A method and system for optimizing management of test data for a winding machine

Abstract

The application discloses a kind of optimization management method and system of winding machine test data, belong to computer technology field, by acquiring multi-source real-time data and analyzing generation multimodal collaborative control instruction set, realize multi-parameter dynamic collaborative adjustment, solve the data fragmentation and control lag problem in related technology, with the advantages of improving winding process stability and optimization efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, in particular to a kind of winding machine test data optimization management method and system. BACKGROUND

[0002] Winding machine is the core equipment in the field of electronic components manufacturing, and a large amount of multi-source heterogeneous test data will be generated during operation, such as process parameters, visual inspection data and electrical performance data. Management and utilization of these data are crucial for improving product quality and production efficiency.

[0003] Currently, the management of winding machine data is mostly limited to the monitoring of a single indicator or the optimization of a local process. Various types of data are usually processed by independent subsystems, lacking integration and linkage mechanisms. This fragmented state of data makes it difficult to comprehensively perceive and optimize complex winding process from a system level. When production anomalies occur, it is often difficult to quickly locate the root cause, restricting the improvement of production process transparency and fine control level.

[0004] Therefore, how to provide a systematic data management method to break down data barriers and achieve multi-dimensional, full-factorial collaborative analysis and optimization of the winding process is a research hotspot. SUMMARY

[0005] The embodiments of the present application provide a winding machine test data optimization management method and system, which can realize collaborative analysis of multi-source data, improve the dynamic regulation and control capability of the winding process, and optimize production stability. The technical solutions are as follows:

[0006] On the one hand, a winding machine test data optimization management method is provided, the method comprising:

[0007] In response to a start instruction of a target winding machine, the specification parameters and material parameters of the target coil of the target winding machine are obtained, and multi-source real-time data in the winding production process are collected, including process parameter sequence, audio stream data, vibration sequence and visual size parameters. Based on the specification parameters, material parameters and multi-source real-time data, the winding process state level of the target winding machine and the optimization priority sequence are determined. The winding process state level is used to indicate the stability stage of the current production process, and the optimization priority sequence is used to determine the execution order when multiple parameters are regulated. Based on the winding process state level, the optimization priority sequence and the parameter deviation degree, a multi-modal collaborative regulation instruction set of the target winding machine is generated, which is used to guide the multi-parameter collaborative adjustment of the winding machine actuator. The parameter deviation degree is determined based on the process parameter sequence and the visual size parameters.

[0008] Further, the application also proposes that based on the specification parameters, material parameters and multi-source real-time data, a material adaptation coefficient, a process stability index and a parameter coupling degree matrix of the target winding machine are determined, the material adaptation coefficient is used to represent the matching degree between the current wire physical characteristics and the structure requirements of the target coil, the process stability index is used to quantify the closeness of the winding process dynamic characteristics to the ideal state, and the parameter coupling degree matrix is used to represent the interaction strength and direction relationship between different process parameters; based on the material adaptation coefficient, the process stability index and the parameter coupling degree matrix, a winding process state level and an optimization priority sequence are determined.

[0009] Further, the application also proposes that dynamic error envelope analysis is performed on the process parameter sequence and the visual size parameter to obtain a quality confidence weight and a parameter deviation degree, the quality confidence weight is used to represent the confidence degree of the consistency between the real-time data and the historical high-quality data distribution, and the parameter deviation degree is used to indicate the deviation amplitude and direction of the real-time parameter value relative to the ideal state reference; time-frequency domain joint analysis is performed on the audio stream data and the vibration sequence to extract multi-modal audio print features, the multi-modal audio print features are used to represent the joint frequency spectrum characteristics of mechanical vibration and acoustic signals in the winding process; based on the specification parameters and the material parameters, a material adaptation coefficient is generated; based on the multi-modal audio print features, the quality confidence weight and the parameter deviation degree, a process stability index and a parameter coupling degree matrix are determined.

[0010] Further, the application also proposes that stability contribution degree analysis is performed on the multi-modal audio print features to obtain stability weight coefficients of the audio print features, the stability weight coefficients are used to represent the influence degree of different audio print features on the winding process stability; based on the quality confidence weight and the parameter deviation degree, a comprehensive process deviation index is generated, the comprehensive process deviation index is used to quantify the degree of deviation of the current winding process from the ideal state; based on the stability weight coefficients, the multi-modal audio print features are weighted and fused to generate an audio print stability index, the audio print stability index is used to represent the dynamic stability of the winding process from the perspective of acoustic vibration characteristics; based on the audio print stability index and the comprehensive process deviation index, multi-dimensional feature fusion is performed through a stability mapping model to obtain a process stability index; based on the correlation relationship between the multi-modal audio print features and the parameter deviation degree, a parameter interaction graph is constructed, the parameter interaction graph is used to visually represent the correlation strength between different process parameters and audio print features; graph feature extraction is performed on the parameter interaction graph to obtain a parameter coupling degree matrix, each element in the parameter coupling degree matrix is used to quantitatively represent the coupling relationship strength between the corresponding process parameters.

[0011] Further, the application also proposes that the material adaptation coefficient, the process stability index and the parameter coupling degree matrix are input into a multi-dimensional evaluation space for coordinate mapping to obtain a state coordinate vector of the current winding process in the multi-dimensional evaluation space, the multi-dimensional evaluation space being a standardized feature space constructed based on historical high-quality production process data; based on the positional relationship between the state coordinate vector and the boundaries of each stability level region, the winding process state level is determined through a clustering analysis algorithm, the boundaries of each stability level region being obtained through density clustering analysis on the historical production process data; the parameter coupling degree matrix is subjected to eigenvalue decomposition to obtain a dominant eigenvector, and based on the dominant eigenvector and the material adaptation coefficient, the sensitivity weight of each process parameter to the overall stability of the target winding machine is determined; based on the size order of the sensitivity weight and the coupling strength between parameters represented by the parameter coupling degree matrix, an optimization priority sequence is generated.

[0012] Further, the application also proposes that based on the winding process state level and the optimization priority sequence, a regulation sensitivity factor of each actuator is determined, the regulation sensitivity factor being used to quantify the response gain of each actuator to the control instruction under different process states; based on the regulation sensitivity factor and the parameter deviation degree, a coupling compensation coefficient between multiple parameters is determined, the coupling compensation coefficient being used to represent the dynamic compensation relationship required between parameters during collaborative adjustment; based on the regulation sensitivity factor, the coupling compensation coefficient and the parameter deviation degree, a multi-modal collaborative regulation instruction set is determined.

[0013] Further, the application also proposes that based on the winding process state level, a reference response parameter of each actuator is matched from a pre-set response characteristic library, the reference response parameter including the acceleration response time of the spindle servo unit, the stiffness coefficient of the tension control unit and the positioning accuracy parameter of the wire winder; based on the optimization priority sequence, the reference response parameter is dynamically weighted and corrected to obtain a priority weighting coefficient, the priority weighting coefficient being used to reflect the importance of each actuator of the target winding machine in the current optimization sequence; by fusing the reference response parameter and the priority weighting coefficient, the regulation sensitivity factor of each actuator is determined, the regulation sensitivity factor of a high-priority actuator being enhanced and the regulation sensitivity factor of a low-priority actuator being attenuated.

[0014] Further, the application also proposes that a control system transfer function matrix with the regulation sensitivity factor as the diagonal element is established to construct a multi-variable control system model; based on the direction and amplitude of the parameter deviation degree, a feedforward compensation vector is determined, the feedforward compensation vector being used to predict the coupling disturbance generated in the parameter adjustment process; the inverse matrix of the multi-variable control system model is determined, and based on the inverse matrix and the feedforward compensation vector, a coupling compensation coefficient matrix is determined; the coupling compensation coefficient matrix is subjected to singular value decomposition to obtain the coupling compensation coefficient.

[0015] Further, the application also proposes that, based on the regulation sensitive factor, the parameter deviation is converted into the initial instruction amplitude of each actuator to generate an initial instruction set; based on the coupling compensation coefficient, the initial instruction set is pre-compensated for coupling interference to form a post-compensation instruction set; and the post-compensation instruction set is subjected to amplitude limiting processing with the physical limit constraint of the actuator of the target winding machine to obtain a multi-modal collaborative control instruction set.

[0016] Further, the application also proposes that, the multi-modal collaborative control instruction set is executed, a new round of multi-source real-time data after instruction execution is collected, and a new parameter deviation is re-determined; based on the new parameter deviation and historical multi-round data, an evaluation benchmark for determining the winding process state level is updated, wherein if the new parameter deviation of continuous multiple production cycles is continuously better than the historical level, the evaluation benchmark is adaptively tightened; based on the updated evaluation benchmark and the new parameter deviation, the optimization priority sequence is dynamically corrected, and the process parameters for which the new parameter deviation is not improved have a corresponding reduced optimization priority; and the dynamically corrected optimization priority sequence is used for generating the control instruction of the subsequent production cycle.

[0017] In one aspect, a winding machine test data optimization management device is provided, and the device comprises:

[0018] An acquisition module is configured to, in response to a start instruction of a target winding machine, acquire specification parameters and material parameters of a target coil of the target winding machine, and collect multi-source real-time data in a winding production process, wherein the multi-source real-time data comprises a process parameter sequence, audio stream data, a vibration sequence, and visual size parameters.

[0019] A determination module is configured to determine a winding process state level and an optimization priority sequence of the target winding machine based on the specification parameters, the material parameters, and the multi-source real-time data, wherein the winding process state level is used to indicate a stability stage in which the current production process is located, and the optimization priority sequence is used to determine an execution order in multi-parameter control.

