Textile production management system

Abstract

The invention, which relates to the technical field of textile product production line management, discloses a textile production management system comprising a textile production parameter real-time monitoring and acquisition module. Production parameters of machine speed, temperature and humidity on a textile production line, operation state parameters of textile equipment and energy consumption information are monitored in real time, original data are obtained, and the original data acquired by monitoring are sent to an energy consumption monitoring optimization management module and a production scheduling dynamic management module for analysis and processing; in the textile production line, the working state information of the production equipment is collected, the production parameters on the textile production line are collected, the energy consumption information of the textile production line is monitored, and the energy consumption of the textile production line is optimized and managed through the energy consumption monitoring optimization management module; and the production scheduling dynamic management module dynamically adjusts equipment work and spinning operation.

Description

Technical Field

[0001] This invention relates to the field of textile production line management technology, specifically to a textile production management system. Background Technology

[0002] Textile production management systems cover the entire production process from raw material procurement to finished product delivery, including raw material management, production planning, production process control, quality management, inventory management, and sales management. The purpose of building a textile production management system is to optimize the production process, improve efficiency, reduce costs, and ensure product quality. However, in the actual production management process, some less common economic losses may occur due to slow technological updates, improper management and control, or human factors.

[0003] For example, application number CN202011598185.2 discloses a textile process management system. By setting up intelligent process generation units, production line management modules, and process management modules in the textile process management system, it achieves the effect of facilitating the adjustment and improvement of textile processes and realizes the function of effectively monitoring and managing the process flow, equipment information, and product quality under different textile processes. However, it lacks negative feedback regulation for the coordinated control of various unit modules of the textile process, and lacks abnormal monitoring and analysis reminders for the working status of various production process modules, resulting in incomplete management of textile equipment. In the existing textile production management system, if IoT, AI, big data analysis, and cloud computing technologies are introduced to monitor and analyze various units of the production line in real time and establish an AI-based intelligent feedback mechanism, although it can analyze production data in real time and automatically adjust production parameters to optimize the production process and ensure quality and efficiency, the system will face problems such as high operating costs, maintenance costs, and product upgrades, poor economic benefits, low market demand, and management difficulties.

[0004] Existing technologies have the following shortcomings: For current textile management systems, the optimization of production parameters such as machine speed, temperature, and humidity in automated and semi-automated textile production lines is insufficient. These parameters cannot be adjusted in real time to adapt to minor changes in production conditions, affecting product quality and consistency and increasing scrap rates. Furthermore, the lack of real-time monitoring and optimization parameters for energy consumption and resource utilization efficiency of textile equipment results in an inability to effectively reduce production costs and consequently, reduced profits for textile products. Finally, the insufficient ability to collect and analyze relevant parameters regarding the maintenance and fault diagnosis of textile production equipment leads to the inability to promptly detect or prevent equipment failures, resulting in production interruptions, increased maintenance costs, and decreased production efficiency.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a textile production management system. This invention optimizes the energy consumption of the textile production line through an energy consumption monitoring and optimization management module, dynamically adjusts equipment operation and textile operation through a production scheduling dynamic management module, and performs fault diagnosis and predictive early warning processing for production equipment through a textile production equipment fault diagnosis module, thereby solving the problems in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a textile production management system, including a real-time monitoring and acquisition module for textile production parameters: by deploying intelligent sensors on the textile production line, the system monitors the production parameters such as machine speed, temperature, and humidity on the textile production line, the operating status parameters of textile equipment, and energy consumption information in real time, obtains raw data, and sends the raw data collected by the monitoring to the energy consumption monitoring and optimization management module and the production scheduling dynamic management module for analysis and processing;

[0008] Energy Consumption Monitoring and Optimization Management Module: The module monitors the energy consumption information collected by the real-time monitoring and acquisition module for textile production parameters. It uses big data analysis technology and particle swarm optimization (PSO) intelligent algorithm to analyze the energy consumption information of the monitored textile equipment. It also uses the energy consumption analysis data to identify and optimize management, and transmits the analyzed energy consumption information to the textile production equipment fault diagnosis module.

[0009] Production scheduling dynamic management module: Receives real-time monitoring parameters of production parameters such as machine speed, temperature, and humidity on the textile production line, as well as the operating status parameters of textile equipment, collected by the real-time monitoring and acquisition module of textile production parameters. It uses mixed integer programming scheduling algorithm and simulated annealing algorithm to optimize production scheduling, and realizes dynamic adjustment of production plan and production line configuration according to market demand, raw material supply and production capacity.

[0010] Textile production equipment fault diagnosis module: Receives raw data and energy consumption analysis data collected by the real-time monitoring and acquisition module of textile production parameters, and uses a deep learning model to monitor the operating status of the production equipment, that is, uses the Long Short-Term Memory (LSTM) network diagnostic algorithm to predict and diagnose faults and obtain fault diagnosis information.

[0011] Comprehensive Analysis and Decision-Driven Module: This module uses data analysis and business intelligence (BI) software tools for data visualization and employs machine learning prediction models to comprehensively predict and analyze dynamic control data from the production scheduling dynamic management module, thereby obtaining comprehensive data visualization and analysis results.

[0012] Alarm Response Module: Based on the comprehensive analysis results of data visualization and fault diagnosis information, the module uses the decision tree classification algorithm in machine learning to determine the conditions for triggering an alarm and obtain the alarm signal to be activated. This enables the module to issue an alarm and notify relevant personnel or systems to take response measures when production anomalies, equipment failure warnings, or excessive energy consumption are detected.

[0013] Optionally, the steps for obtaining the raw data are as follows:

[0014] The production parameters on the textile production line are monitored in real time by deploying smart sensors on the textile production line and calibrated as MP. The production parameters include machine speed, temperature and humidity, which are calibrated as Ms, Tp and Hm respectively, i.e. MP = {Ms, Tp and Hm}.

