Search

CN-121995882-A - Production line debugging method based on digital twin and provider hierarchical modularization

CN121995882ACN 121995882 ACN121995882 ACN 121995882ACN-121995882-A

Abstract

The invention discloses a production line debugging method based on digital twin and provider hierarchical modularization, which relates to the field of production line debugging and comprises the following steps of S1, acquiring multi-source data, presetting a database and synchronously finishing double verification; S2, three-dimensional classification labeling, forming and storing a data set, S3, extracting core parameters to obtain a twin basic model, S4, constructing a virtual-real cooperation debugging model by combining a double algorithm, S5, setting a debugging optimization rule to generate a layered modularized virtual debugging scheme, S6, generating a personalized layered scheme, constructing a coupling feedback correction model based on a weighted iterative algorithm, S7, inputting the personalized scheme into the correction model for verification and adjustment, and generating a six-dimensional debugging verification report. According to the invention, through acquiring the whole flow data of the production line, the accurate and efficient collaborative debugging of the multi-supplier multi-module production line is realized, and the design of no layering adaptation of the twin model in the debugging method is avoided.

Inventors

  • JIANG ZHICHAO
  • YE BO
  • GAO ZHIHONG

Assignees

  • 武汉工学智联科技有限公司

Dates

Publication Date
20260508
Application Date
20260129

Claims (10)