[0020] A generation module is configured to generate a multi-modal collaborative control instruction set of the target winding machine based on the winding process state level, the optimization priority sequence, and a parameter deviation, wherein the multi-modal collaborative control instruction set is used to guide the actuator of the winding machine to perform multi-parameter collaborative adjustment, and the parameter deviation is obtained by analyzing the process parameter sequence and the visual size parameters.

[0021] In one aspect, a computer device is provided, which comprises one or more processors and one or more memories, and at least one computer program is stored in the one or more memories, and the computer program is loaded and executed by the one or more processors to implement the winding machine test data optimization management method.

[0022] In an aspect, a computer readable storage medium is provided, and the computer readable storage medium stores at least one computer program, the computer program is loaded and executed by a processor to implement the optimization management method of the winding machine test data.

[0023] In an aspect, a computer program product or computer program is provided, and the computer program product or computer program includes program code stored in a computer readable storage medium, and a processor of a computer device reads the program code from the computer readable storage medium, and the processor executes the program code to enable the computer device to execute the optimization management method of the winding machine test data.

[0024] As can be seen from the above, the optimization management method and system of the winding machine test data provided in the present application can achieve multi-parameter dynamic collaborative adjustment by acquiring multi-source real-time data and analyzing and generating a multi-modal collaborative control instruction set, solve the data fragmentation and control lag problems in related technologies, and have the advantages of improving the stability and optimization efficiency of the winding process. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0026] Figure 1 is a schematic diagram of an implementation environment of the optimization management method of the winding machine test data provided by the embodiments of the present application;

[0027] Figure 2 is a flowchart of the optimization management method of the winding machine test data provided by the embodiments of the present application;

[0028] Figure 3 is a partial flowchart of the optimization management method of the winding machine test data provided by the embodiments of the present application;

[0029] Figure 4 is a partial flowchart of another optimization management method of the winding machine test data provided by the embodiments of the present application;

[0030] Figure 5 is a structural schematic diagram of an optimization management device of the winding machine test data provided by the embodiments of the present application;

[0031] Figure 6 is a structural schematic diagram of a server provided by the embodiments of the present application. DETAILED DESCRIPTION

[0032] For the purposes of the present application, the technical solutions and advantages, the embodiments of the present application will be described in further detail below with reference to the drawings.

[0033] The terms "first", "second" and the like in the present application are used to distinguish the same items or similar items with substantially the same function and action, and it should be understood that there is no logical or time sequence dependency between "first", "second", "n", and the number and execution order are not limited.

[0034] Artificial intelligence (AI) is to use digital computers or digital computer controlled machine simulation, extension and expansion of human intelligence, perception of the environment, knowledge acquisition and use of knowledge to obtain better results of theory, method, technology and application system. In other words, artificial intelligence is a comprehensive technology of computer science, which tries to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence.

[0035] Machine learning (ML) is a multi-disciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It is a branch of computer science that studies how computers simulate or implement human learning behavior to acquire new knowledge or skills, and reorganize existing knowledge sub-models to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental approach to making computers intelligent, and its applications are widespread in various fields of artificial intelligence.

[0036] Normalization: mapping the number series with different value ranges to the interval (0, 1) for data processing. In some cases, the normalized value can be directly implemented as a probability.

[0037] Winding machine: refers to an automatic industrial equipment that winds a specific specification of wire (such as enameled copper wire) on a specific skeleton or core according to a predetermined path and rule to produce inductance coil, transformer, motor winding and other electronic components. Its core components usually include: spindle system: used to hold and drive the rotation of the coil skeleton or core, and the rotation speed and angle are accurately controlled by the spindle servo unit. Wire arranging system: composed of wire arranging device, used to guide the wire to move orderly on the skeleton, ensuring uniform and smooth winding. Tension control system: composed of tension control unit and sensor, used to maintain and adjust the tension of the wire during winding process, preventing over-tightening or over-loosening. Control system: programmable logic controller (PLC) or industrial computer, used to receive operation instructions and sensor data, and send control signals to each actuator. Detection system: including but not limited to tension sensor, visual sensor, acoustic sensor, etc., used to collect real-time multi-source data such as process parameter sequence, visual size parameter, audio stream data, vibration sequence.

[0038] Target coil: refers to the coil product to be processed in the current winding machine production task, with specific design requirements (specification parameters) and material requirements (material parameters).

[0039] Specification parameters: refer to the design target values of the target coil, including but not limited to coil turns, winding diameter, winding width, wire diameter, etc.

[0040] Material parameters: refer to the physical properties of the wire, including but not limited to material (such as copper, aluminum), yield strength, elastic modulus, surface friction coefficient, etc.

[0041] Process parameter sequence: refers to the set of process parameters recorded in time sequence during winding process, mainly including winding tension, winding speed, wire arranging device position, etc.

[0042] Audio stream data: refers to the original time domain signal data containing the sound characteristics of the winding machine during winding process, continuously collected by acoustic sensors.

[0043] Vibration sequence: refers to the original time domain signal data representing the mechanical vibration state, continuously collected by vibration sensors (such as accelerometers) during winding process.

[0044] Visual size parameter: refers to the size data obtained by measuring the winding coil through visual detection system (such as CCD camera), including but not limited to coil outer diameter, wire arranging spacing, flatness, etc.

[0045] Winding process state level: refers to the discrete level of the stability of the current winding process after analyzing multi-source data, such as "stable", "sub-stable", "critical", "unstable".

[0046] Optimization priority sequence: refers to the list of parameter adjustment order determined according to the degree of influence of each process parameter on the final product quality and the coupling relationship between parameters.

[0047] Stability mapping model: a trained calculation model for mapping the multi-element feature of the winding process into a comprehensive index (i.e. process stability index). The core function of the model is to establish a nonlinear mapping relationship from the multi-dimensional feature space to the one-dimensional stability scale.

[0048] Stability phase: a synonymous expression of "winding process state level", refers to the dynamic stability state category of the winding process.

[0049] Multi-parameter control: refers to the operation mode of simultaneously issuing control instructions to multiple actuators of the winding machine (such as spindle servo, tension controller, wire dispenser) for coordinated adjustment.

[0050] Parameter deviation degree: refers to the quantitative representation of the deviation amplitude of the real-time process parameter value from its preset target value or ideal reference value.

[0051] Multi-modal coordinated control instruction set: refers to a set of control instructions for multiple actuators that are coordinated in timing and amplitude to achieve the overall optimization goal.

[0052] Material adaptation coefficient: refers to a quantitative value representing the matching degree between the physical characteristics of the current wire and the specification parameters of the target coil and the process requirements.

[0053] Process stability index: refers to a comprehensive quantitative index representing the closeness of the winding process dynamic characteristics to an ideal stable state.

[0054] Parameter coupling degree matrix: refers to a mathematical matrix whose matrix elements quantitatively represent the strength and direction of mutual influence between any two process parameters in the system.

[0055] Dynamic error envelope analysis: refers to an analysis method that establishes a reasonable range boundary that changes dynamically with time / process for key parameters in different production stages, and analyzes the deviation of real-time data from the boundary.

[0056] Quality confidence weight: refers to a weight value calculated based on the probability of real-time data falling within the "dynamic error envelope" constructed by historical high-quality data, used to evaluate the reliability of current data quality.

[0057] Time-frequency domain joint analysis: refers to the simultaneous analysis of the time domain characteristics (such as amplitude, time point) and frequency domain characteristics (such as frequency, spectral energy) of the signal to extract more rich information signal processing method.

[0058] Stability contribution analysis: refers to an analysis method for evaluating the influence weight of each specific feature (such as fundamental frequency stability, harmonic distortion rate) in the multi-modal audio feature on the overall process stability.

[0059] Ideal state: refers to the benchmark state of the best operating interval of each parameter based on the production data of a large number of historical high-quality coils.

[0060] Parameter interaction graph: refers to a graph model whose nodes represent different process parameters or features, and edges represent the interaction between parameters, used to visualize and represent the complex association network between parameters.

[0061] Multi-dimensional evaluation space: refers to an abstract mathematical space formed by multiple key features (such as stability index, material adaptation coefficient, etc.) as coordinate axes, used for positioning and evaluating the winding process state.

[0062] State coordinate vector: refers to a spatial coordinate point representing the state of a specific winding process instance after mapping its feature values to the "multi-dimensional evaluation space".

[0063] Stability level region boundary: refers to the boundary between regions corresponding to different "winding process state levels" in the "multi-dimensional evaluation space" divided by clustering algorithm.

[0064] Density clustering analysis: refers to a clustering algorithm based on the spatial distribution density of data points, which can find clusters of arbitrary shape and identify low-density regions as boundaries.

[0065] Eigenvalue decomposition: refers to a linear algebra method that decomposes a square matrix into its eigenvectors and eigenvalues to reveal the principal components and structural characteristics of the matrix.

[0066] Dominant eigenvector: refers to the eigenvector corresponding to the largest eigenvalue after "eigenvalue decomposition", which represents the most important action mode or direction in the system described by the matrix.

[0067] Parameter coupling strength: is the numerical value represented by the matrix elements in the "parameter coupling degree matrix", which quantitatively describes the strength of the mutual influence between two parameters.

[0068] Regulation sensitive factor: refers to a quantitative value that represents the response gain or sensitivity of each actuator to control instructions under different "winding process state levels".

[0069] Dynamic compensation relationship: refers to the real-time and changing compensation relationship established in the coordinated control to offset the impact of the adjustment of one parameter on another parameter.

