-level performance optimization objectives.

[0047] Optionally, based on multiple process-level performance optimization objectives and corresponding production processes, a matching and aggregation process of control optimization algorithms is executed to construct a set of optimization strategies that match the processes of each process. This process first requires identifying the performance objectives required for each process and its controlled process variables, then retrieving optimization algorithms compatible with the control mechanisms of these variables from the optimization algorithm library, and finally, while ensuring the applicability and consistency of the algorithms, completing the fusion and aggregation of process-level optimization objectives.

[0048] In the specific execution process, the system first obtains supporting information on multiple sample production processes by interacting with a knowledge modeling platform, and constructs multiple sets of mapping relationships between supporting optimization performance and multiple sets of supporting optimization algorithms. The sample production processes refer to typical processes used in the historical production of similar silicon steel sheet components. The performance optimization objectives and control algorithms corresponding to these processes have been validated in previous engineering practices. Therefore, the system retrieves this sample data from the production database through a data interface, automatically extracts the parameter tuning objectives (such as target thickness uniformity, surface tension stability, magnetic property control boundaries, etc.) and their associated optimization algorithms (such as multivariable linear programming, genetic algorithms, gradient descent strategies, fuzzy control methods, etc.) corresponding to each process node, and transforms this information into structured knowledge entries for constructing an algorithm knowledge graph.

[0049] Subsequently, a knowledge graph of control optimization algorithms was constructed based on the above information. This graph uses the sample production process as the main entity, optimization performance objectives as intermediary nodes, and optimization algorithms as response objects. These three elements are linked through a multi-level relational storage structure. The knowledge graph not only stores static mapping relationships but also includes dynamic tags such as the difficulty of performance objective control, the applicable scope of the algorithm, the algorithm complexity, and historical adaptation scores, supporting subsequent multi-condition semantic retrieval and intelligent matching. Under this graph structure, each process objective can be mapped to one or more optimization algorithms under different process backgrounds, improving the targeting and adaptability of algorithm calls.

[0050] Next, the system uses the multiple process-level performance optimization objectives and their corresponding production process structures as dual search constraints, and invokes the knowledge graph to perform semantic queries. This retrieves control optimization algorithms matching each objective from the control optimization algorithm library. This search process not only requires the target parameters to match the algorithm input format, but also comprehensively considers multiple factors such as the algorithm's operating efficiency, constraint handling capability, and parameter convergence stability in the current production equipment environment, ensuring the selected algorithm's deployability in real-world engineering scenarios. The system can select multiple candidate algorithms for each process and sort them according to indicators such as matching score and computational complexity, forming a candidate algorithm pool.

[0051] Finally, based on the principle of algorithm consistency, multiple control optimization algorithms are categorized and integrated. Algorithm consistency refers to the common foundations among optimization algorithms used in different processes, such as model framework compatibility, solution strategy synergy, or parameter tuning variable commutativity, enabling each process to operate under a unified optimization system. The system evaluates the structural compatibility and control variable synergy among algorithms, aggregating consistent algorithms into M groups. Each group corresponds to a process segment that can be co-optimized. Furthermore, the system performs target reconstruction or boundary fusion operations on the performance optimization objectives of each group of processes, ensuring that the aggregated optimization objectives possess cross-process adaptability and joint solution characteristics. In other words, among multiple candidate optimization algorithms, those that can consistently achieve process-level performance optimization objectives are selected. For example, for the co-optimization of cold rolling and annealing processes, an algorithm combination that simultaneously satisfies the requirements of cold rolling thickness uniformity and post-annealing magnetic property optimization is selected. The final output of M groups corresponds to the process-level production processes and process-level performance optimization objectives of the actual production path, guiding subsequent cross-process collaborative parameter tuning optimization processes and ensuring that the control strategy has coordination, continuity, and engineering feasibility throughout the entire process.

[0052] P40: Using the performance optimization objectives of the M group of process-level processes as constraints, perform cross-process collaborative parameter tuning optimization on the M group of process-level production processes, and output the collaborative production control parameters of the M group.

[0053] Furthermore, step P40 in this embodiment of the application also includes:

[0054] P41: Traverse the inter-process performance coupling topology to obtain W linkage optimization node groups for W process-level production processes in the first group of process-level production processes; P42: Based on the coupling strength characteristics of the W linkage optimization node groups, decompose the W process-level production processes into multiple cross-process linkage parameter tuning groups and multiple single-process parameter tuning optimization nodes; P43: Use the cross-process linkage parameter tuning optimization results of the multiple cross-process linkage parameter tuning groups as the single-process collaborative parameter tuning optimization boundary conditions of the multiple single-process parameter tuning optimization nodes, perform single-process parameter tuning optimization, and merge and output the first group of collaborative production control parameters; P44: Similarly, through cross-process collaborative parameter tuning optimization, output the M groups of collaborative production control parameters.

[0055] It should be understood that, with the previously obtained M group of process-level performance optimization targets as constraints, and combined with the specific production process of each group of processes, cross-process collaborative parameter tuning optimization is performed to output M group of collaborative production control parameters, which are used to guide the parameter setting and operation control of each group of processes in the actual production process.