[0015] The sensor monitors the operating status parameters of the textile equipment in real time and is calibrated as TSP. The operating status parameters include the equipment vibration characteristics, operating electrical parameters and sound characteristics, which are calibrated as Vfa, Ovc and Stv respectively, i.e. TSP = {Vfa, Ovc, Stv}.

[0016] The sensor monitors the energy consumption information of the textile production line in real time and is calibrated as ECI. The energy consumption information includes the unit product energy consumption rate, energy efficiency ratio and total energy consumption, which are calibrated as Sec, Eer and Tec respectively, i.e. ECI = {Sec, Eer and Tec}.

[0017] The original data, calibrated as OD, includes production parameters MP, equipment operating status parameters TSP, and production line energy consumption information ECI, i.e., OD = {MP, TSP, ECI}.

[0018] 3. The textile production management system according to claim 2, characterized in that the analysis logic steps of the Particle Swarm Optimization (PSO) intelligent algorithm for energy consumption information are as follows:

[0019] The obtained energy consumption information ECI, ECI = {Sec, Eer, Tec} is used to generate a set of particles to initialize the particle swarm.

[0020] The initial particle swarm is evaluated for fitness. Performance is calculated using a predefined fitness function based on unit product energy consumption rate, energy efficiency ratio, and total energy consumption. The formula for the fitness function is as follows: and as well as In the formula, f(Sec), f(Eer), and f(Tes) represent the fitness functions of unit product energy consumption rate, energy efficiency ratio, and total energy consumption Sec, Eer, and Tes, respectively. and Let f(Sec), f(Eer), and f(Tec) represent the unit product energy consumption rate, energy efficiency ratio, and total energy consumption of the textile production line, respectively. P(Sec) represents the total production volume under the unit product energy consumption rate Sec. α and β represent the weight parameters for the transformation of energy efficiency ratio Eer, respectively. C(Eer) represents the total efficiency under energy efficiency ratio Eer. i represents the individual that consumes energy on each textile production line, n represents the amount of energy consumed on the textile production line, and the fitness performance of particle swarm optimization is the optimal position F(ECI) = {f(Sec), f(Eer), f(Tec)}.

[0021] For each initialized particle, the fitness function is iteratively calculated to obtain the individual optimal solution and the population optimal solution, labeled as f′(Tec), f′(Eer), f′(Tes), and F′(ECI). The fitness function value of the current solution is compared with the individual optimal solution and the population optimal solution found after iteration, and the individual and population optimal solutions are updated accordingly.

[0022] The velocity and position of each particle are updated based on the individual optimal solution and the swarm optimal solution, and denoted as v and p respectively. Once the maximum number of iterations is reached, the formula for calculating the particle velocity update is v(ECI). t+1 =w×v(ECI) t +λ1×γ1×(F(ECI)-x(ECI) t )+λ2×γ2×(F′(ECI)-x(ECI) t In the formula, v(ECI) t+1 Let v(ECI) represent the velocity of the particle in the free dimension at time t+1, w represent the inertial weight controlling the particle's velocity, and v(ECI) represent the velocity of the particle in the free dimension. t Let x(ECI) represent the velocity of the particle's ECI in the free dimension at time t, λ1 and λ2 represent the learning factors of individual particle cognition and group cognition, respectively, and γ1 and γ2 represent random numbers in the interval [0,1]. t Let F(ECI) represent the position of particle ECI in the free dimension at time t, F'(ECI) represent the best position found by particle ECI, and F'(ECI) represent the best position of particle ECI in the current global context.

[0023] The formula for calculating the particle position update is x(ECI). t+1 =x(ECI) t +v(ECI) t+1 In the formula, x(ECI) t+1 It represents the position of particle ECI in the free dimension at time t+1.

[0024] Optionally, the steps of the energy consumption optimization management are as follows:

[0025] Sensors are used to monitor energy consumption information of textile production lines in real time, and the energy consumption information is preprocessed by data cleaning, missing value handling, and outlier detection for big data analysis and calculation of energy consumption.

[0026] Define the fitness function for initializing the particle swarm and use the PSO intelligent algorithm to find the optimal solution for energy consumption;

[0027] Based on the optimization results of the PSO algorithm, production parameters such as machine speed, temperature and humidity are adjusted according to the energy consumption information of the textile production line, or equipment operating status parameters such as equipment vibration characteristics, operating electrical parameters and sound characteristics are adjusted, and the operation mode of the textile production line is adjusted to reduce energy consumption by adjusting the batch production, continuous production, start-stop mode, energy recovery and utilization and production scheduling operation mode.

[0028] The optimized energy consumption information is fed back to the equipment fault diagnosis module to help determine whether the cause of the equipment fault is related to poor energy management.

[0029] Based on real-time data and feedback, we continuously adjust and improve our energy consumption optimization management strategies.

[0030] Optionally, the logical steps for optimizing production scheduling are as follows:

[0031] The production parameters MP = {Ms, Tp, Hm} and the textile equipment operating status parameters TSP = {Vfa, Ovc, Stv} on the textile production line will be collected in real time, including machine speed, temperature and humidity.

[0032] Based on the analysis of market demand, raw material supply and production capacity according to textile production volume, production targets and constraints are determined and labeled as Pt and Cc respectively;

[0033] To construct a mixed-integer programming scheduling model, we first use a mixed-integer programming algorithm to find the initial solution. The calculation formula for the mixed-integer programming scheduling algorithm is as follows: In the formula, Z min This can be expressed as an objective function that minimizes the total production time or cost. This is represented as production parameters MP, textile equipment operating status parameters TSP, production target Pt, and constraints Cc, along with the production time or cost on the textile production line. Let i0 and n0 represent the independent decision variables and the total number of decision variables, respectively.

