Data processing method and device, nonvolatile storage medium and electronic equipment

Abstract

The invention discloses a data processing method and device, a nonvolatile storage medium and electronic equipment. The method comprises the following steps: acquiring operation data and production process data of a weaving machine; analyzing the operation data and the production process data by using a deep learning model to obtain broken yarn fault information of the weaving machine; the operation data and the production process data are analyzed through a hybrid model, performance bottleneck information of the weaving machine is obtained, and the hybrid model comprises a reinforcement learning model and a gradient boosting decision tree; and according to the broken yarn fault information and the performance bottleneck information, determining an optimization strategy for the weaving machine. The technical problem that the management efficiency of the weaving machine is low due to the fact that the management mode of the related weaving machine depends on manual inspection and simple sensor detection is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence, in particular, to a data processing method and device, a non-volatile storage medium and an electronic device. BACKGROUND

[0002] In the weaving link of the textile industry, weaving machines as the core equipment of production, its running state is directly related to production efficiency, cost control and product quality. The traditional weaving machine management mode mainly relies on manual inspection and basic sensor technology, although it can preliminarily monitor the basic running parameters of the machine, but this method has obvious limitations and efficiency bottlenecks. Specifically, manual inspection not only consumes time and effort, but also is affected by individual experience and subjective judgment, and it is difficult to guarantee the accuracy and consistency of monitoring. At the same time, simple sensor detection can only provide real-time running data, and lacks deep analysis of data and long-term trend prediction, and cannot effectively identify complex fault modes or carry out preventive maintenance, resulting in frequent equipment failures, and the stability and continuity of the weaving process are seriously affected.

[0003] At present, no effective solution has been proposed for the above problems. SUMMARY

[0004] The present application provides a data processing method and device, a non-volatile storage medium and an electronic device, to at least solve the technical problem of low management efficiency of weaving machines due to the dependence of the management mode of related weaving machines on manual inspection and simple sensor detection.

[0005] According to one aspect of the present application, a data processing method is provided, comprising: obtaining running data and production process data of a weaving machine; using a deep learning model to analyze the running data and the production process data to obtain yarn breakage fault information of the weaving machine; using a hybrid model to analyze the running data and the production process data to obtain performance bottleneck information of the weaving machine, wherein the hybrid model includes a reinforcement learning model and a gradient boosting decision tree; and determining an optimization strategy for the weaving machine according to the yarn breakage fault information and the performance bottleneck information.

[0006] Optionally, the deep learning model is obtained by the following method: obtaining historical running data and historical production process data of the weaving machine; obtaining an initial deep learning model combining a convolutional neural network and a long short-term memory network, wherein the convolutional neural network is used to extract spatial features from the historical running data and the historical production process data, and the long short-term memory network is used to capture the time sequence dependence in the historical running data and the historical production process data; by using a transfer learning technology, the feature extraction layer parameters of a pre-trained model are loaded into the convolutional neural network of the initial deep learning model to initialize the feature extraction part of the initial deep learning model, and a feature initialization model is obtained; the hyperparameters of the feature initialization model are optimized by using a Bayesian optimization method, and the optimized hyperparameters are obtained, and the feature initialization model is trained based on the optimized hyperparameters to obtain the deep learning model.

[0007] Optionally, the performance bottleneck information of the weaving machine is obtained by analyzing the running data and the production process data by using the hybrid model, including: performing regression analysis and feature importance sorting on the running data and the production process data by using a gradient boosting decision tree to identify preset key influencing factors affecting the overall performance of the weaving machine and the contribution of the preset key influencing factors; according to the preset key influencing factors and the contribution of the preset key influencing factors, a state space of a reinforcement learning model is constructed; in the state space, different device control strategies are generated by interacting with the running environment of the simulated weaving machine, and the optimal control strategy meeting the preset requirements is determined according to the performance changes caused by different device control strategies; the preset key influencing factors and the optimal control strategy are combined to obtain the performance bottleneck information of the weaving machine.

[0008] Optionally, the hybrid model is obtained by the following method: obtaining historical running data and historical production process data of the weaving machine; obtaining an initial hybrid model combining a reinforcement learning model and a gradient boosting decision tree; identifying preset key influencing factors in the historical running data and the historical production process data by using the gradient boosting decision tree, and constructing a state space and a reward function of the reinforcement learning model according to the preset key influencing factors; a multi-objective optimization algorithm is used to train the reinforcement learning model in a simulation environment constructed by digital twinning technology, wherein the strategy generated by the reinforcement learning model is verified in real time in the training process; the trained and verified reinforcement learning model and the gradient boosting decision tree are jointly determined as the hybrid model.

[0009] Optionally, a multi-objective optimization algorithm is used to train the reinforcement learning model in a simulation environment constructed using digital twin technology. This includes: constructing a simulation environment for the operation of a weaving machine based on digital twin technology, and determining the parameters to be optimized in the reinforcement learning model as decision variables for a multi-objective optimization problem; using a multi-objective optimization algorithm to generate an initial population of decision variables, where each individual in the initial population represents a set of parameter configurations for the reinforcement learning model; loading the parameter configurations of each individual in the initial population into the reinforcement learning model, performing policy interactions in the simulation environment, and using output, energy consumption, and quality indicators collected in the simulation environment as optimization objectives; calculating the fitness of each individual based on the collected multiple optimization objectives, and iteratively optimizing the initial population using non-dominated sorting and crowding calculations in the multi-objective optimization algorithm to generate a Pareto optimal solution set; and selecting the optimal reinforcement learning model parameter configuration that satisfies preset conditions from the Pareto optimal solution set as the trained reinforcement learning model.

[0010] Optionally, based on yarn breakage fault information and performance bottleneck information, an optimization strategy for the weaving machinery is determined, including: analyzing the yarn breakage fault information to obtain a first parameter set, wherein the first parameter set includes: fault type, occurrence frequency, and production process parameters; analyzing the performance bottleneck information to obtain a second parameter set, wherein the second parameter set includes: limiting factors affecting the overall efficiency of the weaving machinery and the degree of influence of the limiting factors; using predefined optimization decision rules, comprehensively reasoning about the first parameter set and the second parameter set to obtain a preliminary optimization strategy set, wherein the preliminary optimization strategy set includes: adjusting equipment parameters, maintenance scheduling, and modifying production plans; using digital twin technology, simulating the preliminary optimization strategy set in a virtual environment, and evaluating the strategies in the preliminary optimization strategy set based on the comprehensive evaluation indicators in the simulation results, wherein the comprehensive evaluation indicators include: output, quality, and energy consumption indicators; in the evaluated optimization strategy set, the target strategy whose comprehensive evaluation indicators meet the preset conditions is determined as the optimization strategy.

[0011] Optionally, the method further includes: acquiring weaving machinery operation data, wherein the weaving machinery operation data includes: machine start-up time, machine stop time, number of production loops, number of weaving loops, and quality pass rate; performing data cleaning and formatting on the weaving machinery operation data to obtain a standardized operation data sequence; calculating the production performance indicators of the target object based on the standardized operation data sequence, wherein the production performance indicators include: effective working hours, total production workload, and number of qualified products, wherein the effective working hours are calculated based on machine start-up time and machine stop time, and the number of qualified products is calculated based on the number of production loops, number of weaving loops, and quality pass rate; using a preset performance evaluation rule base, performing weighted calculation on the production performance indicators to obtain multi-dimensional scoring results, wherein the multi-dimensional scoring results include: output score, quality score, and comprehensive performance score, wherein the comprehensive performance score is calculated based on the output score and quality score; and filling the multi-dimensional scoring results, the target object's identity information, and time period information into a preset report template to generate a performance report.