[0070] Response characteristic library: refers to a preset database that stores the dynamic response characteristic parameters of each actuator (such as spindle servo and tension controller) of the winding machine in different models or states.

[0071] Actuator: refers to the component of the winding machine that receives control instructions and performs specific actions, mainly including spindle servo unit, tension control unit, wire arranging device, etc.

[0072] Spindle servo unit: refers to the driving system that controls the rotation speed and position of the winding machine spindle.

[0073] Acceleration response time: refers to the time required for the spindle servo unit to reach the target acceleration after receiving a speed change instruction, which represents its dynamic response speed.

[0074] Tension control unit: refers to the device used to maintain and adjust the wire tension during the winding process.

[0075] Stiffness coefficient: in tension control, refers to the ability of the system to resist tension disturbance, the higher the stiffness, the more stable the tension.

[0076] Wire arranging device: refers to the device that guides the wire to arrange on the coil framework according to a specific rule.

[0077] Positioning accuracy parameter: refers to the allowable range of deviation between the actual stop position of the wire arranging device and the required position of the instruction.

[0078] Control system transfer function matrix: refers to the matrix that describes the dynamic relationship between all inputs (control instructions) and all outputs (controlled parameters) in a multivariable control system, and its elements are transfer functions.

[0079] Multivariable control system model: refers to a mathematical model used to describe and analyze a system with multiple inputs and multiple outputs, and the variables have a coupling relationship.

[0080] Feedforward compensation vector: refers to a set of compensation instructions calculated in advance and added in the control system to offset the known disturbance (such as the coupling disturbance caused by adjusting parameter A on parameter B).

[0081] Coupling compensation coefficient matrix: refers to a mathematical matrix whose elements specifically define the size of the feedforward compensation required to offset the coupling effect between parameters.

[0082] Singular value decomposition: refers to a linear algebra method that can decompose any matrix into the product of three matrices, which is commonly used to extract the main mode of the matrix, reduce dimension, or solve the generalized inverse.

[0083] Initial instruction amplitude: refers to the basic adjustment amount of each actuator calculated initially according to the parameter deviation without considering the coupling effect before the cooperative control instruction is generated.

[0084] Coupling interference pre-compensation: refers to the process of calculating the additional instruction required to offset the coupling interference in advance based on the coupling compensation coefficient and superimposing it on the initial instruction amplitude before the control instruction is issued.

[0085] Physical limit constraint: refers to the inherent performance boundary of each actuator of the winding machine, such as maximum speed, maximum tension, maximum stroke, etc., and the control instruction must be within this boundary.

[0086] Amplitude limiting processing: refers to the operation of comparing the calculated control instruction with the physical limit constraint and limiting the instruction value that exceeds the limit within the allowed range.

[0087] Evaluation benchmark: refers to the reference standard for judging the winding process state level or the parameter deviation, which can be set based on historical high-quality data and can be adjusted adaptively according to the improvement of system performance.

[0088] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data for analysis, stored data, displayed data, etc.) and signals involved in the present application are authorized by the user or fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant national and regional laws, regulations and standards.

[0089] Figure 1 is an implementation environment schematic diagram of a winding machine test data optimization management method provided by an embodiment of the present application, referring to Figure 1 The implementation environment can include a winding machine controller 110 and a server 140.

[0090] The winding machine controller 110 is connected to the server 140 through a wireless network or a wired network. The winding machine controller 110 is installed and runs an application program supporting the optimization management of winding machine test data.

[0091] The server 140 is a stand-alone physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The server 140 can provide background services for the application program running on the winding machine controller 110. Since the technical solution provided by the embodiment of the present application involves a large number of operations, in the case that the computing power of the winding machine controller 110 cannot meet the requirements, the technical solution provided by the embodiment of the present application can be executed by the server 140 by sending data to the server 140, and the embodiment of the present application does not limit this.

[0092] In the related art, the winding machine is the core equipment for manufacturing electronic components, and the multi-source heterogeneous data generated during the operation of the winding machine is in a fragmented state for a long time. The traditional management method can only realize the monitoring of a single index or the local process optimization, and various data are processed by independent subsystems, lacking a cross-modal integration mechanism. When tension fluctuation or coil size deviation occurs in the production process, it is difficult for the operator to quickly locate the root cause of the abnormality due to the inability to synchronously analyze the mechanical motion parameters and acoustic vibration characteristics, resulting in a high rate of defective products. For example, during the winding process of a certain type of high-frequency transformer, the misadjustment of the spindle speed and tension control can cause short circuit between coil layers, but the existing system cannot real-time identify the correlation between the changes in acoustic features and the abnormality of electrical parameters.

[0093] In order to solve the above problems, the inventors found that the data island effect is the core bottleneck restricting the optimization of the winding process. By analyzing the operation mechanism of the winding machine, it is realized that the coil forming quality is affected by the coupling of material properties, mechanical motion, vibration noise and other factors. The traditional single-dimensional monitoring cannot reflect the dynamic interaction between parameters, and a comprehensive analysis system covering physical properties, process parameters and multi-modal sensing data needs to be established. Based on this, the specification parameters are taken as the structure reference, the acoustic vibration data and visual size parameters collected in real time are combined, and a multi-dimensional state evaluation model is constructed. By quantifying the parameter deviation and process stability, a multi-actuator cooperative control strategy is designed, forming a closed-loop optimization mechanism from data perception to dynamic adjustment.

[0094] Therefore, the present application provides an optimization management method for winding machine test data, as shown in Figure 2 Taking the winding machine controller as an example, the method includes the following steps.

[0095] 201、in response to the start instruction of the target winding machine, obtaining the specification parameters and material parameters of the target coil of the target winding machine, and collecting multi-source real-time data in the winding production process, the multi-source real-time data including process parameter sequence, audio stream data, vibration sequence and visual size parameter;

[0096] 202、based on the specification parameters, material parameters and multi-source real-time data, determining the winding process state level and optimization priority sequence of the target winding machine, the winding process state level being used to indicate the stability stage of the current production process, and the optimization priority sequence being used to determine the execution order in multi-parameter regulation;

[0097] 203、based on the winding process state level, optimization priority sequence and parameter deviation degree, generating a multi-modal collaborative regulation instruction set of the target winding machine, the multi-modal collaborative regulation instruction set being used to guide the multi-parameter collaborative adjustment of the winding machine actuator, and the parameter deviation degree being determined based on the process parameter sequence and visual size parameter.

[0098] Wherein, the target coil is the winding object of the target winding machine. The specification parameters refer to the geometric dimensions and electrical characteristics specified in the coil design drawing, which can be realized by wire diameter tolerance, interlayer insulation thickness and turn number requirement, and provide structural reference for process parameter setting. The material parameters refer to the physical property data of the winding wire, which can be realized by conductivity, ductility coefficient and surface roughness, and are used to evaluate the adaptability of materials and process equipment. The multi-source real-time data refer to the heterogeneous sensing data collected synchronously in the production process, which can be realized by tension sensor reading, microphone array audio signal, accelerometer vibration waveform and size data of visual measurement system, covering mechanical movement, acoustic vibration and visual detection multidimensional information. The winding process state level refers to the stage division of production stability, which can be realized by clustering algorithm for classification of multi-dimensional feature vector, and is used to identify whether the current is in steady state operation or abnormal transition state. The optimization priority sequence refers to the sequence of parameter adjustment, which can be realized by sensitivity weight ordering and coupling strength analysis, to avoid the chain interference caused by single parameter adjustment. The parameter deviation degree refers to the difference quantization index of real-time parameter value and ideal reference, which can be realized by dynamic error envelope analysis and statistical distribution comparison, to provide deviation correction amount for regulation instruction generation. The multi-modal collaborative regulation instruction set refers to the linkage control strategy of multiple actuators, which can be realized by feedforward compensation mechanism and physical constraint limiting processing, to ensure the space-time consistency of the actions of each mechanism.

[0099] Specifically, when the winding machine starts, first load the target coil wire diameter specification and copper material elongation coefficient, establish a physical reference for subsequent data analysis. Through the synchronous acquisition of distributed sensors, process parameters such as spindle speed and tension value are obtained, and at the same time, the acoustic signals of the microphone array and the size measurement data of the industrial camera are obtained. Match the material parameters with the real-time tension data for matching degree analysis to identify whether the wire tensile strength meets the current process requirements. Perform Mel spectrum transformation on the audio stream data to extract the voiceprint features representing the mechanical friction state. By comparing the dynamic error between the visual size parameters and the specification parameters, the deviation of the coil layer spacing is calculated. Based on the fusion analysis of the above multi-source data, it is determined whether the current production process is in a stable running period or a parameter fluctuation period. When a high-frequency component surge is detected in the vibration sequence, combined with the audio feature, it is identified whether it is caused by abnormal wear of the wire guide rail. According to the coupling strength matrix between parameters, it is determined to adjust the tension parameter instead of directly modifying the spindle speed to avoid wire breakage caused by sudden speed changes. Finally, a collaborative instruction set containing tension compensation and spindle acceleration limit values is generated and synchronized to each execution unit through the control bus.

[0100] Compared with related technologies, the traditional method can only make post-corrections according to visual size deviations and cannot predict potential quality problems caused by abnormal acoustic vibration features. The present scheme can identify early deviations in process parameters before the coil size exceeds the tolerance through time-synchronous acquisition and joint analysis of multi-modal data. For example, when abnormal accumulation of vibration energy in a specific frequency band is detected, combined with the wire fatigue characteristics in the material parameters, the tension relaxation compensation instruction is triggered in advance to avoid wire breakage. In related technologies, the adjustment of spindle speed and wire displacement is often in a serial manner, while the present scheme realizes multi-parameter linkage correction through coupling compensation coefficients, taking mutual interference during the adjustment process into the control model.