[0034] Then, the simulated annealing algorithm is applied to improve and optimize the solution calculated by the mixed integer programming scheduling algorithm to find a better production scheduling scheme. The calculation formula of the simulated annealing algorithm is as follows: and In the formula, P(Z) min) represents the probability of accepting the new solution, e ξ Represented as the objective function value Z min The change Let ρ represent the probability that the objective function value will accept a new solution, and let ρ represent the cooling rate of the simulated annealing algorithm, where ρ∈[0,1]. This represents a new solution improved and optimized using simulated annealing;

[0035] To evaluate whether the optimized production scheduling scheme meets the production target Pt and the constraint Cc, the evaluation calculation formula is as follows: Furthermore, Text0 = δ1 × Pt + δ2 × Cc, where Text represents the optimization evaluation rate, η represents the evaluation parameters of each index of the objective function, Text0 represents the evaluation value of the production target Pt and the constraint Cc, and δ1 and δ2 represent the evaluation index quantities of the production target Pt and the constraint Cc, respectively. By comparing the magnitudes of Text and Text0, it is determined whether the optimized production scheduling scheme meets the predetermined production target and constraint conditions.

[0036] The optimized production plan and production line configuration are applied to actual textile production, and the production plan and configuration are dynamically adjusted based on real-time data and market feedback.

[0037] Optionally, the logical steps of the Long Short-Term Memory (LSTM) network diagnostic algorithm for fault prediction and diagnosis are as follows:

[0038] Features helpful for fault diagnosis are extracted from the original OD data and energy consumption analysis data F′(ECI), and the feature data is scaled to a range of [0,1] to facilitate processing by the Long Short-Term Memory (LSTM) network. The scaling formula for the feature data is as follows: and In the formula, This represents the result of OD normalization of the original data, (OD). min (OD) max These represent the minimum and maximum values ​​in the time series segment of the original data OD, respectively. This represents the result of normalizing the energy consumption analysis data F′(ECI). min F′(ECI) max These represent the minimum and maximum values ​​in the time series segment of the energy consumption analysis data F′(ECI), respectively.

[0039] The LSTM network structure was designed and trained, including an input layer, an LSTM layer, and an output layer. The input, LSTM, and output layers are constructed using a forget gate, an input gate, cell states, and an output gate, respectively. This allows the LSTM network to capture long-term dependencies in time-series data. The calculation formula for the forget gate is as follows: In the formula, y t Let h represent the activation vector of the forget gate, σ represent the activation function used to compress values ​​to between 0 and 1, W1 and W2 represent the weights learned through training for the original data OD and the energy consumption analysis data F′ (ECI) feature data, respectively. t-1 This represents the output at the previous time step t-1. The feature data at time step t is represented as follows. and The training set, b0 represents the bias parameters of the forget gate learned through training;

[0040] The calculation formula for the input gate is: and

[0041] In the formula, I t Let C be the activation vector of the input gate, where b1 and b2 are the bias parameters of the input gate learned through training. t ψ represents the candidate cell state, and ψ represents the output between [-1, 1].

[0042] The formula for calculating cell state is X. t =y t ×X t-1 +I t ×C t In the formula, X t X represents the new cell state at time step t. t-1 This represents the new cell state at the previous time step t-1;

[0043] The calculation formula for the output gate is: And H t =O t ×C t In the formula, O t Let b3 be the activation vector of the output gate, b3 be the bias parameters of the output gate learned through training, and H be the value of H. t This represents the output at the current time step t.

[0044] feature dataset and The dataset is divided into training and test sets. An LSTM model is trained using the training set data, and the mean squared error (MSE) is used as the loss function. The model weights are updated using the backpropagation algorithm. The formula for calculating the MSE loss function is as follows: In the formula, L(MSE) represents the loss value of mean squared error, N represents the total number of samples in the feature dataset, and i′ represents an independent individual in the feature dataset. These are respectively represented as feature data. and The predicted value, and These are represented as the true values ​​of N samples;

[0045] The trained LSTM model is used to make predictions on the newly collected data to identify failure modes;

[0046] The model's accuracy and recall performance metrics are evaluated by comparing it with actual failure data on the test set, and the model structure or performance parameters are adjusted as needed.

[0047] Optionally, the data visualization process includes the following steps:

[0048] Collect data from the production scheduling dynamic management module, and perform data cleaning, transformation, and organization for visualization processing;

[0049] Choose Power BI from the business intelligence (BI) software tools, select a bar chart visualization based on the characteristics of the data and the purpose of analysis, map the organized data onto the selected visualization chart, and display and present it.

[0050] It also designs interactive features to allow users to interact with data and conduct in-depth analysis as needed;

[0051] Interpret and analyze the generated visualization results to discover the correlations, trends, and anomalies between the data.

[0052] Optionally, the predictive analysis logic steps of the machine learning prediction model are as follows:

[0053] A new solution for improvements and optimizations from the production scheduling dynamic management module. Preprocessing is performed, i.e., the data is used as new feature data.

[0054] To build a machine learning prediction model, linear regression is used to predict new feature data. The formula for linear regression is: In the formula, Represented as new feature data The predicted value is denoted by W′, which represents the linear regression parameter in the machine learning prediction model, and ω0 represents the intercept constant.

[0055] Decisions are driven by the prediction results, guiding the dynamic adjustment of the production scheduling dynamic management module.

[0056] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0057] This invention collects operational status information of production equipment in a textile production line, gathers production parameters, and monitors energy consumption. Through an energy consumption monitoring and optimization management module, it optimizes energy consumption; a production scheduling dynamic management module dynamically adjusts equipment operation and textile production; and a textile production equipment fault diagnosis module diagnoses and predicts equipment faults. This achieves optimized energy use, reduced production costs, improved energy efficiency, increased production efficiency and textile output, reduced impact of equipment failures on production, improved equipment reliability and stability, and intelligent management and control of the textile production line. This provides strong support for the sustainable development and industrial upgrading of textile enterprises. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0059] Figure 1 This is a schematic diagram of the modules of the textile production management system of the present invention. Detailed Implementation

[0060] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0061] This invention provides, for example Figure 1 The textile production management system shown includes a real-time monitoring and acquisition module for textile production parameters: by deploying intelligent sensors on the textile production line, it monitors the production parameters such as machine speed, temperature, and humidity, the operating status parameters of textile equipment, and energy consumption information on the textile production line in real time, obtains raw data, and sends the raw data collected by the monitoring to the energy consumption monitoring and optimization management module and the production scheduling dynamic management module for analysis and processing.