[0012] Optionally, the method further includes: acquiring static attribute data and dynamic operation data of the target object, wherein the static attribute data includes at least: skill level certification information, and the dynamic operation data includes at least: historical output target achievement rate, quality pass rate, and equipment operation proficiency indicators; calculating the comprehensive capability index of the target object based on the static attribute data and dynamic operation data using a capability assessment model, wherein the capability assessment model is used to assign different weights to different skill levels and different operation proficiency indicators; determining the capability requirement specifications of the production tasks to be assigned based on the technical complexity, quality requirements, and delivery urgency of the production tasks; calculating the matching degree between the comprehensive capability index of the target object and the capability requirement specifications of the production tasks, and generating a preliminary task allocation plan based on the matching degree calculation results; continuously collecting the operation data of the target object during the execution of the preliminary task allocation plan, and dynamically updating the comprehensive capability index of the target object based on the operation data of the target object; and adaptively adjusting the task allocation based on the updated comprehensive capability index.

[0013] According to another aspect of this application, a data processing apparatus is also provided, comprising: an acquisition module for acquiring operating data and production process data of a weaving machine; a first analysis module for analyzing the operating data and production process data using a deep learning model to obtain yarn breakage fault information of the weaving machine; a second analysis module for analyzing the operating data and production process data using a hybrid model to obtain performance bottleneck information of the weaving machine, wherein the hybrid model includes: a reinforcement learning model and a gradient boosting decision tree; and a determination module for determining an optimization strategy for the weaving machine based on the yarn breakage fault information and performance bottleneck information.

[0014] According to another aspect of this application, a non-volatile storage medium is also provided, the storage medium including a stored program, wherein the program, when running, controls the device where the storage medium is located to execute the above data processing method.

[0015] According to another aspect of this application, an electronic device is also provided, comprising: a memory and a processor, the processor being configured to run a program stored in the memory, wherein the program executes the above-described data processing method during runtime.

[0016] According to another aspect of this application, a computer program product is also provided, comprising a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-described data processing method.

[0017] This application employs the following methods: acquiring operational and production process data of weaving machinery; analyzing the operational and production process data using a deep learning model to obtain yarn breakage fault information of the weaving machinery; and using a hybrid model to analyze the operational and production process data to obtain performance bottleneck information of the weaving machinery, wherein the hybrid model includes a reinforcement learning model and a gradient boosting decision tree. Based on the yarn breakage fault information and performance bottleneck information, an optimization strategy for the weaving machinery is determined, achieving the goal of efficient fault warning and precise performance optimization of the weaving machinery. This results in improved production efficiency, reduced cost consumption, and improved product quality, thereby solving the technical problem of low management efficiency of weaving machinery caused by the reliance on manual inspection and simple sensor detection in the management of related weaving machinery. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a flowchart of a data processing method according to an embodiment of this application;

[0020] Figure 2 This is an architecture diagram of a management system for a weaving machine according to an embodiment of this application;

[0021] Figure 3 This is a deployment architecture diagram according to an embodiment of this application;

[0022] Figure 4 This is a schematic diagram of a cloud platform interface according to an embodiment of this application;

[0023] Figure 5This is a schematic diagram of another cloud platform interface according to an embodiment of this application;

[0024] Figure 6 This is a schematic diagram of another cloud platform interface according to an embodiment of this application;

[0025] Figure 7 This is a structural diagram of a data processing apparatus according to an embodiment of this application;

[0026] Figure 8 This is a hardware structure block diagram of a computer terminal for implementing a data processing method according to an embodiment of this application. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] According to an embodiment of this application, a method embodiment for data processing is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] Figure 1 This is a flowchart of a data processing method according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0031] Step S102: Obtain the operating data and production process data of the weaving machinery.

[0032] Among them, weaving machinery is not limited to traditional looms, but also includes various modern textile equipment, such as high-speed rapier looms, air-jet looms, water-jet looms, shuttleless looms, etc.

[0033] Operating data includes, but is not limited to: Number of weft loops: The number of times the weft yarn passes through the warp yarn to form a complete interlacing unit during the weaving process; this is an important indicator for evaluating weaving efficiency. Rotation speed: The spindle speed of the textile machine, used to reflect the operating speed and efficiency of the equipment. Tension: The tension state of the yarn or fabric during the weaving process, directly affecting product quality and equipment operational stability. Current: The current value of the motor during operation, used to assess the motor's load and operating status. Voltage: The supply voltage, ensuring stable operation of the weaving machinery within a safe voltage range. Power: The power consumption of the weaving machinery, crucial for energy consumption analysis and cost control. Temperature: Temperature data of the weaving environment, affecting equipment operating efficiency and fabric quality. Humidity: The humidity level of the weaving environment, which has a significant impact on yarn strength and fabric properties.

[0034] Production process data includes, but is not limited to: Output: The quantity of yarn or fabric woven per unit time, used to assess production efficiency. Defect Rate: The percentage of non-conforming products generated during production, reflecting the quality control status of the production process. Downtime: Non-operating time of equipment due to malfunctions, maintenance, etc., including statistics on planned and unplanned downtime, which helps in calculating utilization rate and overall equipment efficiency. Fault Type: Detailed records of the types of equipment malfunctions, such as yarn breakage, abnormal tension, mechanical or electrical faults, providing a basis for fault diagnosis and preventive maintenance. Production Tasks: Including task number, task content (such as product type, quantity, process requirements), planned start and finish times, used for production scheduling and resource allocation.

[0035] In step S102, after acquiring the operating data and production process data of the weaving machinery, the collected data is processed in real time on an edge device close to the weaving machinery via an edge computing processing module. This real-time processing includes, but is not limited to, preprocessing operations such as cleaning, filtering, and normalization to remove noise and outliers and extract key features. Real-time processing provides high-quality data support for subsequent data analysis while avoiding the latency of data transmission to the cloud, significantly improving response speed.

[0036] Step S104: Analyze the running data and production process data using a deep learning model to obtain information on yarn breakage faults in the weaving machinery.

[0037] For example, a deep learning model is a hybrid model combining convolutional neural networks and long short-term memory networks. Deep learning models can extract spatial and temporal dependencies from multi-dimensional time-series data, automatically identifying patterns of yarn breakage. By learning from historical failure case libraries and real-time operational data, deep learning models can be trained to recognize yarn breakage faults. When abnormal fluctuations occur in the current or tension data of weaving machinery, deep learning models can respond quickly, accurately diagnose yarn breakage faults caused by yarn quality issues, and promptly send early warning information to on-site operators, thereby avoiding prolonged equipment downtime and reducing production losses.

[0038] Step S106: Analyze the operational data and production process data using a hybrid model to obtain performance bottleneck information of the weaving machinery. The hybrid model includes a reinforcement learning model and a gradient boosting decision tree.

[0039] Step S106 employs a hybrid model to analyze operational and production process data, where reinforcement learning and gradient boosting decision trees work together. Reinforcement learning models, such as the PPO algorithm, learn optimal strategies for adjusting weaving machinery parameters (e.g., speed, tension) through interaction with the environment (real-time operational data) to adapt to constantly changing production conditions and environmental factors. Gradient boosting decision trees are used to analyze historical data, identify key factors affecting weaving machinery performance, such as the correlation between energy consumption and yarn breakage rate, helping to understand the underlying causes of equipment operation. By combining the advantages of both models, performance bottlenecks in the weaving machinery can be accurately located.

[0040] Step S108: Based on the yarn breakage fault information and performance bottleneck information, determine the optimization strategy for the weaving machinery.

[0041] Step S108 involves comprehensively analyzing the type, frequency, and causes of yarn breakage faults, as well as the specific manifestations of performance bottlenecks (such as excessively low rotation speed and unstable tension), to determine the most effective optimization measures. Optimization measures include adjusting equipment operating parameters, optimizing production plans, and improving yarn quality control processes, aiming to improve the operating efficiency and product quality of weaving machinery while reducing energy consumption and maintenance costs.

