Hot air setting machine process parameter optimization method based on big data analysis

Abstract

The invention relates to the field of process parameter optimization, and discloses a hot air setting machine process parameter optimization method based on big data analysis, and the system comprises a data collection module, a data intelligent storage module, a network optimization module, and an intelligent decision module. The data acquisition module is used for respectively acquiring characteristic data of the shaped fabric and various equipment operation parameters in the shaping process by using CT scanning equipment and an intelligent sensor. Yarn-level CT scanning equipment and data acquisition equipment are utilized to respectively obtain feature data of the shaped fabric and real-time operation data of the equipment, the feature data and the real-time operation data are uploaded to a cloud platform center through a communication line, and a process optimization database is established by means of a big data platform. And meanwhile, generic learning is carried out by utilizing a neural network optimized by a genetic algorithm, according to a feature database and a process database of the fabric, hot air setting machine operation parameters with energy consumption and efficiency are matched for enterprises, and the problems of energy consumption waste and optimization over-fitting are solved as much as possible to a greater extent.

Description

Technical Field

[0001] The present invention relates to the technical field of process parameter optimization, and specifically to a method for optimizing the process parameters of a hot air setting machine based on big data analysis. Background Art

[0002] With the development of technology, textiles have become an indispensable item in people's lives, and the textile processing industry has formed a complete production and processing system with upstream and downstream industrial chain support. In recent years, with the continuous development of computer technology, large-scale digital and information-based production in textile enterprises has become very necessary.

[0003] The hot air setting machine plays an important role in the heat setting equipment. It can usually provide relatively stable temperature and relatively uniform hot air, which is important for the drying and setting of textiles, and can improve product quality and production efficiency to a certain extent. However, there is also a certain amount of energy consumption waste. The current big data optimization method for the parameters of the hot air setting machine can achieve energy conservation and emission reduction within a certain range, but it is difficult to achieve the optimal process parameter adjustment effect, mainly limited to learning specific fabric setting processes. The method for optimizing the process parameters of the hot air setting machine involved in the present invention introduces a genetic algorithm on the basis of a traditional neural network to improve the weights and thresholds of the neural network, so as to make the neural network avoid falling into the local optimum problem as much as possible, and realizes the generalization learning and promotion of fabric setting. This method, based on the feature database of fabrics, matches the operating parameters of the hot air setting machine that can achieve both energy consumption and efficiency for enterprises, and solves the problems of energy consumption waste and overfitting optimization to the greatest extent possible. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a method for optimizing the process parameters of a hot air setting machine based on big data analysis. This method uses a yarn-level CT scanning device and intelligent sensors to respectively obtain the characteristic data of the setting fabric and the real-time operating data of the equipment, uploads them to the cloud platform center through a communication line, and borrows the big data platform to establish a process optimization database. At the same time, it uses a neural network optimized by a genetic algorithm for generalization learning, realizes the optimal matching of process parameters for different fabrics, achieves the advanced optimization of equipment parameters, reduces the energy consumption of the setting process, and improves the efficiency of the setting process.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A method for optimizing the process parameters of a hot air setting machine based on big data analysis, which adjusts the process parameters in real time according to the fabric characteristics and the working state of the hot air setting machine, uploads them to the server terminal through a wireless network, borrows the big data platform to establish a process parameter optimization library, and uses a neural network system optimized by a genetic algorithm to match the process parameters of the setting fabric, and sends the matching result to the production equipment.

[0006] This method adopts four modules: a data acquisition module, a data intelligent storage module, a network optimization module, and an intelligent decision-making module.

[0007] The data acquisition module uses a CT scanning device and intelligent sensors to collect the characteristic data of the shaped fabric and the operating parameters of various devices during the shaping process respectively, and uploads them to the cloud platform center in real time using big data technology.

[0008] The CT scanning device is a yarn-level CT scanning device, which can obtain accurate fabric microsections and provide effective shaped fabric characteristic data for the network optimization module; the intelligent sensor can accurately and reliably obtain the operating data of the hot air shaping machine and send it to the cloud platform center through wireless communication technology, providing effective hot air shaping machine data for the network prediction module.