[0062] Specifically, the steps for obtaining the raw data are as follows:

[0063] The production parameters on the textile production line are monitored in real time by deploying smart sensors on the textile production line and calibrated as MP. The production parameters include machine speed, temperature and humidity, which are calibrated as Ms, Tp and Hm respectively, i.e. MP = {Ms, Tp and Hm}.

[0064] The sensor monitors the operating status parameters of the textile equipment in real time and is calibrated as TSP. The operating status parameters include the equipment vibration characteristics, operating electrical parameters and sound characteristics, which are calibrated as Vfa, Ovc and Stv respectively, i.e. TSP = {Vfa, Ovc, Stv}. Among them, the equipment vibration characteristics are the vibration frequency and amplitude of the textile equipment during operation, the operating electrical parameters are the voltage and current of the textile equipment during operation, and the sound characteristics are the timbre and volume of the textile equipment during operation.

[0065] The sensor monitors the energy consumption information of the textile production line in real time and is calibrated as ECI. The energy consumption information includes the unit product energy consumption rate, energy efficiency ratio and total energy consumption, which are calibrated as Sec, Eer and Tec respectively, i.e. ECI = {Sec, Eer and Tec}.

[0066] The raw data, calibrated as OD, includes production parameters MP, equipment operating status parameters TSP, and production line energy consumption information ECI, i.e., OD = {MP, TSP, ECI}.

[0067] The initial data in the form of electrical signals collected by the smart sensor is sent to the processor via wireless network communication technology for preliminary processing such as format conversion and noise filtering to obtain raw data OD, which is then used for subsequent analysis and processing.

[0068] Energy Consumption Monitoring and Optimization Management Module: The module monitors the energy consumption information collected by the real-time monitoring and acquisition module for textile production parameters. It uses big data analysis technology and particle swarm optimization (PSO) intelligent algorithm to analyze the energy consumption information of the monitored textile equipment. The module uses the energy consumption analysis data to identify and optimize management, implement energy-saving measures, reduce energy waste, and transmit the analyzed energy consumption information to the textile production equipment fault diagnosis module to further determine the cause of equipment failure.

[0069] Specifically, the logical steps of the Particle Swarm Optimization (PSO) intelligent algorithm for analyzing energy consumption information are as follows:

[0070] The obtained energy consumption information ECI, ECI = {Sec, Eer, Tec} is used to generate a set of particles to initialize the particle swarm.

[0071] The initial particle swarm is evaluated for fitness. Performance is calculated using a predefined fitness function based on unit product energy consumption rate, energy efficiency ratio, and total energy consumption. The formula for the fitness function is as follows: and as well as In the formula, f(Sec), f(Eer), and f(Tes) represent the fitness functions of unit product energy consumption rate, energy efficiency ratio, and total energy consumption Sec, Eer, and Tes, respectively. and Let f(Sec), f(Eer), and f(Tec) represent the unit product energy consumption rate, energy efficiency ratio, and total energy consumption of the textile production line, respectively. P(Sec) represents the total production volume under the unit product energy consumption rate Sec. α and β represent the weight parameters for the transformation of energy efficiency ratio Eer, respectively. C(Eer) represents the total efficiency under energy efficiency ratio Eer. i represents the individual that consumes energy on each textile production line, n represents the amount of energy consumed on the textile production line, and the fitness performance of particle swarm optimization is the optimal position F(ECI) = {f(Sec), f(Eer), f(Tec)}.

[0072] For each initialized particle, the fitness function is iteratively calculated to obtain the individual optimal solution and the population optimal solution, labeled as f′(Tec), f′(Eer), f′(Tes), and F′(ECI). The fitness function value of the current solution is compared with the individual optimal solution and the population optimal solution found after iteration, and the individual and population optimal solutions are updated accordingly.

[0073] The velocity and position of each particle are updated based on the individual optimal solution and the swarm optimal solution, and denoted as v and p respectively. Once the maximum number of iterations is reached, the formula for calculating the particle velocity update is v(ECI). t+1 =w×v(ECI) t +λ1×γ1×(F(ECI)-x(ECI) t )+λ2×γ2×(F′(ECI)-x(ECI) t In the formula, v(ECI) t+1 Let v(ECI) represent the velocity of the particle in the free dimension at time t+1, w represent the inertial weight controlling the particle's velocity, and v(ECI) represent the velocity of the particle in the free dimension. t Let x(ECI) represent the velocity of the particle's ECI in the free dimension at time t, λ1 and λ2 represent the learning factors of individual particle cognition and group cognition, respectively, and γ1 and γ2 represent random numbers in the interval [0,1]. t Let F(ECI) represent the position of particle ECI in the free dimension at time t, F'(ECI) represent the best position found by particle ECI, and F'(ECI) represent the best position of particle ECI in the current global context.

[0074] The formula for calculating the particle position update is x(ECI). t+1 =x(ECI) t +v(ECI) t+1 In the formula, x(ECI) t+1 It represents the position of particle ECI in the free dimension at time t+1.

[0075] Specifically, the steps for energy consumption optimization management are as follows:

[0076] Sensors are used to monitor energy consumption information of textile production lines in real time, and the energy consumption information is preprocessed by data cleaning, missing value handling, and outlier detection for big data analysis and calculation of energy consumption.