[0042] The above steps involve acquiring operational and production process data of the weaving machinery; analyzing the operational and production process data using a deep learning model to obtain yarn breakage fault information of the weaving machinery; analyzing the operational and production process data using a hybrid model to obtain performance bottleneck information of the weaving machinery, wherein the hybrid model includes a reinforcement learning model and a gradient boosting decision tree; and determining optimization strategies for the weaving machinery based on the yarn breakage fault information and performance bottleneck information. This achieves the goal of efficient fault warning and precise performance optimization of the weaving machinery, thereby realizing the technical effects of improving production efficiency, reducing cost consumption, and improving product quality.

[0043] The following are Figure 1 The steps shown are illustrated and explained by way of example.

[0044] According to some optional embodiments of this application, the deep learning model is trained by the following method: acquiring historical operation data and historical production process data of the weaving machinery; acquiring an initial deep learning model combining a convolutional neural network and a long short-term memory network, wherein the convolutional neural network is used to extract spatial features from the historical operation data and historical production process data, and the long short-term memory network is used to capture the temporal dependencies in the historical operation data and historical production process data; using transfer learning technology, loading the feature extraction layer parameters of the pre-trained model into the convolutional neural network of the initial deep learning model to initialize the feature extraction part of the initial deep learning model, thereby obtaining a feature initialization model; using a Bayesian optimization method to tune the hyperparameters of the feature initialization model to obtain optimized hyperparameters, and training the feature initialization model based on the optimized hyperparameters to obtain a deep learning model.

[0045] In this embodiment, the historical operation data and historical production process data of the weaving machine are first obtained from the historical operation records. The historical operation data includes production cycles, rotation speed, tension, current, voltage, power, temperature and humidity, etc., while the historical production process data includes output, defect rate, downtime, fault type, etc.

[0046] Next, an initial deep learning model is obtained, which consists of two parts: a convolutional neural network (CNN) and a long short-term memory (LSTM) network. The CNN is used to extract spatial features from historical datasets, such as identifying meaningful patterns from the two-dimensional or three-dimensional distribution of sensor data. The LSTM network is used to capture temporal dependencies in the data, remembering historical states and understanding the relationship between the current state and past states. By combining the CNN and LSTM networks, information in both spatial and temporal dimensions can be processed simultaneously, achieving a comprehensive understanding of the operating status and production process of weaving machinery.

[0047] To accelerate the training process and improve model performance, this embodiment employs transfer learning technology for feature initialization. In this embodiment, the feature extraction layer parameters of the pre-trained model are loaded into the convolutional neural network part of the initial deep learning model, thus initializing the model's feature extraction component. This step enables the model to inherit the ability to understand complex data structures from the pre-trained model.

[0048] To find the optimal hyperparameter configuration for model training, this embodiment employs Bayesian optimization. In this embodiment, Bayesian optimization is used to automatically tune the model's hyperparameters, such as learning rate, batch size, layer size, and activation function.

[0049] Finally, based on the optimized hyperparameters, the feature initialization model was further trained. Through training with a large amount of historical operating data and production process data, the model gradually learned the characteristics and occurrence patterns of yarn breakage faults, as well as the temporal dependencies of equipment performance. Through continuous iterative training and validation, the model's performance was significantly improved, enabling it to diagnose yarn breakage faults in weaving machinery with an accuracy rate of over 95% and predict potential performance bottlenecks, providing a scientific basis for preventive maintenance and production efficiency optimization.

[0050] According to some alternative embodiments of this application, the performance bottleneck information of the weaving machine is obtained by analyzing the operating data and production process data using a hybrid model. This can be achieved through the following method: using a gradient boosting decision tree to perform regression analysis and feature importance ranking on the operating data and production process data to identify preset key influencing factors affecting the overall performance of the weaving machine and their contribution; constructing the state space of a reinforcement learning model based on the preset key influencing factors and their contribution; generating different equipment control strategies within the state space through interaction with the simulated operating environment of the weaving machine, and determining the optimal control strategy that meets preset requirements based on the performance changes caused by different equipment control strategies; and combining the preset key influencing factors with the optimal control strategy to obtain the performance bottleneck information of the weaving machine.

[0051] In this embodiment, a gradient boosting decision tree is first used to perform detailed regression analysis on the operating data and production process data of the weaving machinery. This allows for accurate prediction of equipment performance indicators such as output, efficiency, energy consumption, and quality. Based on the regression analysis, the gradient boosting decision tree further ranks the features by importance, meaning it can quantify the impact of each feature in the operating and production process data on the overall performance of the weaving machinery. For example, it can identify which factors, such as yarn tension, rotational speed, and equipment temperature, have a decisive effect on yarn breakage rate, production speed, or energy consumption, and the complex interaction effects between them. The gradient boosting decision tree can identify preset key influencing factors affecting the overall performance of the weaving machinery and determine the contribution of each factor through feature importance ranking.

[0052] Then, based on the preset key influencing factors and their contribution, the state space of the reinforcement learning model is constructed. This state space includes the operating parameters of the weaving machinery, external environmental variables, and other key production factors. By combining the analysis results of the gradient boosting decision tree, the state space not only comprehensively covers all aspects of the weaving process but also accurately reflects the degree of influence of each factor on equipment performance. This allows the reinforcement learning model to more accurately understand the current state and predict future states under different control strategies.

[0053] Within the state space, the reinforcement learning model continuously attempts to generate and evaluate different equipment control strategies through interaction with the simulated operating environment of the weaving machinery. These strategies include specific operations such as adjusting rotational speed, optimizing tension, and controlling temperature, as well as more advanced production planning and resource allocation decisions. Through continuous interaction with the simulated environment, the model can observe changes in mechanical performance under different strategies, such as increased output, reduced energy consumption, or improved product quality, thereby learning which strategies yield the best performance improvements under specific conditions.

[0054] Furthermore, the reinforcement learning model iteratively evaluates the long-term effects of each strategy, that is, the impact of the strategy on improving the overall performance of the weaving machinery after considering long-term benefits and costs. Based on the evaluation results, the optimal control strategy that meets preset requirements (such as maximizing output, minimizing costs, or optimizing product quality) can be determined. The optimal control strategy not only effectively improves the equipment performance in the current state but also considers long-term stability and efficiency.

[0055] Finally, by combining the identified key influencing factors with the optimal control strategy, performance bottleneck information for the weaving machinery can be obtained. For example, if data indicates that tension control is the most significant factor affecting product quality, and the optimal control strategy suggests reducing yarn breakage by adjusting tension, then tension control will be clearly identified as the equipment's performance bottleneck. Corresponding optimization suggestions will be provided to guide on-site operators or managers in specific operations, thereby improving the overall performance of the weaving machinery, increasing production efficiency and product quality, and reducing operating costs.

[0056] In some optional embodiments of this application, the hybrid model is trained by the following method: acquiring historical operation data and historical production process data of the weaving machinery; acquiring an initial hybrid model combining a reinforcement learning model and a gradient boosting decision tree; using the gradient boosting decision tree to identify preset key influencing factors in the historical operation data and historical production process data, and constructing the state space and reward function of the reinforcement learning model based on the preset key influencing factors; training the reinforcement learning model in a simulation environment constructed by digital twin technology using a multi-objective optimization algorithm, wherein the strategy generated by the reinforcement learning model is verified by real-time simulation during the training process; and jointly determining the trained and verified reinforcement learning model and the gradient boosting decision tree as the hybrid model.

[0057] In this embodiment, historical data is first analyzed to identify preset key influencing factors. Gradient boosting decision trees can effectively handle complex and nonlinear data relationships, identifying which operating parameters and production settings have a significant impact on equipment performance, such as energy consumption, yarn breakage rate, and production speed.