[0101] Through the above technical solutions, the present application realizes the deep fusion and dynamic collaborative optimization of multi-source data in the winding process. By establishing a multi-dimensional analysis model covering material characteristics, mechanical motion and acoustic vibration features, early process deviations that cannot be detected by traditional single data sources are identified. The optimization priority sequence generated based on the parameter coupling relationship can guide the execution mechanism to adjust in the optimal order of the system to avoid global performance degradation caused by local optimization. The multi-modal collaborative control instruction set ensures the coordinated action of each execution unit through the space-time synchronization mechanism, improving the accuracy and reliability of complex process parameter adjustment.

[0102] The present application further proposes the following technical solutions, taking the winding machine controller as an example, referring to Figure 3 , comprising the following steps.

[0103] 301、based on the specification parameters, material parameters and multi-source real-time data to determine the material adaptation coefficient, process stability index and parameter coupling degree matrix of the target winding machine;

[0104] 302、determine the winding process state level and the optimization priority sequence based on the material adaptation coefficient, the process stability index and the parameter coupling degree matrix.

[0105] The material adaptation coefficient refers to the matching degree between the current wire physical characteristics and the target coil structure requirement, which can be specifically realized by using a matching degree calculation model of the elastic modulus, ductility of the wire and the size of the coil framework, and a quantitative index is generated by inputting the wire characteristic parameters and the coil specification parameters into a preset matching degree function. The process stability index refers to the closeness of the winding process dynamic characteristics to the ideal state, which can be specifically realized by using a fusion analysis model of multi-modal audio fingerprint features and parameter deviation degree, and the dynamic stability is evaluated by combining the parameter deviation amplitude through joint time-frequency domain analysis to extract the sound and vibration characteristics. The parameter coupling degree matrix refers to the action strength and direction relationship of mutual influence between different process parameters, which can be specifically realized by using a covariance analysis and mutual information calculation model between process parameters, and the non-linear correlation strength between parameters is quantified by constructing a parameter interaction atlas.

[0106] Specifically, the technical scheme first generates quality confidence weight and parameter deviation degree by analyzing the process parameter sequence and visual size parameter through dynamic error envelope, wherein the quality confidence weight reflects the distribution consistency of real-time data and historical high-quality data, and the parameter deviation degree reflects the deviation amplitude of real-time parameters relative to the ideal reference. At the same time, the multi-modal audio fingerprint features are extracted by performing joint time-frequency domain analysis on the audio stream data and vibration sequence, which represent the joint frequency spectrum characteristics of mechanical vibration and acoustic signal. Based on the specification parameters and material parameters, the material adaptation coefficient is calculated by using the matching degree function to quantify the adaptation degree of wire characteristics and coil structure. Further, the influence weight of each audio fingerprint feature on stability is determined by stability contribution degree analysis, the comprehensive process deviation index is generated by combining the quality confidence weight and the parameter deviation degree, and the multi-dimensional feature fusion is performed based on the weighted fusion of the audio fingerprint stability index and the comprehensive process deviation index, and finally the process stability index is obtained. In addition, by analyzing the correlation between multi-modal audio fingerprint features and parameter deviation degree, a parameter interaction atlas is constructed and a parameter coupling degree matrix is extracted, so as to establish a coupling relationship model between process parameters. Finally, the material adaptation coefficient, the process stability index and the parameter coupling degree matrix are input into the multi-dimensional evaluation space for coordinate mapping, and the winding process state level is determined by combining the clustering analysis algorithm, and the optimization priority sequence is generated based on the eigenvalue decomposition result and the sensitivity weight of the parameter coupling degree matrix.

[0107] Compared with the related art, the traditional method relies on a single parameter threshold judgment process state, cannot identify the adaptation relationship between material characteristics and process parameters, and the coupling effect between parameters is only determined by artificial experience. The scheme realizes the quantitative matching of material properties and process requirements by constructing a material adaptation coefficient, improves the dynamic stability evaluation precision by fusing multi-modal audio fingerprint features, and reveals the nonlinear interaction mechanism between parameters by modeling the parameter coupling degree matrix, thereby to a certain extent, solving the problem of difficult identification of process parameter coupling relationship caused by data fragmentation.

[0108] Through the above technical scheme, the material adaptability, dynamic stability and parameter coupling effect in the winding process can be accurately quantified, the objective evaluation of the winding process state is realized based on multi-dimensional index fusion, and the optimization priority sequence is intelligently generated through the parameter coupling degree matrix, thereby to a certain extent, solving the problem of unclear optimization direction in the traditional method, and providing a scientific decision basis for multi-parameter collaborative regulation.

[0109] The application further proposes a method for determining the material adaptation coefficient, process stability index and parameter coupling degree matrix of the target winding machine based on specification parameters, material parameters and multi-source real-time data, including: performing dynamic error envelope analysis on the process parameter sequence and visual size parameters to obtain quality confidence weight and parameter deviation degree; performing time-frequency domain joint analysis on the audio stream data and vibration sequence to extract multi-modal audio fingerprint features; generating a material adaptation coefficient based on specification parameters and material parameters; determining the process stability index and the parameter coupling degree matrix based on the multi-modal audio fingerprint features, the quality confidence weight and the parameter deviation degree.

[0110] The dynamic error envelope analysis refers to the similarity between the real-time data and the historical high-quality data distribution through the sliding window statistics, and the quality confidence weight is calculated, which can be realized by using the kernel density estimation method, and is used to reflect the current data confidence. The parameter deviation degree refers to the standardized distance between the real-time parameter value and the ideal reference value, which can be realized by using the Mahalanobis distance algorithm, and is used to quantify the parameter deviation degree. The multi-modal audio fingerprint feature refers to the joint spectrum feature extracted after time-frequency transformation of the sound and vibration signal, which can be realized by using the wavelet packet decomposition and mel frequency cepstral coefficient fusion method, and is used to represent the mechanical operating state. The material adaptation coefficient refers to the material applicability index calculated according to the matching degree of the wire elastic modulus and the coil structure size, which can be realized by using the linear regression model, and is used to evaluate the compatibility of materials and structures. The process stability index refers to a comprehensive stability quantitative index that fuses sound and vibration features and parameter deviations, which can be realized by using the principal component analysis method, and is used to reflect the overall process stability. The parameter coupling degree matrix refers to a mathematical matrix representing the mutual influence relationship between process parameters, which can be realized by using the mutual information entropy algorithm, and is used to guide the multi-parameter collaborative adjustment.

[0111] Specifically, the dynamic error envelope analysis collects real-time data of the process parameter sequence and visual size parameter through a sliding window, constructs a historical high-quality data distribution model using kernel density estimation, calculates the probability of the current data falling into the distribution as the quality confidence weight, and simultaneously calculates the magnitude and direction of each parameter deviation from the reference value through Mahalanobis distance to generate the parameter deviation degree. The time-frequency domain joint analysis extracts the mel-frequency cepstral coefficient of the audio signal and performs wavelet packet decomposition on the vibration signal, fuses the spectral features of the two to generate multi-modal audioprint features. The material adaptation coefficient calculates the matching degree of the material and structure according to the Young's modulus and yield strength of the wire and the number of turns and layers of the coil through a regression model. The process stability index fuses the audioprint features and the quality confidence weight, and constructs a multi-dimensional evaluation model combined with the parameter deviation degree, and finally generates a comprehensive stability index through principal component analysis. The parameter coupling degree matrix calculates the mutual information entropy of the process parameters and the audioprint features, establishes the interaction relationship map between the parameters, and then converts it into a numerical matrix.

[0112] Compared with related technologies, the traditional method only uses single sensor data for threshold alarm and cannot reflect the dynamic correlation characteristics of multi-source data. The present application realizes dynamic evaluation of data reliability through dynamic error envelope analysis, overcoming the limitations of fixed threshold judgment; through time-frequency domain joint analysis, the joint spectral features of sound and vibration signals are extracted, breaking the one-sidedness of single signal analysis; through the material adaptation coefficient, the quantitative matching relationship between material and structure is established, making up for the shortcomings of traditional empirical material selection; through the fusion of multi-dimensional data, the process stability index and the parameter coupling degree matrix are generated, and a comprehensive evaluation system covering mechanical state, process parameters and material characteristics is constructed.

[0113] Through the above technical solutions, the present application realizes the integration and collaborative analysis of multi-source heterogeneous data, and to some extent solves the problem of non-systematic process optimization caused by data isolation in traditional methods. The dynamic error envelope analysis improves the dynamic adaptability of quality evaluation, the time-frequency domain joint analysis enhances the comprehensiveness of mechanical state monitoring, the material adaptation coefficient provides a material basis for process optimization, and the process stability index and the parameter coupling degree matrix establish a quantitative guidance benchmark for multi-parameter collaborative adjustment, thereby improving the accuracy of winding process state evaluation and the systematization of optimization decision-making.

[0114] The application further proposes a method for determining process stability index and parameter coupling degree matrix based on multi-modal audio print features, quality confidence weight and parameter deviation degree, comprising: performing stability contribution degree analysis on the multi-modal audio print features to obtain stability weight coefficients of each audio print feature; generating a comprehensive process deviation index based on the quality confidence weight and the parameter deviation degree; performing weighted fusion on the multi-modal audio print features based on the stability weight coefficients to generate an audio print stability index; performing multi-dimensional feature fusion through a stability mapping model based on the audio print stability index and the comprehensive process deviation index to obtain a process stability index; constructing a parameter interaction graph based on the correlation between the multi-modal audio print features and the parameter deviation degree; and performing graph feature extraction on the parameter interaction graph to obtain a parameter coupling degree matrix.