  1. 1. The production line debugging method based on digital twin and provider hierarchical modularization is characterized by comprising the following steps of: S1, acquiring production line whole-flow data, provider technical data, production line real-time data and production line debugging historical data, presetting provider layering rules, module interface compatible constraint rules and a multi-mode debugging database, and carrying out double verification on the production line whole-flow data and the provider technical data based on the provider layering rules and the module interface compatible constraint rules; s2, carrying out three-dimensional classification labeling on the production line full-flow data and the supplier technical data after double verification according to supplier levels, module function types and debugging priorities in the supplier technical data to form a production line supplier parameter labeling data set, and carrying out associated structural storage with a multi-mode debugging database; S3, carrying out standardized pretreatment on the production line provider parameter labeling data set, extracting core association parameters by adopting a module association parameter extraction algorithm, presetting a digital twin model level self-adaptive generation rule, constructing a four-dimensional dynamic twin mapping model, and inputting the core association parameters into the four-dimensional dynamic twin mapping model to obtain a provider layered modularized digital twin basic model; S4, presetting a virtual-real synchronous mapping strategy and a production line debugging index system, setting real-time corresponding relations and synchronous weights of module parameters and production line real-time data in a provider layered modular digital twin basic model through the virtual-real synchronous mapping strategy, and constructing a production line virtual-real collaborative debugging model by combining the real-time corresponding relations and the synchronous weights with a self-adaptive particle filtering algorithm and a sliding window smoothing algorithm; S5, setting a debugging dynamic optimization rule, introducing a fault characteristic map, a fuzzy logic reasoning and a particle swarm optimization algorithm, defining a multi-objective optimization function, and carrying out fusion iterative optimization on the fault characteristic map, the fuzzy logic reasoning and the particle swarm optimization algorithm to obtain debugging reference parameters, a fault tolerance interval and a debugging effect prediction curve, so as to generate a provider layered modular virtual debugging scheme; s6, extracting a historical candidate debugging parameter set from the multi-mode debugging database, carrying out combined optimization on the provider layered modular virtual debugging scheme and the historical candidate debugging parameter set to generate a personalized layered debugging scheme, and constructing a debugging coupling feedback correction model based on a weighted iterative algorithm; and S7, inputting the personalized hierarchical debugging scheme into a debugging coupling feedback correction model to verify and adjust the provider hierarchical modular virtual debugging scheme, and generating a six-dimensional debugging verification report.
  2. 2. The production line debugging method based on digital twin and provider hierarchical modularization of claim 1, wherein the standardized pretreatment is performed on a production line provider parameter labeling data set, a module association parameter extraction algorithm is adopted to extract core association parameters, a digital twin model level self-adaptive generation rule is preset, a four-dimensional dynamic twin mapping model is constructed, the core association parameters are input into the four-dimensional dynamic twin mapping model, and the provider hierarchical modularization digital twin basic model is obtained, and the method comprises the following steps: S31, carrying out data format unification, missing value completion, outlier rejection and dimension normalization operation on the production line provider parameter labeling data set, and dividing the data set according to provider levels and module function types to form a standardized parameter data set; S32, calculating the inter-parameter association strength in the standardized parameter data set by adopting a module association parameter extraction algorithm and combining the mutual information entropy and the Pearson correlation coefficient, and screening parameters related to the module function type and the debugging priority to obtain a core association parameter set; S33, integrating a corresponding rule of a preset model level and a provider level, a mapping rule of a module function and a virtual unit, a parameter dynamic update triggering rule and a model precision calibration rule to form a digital twin model level self-adaptive generation rule, and constructing a four-dimensional dynamic twin mapping model comprising a provider level-module function-physical parameter-time sequence based on the digital twin model level self-adaptive generation rule; S34, inputting the core association parameter set according to the dimension classification of the four-dimensional dynamic twin mapping model, matching the corresponding virtual module unit through the digital twin model level self-adaptive generation rule, completing parameter initialization configuration and virtual-real mapping calibration, and generating the provider layered modular digital twin basic model.
  3. 3. The production line debugging method based on digital twinning and provider layering modularization of claim 1, wherein the preset virtual-real synchronous mapping strategy and production line debugging index system is characterized in that the real-time corresponding relation and synchronous weight of each module parameter in the provider layering modularized digital twinning basic model and the real-time data of the production line are set through the virtual-real synchronous mapping strategy, and the real-time corresponding relation and synchronous weight are combined with the adaptive particle filtering algorithm and the sliding window smoothing algorithm to construct the virtual-real collaborative debugging model of the production line, and the method comprises the following steps: S41, presetting a virtual-real synchronous mapping strategy and a production line debugging index system, wherein the virtual-real synchronous mapping strategy comprises parameter mapping rules, a data transmission protocol and state synchronous triggering conditions, and the production line debugging index system covers parameter standard reaching rate, response speed, stability and compatibility dimension; S42, based on a virtual-real synchronous mapping strategy, combing the corresponding relation between each module parameter in the provider layered modular digital twin basic model and real-time data of the production line, and quantizing synchronous weight according to the provider level and the functional importance of the modules to generate a parameter corresponding relation table and a synchronous weight matrix; S43, combining the parameter corresponding relation table and the synchronous weight matrix with the real-time corresponding relation and the synchronous weight of the real-time data of the production line, denoising the high-frequency acquisition data through a sliding window smoothing algorithm, and dynamically compensating the data transmission delay and the error by utilizing a self-adaptive particle filtering algorithm to obtain a calibrated parameter data set; S44, setting up a data input layer-mapping processing layer-algorithm optimizing layer-model output layer four-layer framework by taking the calibrated parameter data set as input, and integrating a virtual-real synchronous mapping strategy and a production line debugging index system to generate a production line virtual-real cooperative debugging model.