[0058] Secondly, based on the key influencing factors identified by the gradient boosting decision tree, the state space and reward function of the reinforcement learning model are constructed. The state space defines all possible environmental states during the operation of the weaving machinery, where environmental states include the equipment's operating parameters, external environmental conditions, and detailed information about the production task. The reward function guides the learning direction of the reinforcement learning model, ensuring that the model can identify and execute control strategies that improve equipment performance, reduce energy consumption, and enhance product quality. The reward function needs to comprehensively consider multiple objectives, such as maximizing output, minimizing energy consumption, and optimizing product quality, enabling the reinforcement learning model to learn to balance multiple objectives to find the optimal control strategy.

[0059] Then, a multi-objective optimization algorithm is used to train the reinforcement learning model in a simulation environment constructed using digital twin technology. In the simulation environment, the reinforcement learning model learns how to take optimal action under various conditions by continuously trying different control strategies and observing their effects on the virtual device. The multi-objective optimization algorithm ensures that the model can simultaneously consider multiple objectives such as output, energy consumption, and quality, rather than optimizing only a single metric.

[0060] During training, the strategies generated by the reinforcement learning model are validated through real-time simulation. That is, each strategy proposed by the model is immediately tested in a digital twin environment to assess its actual impact on device performance. This process not only accelerates the model's learning speed but also improves the effectiveness of the strategies, ensuring that the model can generate optimized strategies truly applicable to real-world production scenarios.

[0061] Finally, the trained and validated reinforcement learning model, together with the gradient boosting decision tree, constitutes a hybrid model. This hybrid model fully leverages the gradient boosting decision tree's ability to identify key factors and the reinforcement learning model's strategy optimization capabilities to accurately pinpoint and dynamically optimize the performance bottlenecks of weaving machinery. When applied to actual production, the hybrid model can automatically generate and execute optimal control strategies based on real-time operational and production process data, significantly improving equipment operating efficiency and production quality, reducing energy consumption and operating costs, and providing strong technical support for the intelligent upgrading of textile enterprises.

[0062] As some optional embodiments of this application, training a reinforcement learning model in a simulation environment constructed using digital twin technology with a multi-objective optimization algorithm can be achieved through the following method: Constructing a simulation environment for the operation of a weaving machine based on digital twin technology, and determining the parameters to be optimized in the reinforcement learning model as decision variables for a multi-objective optimization problem; using a multi-objective optimization algorithm to generate an initial population of decision variables, where each individual in the initial population represents a set of parameter configurations for the reinforcement learning model; loading the parameter configurations of each individual in the initial population into the reinforcement learning model, performing policy interactions in the simulation environment, and using the output, energy consumption, and quality indicators collected in the simulation environment as optimization objectives; calculating the fitness of each individual based on the collected multiple optimization objectives, and iteratively optimizing the initial population through non-dominated sorting and crowding calculations in the multi-objective optimization algorithm to generate a Pareto optimal solution set; selecting the optimal reinforcement learning model parameter configuration that satisfies preset conditions from the Pareto optimal solution set as the trained reinforcement learning model.

[0063] It should be noted that in the context of weaving machinery, digital twin models can accurately reproduce the internal structure and operating mechanism of the equipment, as well as various external variables involved in the production process, such as raw material conditions, production tasks, ambient temperature and humidity.

[0064] At the start of the training process, an initial population of decision variables is generated based on a multi-objective optimization algorithm. This initial population actually consists of many individuals representing different parameter configurations. Each individual contains a specific set of reinforcement learning model parameters, which are considered as different strategies controlling the weaving machinery. In the initial population, the parameter configuration of each individual is randomly generated.

[0065] The parameter configurations of each individual in the population are loaded into a reinforcement learning model, which then interacts with the model in a simulation environment based on digital twin technology. During this interaction, the model takes actions based on the current state input and receives feedback signals (rewards or penalties), such as output, energy consumption, and quality indicators collected in the simulation environment. Through repeated trials, the model begins to learn the impact of different parameter configurations on the weaving process and how to adjust these parameters to achieve the optimization goal.

[0066] After each interaction, the fitness of each individual is calculated based on multiple optimization objectives collected in the simulation environment. Fitness measures the combined impact of model parameter configuration on output, energy consumption, and quality indicators, with the goal of minimizing energy consumption while ensuring output and quality. Fitness calculations allow for the evaluation of the effectiveness of each strategy.

[0067] This algorithm utilizes non-dominated sorting and crowding calculation from multi-objective optimization algorithms to iteratively optimize the initial population based on the fitness of each individual. Non-dominated sorting allows the algorithm to consider multiple optimization objectives simultaneously, thereby finding individuals that are no less than any other solution across all objectives; these individuals form the so-called Pareto front. Crowding calculation is used to assess the distribution density of individuals in the population, preventing the algorithm from getting trapped in local optima. Through iterative optimization, poorly performing individuals are gradually eliminated, while those with better parameter configurations in terms of output, energy consumption, and quality are retained and improved. Ultimately, a Pareto optimal solution set is generated, which is the set of strategies that achieve the best balance among these objectives.

[0068] From the Pareto optimal solution set, one or more optimal reinforcement learning model parameter configurations that meet preset conditions are selected. These preset conditions include minimum output requirements, maximum tolerance for energy consumption, and high standards for fabric quality. The selected parameter configurations are then used as the trained reinforcement learning model. This model guides the actual operation of the weaving machinery to achieve multiple objectives: maximizing output, minimizing energy consumption, and optimizing quality. This training method, based on a simulation environment and multi-objective optimization algorithms, enables this embodiment to predict and optimize the performance of the weaving machinery before actual production, significantly improving its intelligence and flexibility in handling complex production environments.

[0069] In some optional embodiments of this application, determining optimization strategies for weaving machinery based on yarn breakage fault information and performance bottleneck information can be achieved through the following methods: parsing the yarn breakage fault information to obtain a first parameter set, wherein the first parameter set includes: fault type, occurrence frequency, and production process parameters; parsing the performance bottleneck information to obtain a second parameter set, wherein the second parameter set includes: limiting factors affecting the overall efficiency of the weaving machinery and indicators of the degree of influence of the limiting factors; using predefined optimization decision rules, comprehensively reasoning on the first and second parameter sets to obtain a preliminary optimization strategy set, wherein the preliminary optimization strategy set includes: adjusting equipment parameters, maintenance scheduling, and modifying production plans; using digital twin technology, simulating the preliminary optimization strategy set in a virtual environment, and evaluating the strategies in the preliminary optimization strategy set based on the comprehensive evaluation indicators in the simulation results, wherein the comprehensive evaluation indicators include: output, quality, and energy consumption indicators; in the evaluated optimization strategy set, the target strategy whose comprehensive evaluation indicators meet preset conditions is determined as the optimization strategy.

[0070] It's worth explaining that the core objective of optimizing the decision-making process is to identify and resolve yarn breakage faults and performance bottlenecks in weaving machinery operation, in order to achieve higher production efficiency, better product quality, and lower energy consumption. In this process, the yarn breakage fault information is first analyzed in depth to extract a first set of parameters. This first set of parameters includes the specific type of fault, its frequency of occurrence during production, and related production process parameters. For example, the fault type could be yarn breakage due to excessively low or high yarn tension; the frequency of occurrence indicates how common this type of fault is throughout the entire production cycle; and the production process parameters include various settings related to the weaving process, such as loom speed, tension control, and raw material characteristics.

[0071] Secondly, the identified performance bottleneck information is analyzed to obtain a second parameter set. This second parameter set includes limiting factors affecting the overall efficiency of the weaving machinery and indicators of the degree of influence of these factors. Limiting factors include equipment aging, improper operation, and unfavorable environmental conditions. The degree of influence indicators are used to quantify the specific impact of these factors on output, quality, and energy consumption. For example, wear and tear on a certain equipment component may lead to a decrease in weaving speed, thereby affecting output; low operator skill levels may lead to a higher defect rate, affecting the overall product quality; and unstable power supply or the presence of high-energy-consuming components will directly increase production costs.