[0056] Specifically, the process first traverses the inter-process performance coupling topology to identify W interconnected optimization node groups for the W process-level production processes in the first group of process-level production processes. The inter-process performance coupling topology is a network structure describing the performance transfer and coupling relationships between processes. Nodes represent processes, edges represent the coupling relationships between processes, and edge weights represent the coupling strength. By traversing this topology, process node groups that influence each other along the performance transfer path and require collaborative optimization can be found. For example, in silicon steel sheet production, there is a strong coupling relationship between the cold rolling and annealing processes; their process parameters influence each other and jointly determine the magnetic properties of the silicon steel sheet. Therefore, these two processes can be identified as an interconnected optimization node group.

[0057] After identifying the node groups, based on the coupling strength characteristics of the W linked optimization node groups, the W process-level production processes are decomposed into multiple cross-process linked parameter tuning groups and multiple single-process parameter tuning optimization nodes. Specifically, process node groups with high coupling strength are classified as cross-process linked parameter tuning groups, where processes within these groups need to adjust parameters synchronously to achieve collaborative optimization; while process nodes with low coupling strength or relatively independent processes are classified as single-process parameter tuning optimization nodes, which can be optimized independently. For example, if the cold rolling process and the annealing process have high coupling strength, they will be classified as a cross-process linked parameter tuning group; while if the coating process has low coupling strength with the preceding process, it can be used as a single-process parameter tuning optimization node.

[0058] Next, the optimization results of cross-process linkage parameter tuning groups are used as boundary conditions for single-process collaborative parameter tuning optimization of multiple single-process parameter tuning optimization nodes. Single-process parameter tuning optimization is then performed, and the first set of collaborative production control parameters is output. Specifically, for cross-process linkage parameter tuning groups, a multi-objective optimization model is established, considering both the performance optimization objectives and coupling relationships of each process within the group, to perform collaborative optimization. The optimization results will provide a set of collaborative process parameters for each process within the group. These parameters not only meet the performance optimization objectives of their respective processes but also consider the coupling effects between processes. Then, these optimization results are used as boundary conditions for single-process parameter tuning optimization nodes, and these nodes are optimized. For example, the linkage parameter tuning optimization results of cold rolling and annealing processes will provide optimized boundary conditions for the coating process, which is then optimized as a single process. Finally, the optimization results of the cross-process linkage parameter tuning groups and the single-process parameter tuning optimization nodes are merged to form the first set of collaborative production control parameters. These parameters will guide the execution of the first set of process-level production processes.

[0059] The optimization results of cross-process linkage parameter tuning groups are used as boundary conditions for single-process collaborative parameter tuning optimization nodes. Single-process parameter tuning optimization is then performed, and the results are combined to output the first set of collaborative production control parameters. Specifically, for cross-process linkage parameter tuning groups, a multi-objective optimization model is established, considering both the performance optimization objectives and coupling relationships of each process within the group, to perform collaborative optimization. The optimization results provide a set of collaborative process parameters for each process within the group. These parameters not only satisfy the performance optimization objectives of their respective processes but also consider the coupling effects between processes. Then, these optimization results are used as boundary conditions for single-process parameter tuning optimization nodes, which are then optimized. For example, the linkage parameter tuning optimization results of cold rolling and annealing processes provide optimized boundary conditions for the coating process, which is then optimized as a single process. Finally, the optimization results of the cross-process linkage parameter tuning groups and the single-process parameter tuning optimization nodes are combined to form the first set of collaborative production control parameters. These parameters guide the execution of the first set of process-level production processes.

[0060] Similarly, the above collaborative optimization process is repeated for the remaining M-1 groups of process-level production processes. Each group is based on its corresponding performance optimization objective, and topology analysis, group identification, linkage optimization, and boundary parameter tuning are performed in conjunction with the performance coupling relationship between processes. Finally, M groups of collaborative production control parameters are output sequentially. Each group of parameters not only adapts to the actual operating requirements of the current process, but also ensures coordination with the preceding and following processes, forming a set of optimal control parameter systems that cover the entire process, hierarchical levels, and cross processes.

[0061] By implementing this step, coordinated optimization control of multiple processes throughout the production of silicon steel sheets for generators is achieved, significantly improving the coordination between processes and overall production efficiency. At the same time, it provides a stable parameter basis and decision-making logic for the subsequent deployment of an intelligent closed-loop control system.

[0062] Furthermore, step P43 in this embodiment of the application also includes:

[0063] P43-1: Decompose the first set of process-level performance optimization objectives into multiple cross-process performance optimization objective subsets and multiple single-process performance optimization objectives; P43-2: Based on the multiple cross-process performance optimization objective subsets, run a first type of control optimization algorithm to perform process parameter tuning optimization on the multiple cross-process linkage parameter tuning groups, and output the cross-process linkage parameter tuning optimization results; P43-3: Based on the weight coefficients of the inter-process performance coupling topology, transform the cross-process linkage parameter tuning optimization results into the single-process collaborative parameter tuning optimization boundary conditions; P43-4: Under the constraints of the single-process collaborative parameter tuning optimization boundary conditions, based on the multiple single-process performance optimization objectives, run a first type of control optimization algorithm to perform process parameter tuning optimization on the multiple single-process parameter tuning optimization nodes, and output the single-process parameter tuning optimization results; P43-5: Merge the cross-process linkage parameter tuning optimization results and the single-process parameter tuning optimization results to generate the first set of collaborative production control parameters.