[0115] The stability contribution degree analysis refers to determining the influence degree of different sound vibration features on stability through statistical methods, which can be realized by principal component analysis or random forest feature importance evaluation algorithm, and is used for screening key audio print features affecting stability. The comprehensive process deviation index refers to a composite index that fuses data reliability and parameter deviation degree, which can be realized by weighted Euclidean distance or Mahalanobis distance algorithm, and is used to eliminate the influence of single data source error on overall evaluation. The parameter interaction graph refers to a correlation network model between process parameters and sound vibration features, which can be constructed by graph neural network or association rule mining algorithm, and is used to reveal the dynamic coupling relationship between parameters. The graph feature extraction refers to extracting the action strength features between parameters from the network structure, which can be realized by graph convolutional neural network or node centrality index calculation, and is used to convert the topological relationship into a quantifiable coupling degree matrix.

[0116] Specifically, first, the principal component analysis is used to determine the variance contribution rate of each audio print feature in the historical stable state, and the contribution rate is normalized to a stability weight coefficient. Then, the quality confidence weight is taken as a reliability coefficient, and is weighted and summed with the parameter deviation degree to generate a comprehensive process deviation index reflecting the overall deviation degree. Then, the weighted audio print features are nonlinearly fused through a deep neural network to output an index value representing sound vibration stability. The index and the comprehensive process deviation index are jointly input into a pre-trained stability mapping model to obtain a comprehensive process stability index through feature concatenation and full connection layer calculation. At the same time, the mutual information algorithm is used to calculate the correlation strength between the audio print features and the process parameters to construct a parameter interaction graph containing node attributes and edge weights. Finally, the graph attention network is used to extract the potential coupling relationship between nodes in the graph to generate a coupling degree matrix quantitatively describing the mutual influence strength between parameters.

[0117] Compared with the related art, the traditional method usually only analyzes a single type of data, such as separately analyzing vibration spectrum or process parameter curve, resulting in the inability to capture the dynamic coupling relationship between the sound-vibration characteristics and the process parameters. In the related art, parameter coupling analysis is mostly based on empirical formulas or static correlation tables, which are difficult to adapt to the time-varying interaction relationship in the winding process. The present scheme converts the dynamic correlation between sound-vibration characteristics and process parameters into a visual network structure by constructing a parameter interaction map, and realizes adaptive extraction of the coupling relationship by combining a graph neural network, which to some extent solves the problem of correlation modeling between multi-source heterogeneous data.

[0118] Through the above technical scheme, the present application realizes cross-modal correlation analysis of sound-vibration characteristics and process parameters, and can accurately identify key acoustic characteristics affecting stability and their corresponding process parameter combinations. Through the dynamically updated parameter coupling degree matrix, accurate interaction strength data is provided for multi-parameter collaborative regulation, avoiding the regulation conflict problem caused by ignoring parameter coupling in the traditional method. The fusion calculation of the comprehensive process deviation index and the audio fingerprint stability index improves the robustness of process state evaluation and reduces the interference of single sensor data anomaly on overall evaluation.

[0119] The present application further proposes a method for determining the winding process state level and the optimization priority sequence based on the material adaptation coefficient, the process stability index and the parameter coupling degree matrix, including inputting the material adaptation coefficient, the process stability index and the parameter coupling degree matrix into a multi-dimensional evaluation space to obtain a state coordinate vector through coordinate mapping, determining the winding process state level by a clustering analysis algorithm based on the positional relationship between the state coordinate vector and the boundary of the pre-defined stability level area, performing eigenvalue decomposition on the parameter coupling degree matrix to obtain a dominant eigenvector and determining the sensitivity weight of each process parameter based on the dominant eigenvector and the material adaptation coefficient, and generating the optimization priority sequence based on the size order of the sensitivity weight and the coupling strength between parameters represented by the parameter coupling degree matrix.

[0120] The multi-dimensional evaluation space refers to a standardized feature space constructed by historical high-quality production process data. Specifically, principal component analysis can be used to reduce the dimensionality of multi-dimensional parameters and map them into an orthogonal coordinate system to eliminate the incomparability between parameters of different dimensions. The state coordinate vector refers to the coordinate position of the current winding process in the evaluation space after mapping the multi-dimensional parameters. Specifically, a linear transformation algorithm can be used to convert the material adaptation coefficient, process stability index, and parameter coupling degree matrix into a standardized vector to quantitatively represent the comprehensive state of the current production process. The stability level region boundary refers to the boundary region of different stability levels in the evaluation space formed by clustering historical data. Specifically, the DBSCAN density clustering algorithm can be used to divide the historical high-quality data samples into regions to dynamically match the stability stage of the current state. The parameter coupling degree matrix is a matrix that quantitatively represents the mutual influence relationship between process parameters. Specifically, the Pearson correlation coefficient can be used to calculate the linear correlation between parameters to identify the chain reaction during parameter adjustment. The dominant eigenvector is the eigenvector corresponding to the largest eigenvalue of the parameter coupling degree matrix after eigenvalue decomposition. Specifically, the power iteration method can be used to solve the principal feature direction of the matrix to extract the core parameter combination that has the greatest impact on system stability. The sensitivity weight is an index that represents the influence of each process parameter on the overall stability. Specifically, the product of the absolute value of the eigenvector element and the material adaptation coefficient can be calculated to determine the priority order of parameter optimization. The optimization priority sequence refers to the execution order list of process parameter adjustment. Specifically, the sensitivity weight can be sorted in descending order and dynamically adjusted based on the coupling strength between parameters to guide the operation order during multi-parameter collaborative optimization.

[0121] Specifically, the material adaptation coefficient, process stability index, and parameter coupling degree matrix are first standardized and input into the multi-dimensional evaluation space, and are converted into a state coordinate point in a three-dimensional space through coordinate mapping. The evaluation space is constructed based on historical high-quality production data, and the original multi-dimensional parameters are reduced to a three-dimensional orthogonal coordinate system using principal component analysis to eliminate the incomparability between parameters of different dimensions. Then, the historical data samples are clustered using the DBSCAN density clustering algorithm to form the region boundaries of different stability levels. The positional relationship between the current state coordinate point and these regions is analyzed, and if the coordinate point falls into a high-density clustering region, it is determined to be the corresponding stability level. The dominant eigenvector is extracted from the parameter coupling degree matrix through eigenvalue decomposition, which reflects the main direction of parameter coupling. The sensitivity weight of each parameter is calculated by weighting the elements of the eigenvector with the material adaptation coefficient. Finally, the optimization priority sequence is generated according to the weight value size, and the sorting is dynamically adjusted considering the coupling strength between parameters, for example, when two parameters have a strong coupling relationship, the parameter with higher sensitivity weight is adjusted first to avoid chain interference.

[0122] Compared with the related art, the traditional method usually uses a single index threshold to judge the production state, and cannot integrate multi-dimensional data such as material characteristics, process stability and parameter coupling relationship. In the related art, the determination of the optimization sequence depends on artificial experience or fixed rules, and it is difficult to dynamically adapt to different production scenes. The scheme realizes unified quantitative representation of multi-source data by constructing a multi-dimensional evaluation space, uses the dynamic region boundary formed by density clustering to replace the fixed threshold, and improves the adaptability of state evaluation. By extracting the dominant influence parameters through feature decomposition of the parameter coupling degree matrix, and combining the sensitivity weight calculated by the material adaptation coefficient, the generation of the optimization sequence considers both the importance of parameters and the coupling characteristics of the system.

[0123] Through the above technical scheme, the application solves the problems of inaccurate production process stability evaluation and unreasonable optimization sequence generation caused by the fragmentation of multi-source data of the winding machine to some extent. The construction of the multi-dimensional evaluation space realizes unified quantitative analysis of multi-dimensional parameters, and the region boundary formed based on dynamic clustering improves the accuracy of state level determination. The feature decomposition of the parameter coupling degree matrix combined with the weighted calculation of the material adaptation coefficient can accurately identify the core parameters that have the greatest impact on system stability, and the generated optimization priority sequence can guide the coordinated adjustment of multiple parameters, avoid adjustment conflicts caused by strong coupling between parameters, and thus improve the stability and efficiency of the winding process control.

[0124] The application further proposes the following technical scheme, see Figure 4 Taking the winding machine controller as an example, the execution subject, the following steps are included.

[0125] 401. Determine the control sensitive factor of each actuator based on the winding process state level and the optimization priority sequence;

[0126] 402. Determine the coupling compensation coefficient between multiple parameters based on the control sensitive factor and the parameter deviation degree;

[0127] 403. Determine the multi-modal coordinated control instruction set based on the control sensitive factor, the coupling compensation coefficient and the parameter deviation degree.