[0077] Define the fitness function for initializing the particle swarm and use the PSO intelligent algorithm to find the optimal solution for energy consumption;

[0078] Based on the optimization results of the PSO algorithm, production parameters such as machine speed, temperature and humidity are adjusted according to the energy consumption information of the textile production line, or equipment operating status parameters such as equipment vibration characteristics, operating electrical parameters and sound characteristics are adjusted, and the operation mode of the textile production line is adjusted to reduce energy consumption and implement energy-saving measures.

[0079] The optimized energy consumption information is fed back to the equipment fault diagnosis module to help determine whether the cause of the equipment fault is related to poor energy management.

[0080] Based on real-time data and feedback, we continuously adjust and improve energy consumption optimization management strategies. These strategies include energy consumption benchmarks for the production process, predictive analysis of energy consumption data, adjustment of production parameters and operating modes, and updating and upgrading textile production line equipment.

[0081] Production scheduling dynamic management module: Receives real-time monitoring parameters of production parameters such as machine speed, temperature, and humidity on the textile production line, as well as the operating status parameters of textile equipment, collected by the real-time monitoring and acquisition module of textile production parameters. It uses mixed integer programming scheduling algorithm and simulated annealing algorithm to optimize production scheduling, realizes dynamic adjustment of production plan and production line configuration according to market demand, raw material supply and production capacity, and improves production efficiency and response speed.

[0082] Specifically, the logical steps for optimizing production scheduling are as follows:

[0083] The production parameters MP = {Ms, Tp, Hm} and the textile equipment operating status parameters TSP = {Vfa, Ovc, Stv} on the textile production line will be collected in real time, including machine speed, temperature and humidity.

[0084] Based on the analysis of market demand, raw material supply and production capacity according to textile production volume, production targets and constraints are determined and labeled as Pt and Cc respectively;

[0085] To construct a mixed-integer programming scheduling model, we first use a mixed-integer programming algorithm to find the initial solution. The calculation formula for the mixed-integer programming scheduling algorithm is as follows: In the formula, Z min This can be expressed as an objective function that minimizes the total production time or cost. This is represented as production parameters MP, textile equipment operating status parameters TSP, production target Pt, and constraints Cc, along with the production time or cost on the textile production line. Let i0 and n0 represent the independent decision variables and the total number of decision variables, respectively.

[0086] Then, the simulated annealing algorithm is applied to improve and optimize the solution calculated by the mixed integer programming scheduling algorithm to find a better production scheduling scheme. The calculation formula of the simulated annealing algorithm is as follows: and In the formula, P(Z) min ) represents the probability of accepting the new solution, e ξ Represented as the objective function value Z min The change Let ρ represent the probability that the objective function value will accept a new solution, and let ρ represent the cooling rate of the simulated annealing algorithm, where ρ∈[0,1]. This represents a new solution improved and optimized using simulated annealing;

[0087] To evaluate whether the optimized production scheduling scheme meets the production target Pt and the constraint Cc, the evaluation calculation formula is as follows: Furthermore, Text0 = δ1 × Pt + δ2 × Cc, where Text represents the optimization evaluation rate, η represents the evaluation parameters of each index of the objective function, Text0 represents the evaluation value of the production target Pt and the constraint Cc, and δ1 and δ2 represent the evaluation index quantities of the production target Pt and the constraint Cc, respectively. By comparing the magnitudes of Text and Text0, it is determined whether the optimized production scheduling scheme meets the predetermined production target and constraint conditions. When Text ≥ Text0, the optimized production scheduling scheme meets the predetermined production target and constraint conditions, and vice versa.

[0088] The optimized production plan and production line configuration are applied to actual textile production, and the production plan and configuration are dynamically adjusted based on real-time data and market feedback.

[0089] Textile production equipment fault diagnosis module: Receives raw data and energy consumption analysis data collected by the real-time monitoring and acquisition module of textile production parameters, and uses a deep learning model to monitor the operating status of the production equipment. Specifically, it uses the Long Short-Term Memory (LSTM) network diagnostic algorithm to predict and diagnose faults, obtain fault diagnosis information, and use it for timely repair or replacement to reduce downtime.

[0090] Specifically, the logical steps of the Long Short-Term Memory (LSTM) network diagnostic algorithm for fault prediction and diagnosis are as follows:

[0091] Features helpful for fault diagnosis are extracted from the original OD data and energy consumption analysis data F′(ECI), and the feature data is scaled to a range of [0,1] to facilitate processing by the Long Short-Term Memory (LSTM) network. The scaling formula for the feature data is as follows: and In the formula, This represents the result of OD normalization of the original data, (OD). min (OD) max These represent the minimum and maximum values ​​in the time series segment of the original data OD, respectively. This represents the result of normalizing the energy consumption analysis data F′(ECI). min F′(ECI) max These represent the minimum and maximum values ​​in the time series segment of the energy consumption analysis data F′(ECI), respectively.

[0092] The LSTM network structure was designed and trained, including an input layer, an LSTM layer, and an output layer. The input, LSTM, and output layers are constructed using a forget gate, an input gate, cell states, and an output gate, respectively. This allows the LSTM network to capture long-term dependencies in time-series data. The calculation formula for the forget gate is as follows: In the formula, y t Let h represent the activation vector of the forget gate, σ represent the activation function used to compress values ​​to between 0 and 1, W1 and W2 represent the weights learned through training for the original data OD and the energy consumption analysis data F′ (ECI) feature data, respectively. t-1 This represents the output at the previous time step t-1. The feature data at time step t is represented as follows. and The training set, b0 represents the bias parameters of the forget gate learned through training;

[0093] The calculation formula for the input gate is: and In the formula, I t Let C be the activation vector of the input gate, where b1 and b2 are the bias parameters of the input gate learned through training. t ψ represents the candidate cell state, and ψ represents the output between [-1, 1].

[0094] The formula for calculating cell state is X. t =y t ×X t-1 +I t ×C t In the formula, X t X represents the new cell state at time step t. t-1This represents the new cell state at the previous time step t-1;

[0095] The calculation formula for the output gate is: And H t =O t ×C t In the formula, O t Let b3 be the activation vector of the output gate, b3 be the bias parameters of the output gate learned through training, and H be the value of H. t This represents the output at the current time step t.