[0064] In one possible embodiment of this application, the collaborative process of cross-process linkage parameter tuning optimization and single-process parameter tuning optimization can be further refined, and the first group of process-level performance optimization objectives can be decomposed, classified and collaboratively solved according to their influence scope and structural characteristics.

[0065] First, the first set of process-level performance optimization objectives is decomposed into multiple cross-process performance optimization objective subsets and multiple single-process performance optimization objectives. Specifically, based on the coupling relationships between processes in the inter-process performance coupling topology, the overall performance optimization objectives are broken down into sub-objectives suitable for cross-process linkage parameter tuning groups and single-process parameter tuning optimization nodes. For example, if the first set of process-level performance optimization objectives is to improve the magnetic properties and dimensional accuracy of silicon steel sheets, then the magnetic property optimization objective can be decomposed into two cross-process performance optimization objective subsets: flatness optimization in the cold rolling process and grain size optimization in the annealing process; while the dimensional accuracy optimization objective is assigned to the cold rolling process as a single-process performance optimization objective.

[0066] Based on multiple cross-process performance optimization objective subsets, a first-type control optimization algorithm is run to perform process parameter tuning optimization on multiple cross-process linkage parameter tuning groups, outputting the cross-process linkage parameter tuning optimization results. The first-type control optimization algorithm can be based on Model Predictive Control (MPC), Genetic Algorithm (GA), or other algorithms suitable for multi-objective optimization. These algorithms can simultaneously consider the performance optimization objectives and coupling relationships of multiple processes. By establishing mathematical models and optimization models, they adjust the process parameters of each process to achieve the cross-process performance optimization objective subsets. For example, for the linkage parameter tuning optimization of cold rolling and annealing processes, the optimization algorithm will adjust parameters such as the rolling force of the cold rolling process and the temperature of the annealing process according to the magnetic performance optimization objective subset, outputting a set of optimized process parameters as the cross-process linkage parameter tuning optimization results.

[0067] Based on the weighting coefficients of the inter-process performance coupling topology, the results of cross-process linkage parameter tuning optimization are transformed into boundary conditions for single-process collaborative parameter tuning optimization. The weighting coefficients reflect the importance and mutual influence of each process within the performance coupling topology. By analyzing the weighting coefficients, the influence range and constraints of the cross-process linkage parameter tuning optimization results on single-process parameter tuning optimization nodes can be determined. For example, if the weighting coefficient between the cold rolling and annealing processes is high, it indicates that the optimization results of the cold rolling process have a significant impact on the annealing process. Therefore, the optimization parameters of the cold rolling process (such as rolling force) can be used as one of the boundary conditions for the parameter tuning optimization of the annealing process, ensuring that the influence of preceding processes is considered during single-process optimization.

[0068] Under the constraint of boundary conditions for single-process collaborative parameter tuning optimization, based on multiple single-process performance optimization objectives, the first type of control optimization algorithm is run to perform process parameter tuning optimization on multiple single-process parameter tuning optimization nodes, and outputs the single-process parameter tuning optimization results. During the single-process parameter tuning optimization process, the optimization algorithm considers the single-process performance optimization objectives and the boundary conditions derived from the cross-process linkage parameter tuning optimization results. For example, for the single-process parameter tuning optimization of the coating process, the optimization algorithm will adjust the process parameters (such as coating thickness and coating speed) of the coating process according to the dimensional accuracy optimization objective and the boundary conditions derived from the cold rolling process (such as the flatness requirements of silicon steel sheets), and output the single-process parameter tuning optimization results.

[0069] Finally, the results of cross-process linkage parameter tuning and optimization and the results of single-process parameter tuning and optimization are merged to generate the first set of collaborative production control parameters. Specifically, the process parameters of each process obtained from cross-process linkage parameter tuning and optimization are integrated with the process parameters obtained from single-process parameter tuning and optimization to form a complete set of collaborative production control parameters. These parameters not only meet the requirements of the cross-process performance optimization target subset, but also consider the single-process performance optimization target and the coupling relationship between processes, and can guide the collaborative execution of the first set of process-level production processes. For example, the first set of collaborative production control parameters may include optimized rolling force for the cold rolling process, optimized temperature for the annealing process, and optimized coating thickness for the coating process. These parameters work together to achieve the first set of process-level performance optimization targets.

[0070] Through this series of refined and hierarchical parameter tuning and optimization steps, the system not only ensures consistent performance and orderly linkage among multiple processes, but also guarantees the adjustable precision and target adaptability of local process details, forming a set of highly robust control parameters that can be directly deployed and executed in actual silicon steel sheet production.