  4. 4. The production line debugging method based on digital twinning and provider hierarchical modularization according to claim 1, wherein the setting of debugging dynamic optimization rules, the introduction of fault feature patterns, fuzzy logic reasoning and particle swarm optimization algorithms, the definition of multi-objective optimization functions, and the fusion iterative optimization of the fault feature patterns, the fuzzy logic reasoning and the particle swarm optimization algorithms to obtain debugging reference parameters, fault tolerance intervals and debugging effect prediction curves, the generation of provider hierarchical modularization virtual debugging scheme comprises the following steps: S51, setting a debugging dynamic optimization rule comprising a parameter adjustment threshold value, an iteration termination condition, a provider level optimization priority and a module function constraint standard, and defining debugging parameter adjustment authorities and linkage logics of different levels of provider modules to form a standardized optimization rule document; S52, introducing a fault feature map, a fuzzy logic reasoning and a particle swarm optimization algorithm, constructing a three-dimensional fault feature map comprising module-fault-features, defining membership functions and rule bases of the fuzzy logic reasoning, reasoning the three-dimensional fault feature map to obtain a trusted parameter interval, and then configuring inertia weight, learning factors and population scale parameters of the particle swarm optimization algorithm; S53, defining a multi-objective optimization function with debugging efficiency, parameter standard reaching rate, cost control and compatibility as objective dimensions based on a production line debugging index system and a debugging dynamic optimization rule, and taking effective features and a trusted parameter interval of a three-dimensional fault feature map as input constraints of a particle swarm optimization algorithm to perform fusion iterative optimization; S54, the parameter convergence condition is monitored in real time in the iterative optimization process, when the iterative termination condition is met, the debugging reference parameter, the fault tolerance interval and the debugging effect prediction curve are output, and the vendor hierarchical modularization characteristic and the module function requirement are combined to generate the vendor hierarchical modularization virtual debugging scheme.
  5. 5. The production line debugging method based on digital twinning and provider hierarchical modularization according to claim 1, wherein the steps of extracting a historical candidate debugging parameter set from a multi-mode debugging database, performing combined optimization on a provider hierarchical modularization virtual debugging scheme and the historical candidate debugging parameter set to generate a personalized hierarchical debugging scheme, and constructing a debugging coupling feedback correction model based on a weighted iteration algorithm comprise the following steps: S61, extracting a historical debugging parameter set matched with the current production line scene, the supplier level and the module function type from a multi-mode debugging database, and obtaining a historical candidate debugging parameter set through the processes of weight elimination, outlier elimination and parameter format standardization; s62, based on the supplier hierarchy optimization priority and the module function constraint standard, establishing a combined optimizing rule, presetting a matching degree weight distribution rule, and matching the debugging reference parameters in the supplier hierarchy modularized virtual debugging scheme with the historical candidate debugging parameter sets according to the matching degree weight distribution rule; S63, selecting historical candidate debugging parameters in a historical candidate debugging parameter set according to a similarity matching result, and verifying the validity of the provider layered modular virtual debugging scheme according to the historical candidate debugging parameters, and eliminating invalid combinations beyond module function constraint and debugging dynamic optimization rules to generate a personalized layered debugging scheme; S64, defining an iteration weight coefficient and a convergence threshold based on a weighted iteration algorithm, constructing a four-layer architecture comprising a parameter input layer, a weight distribution layer, an iterative calculation layer and a model output layer, integrating verification logic and a feedback mechanism of a personalized hierarchical debugging scheme, and constructing a debugging coupling feedback correction model.
  6. 6. The production line debugging method based on digital twin and provider hierarchical modularization according to claim 2, wherein the method is characterized in that a module association parameter extraction algorithm is adopted, the inter-parameter association strength in a standardized parameter data set is calculated by combining mutual information entropy and pearson correlation coefficient, and parameters related to module function types and debugging priorities are screened, and a core association parameter set is obtained, and the method comprises the following steps: s321, dividing a standardized parameter data set into parameter data subsets according to provider levels and module function types, and defining parameter dimensions and calculation objects in each parameter data subset to form a parameter calculation subset; S322, calculating mutual information entropy values among parameters in each parameter calculation subset based on a module association parameter extraction algorithm, quantifying nonlinear association strength among parameters, presetting a mutual information entropy threshold, and screening a strong association parameter pair with the mutual information entropy value higher than the mutual information entropy threshold; S323, calculating a Pearson correlation coefficient for the strong correlation parameters in the strong correlation parameter pair, quantifying the linear correlation degree among the parameters, presetting a correlation coefficient threshold, and integrating the strong correlation parameters meeting the mutual information entropy threshold and the correlation coefficient threshold simultaneously to generate an initial core candidate set; s324, performing secondary screening on the initial core candidate set according to the module function type and the debugging priority to obtain a core association parameter set.