[0009] The fabric characteristic data includes fabric micro data and fabric macro data. The micro data includes fabric fiber volume fraction, yarn cross-sectional area, and porosity. The macro data includes the thickness, gram weight, and loom width of the shaped fabric; the operating data of the hot air shaping machine mainly includes vehicle speed, oven data, and overfeed rate. The oven data includes oven temperature, oven humidity, hot air speed, and steam valve opening.

[0010] The data intelligent storage module uses an advanced distributed platform to connect to the enterprise's big data center. Using clustering analysis technology, it first establishes a fabric characteristic database for different fabrics, and then groups the data of different fabric shaping processes, extracts features to analyze similarity, and then divides them into different clusters, providing a reliable analysis basis for the subsequent network optimization model; at the same time, according to the results of the network optimization model, it establishes a specific hot air shaping machine process database for different fabrics and stores these databases in the cloud platform center.

[0011] The network optimization module mainly includes a neural network improved based on the genetic algorithm. The neural network includes an input layer, a hidden layer, and an output layer. The input layer is used to receive the characteristic data of the shaped fabric and the operating data of the hot air shaping machine. The output layer is used to output the predicted shaping energy consumption, predicted shaping time, and predicted fabric gram weight. The optimal weights and thresholds of the hidden layer come from the genetic algorithm, and all optimization results will be uploaded to the cloud platform center.

[0012] The operation of the neural network: Preprocess all the collected data and divide the processed data into two groups, one group as the training set and the other group as the test set. After initializing the weights and thresholds of the neural network, import the training set into the input layer, then perform forward propagation calculation of the data, then calculate the predicted value of the output layer, select the Sigmoid function as the activation function, and calculate the error using the E(j) function after obtaining the predicted value.

[0013] Among them, the Sigmoid function includes:

[0014] S(x) = 1 / (1 + e -x )

[0015] Among them, x in e -x is the sum of the product of the input value and the weight value and the threshold.

[0016] Among them, the E(j) function includes:

[0017] E(j) = (1 / 2) * Sum[(p j (w, x) - y j ) 2

[0018] Among them, Sum[(p j (w, x) - y j ) 2 is the summation function, p is the overall output, w is the weight vector, x is the input vector, and j is the number of iterations.

[0019] Furthermore, the operation of the genetic algorithm: obtaining the initial weights and thresholds of the neural network, performing binary encoding, selecting the reciprocal of the square of the network output error as the fitness function, then performing selection operation, crossover operation, and mutation operation, and calculating the fitness value, and outputting the optimal weights and thresholds of the neural network after satisfying the termination condition.

[0020] Among them, the fitness function includes:

[0021] F(j) = 1 / E(j)

[0022] Among them, F(j) is the fitness and j is the number of iterations.

[0023] After the neural network obtains the weights and thresholds optimized by the genetic algorithm, it performs forward propagation and error backpropagation, updates the weights, and performs iterative calculation. When the dual objectives of energy consumption and efficiency are satisfied, it outputs the predicted final energy consumption, predicted finalization time, and predicted fabric grammage, and uploads the output results and the corresponding input variables to the cloud platform center.

[0024] The intelligent decision-making module will perform heat setting on the fabric whose characteristics conform to the fabric characteristic database on the enterprise's hot air setting machine. When the enterprise's big data center sends an application to the cloud platform center, it accepts the application, retrieves the process database, matches the optimal process parameters for the current model of hot air setting machine, fabric, and production requirements, and returns the matching results to the enterprise to adjust the equipment operation parameters in real time.

[0025] The present invention provides a method for optimizing the process parameters of a hot air setting machine based on big data analysis. It has the following beneficial effects:​

[0026] 1. The present invention realizes the advance prediction of the energy consumption, setting time, and gram weight of fabric setting, thereby adjusting the equipment in advance, reducing the energy consumption of the setting process, and improving the efficiency of the setting process.

[0027] 2. The neural network improved by the genetic algorithm is adopted in the present invention, which can continuously learn through daily prediction training to improve the prediction accuracy.