[0096] feature dataset and The dataset is divided into training and test sets. An LSTM model is trained using the training set data, and the mean squared error (MSE) is used as the loss function. The model weights are updated using the backpropagation algorithm. The formula for calculating the MSE loss function is as follows: In the formula, L(MSE) represents the loss value of mean squared error, N represents the total number of samples in the feature dataset, and i′ represents an independent individual in the feature dataset. These are respectively represented as feature data. and The predicted value, and These are represented as the true values ​​of N samples;

[0097] The trained LSTM model is used to predict newly collected data to identify failure modes, and corresponding maintenance or textile production line operation adjustment measures are taken based on the model prediction results to reduce downtime.

[0098] The model's accuracy and recall performance metrics are evaluated by comparing it with actual fault data on the test set, and the model structure or performance parameters are adjusted as needed to improve diagnostic accuracy.

[0099] The comprehensive analysis and decision-driven module uses data analysis and business intelligence (BI) software tools for data visualization and employs machine learning prediction models to comprehensively predict and analyze dynamic control data from the production scheduling dynamic management module. This yields comprehensive data visualization and analysis results to provide decision support, optimize production management and operational efficiency, and meet more accurate market demand forecasting and resource planning needs. The data analysis and BI software tools selected are any one of the following: Python's Pandas library, R language, Tableau, Power BI, or Looker.

[0100] Specifically, the steps for data visualization processing are as follows:

[0101] Collect data from the production scheduling dynamic management module, and perform data cleaning, transformation, and organization for visualization processing;

[0102] Choose Power BI from the business intelligence (BI) software tools, select a bar chart visualization based on the characteristics of the data and the purpose of analysis, map the organized data onto the selected visualization chart, and display and present it. Add necessary labels, titles, and legends to enhance readability.

[0103] It also designs interactive features to allow users to interact with data and conduct in-depth analysis as needed;

[0104] The generated visualizations are interpreted and analyzed to identify correlations, trends, and anomalies among the data, providing support for decision-making.

[0105] Specifically, the predictive analysis logic steps of the machine learning prediction model are as follows:

[0106] A new solution for improvements and optimizations from the production scheduling dynamic management module. Preprocessing is performed, i.e., the data is used as new feature data.

[0107] To build a machine learning prediction model, linear regression is used to predict new feature data. The formula for linear regression is: In the formula, Represented as new feature data The predicted value is denoted by W′, which represents the linear regression parameter in the machine learning prediction model, and ω0 represents the intercept constant.

[0108] Decision-making is driven by the forecast results, which guides the dynamic control of the production scheduling dynamic management module. The dynamic control strategies include production plan adjustment, equipment operation control, material supply allocation, energy consumption control, data analysis and optimization, quality control and fault prevention.

[0109] Alarm Response Module: Based on the comprehensive analysis results of data visualization and fault diagnosis information, the module uses the decision tree classification algorithm in machine learning to determine the conditions for triggering an alarm and obtain the alarm signal to be activated. This enables the module to issue an alarm and notify relevant personnel or systems to take response measures when production abnormalities, equipment failure warnings, or excessive energy consumption are detected.

[0110] The modules are then closely linked through data sharing and information feedback mechanisms, forming a closed-loop management system.

[0111] Specifically, the logic steps for obtaining the alarm signal are as follows:

[0112] A decision tree model is established, and the model is trained, learned, evaluated, and adjusted based on the comprehensive analysis results of data visualization and fault diagnosis information to obtain a well-trained decision tree model. The establishment of the decision tree model is achieved by selecting the optimal splitting feature by calculating the Gini coefficient or information gain of different features.

[0113] Then, the data visualization and comprehensive analysis results and fault diagnosis information are input into the trained decision tree model, and the model's output results are used to determine whether to trigger an alarm.

[0114] When the alarm triggering conditions are met, an alarm signal is generated to notify relevant personnel or the system to take response measures.

[0115] The specific methods and processes of the textile production management system provided in this embodiment of the invention are detailed in the above-described embodiment of the textile production management system, and will not be repeated here.

[0116] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0117] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0118] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0119] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0120] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A textile production management system, characterized in that, This includes a real-time monitoring and acquisition module for textile production parameters: by deploying intelligent sensors on the textile production line, it monitors production parameters such as machine speed, temperature, and humidity, as well as the operating status parameters of textile equipment and energy consumption information in real time, obtains raw data, and sends the raw data collected by the monitoring to the energy consumption monitoring and optimization management module and the production scheduling dynamic management module for analysis and processing. Energy Consumption Monitoring and Optimization Management Module: The module monitors the energy consumption information collected by the real-time monitoring and acquisition module for textile production parameters. It uses big data analysis technology and particle swarm optimization (PSO) intelligent algorithm to analyze the energy consumption information of the monitored textile equipment. It also uses the energy consumption analysis data to identify and optimize management, and transmits the analyzed energy consumption information to the textile production equipment fault diagnosis module. Production scheduling dynamic management module: Receives real-time monitoring parameters of production parameters such as machine speed, temperature, and humidity on the textile production line, as well as the operating status parameters of textile equipment, collected by the real-time monitoring and acquisition module of textile production parameters. It uses mixed integer programming scheduling algorithm and simulated annealing algorithm to optimize production scheduling, and realizes dynamic adjustment of production plan and production line configuration according to market demand, raw material supply and production capacity. Textile production equipment fault diagnosis module: Receives raw data and energy consumption analysis data collected by the real-time monitoring and acquisition module of textile production parameters, and uses a deep learning model to monitor the operating status of the production equipment, that is, uses the Long Short-Term Memory (LSTM) network diagnostic algorithm to predict and diagnose faults and obtain fault diagnosis information. Comprehensive Analysis and Decision-Driven Module: This module uses data analysis and business intelligence (BI) software tools for data visualization and employs machine learning prediction models to comprehensively predict and analyze dynamic control data from the production scheduling dynamic management module, thereby obtaining comprehensive data visualization and analysis results. Alarm Response Module: Based on the comprehensive analysis results of data visualization and fault diagnosis information, the module uses the decision tree classification algorithm in machine learning to determine the conditions for triggering an alarm and obtain the alarm signal to be activated. This enables the module to issue an alarm and notify relevant personnel or systems to take response measures when production anomalies, equipment failure warnings, or excessive energy consumption are detected.