[0072] Then, using predefined optimization decision rules, a preliminary optimization strategy set is generated by comprehensively reasoning about the first and second parameter sets. These optimization decision rules, based on industry experience and expert knowledge, aim to resolve identified yarn breakage issues and performance bottlenecks by adjusting equipment parameters, developing maintenance schedules, and modifying production plans. For example, for frequent yarn breakage, suggestions might include adjusting the yarn tension range or improving the raw material pretreatment process; for performance bottlenecks, suggestions might include updating or maintaining key equipment components, optimizing production schedules to maximize the use of high-efficiency periods, or adjusting loom operating parameters, such as speed or temperature settings, to reduce energy consumption and maintain product quality.

[0073] Digital twin technology was used again to simulate and validate these preliminary optimization strategy sets in a virtual environment. By executing the optimization strategies in the digital twin environment, their performance on three indicators—output, quality, and energy consumption—could be evaluated without physically operating the weaving machinery, thus avoiding the risks and costs that could result from directly changing production settings. A comprehensive evaluation metric was used to measure the overall effectiveness of the strategies across these three dimensions, identifying which strategies are most likely to have a positive impact on actual production.

[0074] Finally, from the set of optimized strategies evaluated through simulation, those strategies that meet the preset conditions in terms of comprehensive evaluation indicators are selected as the optimized strategies. These preset conditions include, for example, output not falling below a certain threshold, maintaining stable high quality, and energy consumption not exceeding budget limits. The optimized strategies are then recommended to production managers for implementation, ensuring product quality and production efficiency while maximizing energy savings, thus achieving intelligent management and sustainable development in the textile industry.

[0075] As some optional embodiments of this application, the data processing method further includes the following steps: acquiring weaving machinery operation data, wherein the weaving machinery operation data includes: start-up time, downtime, number of production loops, number of weaving loops, and quality pass rate; performing data cleaning and formatting on the weaving machinery operation data to obtain a standardized operation data sequence; calculating the production performance indicators of the target object based on the standardized operation data sequence, wherein the production performance indicators include: effective working hours, total production workload, and number of qualified products, wherein the effective working hours are calculated based on start-up time and downtime, and the number of qualified products is calculated based on the number of production loops, number of weaving loops, and quality pass rate; using a preset performance evaluation rule base, performing weighted calculation on the production performance indicators to obtain multi-dimensional scoring results, wherein the multi-dimensional scoring results include: output score, quality score, and comprehensive performance score, wherein the comprehensive performance score is calculated based on the output score and quality score; filling the multi-dimensional scoring results, the target object's identity information, and time period information into a preset report template to generate a performance report.

[0076] In this embodiment, weaving machine operation data is acquired in real-time or periodically. This data includes machine start-up time, downtime, number of production turns, number of weaving turns, and quality pass rate. The weaving machine operation data is cleaned to remove outliers, missing values, or duplicate data, ensuring the accuracy and completeness of the data. Based on the standardized operation data sequence, specific production performance indicators for the target object are calculated. The specific calculation method is as follows:

[0077] Effective working hours: Calculate the effective working time of the target object within a given time period based on the startup time and downtime. Effective working hours = Startup time - Downtime.

[0078] Total production workload: Based on the number of production loops and weaving loops, calculate the production workload of the target object per unit time. The specific algorithm is as follows: Total production workload = Number of production loops / Number of weaving loops.

[0079] Quantity of qualified products: Calculate the quantity of qualified products during the production process using the number of production turns, the number of weaving turns, and the quality pass rate. Quantity of qualified products = Number of production turns * Quality pass rate.

[0080] Performance Appraisal Rules: A pre-defined performance appraisal rule base is used to weight and calculate the production performance indicators, resulting in multi-dimensional scores. The rule base includes various weight parameters to balance factors such as output, quality, and working time. The specific calculation steps are as follows: Output Score: The output score is calculated based on the total production workload and a pre-defined output benchmark, reflecting production efficiency. Quality Score: The quality score is calculated based on the number of qualified products and pre-defined quality standards, examining production quality. Overall Performance Score: The output score and quality score are combined, and an overall performance score is calculated using pre-defined overall performance scoring rules.

[0081] The multi-dimensional scoring results obtained from the above calculations, the target object's identity information, and the time period information are then populated into a pre-defined performance report template. The report template includes charts, tables, and text descriptions to visually present the performance data. Finally, a detailed performance report is generated, listing the target object's productivity, quality performance, and overall performance score within the specified time period.

[0082] In some optional embodiments, the data processing method further includes the following steps: acquiring static attribute data and dynamic operation data of the target object, wherein the static attribute data includes at least: skill level certification information, and the dynamic operation data includes at least: historical output target achievement rate, quality pass rate, and equipment operation proficiency index; calculating the comprehensive capability index of the target object based on the static attribute data and dynamic operation data through a capability assessment model, wherein the capability assessment model is used to assign different weights to different skill levels and different operation proficiency indicators; determining the capability requirement specifications of the production task according to the technical complexity, quality requirements, and delivery urgency of the production task to be assigned; calculating the matching degree between the comprehensive capability index of the target object and the capability requirement specifications of the production task, and generating a preliminary task allocation plan based on the matching degree calculation result; continuously collecting the operation data of the target object during the execution of the preliminary task allocation plan, and dynamically updating the comprehensive capability index of the target object based on the operation data of the target object; and adaptively adjusting the task allocation based on the updated comprehensive capability index.

[0083] In this embodiment, static attribute data of the target object, such as skill level certification information, is obtained from an enterprise resource planning system or other human resource management system. This static attribute data reflects the employee's basic qualities and professional skills. Then, dynamic operational data of the target object is collected in real time or on demand. This dynamic operational data includes historical output target achievement rate, quality pass rate, equipment operation proficiency indicators, etc., and reflects the employee's performance in actual production. The static attribute data and dynamic operational data are cleaned and formatted to ensure data consistency and accuracy.

[0084] Different skill levels and operational proficiency indicators are combined and assigned different weights. Static attribute data and dynamic operational data of the target object are input into the capability assessment model to calculate the target object's comprehensive capability index. This comprehensive capability index represents the operator's overall operational skills and work performance level. Based on the technical complexity, quality requirements, and delivery urgency of the production task to be assigned, the task's requirements are determined. The minimum operational skills and performance levels required to complete the task are listed in detail to ensure that task assignment aligns with the operator's comprehensive capabilities.

[0085] The overall capability index of each target entity is matched with the capability requirements of the production task to assess the fit between operators and tasks. Based on the matching results, a preliminary task allocation plan is generated. This plan aims to assign tasks to operators whose capability indices match the requirements as closely as possible, thereby improving production efficiency and product quality. While implementing the preliminary task allocation plan, operational data of the target entities continues to be collected in real time. The overall capability index of the target entities is updated periodically or in real time. The update process requires re-entering the latest collected dynamic operational data into the capability assessment model to ensure that the index reflects the current skill level and work performance of employees.

[0086] Based on the updated comprehensive capability index, the initial task allocation plan is adaptively adjusted. Adjustments may include reassigning tasks, increasing auxiliary training, and adjusting production plans to better match the current capabilities of operators with the demands of production tasks. Furthermore, the adaptive adjustment strategy should consider multiple factors, such as employee workload, equipment load status, and the urgency of production tasks, to optimize production efficiency and maximize employee satisfaction.