  7. 7. The production line debugging method based on digital twin and provider hierarchical modularization according to claim 3, wherein the combination of the real-time correspondence of the parameter correspondence table and the synchronous weight matrix with the real-time data of the production line and the synchronous weight, the denoising treatment of the high-frequency acquisition data by the sliding window smoothing algorithm, the dynamic compensation of the data transmission delay and the error by the adaptive particle filtering algorithm, and the obtaining of the calibrated parameter data set comprise the following steps: s431, carrying out association fusion on the parameter corresponding relation table, the synchronous weight matrix, the real-time corresponding relation of the real-time data of the production line and the synchronous weight, unifying the data format and removing redundant fields to form a virtual-real fusion parameter data set of the production line; s432, based on a sliding window smoothing algorithm, presetting window size and sliding step length, carrying out section-by-section smoothing denoising on high-frequency acquisition data in the virtual-real fusion parameter data set of the production line, and filtering random interference and abnormal fluctuation to obtain a denoised parameter data set; s433, presetting particle number and resampling threshold, constructing a self-adaptive particle filtering model, inputting the denoised parameter data set into the self-adaptive particle filtering model, and dynamically estimating and compensating delay deviation and acquisition error to obtain a compensated parameter data set; s434, carrying out integrity check on the compensated parameter data set to obtain a calibrated parameter data set.
  8. 8. The production line debugging method based on digital twinning and provider hierarchical modularization according to claim 4, wherein the fault feature map, the fuzzy logic reasoning and the particle swarm optimization algorithm are introduced to construct a three-dimensional fault feature map comprising module-fault-feature, membership functions and rule bases of the fuzzy logic reasoning are defined, the three-dimensional fault feature map is deduced to obtain a trusted parameter interval, and inertia weight, learning factors and population scale parameters of the particle swarm optimization algorithm are reconfigured, and the method comprises the following steps: S521, module information, typical fault types and fault characteristic parameters corresponding to the levels and the function types of the modules are extracted from a multi-mode debugging database, production line debugging historical data and provider technical data, and a module-fault-characteristic association mapping table is established; s522, constructing a module-fault-feature three-dimensional fault feature map containing provider hierarchy dimensions based on the association mapping table, and determining association weights and search rules of all nodes in the three-dimensional fault feature map to finish initialization and verification of the three-dimensional fault feature map; S523, taking fault characteristic parameters and debugging parameter intervals as input variables and output variables of fuzzy logic reasoning, constructing membership functions adapting to different module function types, establishing a fuzzy reasoning rule base by combining with debugging dynamic optimization rules, matching the fuzzy reasoning rule base with a three-dimensional fault characteristic map, and reasoning the fault characteristic parameters in the three-dimensional fault characteristic map through the matched fuzzy reasoning rule to obtain a trusted parameter interval; S524, combining the production line debugging index system and the supplier level optimization priority, configuring an inertia weight initial value, a learning factor and population scale parameters of the particle swarm optimization algorithm, and defining a dynamic adjustment range of the parameters to form a particle swarm optimization algorithm parameter configuration scheme.
  9. 9. The production line debugging method based on digital twin and provider hierarchical modularization according to claim 5, wherein the steps of establishing a combined optimizing rule based on the provider hierarchical optimizing priority and module function constraint standard, presetting a matching degree weight distribution rule, and matching the debugging reference parameters in the provider hierarchical modularization virtual debugging scheme with the historical candidate debugging parameter sets according to the matching degree weight distribution rule, wherein the steps of: S621, sorting the hierarchical optimization priorities of the providers and module function constraint standards corresponding to each hierarchy to form a hierarchical function constraint basis list; S622, based on the hierarchical functional constraint basis list, establishing a combination optimizing rule, and defining screening conditions of parameter combinations, judging standards of invalid combinations and parameter combination linkage logics of different hierarchical modules to form a standardized combination optimizing rule document; S623, presetting a matching degree weight distribution rule, and generating a matching degree weight distribution table according to the supplier level optimization priority and the module functional importance definition weight dynamic adjustment range; S624, performing feature alignment on debugging reference parameters in the provider hierarchical modular virtual debugging scheme and the historical candidate debugging parameter set by adopting an Euclidean distance algorithm, quantifying matching contribution degrees of each dimension according to weights in a matching degree weight distribution table, calculating comprehensive matching degrees of every two parameters, and outputting a similarity matching result table.
  10. 10. The production line debugging method based on digital twin and provider hierarchical modularization according to claim 7, wherein the preset particle number and resampling threshold value are used for constructing an adaptive particle filtering model, the denoised parameter data set is input into the adaptive particle filtering model, the delay deviation and the acquisition error are dynamically estimated and compensated, and the compensated parameter data set comprises the following steps: S4331, presetting particle number ranges, resampling thresholds and state estimation initial error thresholds corresponding to different hierarchy modules by combining the hierarchy importance of a provider and the module function data acquisition frequency to form a particle filtering parameter configuration table; S4332, constructing an adaptive particle filter model comprising a state initialization module, a particle sampling module, a weight updating module and an adaptive adjustment module based on a particle filter parameter configuration table; S4333, performing time stamp alignment and data format standardization pretreatment on the denoised parameter data set, inputting the parameter data set into a self-adaptive particle filter model in groups according to a provider level and a module function type, completing parameter state initialization through a state initialization module, and generating a corresponding initial particle set through a particle sampling module according to a particle number range in a particle filter parameter configuration table; S4334, calculating a state estimation error value and a particle weight of an initial particle set through a weight updating module based on a particle filtering parameter configuration table, and dynamically estimating and compensating delay deviation and acquisition error through a self-adaptive adjusting module by combining the state estimation error value and the particle weight to obtain a compensated parameter data set.