[0128] The control sensitive factor is a weight coefficient quantifying the response gain of the execution mechanism to the control instruction under different process states, and can be realized by fusing a benchmark response parameter and a priority weighting coefficient. The benchmark response parameter includes an acceleration response time of the main shaft servo unit, a stiffness coefficient of the tension control unit, and a positioning accuracy parameter of the wire winder. The priority weighting coefficient is obtained by dynamically modifying the benchmark parameter through optimization of the priority sequence. The factor is used to dynamically adjust the control weight of different mechanisms in parameter adjustment. The coupling compensation coefficient is a quantitative index representing the dynamic compensation relationship between parameters during collaborative adjustment, and can be obtained by singular value decomposition of the inverse matrix of the transfer function matrix of the multivariable control system combined with the feedforward compensation vector. The coefficient is used to predict and offset the mutual interference generated during parameter adjustment. The multi-modal collaborative control instruction set is a set of operation instructions for guiding the execution mechanism of the winding machine to perform multi-parameter collaborative adjustment, and can be generated by coupling interference pre-compensation of the initial instruction set combined with physical limit constraint amplitude limiting processing, such as simultaneously pre-adjusting the tension parameter when increasing the main shaft speed instruction. The instruction set is used to ensure the overall coordination and device safety of multi-parameter adjustment.

[0129] Specifically, when the winding process is in a high vibration level state, the optimization priority sequence indicates that the tension control is the highest priority. At this time, the control sensitive factor calculation module fuses the benchmark stiffness coefficient of the tension mechanism with the priority weighting coefficient to enhance the control sensitive factor. The parameter deviation analysis shows that the main shaft speed exists a positive deviation. According to the coupling compensation coefficient calculation model, it is predicted that the speed increase will lead to an increase in tension demand, so the generated main shaft acceleration instruction is synchronized with the tension compensation instruction. The initial instruction set is processed by the physical limit constraint module to ensure that the compensated tension instruction does not exceed the maximum torque of the servo motor. Finally, a multi-modal instruction set containing speed-tension collaborative adjustment parameters is generated to drive the execution mechanism to realize collaborative control within a safe range.

[0130] Compared with related technologies, the traditional method usually uses fixed weight coefficients for parameter adjustment, which cannot dynamically adjust the control strategy according to the real-time process state, and lacks modeling and compensation of the dynamic coupling relationship between parameters. For example, the existing tension control system only adjusts feedback according to a preset threshold, and cannot timely predict tension fluctuations when the main shaft speed suddenly changes. The present scheme introduces a control sensitive factor dynamic modification mechanism driven by a process state level, and a coupling compensation coefficient calculation based on a multivariable control model, realizes real-time optimization of the parameter adjustment weight and interference compensation relationship, and to some extent solves the problems of adjustment lag and insufficient precision caused by static control model in traditional methods.

[0131] By the technical solution, the application solves the problems of low adjustment efficiency and insufficient precision caused by parameter coupling effect in the process of multi-parameter collaborative adjustment of the winding machine. The adaptive allocation of control weights in different process states is realized by dynamically regulating the sensitive factor, avoiding excessive adjustment of low-priority parameters. The pre-judgment and offset mechanism of dynamic interference between parameters is established by the coupling compensation coefficient, such as synchronously compensating the tension parameter in the spindle acceleration stage to offset the centrifugal effect, improving the nature and response speed of multi-parameter collaborative adjustment. The physical limit constraint processing ensures that the generated instruction set meets the safe operation boundary of the equipment, improving the control precision while ensuring the stability of the production process.

[0132] The application further proposes a method for determining the regulation sensitive factor of each actuator based on the winding process state level and the optimization priority sequence. Specifically, it includes: matching the baseline response parameters of each actuator from the pre-set response characteristic library based on the winding process state level; dynamically weighting and correcting the baseline response parameters to obtain the priority weighting coefficient based on the optimization priority sequence; determining the regulation sensitive factor of each actuator by fusing the baseline response parameters and the priority weighting coefficient, so that the sensitive factor of high-priority mechanism is enhanced and the sensitive factor of low-priority mechanism is attenuated.

[0133] Among them, the baseline response parameter refers to the inherent dynamic characteristic parameter of the actuator under standard working conditions, which can be realized by using the step response curve fitting parameter of the servo system, for example, the acceleration response time of the spindle servo unit is determined by the first derivative extreme point of the measured speed following curve, and the stiffness coefficient of the tension control unit is obtained by calculating the slope of the spring deformation and tension relationship curve. This parameter provides a physical characteristic benchmark for regulation sensitivity calculation.

[0134] Among them, the priority weighting coefficient refers to a dynamic correction factor reflecting the optimization importance of the actuator, which can be realized by using the exponential weighted moving average algorithm, for example, the optimization priority sequence is converted into a normalized weight vector, and the historical priority influence weight is adjusted by a time decay factor. This coefficient quantifies the process optimization target into a mathematical weight parameter.

[0135] Among them, the regulation sensitive factor refers to a gain adjustment parameter that integrates device characteristics and process requirements, which can be realized by using a parameter fusion formula, for example, the Hadamard product operation is performed on the baseline response parameter and the priority weighting coefficient, and then the nonlinear mapping is performed through the sigmoid function. This factor realizes the collaborative adaptation of device inherent characteristics and dynamic optimization requirements.

[0136] Specifically, when the winding process is in a high-speed precision winding state, the system first calls the corresponding spindle servo acceleration parameters and tension stiffness parameters in the response characteristic library as reference values. Subsequently, according to the sequence of the winding precision priority in the optimization priority sequence, a higher weight coefficient is applied to the winding positioning precision parameter. By performing matrix point multiplication operation on the reference parameters and the weight coefficient, a combination of winding machine gain promotion and tension mechanism response moderate control sensitive factors is generated. This dynamic adjustment mechanism enables the fine tuning sensitivity of the winding mechanism to be automatically enhanced when winding thin wire diameter coils, while the response amplitude of the tension mechanism remains stable, avoiding the conflict of multiple mechanism actions.

[0137] Compared with related technologies, the conventional method using fixed gain coefficient cannot adapt to different process states, which is easy to cause insufficient response of high priority mechanism or excessive regulation of low priority mechanism. The scheme realizes dynamic gain distribution by state level matching and priority weighting dual mechanism, which not only retains the physical characteristics of the equipment but also realizes dynamic gain distribution. For example, the existing tension control adopts constant stiffness coefficient, which causes wire shaking at high speed winding, and the scheme automatically reduces the gain weight of the tension mechanism when the high speed state is detected, which suppresses mechanical resonance.

[0138] Through the above technical scheme, the application realizes dynamic optimization configuration of the control sensitivity of multiple execution mechanisms, and to some extent solves the control oscillation problem caused by the mismatch of priority and response gain in the conventional method. When the winding machine switches different wire diameter specifications, the system automatically adjusts the sensitive factor proportion of the spindle and the winding machine, ensures that the winding precision is given higher control weight when precision winding, and maintains balanced response of each mechanism when regular winding. This adaptive adjustment mechanism improves the cooperative control precision of multiple mechanisms and avoids the efficiency loss caused by manual parameter setting.

[0139] The application further proposes a technical scheme for determining a coupling compensation coefficient between multiple parameters based on the control sensitive factor and the parameter deviation degree, including: establishing a control system transfer function matrix with the control sensitive factor as the diagonal element to construct a multivariable control system model; determining a feedforward compensation vector based on the direction and amplitude of the parameter deviation degree; determining the inverse matrix of the multivariable control system model and generating a coupling compensation coefficient matrix in combination with the feedforward compensation vector; singular value decomposition is performed on the coupling compensation coefficient matrix to obtain the coupling compensation coefficient.

[0140] The control sensitive factor refers to a coefficient quantifying the response gain of the execution mechanism to the control instruction, which can be specifically implemented by dynamically weighting and correcting the reference response parameter, and is used to reflect the dynamic response difference of different mechanisms in collaborative adjustment. The parameter deviation degree refers to the deviation amplitude and direction of the real-time parameter value relative to the ideal state reference, which can be specifically calculated by using a dynamic error envelope analysis method, and is used to represent the deviation characteristics of the current process parameter. The feedforward compensation vector refers to a set of compensation parameters generated in the prediction parameter adjustment process to couple the interference, which can be specifically generated by analyzing the spatio-temporal distribution characteristics of the parameter deviation degree, and is used to predict and offset the cross-parameter interference before the control instruction is executed. The singular value decomposition refers to the mathematical dimension reduction operation on the coupling compensation coefficient matrix, which can be specifically implemented by using a matrix eigenvalue decomposition algorithm, and is used to extract the core compensation relationship corresponding to the key coupling mode.

[0141] Specifically, the scheme first establishes a multivariable control model by constructing a transfer function matrix, and quantifies the dynamic response characteristics of each execution mechanism into a mathematical model. Then, the feedforward compensation vector is generated according to the direction and amplitude of the parameter deviation degree, and the cross-parameter interference possibly caused by parameter adjustment is predicted. Subsequently, the dynamic coupling relationship between parameters is derived by solving the inverse matrix of the control model, and the compensation coefficient matrix is generated in combination with the feedforward compensation vector. Finally, the compensation coefficient matrix is subjected to singular value decomposition, and the compensation coefficient corresponding to the key coupling mode is extracted, thereby avoiding the over-parameterization problem while retaining the core compensation relationship. Through the mathematical modeling and dynamic compensation mechanism, the interference prediction and active offsetting during multi-parameter adjustment are realized.

[0142] Compared with related technologies, the traditional method usually adopts fixed compensation coefficients or single feedback control, and cannot handle the dynamic coupling effect between multiple parameters of the winding machine. In related technologies, parameter adjustment often causes a chain deviation, which needs to be iteratively corrected multiple times, resulting in low control efficiency. The present scheme predicts and compensates the dynamic coupling effect between parameters in the control instruction generation stage through feedforward compensation and matrix decomposition technology, thereby improving the precision and response speed of multivariable collaborative control.