[0071] P50: Using the aforementioned M-group collaborative production control parameters, execute the entire process of intelligent collaborative production of the silicon steel sheet components for generators.

[0072] Furthermore, step P50 in this embodiment of the application also includes:

[0073] P51: Decompose the M sets of collaborative production control parameters into collaborative production control sequences based on the component production process sequence; P52: Using the collaborative production control sequences, during the full-process intelligent collaborative production of the silicon steel sheet component for the generator, simultaneously collect multi-modal production monitoring sequences; P53: Perform parameter deviation analysis on the collaborative production control sequences and multi-modal production monitoring sequences to locate the first production risk node; P54: Perform temperature gradient calculation on the multi-modal production monitoring sequences to locate the second production risk node; P55: Link the first and second production risk nodes to perform production risk event backtracking and locate real-time risk compensation strategies.

[0074] It should be understood that by adopting the M-group collaborative production control parameters, the entire process of intelligent collaborative production of silicon steel sheet components for generators is executed, so as to achieve real-time monitoring and risk control of the production process.

[0075] First, based on the component manufacturing process sequence, the M sets of collaborative production control parameters are decomposed into a collaborative production control sequence. Specifically, the manufacturing process sequence defines the execution order and interrelationships of each process in the silicon steel sheet production process. According to this sequence, the optimized collaborative production control parameters are decomposed according to the process sequence, forming a control parameter sequence arranged in chronological order. For example, if the manufacturing process sequence includes three processes: cold rolling, annealing, and coating, then the collaborative production control sequence will sequentially include parameters such as the optimized rolling force for the cold rolling process, the optimized temperature for the annealing process, and the optimized coating thickness for the coating process. This decomposition method ensures that each process obtains the corresponding optimized control parameters during execution, thereby achieving collaborative production throughout the entire process.

[0076] The collaborative production control sequence is used to execute intelligent collaborative production throughout the entire process, simultaneously collecting multimodal production monitoring sequences. Multimodal production monitoring sequences refer to data collected during production through various sensors and monitoring equipment, including but not limited to temperature, pressure, speed, vibration, and current. This data can reflect the actual operating status of equipment and process parameters in real time. For example, in the cold rolling process, temperature sensors collect the roll temperature, and in the annealing process, pressure sensors collect the furnace pressure. By simultaneously collecting this multimodal data, the production process can be monitored in real time, providing data support for subsequent risk analysis.

[0077] Parameter deviation analysis is performed on the collaborative production control sequence and multimodal production monitoring sequence to pinpoint the first production risk node. Specifically, the actual collected multimodal production monitoring data is compared with the optimized parameters in the collaborative production control sequence to calculate the deviation value. If the deviation between the actual and optimized parameters of a certain process exceeds a preset threshold, the process is considered to have a production risk and is identified as the first production risk node. For example, in the annealing process, if the deviation between the actual furnace temperature and the optimized temperature exceeds the allowable range, the annealing process is identified as the first production risk node. This deviation analysis can promptly detect potential anomalies in the production process, providing a basis for risk prevention and control.

[0078] Temperature gradient calculations are performed on multimodal production monitoring sequences to pinpoint secondary production risk nodes. Temperature gradient refers to the rate of temperature change over time and space, and is a crucial indicator reflecting heat transfer and heat treatment effectiveness during production. By calculating the gradient changes in temperature data within the multimodal production monitoring sequences, nodes exhibiting abnormal temperature changes can be identified. For example, in the annealing process, a sudden increase in the temperature gradient may indicate uneven temperature distribution within the furnace, posing a risk of localized overheating or excessively rapid cooling, thus identifying this node as a secondary production risk node. This temperature gradient calculation allows for the identification of potential production risks from a heat treatment perspective, further improving the risk control system.

[0079] This approach links the first and second production risk nodes to trace production risk events and pinpoint real-time risk compensation strategies. Specifically, it involves a comprehensive analysis of the first and second production risk nodes (based on parameter deviation analysis) to trace the propagation path and impact range of production risk events. By analyzing the root causes and propagation paths of risk events, corresponding real-time risk compensation strategies are formulated. For example, if the annealing process is identified as the first or second production risk node, and the retrospective analysis reveals that the uneven temperature distribution is caused by unstable rolling force in the cold rolling process, then the rolling force control strategy for the cold rolling process can be adjusted, while temperature compensation adjustments can be made for the annealing process. This linked analysis and real-time compensation strategy effectively reduces production risks, ensuring the stability of the production process and the reliability of product quality.

[0080] Through the above steps, the entire production process can be monitored and risk managed, which not only ensures the coordinated operation between various processes, but also enables timely compensation measures to be taken when potential risks occur, thus guaranteeing the stability, efficiency and quality of the production process.

[0081] Furthermore, the embodiments of this application also include step P60, which further includes:

[0082] P61: After performing online process compensation using the real-time risk compensation strategy, collect multimodal production tracking sequences within a preset tracking time window; P62: If the analysis results of the multimodal production tracking sequences still include risk nodes, iteratively update the compensation strategy until the analysis results of the updated multimodal production tracking sequences are empty sets.