[0087] In some optional embodiments, the data processing method further includes the following steps: real-time collection of loom data operated by workers, including key indicators such as machine start-up time, downtime, number of production loops, number of weaving loops, and quality pass rate; automatic calculation of each worker's working hours and workload; and generation of detailed performance reports based on preset performance appraisal rules. Simultaneously, production task allocation is intelligently adjusted based on workers' skill levels and operational proficiency to ensure maximum production efficiency. Furthermore, the platform provides multiple salary templates for flexible configuration, and performance data is synchronized to the salary calculation module to achieve accurate payroll calculation, reduce manual intervention, and improve the efficiency and transparency of performance management. The production management model algorithm is as follows:

[0088] (1) Worker's workload = Σ (workload of small tasks);

[0089] Worker_workload=Σ((Production_cycles / Weaving_cycles)*Actual_value;

[0090] The workload of a small task = output * actual value;

[0091] Sub_task_workload=Output*Actual_value;

[0092] Output = Production cycles / Weaving cycles.

[0093] The parameters are explained below:

[0094]

[0095] (2) Individual output = Σ (output per machine);

[0096] Total_Output=Σ(Production_cycles / Weaving_cycles);

[0097] Total_Output=∑(Individual_Machine_Output);

[0098] Output per machine = Number of production loops / Number of weaving loops;

[0099]

[0100] The parameters are explained below:

[0101]

[0102] (3) Actual balance rate = [∑process CT × number of workers in each process / (total number of workers * bottleneck process CT)] × 100%

[0103]

[0104] Typical processes in a loom factory include: warping, drafting and denting, weaving, fabric inspection, and packing. The cycle time (CT) and number of operators for each process need to be measured in practice.

[0105] The parameters are explained below:

[0106]

[0107] (4) Salary = :

[0108]

[0109] Parameter description:

[0110]

[0111] There are multiple ways to calculate dynamic daily wages.

[0112] Overproduction wage scheme: Dynamic daily wage = number of overproduction rounds * piece rate unit price / 1000;

[0113] DailySalaryltem=OverCycles*PieceRate / 1000;

[0114] Overproduction cycle count = Production cycle count - Overproduction quota;

[0115] OverCycles=Production_cycles–OverProductionQuota;

[0116] Pure rotation count scheme: Dynamic daily wage = number of production rotations * piece rate unit price / 1000;

[0117] DailySalaryltem=Production_cycles*PieceRate / 1000.

[0118] Custom solution: Assemble the calculation formula yourself based on the parameters provided by the system.

[0119] Optional parameter descriptions:

[0120]

[0121] Figure 2 This is an architecture diagram of a management system for a weaving machine according to an embodiment of this application. Figure 2 The management system shown is Figure 1 The system upon which the method shown depends. For example... Figure 2As shown, the perception layer includes sensor devices for data collection; the network layer includes the use of different communication protocols, such as MQTT and HTTP / HTTPS, for data transmission; the access layer is used for device access management; the data layer includes storage units for static and dynamic data, including relational databases (MySQL), cache databases (Redis), and other storage methods; the business layer includes core functions such as system management, permission management, log management, device access, data analysis, and operation control; the application layer is user-facing, providing a PC-based backend management system and a mobile application to ensure comprehensive system functionality and ease of operation. This layered architecture clearly demonstrates the complete process from data acquisition to user interaction, highlighting its high integration, high availability, and high scalability.

[0122] Figure 3 This is a deployment architecture diagram according to an embodiment of this application. Figure 3 This demonstrates the deployment path from code development, version control repository, build server (Jenkins), to container orchestration platform (Kubernetes). Specifically, after code development is complete, developers commit the code to the repository. Jenkins administrators manually trigger the build process or automatically via a webhook in the code repository. The Jenkins platform then automatically performs code fetching, unit testing, compilation, and packaging until deployable files are generated. These files are pushed to Kubernetes development, testing, pre-release, and production environments, enabling rapid deployment and iterative updates, ensuring software stability and security.

[0123] Figure 4 This is a schematic diagram of a cloud platform interface according to an embodiment of this application. Figure 4 The user interface of the loom management cloud platform is shown. The left side of the interface features a navigation bar with options such as "Basic Information," "Reports," and "Host Management," allowing users to quickly locate functional modules. The top navigation bar provides entry points for "Home," "Equipment Management," and "Process Monitoring," helping users overview and operate the loom status. The main display area presents the overall operational status of the loom in real time, including key indicators such as the number of online machines, real-time output, and process efficiency rankings. The work hour statistics and copyright information below make it an important tool for production management decisions.

[0124] Figure 5 This is a schematic diagram of another cloud platform interface according to an embodiment of this application. Figure 5This page displays a detailed view of the loom's operational status, including each machine's number, current status (running, stopped, offline), number of runs, and utilization rate. The page uses a rich color-coding system to visually represent the machine's real-time status: green indicates good operation, red indicates a malfunction, and gray indicates offline, allowing operators to quickly identify problematic machines. Meanwhile, the navigation bar on the left provides options such as "Home" and "Equipment Monitoring," ensuring users can easily access other functions.

[0125] Figure 6 This is a schematic diagram of another cloud platform interface according to an embodiment of this application. Figure 6 This page demonstrates the daily wage calculation and management functions for loom operators. It lists each employee's daily data, including fabric type name, fabric number, get off work hours, number of loops not included in output, actual number of slivers produced, production weight, number of loom loops, and overproduction quota. Through a series of preset algorithms, the daily wage for each worker is automatically calculated, taking into account multiple dimensions such as output, quality, and additional subsidies. The "Export," "Set Time," and "Interface Information" buttons on the right side of the page provide further data processing and system configuration options.

[0126] Figure 7 This is a structural diagram of a data processing apparatus according to an embodiment of this application, such as... Figure 7 As shown, the device includes:

[0127] The acquisition module 70 is used to acquire the operating data and production process data of the weaving machinery.

[0128] The first analysis module 72 is used to analyze the running data and production process data using a deep learning model to obtain information on yarn breakage faults in the weaving machinery.

[0129] The second analysis module 74 is used to analyze the operating data and production process data using a hybrid model to obtain the performance bottleneck information of the weaving machinery. The hybrid model includes a reinforcement learning model and a gradient boosting decision tree.

[0130] The determination module 76 is used to determine the optimization strategy for the weaving machinery based on the yarn breakage fault information and performance bottleneck information.

[0131] Optionally, the deep learning model is trained using the following method: acquiring historical operation data and historical production process data of the weaving machinery; acquiring an initial deep learning model combining a convolutional neural network and a long short-term memory network, wherein the convolutional neural network is used to extract spatial features from the historical operation data and historical production process data, and the long short-term memory network is used to capture temporal dependencies in the historical operation data and historical production process data; using transfer learning techniques, loading the feature extraction layer parameters of the pre-trained model into the convolutional neural network of the initial deep learning model to initialize the feature extraction part of the initial deep learning model, thus obtaining a feature initialization model; using Bayesian optimization to tune the hyperparameters of the feature initialization model, obtaining optimized hyperparameters, and training the feature initialization model based on the optimized hyperparameters to obtain the deep learning model.

[0132] Optionally, a hybrid model can be used to analyze operational and production process data to obtain performance bottleneck information of the weaving machinery. Specifically, this includes the following steps: using a gradient boosting decision tree to perform regression analysis and feature importance ranking on the operational and production process data to identify preset key influencing factors affecting the overall performance of the weaving machinery and their contribution levels; constructing the state space of a reinforcement learning model based on the preset key influencing factors and their contribution levels; generating different equipment control strategies within the state space through interaction with the simulated operating environment of the weaving machinery, and determining the optimal control strategy that meets preset requirements based on the performance changes caused by different equipment control strategies; and combining the preset key influencing factors with the optimal control strategy to obtain performance bottleneck information of the weaving machinery.