[0028] 3. Based on the fabric feature database and process database, the present invention matches the operating parameters of the hot air setting machine that can achieve both energy consumption and efficiency for enterprises, and solves the problems of energy consumption waste and overfitting optimization to the greatest extent possible.

[0029] 4. After each round of training, the present invention generates a cluster of equipment operating parameters that can achieve both energy consumption and efficiency for each type of fabric. Thus, when fabrics with a high degree of similarity are detected next time, they can be used as the initial equipment operating parameters; subsequently, the process is adjusted in real time according to the operating conditions of the equipment, and the adjustment results will be added to the process database to achieve the iterative update of the process database. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a flowchart of a method for optimizing the process parameters of a hot air setting machine based on big data analysis provided by an embodiment of the present invention;

[0031] Figure 2 is a module diagram of the optimization of the process parameters of a hot air setting machine based on big data analysis provided by an embodiment of the present invention;

[0032] Figure 3 is a flowchart of the genetic algorithm for optimizing the process parameters of a hot air setting machine based on big data analysis provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0034] Embodiment:

[0035] Please refer to the attached Figure 1 - attached Figure 3 , in an embodiment of the system for optimizing the process parameters of a hot air setting machine based on big data analysis:

[0036] S1. The CT scanning device and intelligent sensors acquire fabric characteristic data and equipment operation data, and use wireless communication technology to transmit them to the enterprise database in real time, and then upload them to the cloud platform center.

[0037] The CT scanning device mentioned above is a yarn-level CT scanning device, which can obtain accurate fabric microsections and obtain effective shaped fabric characteristic data.

[0038] The intelligent sensors mainly include industrial cameras, speed measuring instruments, voltage and current measuring instruments, humidity detectors, temperature detectors, and steam valve opening detectors.

[0039] The fabric characteristic data includes fabric micro data and fabric macro data. The micro data includes fabric fiber volume fraction, yarn cross-sectional area, and porosity. The macro data includes shaped fabric thickness, gram weight, and loom width.

[0040] The production equipment operation parameters include vehicle speed, oven data, and overfeed rate. The oven data includes oven temperature, oven humidity, hot air speed, and steam valve opening.

[0041] S2. The data intelligent storage module uses an advanced distributed platform to connect to the enterprise's big data center. Using clustering analysis technology, first establish a fabric characteristic database for different fabrics, then group the data of different fabric shaping processes, and perform feature extraction to analyze similarities, and then divide them into different clusters, providing a reliable analysis basis for the subsequent network optimization model; at the same time, establish a specific hot air shaping machine process database for different fabrics according to the results of the network optimization model, and store these databases in the cloud platform center.

[0042] Among them, the feature process matching algorithm is:

[0043] Let C1 and C2 be two propositions under the recognition framework, and n(C1) and n(C2) be their basic probability assignments, satisfying:

[0044] P(C1) = max{n(Ci), Ci ∈ U}

[0045] P(C2) = max{n(Ci), Ci ∈ U, and Ci ≠ C1}

[0046] If there is:

[0047] n(A1) - n(A2) ≥ y1

[0048] n(U) ≤ y2

[0049] n(C1) ≥ n(U)

[0050] Then C1 is the judgment result, where y1 and y2 are preset judgment values.

[0051] S3. The genetic algorithm module processes the data to obtain the neural network weights and thresholds required for each cluster and the predicted setting energy consumption, predicted setting time, and predicted fabric grammage, and forms a corresponding characteristic information database.

[0052] The steps for constructing the genetic algorithm are as follows:

[0053] S31. Initialize parameters;

[0054] Set the parameters of the genetic algorithm, including the DNA length, population size, crossover rate, mutation rate, number of generations of evolution, and hyperparameter bounds;

[0055] S32. Initialize the population;

[0056] Input the independent variable data into the genetic algorithm model, transform it into an independent variable matrix through data processing, and then transform it into an initial population through binary encoding;

[0057] Among them, the independent variable matrix is an initial population, and each row in the matrix represents a population individual, corresponding to a set of working state parameters of the dry heat setting machine, and generates the corresponding DNA after binary conversion;

[0058] Among them, the data of the independent variable matrix includes the initial weights and thresholds of the neural network;

[0059] S33. Crossover and mutation;

[0060] Perform crossover and mutation operations on the population to generate a new generation of population;

[0061] S34. Calculate the fitness;

[0062] First, define a fitness function for calculating the reciprocal of the square of the network output error under given parameters, and then calculate the fitness of each individual in the new population;

[0063] S35. Natural selection;

[0064] Consider the fitness and the "roulette principle" to select new individuals to generate a new population;

[0065] S36. Repeat S2 to S5 until the fitness meets the requirements;

[0066] S37. Output the optimal weights and thresholds of the neural network.