[0083] Optionally, risk response and optimization mechanisms in the production process can be further strengthened. Through real-time monitoring and feedback, continuous adjustments and compensation can be made when risks occur in the production process, thereby maintaining the stability and efficiency of the production process.

[0084] First, a real-time risk compensation strategy is employed to implement online process compensation. During production, once a risk node is identified and a real-time risk compensation strategy is established, the process parameters of the relevant steps are immediately adjusted according to that strategy. For example, if the annealing process is identified as a risk node due to uneven temperature distribution, the compensation strategy might be to adjust the heating power of the annealing furnace or change the flow rate of the cooling medium. After implementing the compensation strategy, a preset tracking time window is entered to collect multimodal production tracking sequences. The tracking time window is a preset time interval used to evaluate the effectiveness of the compensation strategy. Within this time window, multimodal data from the production process, including key parameters such as temperature, pressure, and speed, are collected using various sensors and monitoring devices. This data will be used for subsequent risk analysis to verify the effectiveness of the compensation strategy.

[0085] Next, the collected multimodal production tracking sequences are analyzed to determine if any risk nodes still exist. If the analysis results still include risk nodes, it indicates that the current compensation strategy has not completely eliminated the risk and requires further iteration and updating. Specifically, the compensation strategy is adjusted based on the manifestation and characteristics of risk nodes in the multimodal production tracking sequences. For example, if the temperature distribution problem in the annealing process remains unresolved, it may be necessary to further adjust the control logic of the heating power or optimize the flow distribution of the cooling medium. After updating the compensation strategy, online process compensation is executed again, and multimodal production tracking sequences are collected within the new tracking time window, repeating the above analysis process. By continuously iterating and updating the compensation strategy until the analysis results of the updated multimodal production tracking sequences are empty sets, i.e., no longer containing risk nodes, the stability of the production process and the reliability of product quality are ensured.

[0086] Through this continuous iterative update process, real-time response and adjustment to dynamic risks in the production process can be achieved, ensuring that normal production can be automatically adapted to and restored in the event of process changes or fluctuations in the production environment. This improves the stability of the production process, reduces the need for human intervention, and thus enhances the automation and efficiency of production.

[0087] In summary, the embodiments of this application have at least the following technical effects:

[0088] This application establishes a performance coupling decomposition mechanism at the process chain level to achieve structured mapping of design requirement parameters across multiple processes, enhancing the systematic nature and traceability of process control. It introduces a control optimization algorithm library and aggregates it based on the principle of algorithm consistency, improving the matching accuracy between process-level performance targets and control strategies. It constructs an inter-process performance coupling topology, implementing cross-process parameter tuning optimization, breaking the limitations of independent adjustment of traditional process links, and forming a multi-process parameter linkage control mechanism. It integrates multi-modal production monitoring data to conduct online risk identification and deviation backtracking, improving the rapid response and compensation capabilities for abnormal states. Through iterative optimization and feedback, it establishes a closed-loop adaptive control system, enhancing the stability and robustness of the production process.

[0089] This technology has achieved the goal of intelligent collaborative production throughout the entire process through cross-process collaborative parameter adjustment and optimization, thereby improving production efficiency and product quality stability, and providing key support for the efficient manufacturing of energy-saving generators and the improvement of material energy efficiency.

[0090] Example 2, based on the same inventive concept as the intelligent control optimization method for producing silicon steel sheets for generators in the foregoing examples, such as... Figure 2 As shown, this application provides an intelligent control optimization device for the production of silicon steel sheets for generators. The device and method embodiments in this application are based on the same inventive concept. The device includes:

[0091] The basic parameter acquisition module 11 is used to interactively obtain the design requirement parameters and component manufacturing process sequence of the silicon steel sheet component.

[0092] The process chain performance decomposition module 12 is used to perform process chain performance coupling decomposition on the design requirement parameters based on the component manufacturing process sequence, and map the design requirement parameters to multiple process-level performance optimization targets corresponding to multiple process-level production processes in the component manufacturing process sequence.

[0093] The performance optimization target aggregation module 13 is used to perform control optimization algorithm matching based on the multiple process-level performance optimization targets and multiple process-level production processes, and then aggregate and output M sets of process-level performance optimization targets corresponding to M sets of process-level production processes based on the consistency of the optimization algorithms.

[0094] The cross-process collaborative parameter tuning and optimization module 14 is used to perform cross-process collaborative parameter tuning and optimization on the M group of process-level production processes with the M group of process-level performance optimization targets as constraints, and output the M group of collaborative production control parameters.

[0095] The intelligent collaborative production module 15 is used to execute the entire process of intelligent collaborative production of the silicon steel sheet component for generators by adopting the M sets of collaborative production control parameters.