[0133] Optionally, the hybrid model is trained using the following method: acquiring historical operating data and historical production process data of the weaving machinery; acquiring an initial hybrid model combining a reinforcement learning model and a gradient boosting decision tree; using the gradient boosting decision tree to identify preset key influencing factors in the historical operating data and historical production process data, and constructing the state space and reward function of the reinforcement learning model based on the preset key influencing factors; training the reinforcement learning model in a simulation environment constructed using digital twin technology using a multi-objective optimization algorithm, wherein the strategy generated by the reinforcement learning model is verified by real-time simulation during the training process; and jointly determining the trained and verified reinforcement learning model and the gradient boosting decision tree as the hybrid model.

[0134] Optionally, a multi-objective optimization algorithm is used to train the reinforcement learning model in a simulation environment constructed using digital twin technology. Specifically, this includes the following steps: constructing a simulation environment for the operation of a weaving machine based on digital twin technology, and determining the parameters to be optimized in the reinforcement learning model as decision variables for a multi-objective optimization problem; using a multi-objective optimization algorithm to generate an initial population of decision variables, where each individual in the initial population represents a set of parameter configurations for the reinforcement learning model; loading the parameter configurations of each individual in the initial population into the reinforcement learning model, performing policy interactions in the simulation environment, and using the output, energy consumption, and quality indicators collected in the simulation environment as optimization objectives; calculating the fitness of each individual based on the collected multiple optimization objectives, and iteratively optimizing the initial population using non-dominated sorting and crowding calculations in the multi-objective optimization algorithm to generate a Pareto optimal solution set; selecting the optimal reinforcement learning model parameter configuration that satisfies preset conditions from the Pareto optimal solution set as the trained reinforcement learning model.

[0135] Optionally, based on yarn breakage fault information and performance bottleneck information, an optimization strategy for the weaving machinery is determined, specifically including the following steps: Analyzing the yarn breakage fault information to obtain a first parameter set, wherein the first parameter set includes: fault type, occurrence frequency, and production process parameters; analyzing the performance bottleneck information to obtain a second parameter set, wherein the second parameter set includes: limiting factors affecting the overall efficiency of the weaving machinery and the degree of influence of these limiting factors; using predefined optimization decision rules, comprehensively reasoning about the first and second parameter sets to obtain a preliminary optimization strategy set, wherein the preliminary optimization strategy set includes: adjusting equipment parameters, maintenance scheduling, and modifying the production plan; using digital twin technology to simulate the preliminary optimization strategy set in a virtual environment, and evaluating the strategies in the preliminary optimization strategy set based on the comprehensive evaluation indicators in the simulation results, wherein the comprehensive evaluation indicators include: output, quality, and energy consumption indicators; in the evaluated optimization strategy set, the target strategy whose comprehensive evaluation indicators meet the preset conditions is determined as the optimization strategy.

[0136] Optionally, the data processing device is further configured to perform the following steps: acquire weaving machinery operation data, wherein the weaving machinery operation data includes: start-up time, downtime, number of production loops, number of weaving loops, and quality pass rate; perform data cleaning and formatting on the weaving machinery operation data to obtain a standardized operation data sequence; calculate the production performance indicators of the target object based on the standardized operation data sequence, wherein the production performance indicators include: effective working hours, total production workload, and number of qualified products, wherein the effective working hours are calculated based on start-up time and downtime, and the number of qualified products is calculated based on the number of production loops, number of weaving loops, and quality pass rate; use a preset performance evaluation rule base to perform weighted calculation on the production performance indicators to obtain multi-dimensional scoring results, wherein the multi-dimensional scoring results include: output score, quality score, and comprehensive performance score, wherein the comprehensive performance score is calculated based on the output score and quality score; and fill the multi-dimensional scoring results, the target object's identity information, and time period information into a preset report template to generate a performance report.

[0137] Optionally, the data processing device is further configured to perform the following steps: acquiring static attribute data and dynamic operation data of the target object, wherein the static attribute data includes at least: skill level certification information, and the dynamic operation data includes at least: historical output target achievement rate, quality pass rate, and equipment operation proficiency index; calculating the comprehensive capability index of the target object based on the static attribute data and dynamic operation data through a capability assessment model, wherein the capability assessment model is used to assign different weights to different skill levels and different operation proficiency indicators; determining the capability requirement specifications of the production task according to the technical complexity, quality requirements, and delivery urgency of the production task to be assigned; calculating the matching degree between the comprehensive capability index of the target object and the capability requirement specifications of the production task, and generating a preliminary task allocation plan based on the matching degree calculation result; continuously collecting the operation data of the target object during the execution of the preliminary task allocation plan, and dynamically updating the comprehensive capability index of the target object based on the operation data of the target object; and adaptively adjusting the task allocation based on the updated comprehensive capability index.

[0138] It should be noted that the above Figure 7 Each module can be a program module (e.g., a set of program instructions that implements a specific function) or a hardware module. For the latter, it can take the following forms, but is not limited to them: each of the above modules is represented by a processor, or the functions of each of the above modules are implemented by a processor.

[0139] It should be noted that, Figure 7 Preferred embodiments of the examples can be found in [reference needed]. Figure 1 The relevant descriptions of the embodiments shown will not be repeated here.

[0140] Figure 8 A hardware structure block diagram of a computer terminal for implementing a data processing method is shown. Figure 8 As shown, the computer terminal 80 may include one or more processors 802 (shown as 802a, 802b, ..., 802n in the figure) 802 (processor 802 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 804 for storing data, and a transmission module 806 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 8 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, the computer terminal 80 may also include... Figure 8 The more or fewer components shown, or having the same Figure 8 The different configurations shown.

[0141] It should be noted that the aforementioned one or more processors 802 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 80. As involved in the embodiments of this application, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0142] The memory 804 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the data processing method in this embodiment. The processor 802 executes various functional applications and data processing by running the software programs and modules stored in the memory 804, thereby realizing the above-mentioned data processing method. The memory 804 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 804 may further include memory remotely located relative to the processor 802, and these remote memories can be connected to the computer terminal 80 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0143] The transmission module 806 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 80. In one example, the transmission module 806 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission module 806 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0144] The display can be, for example, a touchscreen liquid crystal display (LCD), which allows the user to interact with the user interface of the computer terminal 80.

[0145] It should be noted here that, in some optional embodiments, the above... Figure 8 The computer terminal shown may include hardware elements (including circuitry), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware and software elements. It should be noted that... Figure 8 This is only one instance of a specific particular instance, and is intended to illustrate the types of components that may exist in the aforementioned computer terminal.

[0146] It should be noted that, Figure 8 The computer terminal shown is used to execute Figure 1 The data processing method shown above is also applicable to this electronic device, and will not be repeated here.

[0147] This application also provides a non-volatile storage medium, which includes a stored program, wherein the program controls the device where the storage medium is located to execute the above data processing method when it runs.

[0148] A non-volatile storage medium performs the following functions: acquiring operating data and production process data of the weaving machine; analyzing the operating data and production process data using a deep learning model to obtain yarn breakage fault information of the weaving machine; analyzing the operating data and production process data using a hybrid model to obtain performance bottleneck information of the weaving machine, wherein the hybrid model includes: a reinforcement learning model and a gradient boosting decision tree; and determining an optimization strategy for the weaving machine based on the yarn breakage fault information and performance bottleneck information.

[0149] This application also provides an electronic device, including a memory and a processor, wherein the processor is used to run a program stored in the memory, wherein the program executes the above-described data processing method during runtime.

[0150] The processor is used to run programs that perform the following functions: acquire operating data and production process data of the weaving machine; analyze the operating data and production process data using a deep learning model to obtain yarn breakage fault information of the weaving machine; analyze the operating data and production process data using a hybrid model to obtain performance bottleneck information of the weaving machine, wherein the hybrid model includes: reinforcement learning model and gradient boosting decision tree; and determine optimization strategies for the weaving machine based on the yarn breakage fault information and performance bottleneck information.