2. The textile production management system according to claim 1, characterized in that, The steps for obtaining the raw data are as follows: The production parameters on the textile production line are monitored in real time by deploying smart sensors on the textile production line and calibrated as MP. The production parameters include machine speed, temperature and humidity, which are calibrated as Ms, Tp and Hm respectively, i.e. MP = {Ms, Tp and Hm}. The sensor monitors the operating status parameters of the textile equipment in real time and is calibrated as TSP. The operating status parameters include the equipment vibration characteristics, operating electrical parameters and sound characteristics, which are calibrated as Vfa, Ovc and Stv respectively, i.e. TSP = {Vfa, Ovc, Stv}. The sensor monitors the energy consumption information of the textile production line in real time and is calibrated as ECI. The energy consumption information includes the unit product energy consumption rate, energy efficiency ratio and total energy consumption, which are calibrated as Sec, Eer and Tec respectively, i.e. ECI = {Sec, Eer and Tec}. The original data, calibrated as OD, includes production parameters MP, equipment operating status parameters TSP, and production line energy consumption information ECI, i.e., OD = {MP, TSP, ECI}.

3. A textile production management system according to claim 2, characterized in that, The logical steps of the Particle Swarm Optimization (PSO) intelligent algorithm for analyzing energy consumption information are as follows: The obtained energy consumption information ECI, ECI = {Sec, Eer, Tec} is used to generate a set of particles to initialize the particle swarm. The initial particle swarm is evaluated for fitness. Performance is calculated using a predefined fitness function based on unit product energy consumption rate, energy efficiency ratio, and total energy consumption. The formula for the fitness function is as follows: and as well as In the formula, f(Sec), f(Eer), and f(Tes) represent the fitness functions of unit product energy consumption rate, energy efficiency ratio, and total energy consumption Sec, Eer, and Tes, respectively. and Let f(Sec), f(Eer), and f(Tec) represent the unit product energy consumption rate, energy efficiency ratio, and total energy consumption of the textile production line, respectively. P(Sec) represents the total production volume under the unit product energy consumption rate Sec. α and β represent the weight parameters for the transformation of energy efficiency ratio Eer, respectively. C(Eer) represents the total efficiency under energy efficiency ratio Eer. i represents the individual that consumes energy on each textile production line, n represents the amount of energy consumed on the textile production line, and the fitness performance of particle swarm optimization is the optimal position F(ECI) = {f(Sec), f(Eer), f(Tec)}. For each initialized particle, the fitness function is iteratively calculated to obtain the individual optimal solution and the population optimal solution, labeled as f′(Tec), f′(Eer), f′(Tes), and F′(ECI). The fitness function value of the current solution is compared with the individual optimal solution and the population optimal solution found after iteration, and the individual and population optimal solutions are updated accordingly. The velocity and position of each particle are updated based on the individual optimal solution and the swarm optimal solution, and denoted as v and p respectively. Once the maximum number of iterations is reached, the formula for calculating the particle velocity update is v(ECI). t+1 =w×v(ECI) t +λ1×γ1×(F(ECI)-x(ECI) t )+λ2×γ2×(F′(ECI)-x(ECI) t In the formula, v(ECI) t+1 Let v(ECI) represent the velocity of the particle in the free dimension at time t+1, w represent the inertial weight controlling the particle's velocity, and v(ECI) represent the velocity of the particle in the free dimension. t Let x(ECI) represent the velocity of the particle's ECI in the free dimension at time t, λ1 and λ2 represent the learning factors of individual particle cognition and group cognition, respectively, and γ1 and γ2 represent random numbers in the interval [0,1]. t Let F(ECI) represent the position of particle ECI in the free dimension at time t, F'(ECI) represent the best position found by particle ECI, and F'(ECI) represent the best position of particle ECI in the current global context. The formula for calculating the particle position update is x(ECI). t+1 =x(ECI) t +v(ECI) t+1 In the formula, x(ECI) t+1 It represents the position of particle ECI in the free dimension at time t+1.

4. A textile production management system according to claim 3, characterized in that, The steps for energy consumption optimization management are as follows: Sensors are used to monitor energy consumption information of textile production lines in real time, and the energy consumption information is preprocessed by data cleaning, missing value handling, and outlier detection for big data analysis and calculation of energy consumption. Define the fitness function for initializing the particle swarm and use the PSO intelligent algorithm to find the optimal solution for energy consumption; Based on the optimization results of the PSO algorithm, production parameters such as machine speed, temperature and humidity are adjusted according to the energy consumption information of the textile production line, or equipment operating status parameters such as equipment vibration characteristics, operating electrical parameters and sound characteristics are adjusted, and the operation mode of the textile production line is adjusted to reduce energy consumption by adjusting the batch production, continuous production, start-stop mode, energy recovery and utilization and production scheduling operation mode. The optimized energy consumption information is fed back to the equipment fault diagnosis module to help determine whether the cause of the equipment fault is related to poor energy management. Based on real-time data and feedback, we continuously adjust and improve our energy consumption optimization management strategies.