[0096] Furthermore, the process chain performance decomposition module 12 is also used to perform the following steps:

[0097] Based on the performance regulation effects of multiple silicon steel sheets in the multiple process-level production processes, an inter-process performance transfer topology is constructed; the coupling strength between process nodes is quantified in the inter-process performance transfer topology to output the inter-process performance coupling topology; gradient sensitivity analysis between adjacent downstream nodes is performed on the inter-process performance coupling topology to locate the dominant decomposition path of the process chain; the design requirement parameters are back-mapped along the chain gradient of the dominant decomposition path of the process chain to output multiple process-level performance boundaries as the performance optimization targets of the multiple process-level processes.

[0098] Furthermore, the process chain performance decomposition module 12 is also used to perform the following steps:

[0099] By backtracking local production test data, multiple sets of performance regulation quantification values ​​for the multiple process-level production processes are obtained; the multiple sets of performance regulation quantification values ​​are aggregated by performance regulation type to obtain P performance regulation types; using the component production process sequence as a directed connection constraint, the multiple process-level production processes are used as multiple topology nodes, and the performance transmission path is fitted by traversing the multiple sets of performance regulation quantification values ​​using the P performance regulation types, outputting P one-dimensional performance transmission topologies; the P one-dimensional performance transmission topologies are spatially overlapped to obtain the inter-process performance transfer topology.

[0100] Furthermore, the process chain performance decomposition module 12 is also used to perform the following steps:

[0101] The P performance control types are used to calculate the performance control weights of the multiple sets of performance control quantification values ​​to locate the multiple sets of performance control intensity coefficients; based on the P performance control types, the multiple sets of performance control intensity coefficients are subjected to cross-node serialization processing to obtain P performance control intensity sequences; a cross-stage control intensity threshold is preset; the cross-stage control intensity threshold is used to traverse the P performance control intensity sequences to locate P sets of collaborative control nodes; the multiple sets of performance control intensity coefficients are sorted within the group to locate multiple independent control nodes; the inter-process performance transfer topology is reconstructed based on the P sets of collaborative control nodes and multiple independent control nodes to output the inter-process performance coupling topology, wherein the coupling nodes have performance control intensity weights.

[0102] Furthermore, the performance optimization target aggregation module 13 is also used to perform the following steps:

[0103] The process involves interactively obtaining multiple sets of support optimization performance and multiple sets of support optimization algorithms for various sample production processes; storing these multiple sample production processes, support optimization performance, and support optimization algorithms based on a knowledge graph to obtain a control optimization algorithm library; using the multiple process-level performance optimization objectives and multiple process-level production processes as dual retrieval and matching constraints, retrieving and calling multiple control optimization algorithms from the control optimization algorithm library; and based on the consistency of optimization algorithms, aggregating the multiple process-level performance optimization objectives and multiple process-level production processes according to the multiple control optimization algorithms to obtain the M sets of process-level production processes and M sets of process-level performance optimization objectives.

[0104] Furthermore, the cross-process collaborative parameter tuning and optimization module 14 is also used to perform the following steps:

[0105] Traversing the inter-process performance coupling topology, W linkage optimization node groups of W process-level production processes in the first group of process-level production processes are obtained; based on the coupling strength characteristics of the W linkage optimization node groups, the W process-level production processes are decomposed into multiple cross-process linkage parameter tuning groups and multiple single-process parameter tuning optimization nodes; the cross-process linkage parameter tuning optimization results of the multiple cross-process linkage parameter tuning groups are used as the single-process collaborative parameter tuning optimization boundary conditions of the multiple single-process parameter tuning optimization nodes, and single-process parameter tuning optimization is performed to merge and output the first group of collaborative production control parameters; and so on, through cross-process collaborative parameter tuning optimization, the M groups of collaborative production control parameters are output.

[0106] Furthermore, the cross-process collaborative parameter tuning and optimization module 14 is also used to perform the following steps:

[0107] The first set of process-level performance optimization objectives is decomposed into multiple cross-process performance optimization objective subsets and multiple single-process performance optimization objectives. Based on the multiple cross-process performance optimization objective subsets, a first type of control optimization algorithm is run to perform process parameter tuning optimization on the multiple cross-process linkage parameter tuning groups, and the cross-process linkage parameter tuning optimization results are output. Based on the weight coefficients of the inter-process performance coupling topology, the cross-process linkage parameter tuning optimization results are transformed into the single-process collaborative parameter tuning optimization boundary conditions. Under the constraints of the single-process collaborative parameter tuning optimization boundary conditions, based on the multiple single-process performance optimization objectives, a first type of control optimization algorithm is run to perform process parameter tuning optimization on the multiple single-process parameter tuning optimization nodes, and the single-process parameter tuning optimization results are output. The cross-process linkage parameter tuning optimization results and the single-process parameter tuning optimization results are merged to generate the first set of collaborative production control parameters.