Description

Production line debugging method based on digital twin and provider hierarchical modularization Technical Field The invention relates to the field of production line debugging, in particular to a production line debugging method based on digital twin and provider hierarchical modularization. Background In the era of intellectualized transformation of manufacturing industry and the deepening of cooperation of global supply chains, the technology is based on the debugging technology of production lines with cooperation of multiple suppliers, is used as a core link for connecting the construction of the production lines, the adaptation of the supply chains and the efficient production, is widely applied to key fields such as automobile whole vehicle assembly production lines, 3C electronic flexible intelligent construction lines, engineering equipment general assembly production lines, industrial part lean production lines and the like, and has the advantages of debugging timeliness, parameter precision and compatibility of multiple supplier modules, and is a core technical bottleneck which is needed to break through in flexible upgrading of manufacturing industry, and is directly related with the production line production period, production stability, operation and maintenance cost control and market capacity competitiveness. The deep fusion of the digital twin technology and the provider layered modular technology shows breakthrough value in the field of multi-provider cooperative production line debugging by means of the high-efficiency cooperative advantage of the virtual-real cooperation of the production line and the hierarchical adaptation of the multi-provider, and the method has important significance in breaking through the limitation of the traditional production line debugging method, improving the debugging efficiency of the multi-provider cooperative production line, improving the production line production efficiency, optimizing the supply chain cooperative adaptation, reducing the operation and maintenance cost of enterprises, and enhancing the market competitiveness by integrating core technical elements such as multi-source data fusion verification, layered parameter labeling modeling, virtual-real synchronous dynamic debugging, multi-algorithm fusion optimization, closed loop feedback correction and the like, and constructing a scheme integrated cooperative system. However, when the existing production line debugging method based on digital twin and provider hierarchical modularization is used, most schemes adopt a universal twin modeling debugging mode, dynamic adaptation capability of provider hierarchy difference and module functional characteristics is lacked, personalized debugging requirements of multi-provider multi-module production lines are difficult to match, single virtual-real mapping of a multi-focus physical production line and a twin model in a debugging process is not formed, multi-dimensional coupling mechanisms of provider hierarchical control, module hierarchical debugging, data whole-course calibration and scheme closed-loop iteration are not formed, so that debugging suitability is poor and parameter deviation is large, meanwhile, the existing method is mostly dependent on a single algorithm to perform data processing and parameter optimization, a closed-loop iteration mechanism of multi-mode historical debugging data support is lacked, association rules of provider hierarchy and debugging priority are not established, and collaborative problems such as multi-provider module interface compatibility conflict and core module debugging priority ambiguity are difficult to cope with. For the problems in the related art, no effective solution has been proposed at present. Disclosure of Invention Aiming at the problems in the related art, the invention provides a production line debugging method based on digital twinning and provider hierarchical modularization, which aims to overcome the technical problems in the prior art. In order to achieve the above purpose, the specific technical scheme adopted by the invention is as follows: A production line debugging method based on digital twin and provider hierarchical modularization comprises the following steps: S1, acquiring production line whole-flow data, provider technical data, production line real-time data and production line debugging historical data, presetting provider layering rules, module interface compatible constraint rules and a multi-mode debugging database, and carrying out double verification on the production line whole-flow data and the provider technical data based on the provider layering rules and the module interface compatible constraint rules; s2, carrying out three-dimensional classification labeling on the production line full-flow data and the supplier technical data after double verification according to supplier levels, module function types and debugging priorities in the supplier technical data t