5. A textile production management system according to claim 4, characterized in that, The logical steps for optimizing production scheduling are as follows: The production parameters MP = {Ms, Tp, Hm} and the textile equipment operating status parameters TSP = {Vfa, Ovc, Stv} on the textile production line will be collected in real time, including machine speed, temperature and humidity. Based on the analysis of market demand, raw material supply and production capacity according to textile production volume, production targets and constraints are determined and labeled as Pt and Cc respectively; To construct a mixed-integer programming scheduling model, we first use a mixed-integer programming algorithm to find the initial solution. The calculation formula for the mixed-integer programming scheduling algorithm is as follows: In the formula, Z min This can be expressed as an objective function that minimizes the total production time or cost. This is represented as production parameters MP, textile equipment operating status parameters TSP, production target Pt, and constraints Cc, along with the production time or cost on the textile production line. Let i0 and n0 represent the independent decision variables and the total number of decision variables, respectively. Then, by applying the simulated annealing algorithm to improve and optimize the solution calculated by the mixed integer programming scheduling algorithm, a better production scheduling scheme can be found. The calculation formula of the simulated annealing algorithm is as follows: and In the formula, P(Z) min ) represents the probability of accepting the new solution, e ξ Represented as the objective function value Z min The change Let ρ represent the probability that the objective function value will accept a new solution, and let ρ represent the cooling rate of the simulated annealing algorithm, where ρ∈[0,1]. This represents a new solution improved and optimized using simulated annealing. To evaluate whether the optimized production scheduling scheme meets the production target Pt and the constraint Cc, the evaluation calculation formula is as follows: Furthermore, Text0 = δ1 × Pt + δ2 × Cc, where Text represents the optimization evaluation rate, η represents the evaluation parameters of each index of the objective function, Text0 represents the evaluation value of the production target Pt and the constraint Cc, and δ1 and δ2 represent the evaluation index quantities of the production target Pt and the constraint Cc, respectively. By comparing the magnitudes of Text and Text0, it is determined whether the optimized production scheduling scheme meets the predetermined production target and constraint conditions. The optimized production plan and production line configuration are applied to actual textile production, and the production plan and configuration are dynamically adjusted based on real-time data and market feedback.

6. A textile production management system according to claim 5, characterized in that, The logical steps of the Long Short-Term Memory (LSTM) network diagnostic algorithm for fault prediction and diagnosis are as follows: Features helpful for fault diagnosis are extracted from the original OD data and energy consumption analysis data F′(ECI), and the feature data is scaled to a range of [0,1] to facilitate processing by the Long Short-Term Memory (LSTM) network. The scaling formula for the feature data is as follows: and In the formula, This represents the result of OD normalization of the original data, (OD). min (OD) max These represent the minimum and maximum values ​​in the time series segment of the original data OD, respectively. This represents the result of normalizing the energy consumption analysis data F′(ECI). min F′(ECI) max These represent the minimum and maximum values ​​in the time series segment of the energy consumption analysis data F′(ECI), respectively. The LSTM network structure was designed and trained, including an input layer, an LSTM layer, and an output layer. The input, LSTM, and output layers are constructed using a forget gate, an input gate, cell states, and an output gate, respectively. This allows the LSTM network to capture long-term dependencies in time-series data. The calculation formula for the forget gate is as follows: In the formula, y t Let h represent the activation vector of the forget gate, σ represent the activation function used to compress values ​​to between 0 and 1, W1 and W2 represent the weights learned through training for the original data OD and the energy consumption analysis data F′ (ECI) feature data, respectively. t-1 This represents the output at the previous time step t-1. The feature data at time step t is represented as follows. and The training set, b0 represents the bias parameters of the forget gate learned through training; The calculation formula for the input gate is: and In the formula, I t Let C be the activation vector of the input gate, where b1 and b2 are the bias parameters of the input gate learned through training. t ψ represents the candidate cell state, and ψ represents the output between [-1, 1]. The formula for calculating cell state is X t =y t ×X t-1 +I t ×C t In the formula, X t X represents the new cell state at time step t. t-1 This represents the new cell state at the previous time step t-1; The calculation formula for the output gate is: And H t =O t ×C t In the formula, O t Let b3 be the activation vector of the output gate, b3 be the bias parameters of the output gate learned through training, and H be the value of H. t This represents the output at the current time step t. feature dataset and The dataset is divided into training and test sets. An LSTM model is trained using the training set data, and the mean squared error (MSE) is used as the loss function. The model weights are updated using the backpropagation algorithm. The formula for calculating the MSE loss function is as follows: In the formula, L(MSE) represents the loss value of mean squared error, N represents the total number of samples in the feature dataset, and i′ represents an independent individual in the feature dataset. These are respectively represented as feature data. and The predicted value, and These are represented as the true values ​​of N samples; The trained LSTM model is used to make predictions on the newly collected data to identify failure modes; The model's accuracy and recall performance metrics are evaluated by comparing it with actual failure data on the test set, and the model structure or performance parameters are adjusted as needed.

7. A textile production management system according to claim 6, characterized in that, The steps for data visualization processing are as follows: Collect data from the production scheduling dynamic management module, and perform data cleaning, transformation, and organization for visualization processing; Choose Power BI from the business intelligence (BI) software tools, select a bar chart visualization based on the characteristics of the data and the purpose of analysis, map the organized data onto the selected visualization chart, and display and present it. It also designs interactive features to allow users to interact with data and conduct in-depth analysis as needed; Interpret and analyze the generated visualization results to discover the correlations, trends, and anomalies between the data.

8. A textile production management system according to claim 7, characterized in that, The predictive analysis logic steps of the machine learning prediction model are as follows: A new solution for improvements and optimizations from the production scheduling dynamic management module. Preprocessing is performed, i.e., the data is used as new feature data. To build a machine learning prediction model, linear regression is used to predict new feature data. The formula for linear regression is: In the formula, Represented as new feature data The predicted value is denoted by W′, which represents the linear regression parameter in the machine learning prediction model, and ω0 represents the intercept constant. Decisions are driven by the prediction results, guiding the dynamic adjustment of the production scheduling dynamic management module.

Images