[0108] Furthermore, the intelligent collaborative production module 15 is also used to perform the following steps:

[0109] Based on the component manufacturing process sequence, the M sets of collaborative production control parameters are decomposed into collaborative production control sequences. Using these collaborative production control sequences, the entire intelligent collaborative production process of the silicon steel sheet component for the generator is executed, while simultaneously collecting multi-modal production monitoring sequences. Parameter deviation analysis is performed on the collaborative production control sequences and the multi-modal production monitoring sequences to locate the first production risk node. Temperature gradient calculation is performed on the multi-modal production monitoring sequences to locate the second production risk node. The first and second production risk nodes are linked to perform production risk event backtracking to determine real-time risk compensation strategies.

[0110] Furthermore, the device also includes a policy update module for performing the following steps:

[0111] After implementing online process compensation using the aforementioned real-time risk compensation strategy, a multimodal production tracking sequence is collected within a preset tracking time window. If the analysis results of the multimodal production tracking sequence still include risk nodes, the compensation strategy is iteratively updated until the analysis results of the updated multimodal production tracking sequence are empty.

[0112] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0113] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0114] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for intelligent control optimization in the production of silicon steel sheets for generators, characterized in that, The method includes: Interactively obtain the design requirements parameters and component manufacturing process sequence for silicon steel sheet components; Based on the component manufacturing process sequence, the design requirement parameters are decomposed into process chain performance coupling, and the design requirement parameters are mapped to multiple process-level performance optimization targets corresponding to multiple process-level manufacturing processes in the component manufacturing process sequence. Based on the multiple process-level performance optimization objectives and multiple process-level production processes, after matching the control optimization algorithms, and based on the consistency of the optimization algorithms, aggregate and output the M sets of process-level performance optimization objectives corresponding to the M sets of process-level production processes; With the performance optimization objectives of the M groups of process-level processes as constraints, cross-process collaborative parameter tuning optimization is performed on the M groups of process-level production processes, and the collaborative production control parameters of the M groups are output. The entire process of intelligent collaborative production of the silicon steel sheet components for generators is executed using the aforementioned M sets of collaborative production control parameters. The method involves performing process chain performance coupling decomposition on the design requirement parameters based on the component manufacturing process sequence, mapping the design requirement parameters to multiple process-level performance optimization objectives corresponding to multiple process-level manufacturing processes in the component manufacturing process sequence. The method includes: Based on the performance regulation effects of multiple silicon steel sheets in the aforementioned multi-process production process, an inter-process performance transfer topology is constructed. By quantifying the coupling strength between process nodes in the inter-process performance transfer topology, the inter-process performance coupling topology is output. Perform gradient sensitivity analysis between adjacent downstream nodes on the aforementioned inter-process performance coupling topology to locate the dominant decomposition path of the process chain; The design requirement parameters are back-mapped along the dominant decomposition path of the process chain using a chain gradient to output multiple process-level performance boundaries, which serve as the performance optimization targets for the multiple process levels.

2. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 1, characterized in that, Based on the performance regulation effects of multiple silicon steel sheets in the multiple process-level production processes, an inter-process performance transfer topology is constructed, the method comprising: By backtracking local production test data, multiple sets of performance control quantification values ​​for the various process-level production processes are obtained; The performance regulation types are aggregated from the multiple sets of performance regulation quantification values ​​to obtain P types of performance regulation; Using the component manufacturing process sequence as a directed connection constraint, the multiple process-level manufacturing processes are taken as multiple topological nodes. The performance transmission path is fitted by traversing the multiple sets of performance control quantification values ​​using the P performance control types, and P one-dimensional performance transmission topologies are output. The P single-dimensional performance transfer topologies are spatially overlapped to obtain the inter-process performance transfer topology.

3. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 2, characterized in that, The method includes quantifying the coupling strength between process nodes in the inter-process performance transfer topology and outputting the inter-process performance coupling topology. The performance control weights of the multiple sets of performance control quantification values ​​are calculated using the P types of performance control, and the intensity coefficients of the multiple sets of performance control are located. Based on the P types of performance regulation, the multiple sets of performance regulation intensity coefficients are subjected to cross-node serialization processing to obtain P performance regulation intensity sequences. Preset cross-stage control intensity threshold; The P performance regulation intensity sequences are traversed using the cross-stage regulation intensity threshold to locate the P groups of collaborative regulation nodes; The multiple sets of performance regulation intensity coefficients are sorted within each group to locate multiple independent regulation nodes; The inter-process performance transfer topology is reconstructed based on the P-group of collaborative control nodes and multiple independent control nodes, and the inter-process performance coupling topology is output, wherein the coupling nodes have performance control intensity weights.

4. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 1, characterized in that, Based on the multiple process-level performance optimization objectives and multiple process-level production processes, after matching control optimization algorithms, and based on the consistency of optimization algorithms, aggregate and output M sets of process-level performance optimization objectives corresponding to M sets of process-level production processes. The method includes: Interactively obtain multiple sets of support optimization performance and multiple sets of support optimization algorithms for multiple sample production processes; Based on the knowledge graph, the production processes of the multiple samples, the performance support for optimization, and the algorithms support for optimization are stored together to obtain a control optimization algorithm library. Using the multiple process-level performance optimization objectives and multiple process-level production processes as dual search and matching constraints, multiple control optimization algorithms are retrieved and invoked from the control optimization algorithm library. Based on the consistency of the optimization algorithm, the multiple process-level performance optimization objectives and multiple process-level production processes are aggregated according to the multiple control optimization algorithms to obtain the M sets of process-level production processes and the M sets of process-level performance optimization objectives.

5. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 4, characterized in that, Using the performance optimization objectives of the M groups of process-level processes as constraints, cross-process collaborative parameter tuning optimization is performed on the M groups of process-level production processes, and the collaborative production control parameters of the M groups are output. The method includes: By traversing the inter-process performance coupling topology, we obtain W linkage optimization node groups for W process-level production processes in the first group of process-level production processes; Based on the coupling strength characteristics of the W linkage optimization node groups, the W process-level production processes are decomposed into multiple cross-process linkage parameter tuning groups and multiple single-process parameter tuning optimization nodes; The cross-process linkage parameter tuning optimization results of the multiple cross-process linkage parameter tuning groups are used as the single-process collaborative parameter tuning optimization boundary conditions of the multiple single-process parameter tuning optimization nodes. Single-process parameter tuning optimization is performed, and the first group of collaborative production control parameters is output. Similarly, through cross-process collaborative parameter tuning and optimization, the M sets of collaborative production control parameters are output.

6. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 5, characterized in that, The method includes using the cross-process linkage parameter tuning optimization results of the multiple cross-process linkage parameter tuning groups as the single-process collaborative parameter tuning optimization boundary conditions of the multiple single-process parameter tuning optimization nodes, performing single-process parameter tuning optimization, and merging and outputting the first group of collaborative production control parameters. The first set of process-level performance optimization objectives is decomposed into multiple cross-process performance optimization objective subsets and multiple single-process performance optimization objectives; Based on the multiple cross-process performance optimization target subsets, the first type of control optimization algorithm is run to perform process parameter tuning optimization on the multiple cross-process linkage parameter tuning groups, and the cross-process linkage parameter tuning optimization results are output. Based on the weighting coefficients of the inter-process performance coupling topology, the cross-process linkage parameter tuning optimization results are transformed into the single-process collaborative parameter tuning optimization boundary conditions. Under the boundary condition constraints of the single-process collaborative parameter tuning optimization, and based on the performance optimization objectives of the multiple single processes, the first type of control optimization algorithm is run to perform process parameter tuning optimization on the multiple single-process parameter tuning optimization nodes, and the single-process parameter tuning optimization results are output. The cross-process linkage parameter adjustment optimization results and the single-process parameter adjustment optimization results are combined to generate the first set of collaborative production control parameters.

7. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 1, characterized in that, Using the aforementioned M sets of collaborative production control parameters, the entire process of intelligent collaborative production of the silicon steel sheet components for generators is executed, the method comprising: Based on the component manufacturing process sequence, the M sets of collaborative production control parameters are decomposed into collaborative production control sequences; Using the aforementioned collaborative production control sequence, during the entire intelligent collaborative production process of the silicon steel sheet components for generators, multi-modal production monitoring sequences are simultaneously collected. Parameter deviation analysis is performed on the collaborative production control sequence and the multimodal production monitoring sequence to locate the first production risk node; Perform temperature gradient calculation on the multimodal production monitoring sequence to locate the second production risk node; By linking the first and second production risk nodes, production risk events can be traced back to pinpoint real-time risk compensation strategies.

8. The intelligent control optimization method for producing silicon steel sheets for generators as described in claim 7, characterized in that, The method further includes: After implementing online process compensation using the aforementioned real-time risk compensation strategy, a multimodal production tracking sequence is collected within a preset tracking time window; If the analysis results of the multimodal production tracking sequence still include risk nodes, the compensation strategy is iteratively updated until the analysis results of the updated multimodal production tracking sequence are empty.

9. An intelligent control optimization device manufactured from silicon steel sheets for generators, characterized in that, The apparatus is used to implement the intelligent control optimization method for producing silicon steel sheets for generators as described in any one of claims 1-8, and the apparatus comprises: A basic parameter acquisition module is used to interactively obtain the design requirement parameters and component manufacturing process sequence of silicon steel sheet components. A process chain performance decomposition module is used to perform process chain performance coupling decomposition on the design requirement parameters based on the component manufacturing process sequence, and map the design requirement parameters to multiple process-level performance optimization targets corresponding to multiple process-level manufacturing processes in the component manufacturing process sequence. The performance optimization target aggregation module is used to perform control optimization algorithm matching based on the multiple process-level performance optimization targets and multiple process-level production processes, and then aggregate and output M sets of process-level performance optimization targets corresponding to M sets of process-level production processes based on the consistency of the optimization algorithms. A cross-process collaborative parameter tuning and optimization module is used to perform cross-process collaborative parameter tuning and optimization on the M groups of process-level production processes with the M groups of process-level performance optimization objectives as constraints, and output M groups of collaborative production control parameters. The intelligent collaborative production module is used to execute the entire process of intelligent collaborative production of the silicon steel sheet components for generators by adopting the M sets of collaborative production control parameters.

Images