[0143] Through the above technical scheme, the dynamic coupling interference problem during multi-parameter collaborative adjustment of the winding machine is solved to some extent, and the chain deviation caused by single parameter adjustment is avoided. By constructing a multivariable control model and a dynamic compensation mechanism, decoupling control between multiple execution mechanisms is realized, so that key parameters such as tension control and spindle speed can be adjusted synchronously, thereby improving the control precision and system stability of the winding process.

[0144] The application further proposes a method for determining a multi-modal collaborative control instruction set based on a regulation sensitive factor, a coupling compensation coefficient and a parameter deviation degree, comprising: converting the parameter deviation degree into initial instruction amplitudes of each execution mechanism to generate an initial instruction set; pre-compensating the initial instruction set for coupling interference to form a post-compensation instruction set; and performing amplitude limiting processing on the post-compensation instruction set and physical limit constraints of the execution mechanism to obtain the multi-modal collaborative control instruction set.

[0145] The regulation sensitive factor is a coefficient for quantifying the response gain difference of the execution mechanism to the control instruction, which can be specifically realized by weighted calculation of the acceleration response time of the main shaft servo unit and the stiffness coefficient of the tension control unit, and is used to reflect the dynamic adjustment capability of different mechanisms in the current optimization sequence. The coupling compensation coefficient is a value representing the dynamic compensation relationship between parameters, which can be specifically realized by establishing a multi-variable control system model and decomposing its inverse matrix, and is used to predict and offset the mutual interference generated by multi-parameter linkage adjustment. The parameter deviation degree is the deviation of the real-time parameter value from the ideal state reference, which can be specifically realized by dynamic error envelope analysis of the process parameter sequence and the visual size parameter, and is used to quantify the current parameter adjustment requirement. The physical limit constraint is the safety operation range boundary of the execution mechanism, which can be specifically realized by the maximum speed limit of the servo motor or the range threshold of the tension sensor, and is used to prevent equipment damage caused by instruction overrun.

[0146] Specifically, the scheme realizes multi-parameter collaborative control through a phased instruction generation and correction mechanism. First, the parameter deviation degree is converted into the initial instruction amplitude based on the regulation sensitive factor. In this stage, the response gain difference of each execution mechanism to the control instruction is quantified to ensure that the initial instruction set matches the dynamic characteristics of the mechanism under the current process state. For example, when the regulation sensitive factor of the main shaft servo unit is high, the speed adjustment instruction amplitude will be amplified to match the fast response requirement. Then, the initial instruction is pre-compensated by introducing the coupling compensation coefficient. This compensation mechanism is based on the physical coupling relationship between parameters. For example, when the main shaft speed instruction is increased, the tension instruction is simultaneously pre-increased according to the correlation strength of the centrifugal effect and the tension change to offset the possible wire relaxation. Finally, the amplitude limiting processing of the physical limit constraint maps the theoretical instruction value to the safe operation range of the execution mechanism. For example, when the calculated main shaft speed exceeds the rated maximum value of the motor, it is limited within the safety threshold.

[0147] Compared with the related art, the traditional method generates independent control instructions based on only a single parameter deviation, without considering the coupling effect between parameters. For example, when adjusting the spindle speed, the tension change is not compensated synchronously, resulting in uneven arrangement of the wire. The scheme actively predicts the interaction between parameters through a feedforward compensation mechanism, introduces a compensation coefficient in the control instruction generation stage, and avoids secondary interference. At the same time, combined with the physical limit constraint, the engineering realizability of the control instruction is ensured, and the instruction invalidation problem caused by ignoring the dynamic characteristics of the actuator in the traditional method is overcome.

[0148] Through the above technical scheme, the application solves the problem of control instruction conflict or insufficient compensation caused by the coupling effect between parameters during multi-parameter collaborative adjustment of the winding machine to some extent. Through the pre-compensation mechanism, the tension fluctuation caused by spindle acceleration is offset, and the secondary interference caused by single parameter adjustment is avoided. Combined with the amplitude limiting processing of the physical limit constraint, the execution reliability of the multi-modal regulation and control instruction under complex working conditions is ensured, and the control precision and equipment operation safety of the winding process are improved.

[0149] The application further proposes that after generating the multi-modal collaborative regulation and control instruction set of the target winding machine, the instruction set is executed and a new round of multi-source real-time data is collected, and a new parameter deviation is determined again. Based on the new parameter deviation and the historical data, the evaluation benchmark is updated. If the parameter deviation of the continuous multiple production cycles is continuously better than the historical level, the evaluation benchmark is adaptively tightened. Based on the updated evaluation benchmark and the new parameter deviation, the optimization priority sequence is dynamically corrected. The priority of the process parameters that need to be improved is reduced. The corrected optimization priority sequence is used for the generation of the regulation and control instruction in the subsequent production cycle.

[0150] The new parameter deviation refers to the deviation of the multi-source real-time data collected after the execution of the regulation and control instruction from the ideal state benchmark. It can be realized by a real-time data collection module and a dynamic error envelope analysis algorithm, and is used to verify the actual effect after the execution of the regulation and control instruction. The adaptive update of the evaluation benchmark refers to dynamically adjusting the state level determination standard according to the historical data. It can be realized by a sliding window statistics and threshold adaptive adjustment algorithm, and is used to match the demand for continuous optimization of the production process. The dynamic correction of the optimization priority sequence refers to adjusting the parameter regulation sequence based on the parameter deviation trend. It can be realized by dynamic calculation of sensitivity weight and re-evaluation of coupling strength, and is used to optimize the resource allocation efficiency.

[0151] Specifically, after each execution of the regulation instruction, the process parameters, audio stream, vibration sequence and other data of the winding machine are collected in real time through the sensor network, and the new parameter deviation is calculated by using dynamic error envelope analysis. When the deviation index of continuous multiple production cycles is better than the historical best level, the evaluation benchmark tightening mechanism is triggered, for example, the state level judgment threshold is increased by 5%-10%. Based on the updated benchmark, the sensitivity weight of each process parameter is recalculated, and for the parameters with an improvement amplitude lower than the set threshold (for example, less than 3%), their ranking in the optimization priority sequence is reduced. The revised priority sequence is input into the next round of regulation instruction generation module to form a closed-loop optimization process.

[0152] Compared with the related art, the traditional method uses fixed evaluation benchmark and static priority sequence, which cannot adapt to the dynamic changes of the production process, resulting in that the optimization strategy lags behind the actual working condition. The scheme realizes the dynamic optimization of the benchmark and the continuous adjustment of the priority through the closed-loop feedback mechanism, for example, after the tension parameter deviation of a certain winding machine is reduced by 15% in continuous 10 production cycles, the system automatically promotes the tension control priority from the second to the first, and adjusts the optimization threshold of the wire arranging accuracy parameter from ±0.1mm to ±0.08mm.

[0153] Through the above technical scheme, the application solves the problem of strategy lag caused by lack of dynamic feedback in the optimization and regulation process of the winding machine to a certain extent, and realizes the real-time matching of the evaluation benchmark and the production condition. Through the dynamic revision of the priority sequence, the optimization resources are tilted to the key parameters.

[0154] All the optional technical solutions described above can be combined to form optional embodiments of the application, which will not be described one by one here.

[0155] Figure 5 is a structural schematic diagram of a winding machine test data optimization management device provided by an embodiment of the application, referring to Figure 5 , the device comprises:

[0156] The acquisition module 501 is configured to, in response to a start instruction of a target winding machine, acquire the specification parameters and material parameters of a target coil of the target winding machine, and collect multi-source real-time data in the winding production process, wherein the multi-source real-time data comprises a process parameter sequence, audio stream data, a vibration sequence and a visual size parameter.

[0157] The determination module 502 is configured to determine, based on the specification parameters, the material parameters and the multi-source real-time data, a winding process state level of the target winding machine and an optimization priority sequence, wherein the winding process state level is used to indicate the stability stage of the current production process, and the optimization priority sequence is used to determine the execution order in multi-parameter regulation.

[0158] The generating module 503 is configured to generate a multi-modal cooperative control instruction set of the target winding machine based on the winding process state level, the optimization priority sequence, and the parameter deviation degree, the multi-modal cooperative control instruction set being used to guide the multi-parameter cooperative adjustment of the winding machine actuator, and the parameter deviation degree being obtained by analyzing the process parameter sequence and the visual size parameter.

[0159] It should be noted that the winding machine test data optimization management device provided in the above embodiments is used to perform winding machine optimization management, and only the above functional modules are used as examples for illustration. In actual applications, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the above described functions. In addition, the winding machine test data optimization management device and the winding machine test data optimization management method provided in the above embodiments belong to the same concept, and the specific implementation process is described in the method embodiments, which will not be repeated here.

[0160] Figure 6 FIG. 6 is a structural schematic diagram of a winding machine controller provided in an embodiment of the present application. The winding machine controller 600 can have great differences due to different configurations or performances, and can include one or more processors (Central Processing Units, CPU) 601 and one or more memories 602. The one or more memories 602 store at least one computer program, which is loaded and executed by the one or more processors 601 to implement the method provided in each of the above method embodiments. Of course, the winding machine controller 600 can also have a wired or wireless network interface, a keyboard, an input and output interface, and other components for realizing the functions of the device, and will not be described here.