[0151] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0152] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0153] In the above embodiments of this application, the information collected is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with relevant laws, regulations and standards, take necessary protective measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0154] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0155] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0156] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0157] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0158] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A data processing method, characterized in that, include: Obtain operational data and production process data of weaving machinery; By using a deep learning model to analyze the operational data and the production process data, the yarn breakage fault information of the weaving machine can be obtained. The performance bottleneck information of the weaving machine is obtained by analyzing the operational data and the production process data using a hybrid model, wherein the hybrid model includes: a reinforcement learning model and a gradient boosting decision tree; Based on the yarn breakage fault information and the performance bottleneck information, an optimization strategy for the weaving machine is determined.

2. The method according to claim 1, characterized in that, The deep learning model was trained using the following method: Obtain the historical operating data and historical production process data of the weaving machinery; An initial deep learning model combining a convolutional neural network and a long short-term memory network is obtained, wherein the convolutional neural network is used to extract spatial features from the historical operating data and the historical production process data, and the long short-term memory network is used to capture the temporal dependencies in the historical operating data and the historical production process data; By using transfer learning technology, the feature extraction layer parameters of the pre-trained model are loaded into the convolutional neural network of the initial deep learning model to initialize the feature extraction part of the initial deep learning model, thereby obtaining a feature initialization model. The hyperparameters of the feature initialization model are tuned using the Bayesian optimization method to obtain optimized hyperparameters. Based on the optimized hyperparameters, the feature initialization model is trained to obtain the deep learning model.

3. The method according to claim 1, characterized in that, By analyzing the operational data and production process data using a hybrid model, performance bottleneck information of the weaving machinery is obtained, including: The gradient boosting decision tree is used to perform regression analysis and feature importance ranking on the operational data and the production process data to identify the preset key influencing factors affecting the overall performance of the weaving machinery and the contribution of the preset key influencing factors. Based on the preset key influencing factors and their contribution, the state space of the reinforcement learning model is constructed. Within the state space, different equipment control strategies are generated through interaction with the simulated operating environment of the weaving machine, and the optimal control strategy that meets the preset requirements is determined based on the performance changes caused by the different equipment control strategies. By combining the preset key influencing factors with the optimal control strategy, the performance bottleneck information of the weaving machine is obtained.

4. The method according to claim 1 or 3, characterized in that, The hybrid model was trained using the following method: Obtain the historical operating data and historical production process data of the weaving machinery; Obtain an initial hybrid model that combines the reinforcement learning model and the gradient boosting decision tree; The gradient boosting decision tree is used to identify preset key influencing factors in the historical operating data and the historical production process data, and the state space and reward function of the reinforcement learning model are constructed based on the preset key influencing factors. The reinforcement learning model is trained in a simulation environment constructed using digital twin technology using a multi-objective optimization algorithm, wherein the policy generated by the reinforcement learning model is verified by real-time simulation during the training process. The trained and validated reinforcement learning model and the gradient boosting decision tree are jointly determined as the hybrid model.

5. The method according to claim 4, characterized in that, The reinforcement learning model is trained using a multi-objective optimization algorithm in a simulation environment constructed using digital twin technology, including: Based on the digital twin technology, a simulation environment for the operation of the weaving machinery is constructed, and the parameters to be optimized in the reinforcement learning model are determined as decision variables for a multi-objective optimization problem. A multi-objective optimization algorithm is used to generate an initial population of the decision variables, wherein each individual in the initial population represents a set of parameter configurations of the reinforcement learning model; The parameter configuration of each individual in the initial population is loaded into the reinforcement learning model, policy interaction is performed in the simulation environment, and the output, energy consumption and quality indicators collected in the simulation environment are used as optimization targets. Based on the collected multiple optimization objectives, the fitness of each individual is calculated, and the initial population is iteratively optimized through non-dominated sorting and crowding calculation in the multi-objective optimization algorithm to generate a Pareto optimal solution set; Select the optimal reinforcement learning model parameter configuration that satisfies the preset conditions from the Pareto optimal solution set, and use it as the trained reinforcement learning model.

6. The method according to claim 1, characterized in that, Based on the yarn breakage fault information and the performance bottleneck information, an optimization strategy for the weaving machine is determined, including: The yarn breakage fault information is analyzed to obtain a first parameter set, wherein the first parameter set includes: fault type, occurrence frequency and production process parameters; The performance bottleneck information is analyzed to obtain a second parameter set, wherein the second parameter set includes: limiting factors affecting the overall efficiency of the weaving machinery and the degree of influence of the limiting factors; By using predefined optimization decision rules, a preliminary optimization strategy set is obtained by comprehensively reasoning about the first parameter set and the second parameter set. The preliminary optimization strategy set includes: adjusting equipment parameters, maintenance scheduling, and modifying production plans. Using digital twin technology, the preliminary optimization strategy set is simulated in a virtual environment, and the strategies in the preliminary optimization strategy set are evaluated based on the comprehensive evaluation indicators in the simulation results. The comprehensive evaluation indicators include: output, quality and energy consumption indicators. In the set of optimized strategies that have been evaluated, the target strategy that meets the preset conditions for the comprehensive evaluation index is determined as the optimized strategy.

7. The method according to claim 1, characterized in that, The method further includes: Acquire weaving machinery operation data, which includes: machine start-up time, machine stop time, number of production turns, number of weaving turns, and quality pass rate; The operation data of the weaving machinery is cleaned and formatted to obtain a standardized operation data sequence; Based on the standardized operation data sequence, the production performance indicators of the target object are calculated, wherein the production performance indicators include: effective working hours, total production workload and number of qualified products, the effective working hours are calculated based on the start-up time and the downtime, and the number of qualified products is calculated based on the production cycle number, the weaving cycle number and the quality pass rate. Using a pre-set performance evaluation rule base, the production performance indicators are weighted and calculated to obtain multi-dimensional scoring results. The multi-dimensional scoring results include: output score, quality score, and comprehensive performance score. The comprehensive performance score is calculated based on the output score and the quality score. The multi-dimensional scoring results, the identity information of the target object, and the time period information are filled into a preset report template to generate a performance report.

8. The method according to claim 1, characterized in that, The method further includes: Obtain static attribute data and dynamic operation data of the target object, wherein the static attribute data includes at least: skill level certification information, and the dynamic operation data includes at least: historical output target achievement rate, quality pass rate and equipment operation proficiency index; Based on the static attribute data and dynamic operation data, the comprehensive ability index of the target object is calculated through the ability assessment model, wherein the ability assessment model is used to assign different weights to different skill levels and different operation proficiency indicators. The capacity requirement specifications for the production tasks to be assigned are determined based on their technical complexity, quality requirements, and delivery urgency. The matching degree between the comprehensive capability index of the target object and the capability requirement specification of the production task is calculated, and a preliminary task allocation plan is generated based on the matching degree calculation result. During the execution of the initial task allocation plan, the operation data of the target object is continuously collected, and the comprehensive capability index of the target object is dynamically updated based on the operation data of the target object; Task allocation is adaptively adjusted based on the updated comprehensive capability index.

9. A data processing apparatus, characterized in that, include: The acquisition module is used to acquire the operating data and production process data of the weaving machinery; The first analysis module is used to analyze the operating data and the production process data using a deep learning model to obtain the yarn breakage fault information of the weaving machine. The second analysis module is used to analyze the operating data and the production process data using a hybrid model to obtain the performance bottleneck information of the weaving machinery. The hybrid model includes a reinforcement learning model and a gradient boosting decision tree. The determination module is used to determine an optimization strategy for the weaving machine based on the yarn breakage fault information and the performance bottleneck information.

10. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the non-volatile storage medium to perform the data processing method according to any one of claims 1 to 8.

11. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when executed, performs the data processing method according to any one of claims 1 to 8.

12. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the data processing method according to any one of claims 1 to 8.

Images