[0161] In the exemplary embodiments, a computer readable storage medium, such as a memory including a computer program, is also provided. The computer program can be executed by a processor to complete the winding machine test data optimization management method in the above embodiments. For example, the computer readable storage medium can be a Read-Only Memory (ROM), a Random Access Memory (RAM), a Compact Disc Read-Only Memory (CD-ROM), a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0162] In the example embodiment, a computer program product or computer program is also provided, the computer program product or computer program comprising program code stored in a computer readable storage medium, the program code being read by a processor of a computer device from the computer readable storage medium, the processor executing the program code causing the computer device to perform the above-mentioned method for optimizing management of test data of a winding machine.

[0163] In some embodiments, the computer program related to the embodiments of the present application can be deployed to execute on one computer device, or on multiple computer devices located in one place, or on multiple computer devices distributed in multiple places and interconnected through a communication network, which can constitute a blockchain system.

[0164] Those of ordinary skill in the art can understand that all or part of the steps of the above-mentioned embodiments can be completed by hardware, or by program instructing relevant hardware, and the program can be stored in a computer readable storage medium, such as a read-only memory, a magnetic disk or an optical disk.

[0165] The above is only an optional embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for optimizing management of test data for a wire winding machine, characterized in that, The method comprises: in response to a start instruction of a target winding machine, acquiring specification parameters and material parameters of a target coil of the target winding machine, and collecting multi-source real-time data in a winding production process, the multi-source real-time data comprising a process parameter sequence, audio stream data, a vibration sequence and visual size parameters; based on the specification parameters, the material parameters and the multi-source real-time data, determining a material adaptation coefficient, a process stability index and a parameter coupling degree matrix of the target winding machine, the material adaptation coefficient being used to represent a matching degree between a current wire physical characteristic and a structure requirement of the target coil, the process stability index being used to quantify an approaching degree of a winding process dynamic characteristic to an ideal state, and the parameter coupling degree matrix being used to represent an interaction strength and a direction relationship between different process parameters; based on the material adaptation coefficient, the process stability index and the parameter coupling degree matrix, determining a winding process state level and an optimization priority sequence of the target winding machine, the winding process state level being used to indicate a stability stage of a current production process, and the optimization priority sequence being used to determine an execution order in multi-parameter regulation; based on the winding process state level, the optimization priority sequence and a parameter deviation degree, generating a multi-modal collaborative regulation instruction set of the target winding machine, the multi-modal collaborative regulation instruction set being used to guide a multi-parameter collaborative adjustment of a winding machine actuator, and the parameter deviation degree being determined based on the process parameter sequence and the visual size parameters.

2. The method of claim 1, wherein, The method comprises: performing dynamic error envelope analysis on the process parameter sequence and the visual size parameters to obtain a quality confidence weight and a parameter deviation degree, the quality confidence weight being used to represent a confidence degree of consistency between real-time data and historical high-quality data distribution, and the parameter deviation degree being used to indicate a deviation amplitude and a direction of a real-time parameter value relative to an ideal state reference; performing time-frequency domain joint analysis on the audio stream data and the vibration sequence to extract multi-modal audio print features, the multi-modal audio print features being used to represent joint frequency spectrum characteristics of mechanical vibration and acoustic signals in the winding process; based on the specification parameters and the material parameters, generating the material adaptation coefficient; based on the multi-modal audio print features, the quality confidence weight and the parameter deviation degree, determining the process stability index and the parameter coupling degree matrix.

3. The method of claim 2, wherein, The method comprises: performing stability contribution degree analysis on the multi-modal audio print features to obtain a stability weight coefficient of each audio print feature, the stability weight coefficient being used to represent an influence degree of different audio print features on the winding process stability; generate a comprehensive process deviation index based on the quality confidence weight and the parameter deviation degree, the comprehensive process deviation index being used to quantify the degree of the current winding process deviating from an ideal state as a whole; weight and fuse the multi-modal audio print features based on the stability weight coefficient to generate an audio print stability index, the audio print stability index being used to represent the dynamic stability of the winding process from the perspective of acoustic vibration characteristics; perform multi-dimensional feature fusion through a stability mapping model based on the audio print stability index and the comprehensive process deviation index to obtain the process stability index; construct a parameter interaction atlas based on the correlation between the multi-modal audio print features and the parameter deviation degree, the parameter interaction atlas being used to visually represent the correlation strength between different process parameters and audio print features; perform atlas feature extraction on the parameter interaction atlas to obtain the parameter coupling degree matrix, each element in the parameter coupling degree matrix being used to quantitatively represent the coupling relationship strength between corresponding process parameters.

4. The method of claim 1, wherein, The determination of the winding process state level and the optimization priority sequence based on the material adaptation coefficient, the process stability index and the parameter coupling degree matrix includes: input the material adaptation coefficient, the process stability index and the parameter coupling degree matrix into a multi-dimensional evaluation space for coordinate mapping to obtain a state coordinate vector of the current winding process in the multi-dimensional evaluation space, the multi-dimensional evaluation space being a standardized feature space constructed based on historical high-quality production process data; determine the winding process state level based on the positional relationship between the state coordinate vector and the boundaries of each stability level region through a clustering analysis algorithm, each stability level region boundary being obtained through density clustering analysis on historical production process data; perform eigenvalue decomposition on the parameter coupling degree matrix to obtain a dominant eigenvector, and determine the sensitivity weight of each process parameter to the overall stability of the target winding machine based on the dominant eigenvector and the material adaptation coefficient; generate the optimization priority sequence based on the size order of the sensitivity weight and the coupling strength between parameters represented by the parameter coupling degree matrix.

5. The method of claim 1, wherein, The generation of the multi-modal collaborative regulation instruction set of the target winding machine based on the winding process state level, the optimization priority sequence and the parameter deviation degree includes: determine the regulation sensitive factor of each actuator based on the winding process state level and the optimization priority sequence, the regulation sensitive factor being used to quantitatively represent the response gain of each actuator to the control instruction under different process states; determine the coupling compensation coefficient between multiple parameters based on the regulation sensitive factor and the parameter deviation degree, the coupling compensation coefficient being used to represent the dynamic compensation relationship required between parameters during collaborative adjustment; determine the multi-modal collaborative regulation instruction set based on the regulation sensitive factor, the coupling compensation coefficient and the parameter deviation degree.

6. The method of claim 5, wherein, The determination of the regulation sensitive factor of each actuator based on the winding process state level and the optimization priority sequence includes: Based on the winding process state level, the reference response parameters of each actuator are matched from a preset response characteristic library, the reference response parameters including acceleration response time of a spindle servo unit, stiffness coefficient of a tension control unit and positioning accuracy parameter of a wire winder; Based on the optimization priority sequence, the reference response parameters are dynamically weighted and corrected to obtain priority weighting coefficients, the priority weighting coefficients being used to reflect the importance of each actuator of the target winding machine in the current optimization sequence; By fusing the reference response parameters and the priority weighting coefficients, the regulation and control sensitive factors of each actuator are determined, the regulation and control sensitive factors of high-priority actuators being enhanced and the regulation and control sensitive factors of low-priority actuators being attenuated.

7. The method of claim 5, wherein, Based on the regulation and control sensitive factors and the parameter deviation degrees, coupling compensation coefficients among multiple parameters are determined, including: A control system transfer function matrix is established with the regulation and control sensitive factors as diagonal elements to construct a multivariable control system model; Based on the direction and amplitude of the parameter deviation degree, a feedforward compensation vector is determined, the feedforward compensation vector being used to predict coupling interference generated in the parameter adjustment process; An inverse matrix of the multivariable control system model is determined, and based on the inverse matrix and the feedforward compensation vector, a coupling compensation coefficient matrix is determined; The coupling compensation coefficient matrix is singular value decomposed to obtain the coupling compensation coefficients.

8. The method of claim 5, wherein, Based on the regulation and control sensitive factors, the coupling compensation coefficients and the parameter deviation degrees, the multimodal collaborative regulation and control instruction set is determined, including: Based on the regulation and control sensitive factors, the parameter deviation degrees are converted into initial instruction amplitudes of each actuator to generate an initial instruction set; Based on the coupling compensation coefficients, the initial instruction set is pre-compensated for coupling interference to form a post-compensation instruction set; The post-compensation instruction set is amplitude-limited with physical limit constraints of actuators of the target winding machine to obtain the multimodal collaborative regulation and control instruction set.

9. An optimization management system for winding machine test data, characterized in that, An acquisition module is configured to, in response to a start instruction for a target winding machine, acquire specification parameters and material parameters of a target coil of the target winding machine and collect multi-source real-time data in a winding production process, the multi-source real-time data including a process parameter sequence, audio stream data, a vibration sequence and visual size parameters; A determining module is configured to determine a material adaptation coefficient, a process stability index, and a parameter coupling degree matrix of the target winding machine based on the specification parameter, the material parameter, and the multi-source real-time data. The material adaptation coefficient is used to represent the matching degree between the current wire physical characteristics and the structure requirements of the target coil. The process stability index is used to quantify the closeness of the winding process dynamic characteristics to the ideal state. The parameter coupling degree matrix is used to represent the action strength and direction relationship of the mutual influence between different process parameters. Based on the material adaptation coefficient, the process stability index, and the parameter coupling degree matrix, a winding process state level and an optimization priority sequence of the target winding machine are determined. The winding process state level is used to indicate the stability stage of the current production process. The optimization priority sequence is used to determine the execution order when multiple parameters are regulated. A generating module is configured to generate a multi-modal collaborative regulation instruction set of the target winding machine based on the winding process state level, the optimization priority sequence, and a parameter deviation degree. The multi-modal collaborative regulation instruction set is used to guide the multi-parameter collaborative adjustment of the winding machine actuator. The parameter deviation degree is obtained by analyzing the process parameter sequence and the visual size parameter.

Images