[0067] Among them, the neural network is trained based on the genetic algorithm. By inputting the independent variable data and combining the weight values and thresholds provided by the genetic algorithm, the optimal process parameters required for the set fabric and production target are obtained and output while meeting the requirements of energy consumption and efficiency.

[0068] S4. The intelligent decision-making module will display the optimal process parameters required for specific fabrics and production targets in real time and synchronously adjust the equipment operation parameters.

[0069] The problem solved by the present invention is to design a process optimization system for a hot air setting machine based on big data analysis, which is characterized by using a yarn-level CT scanning device and a data acquisition device to respectively obtain the characteristic data of the setting fabric and the real-time operation data of the equipment, uploading them to the cloud platform center through a communication line, and establishing a process database. At the same time, a neural network optimized by a genetic algorithm is used for general learning to achieve the optimal matching of process parameters for different fabrics, so as to achieve the advanced optimization of equipment parameters, reduce the energy consumption of the setting process, and improve the efficiency of the setting process.

[0070] The advantages and positive effects of this embodiment are as follows: 1. The present invention realizes the advance prediction of the energy consumption, setting time and gram weight of fabric setting, so as to adjust the equipment in advance, reduce the energy consumption of the setting process, and improve the efficiency of the setting process. 2. The neural network improved by the genetic algorithm in the present invention can continuously learn through daily prediction training to improve the prediction accuracy. 3. Based on the characteristic database and process database of the fabric, the present invention matches the operation parameters of the hot air setting machine that can achieve both energy consumption and efficiency for enterprises, and solves the problems of energy consumption waste and overfitting optimization to the greatest extent possible. 4. After each round of training, the present invention generates a cluster of equipment operation parameters that can achieve both energy consumption and efficiency for each type of fabric. Therefore, when fabrics with high similarity are detected next time, they can be used as the initial equipment operation parameters; subsequently, the process is adjusted in real time according to the operation conditions of the equipment, and the adjustment results will be added to the process database to realize the iterative update of the process database.

[0071] Comparative experiment:

[0072] Experiment preparation

[0073] Equipment and materials: Prepare a hot air setting machine, a yarn-level CT scanning device, various intelligent sensors (industrial cameras, speed measuring instruments, voltage and current measuring instruments, humidity detectors, temperature detectors and steam valve opening detectors), and fabric samples of different types (such as pure cotton, chemical fiber, blended, etc., and prepare multiple samples of different thicknesses, gram weights and widths for each type of fabric).

[0074] Build an experimental system: Install the CT scanning device and intelligent sensors on the hot air setting machine and the fabric transmission path to ensure that the fabric characteristic data and equipment operation data can be accurately collected, and connect them to the cloud platform center through wireless communication technology. At the same time, build the software environment and database required for the data intelligent storage module, network optimization module and intelligent decision-making module in the cloud platform center.

[0075] Experiment steps

[0076] Data collection:

[0077] Use a yarn-level CT scanning device to perform microscopic section scanning on each fabric sample to obtain microscopic data such as fabric fiber volume fraction, yarn cross-sectional area, and porosity.

[0078] Use intelligent sensors to collect real-time operation data of the hot air setting machine during the setting process of different fabrics, including vehicle speed, oven temperature, oven humidity, hot air speed, steam valve opening, and overfeed rate, etc. At the same time, record macroscopic data such as the thickness, gram weight, and loom width of the set fabrics.

[0079] Transmit the collected data to the enterprise database in real time using wireless communication technology and then upload it to the cloud platform center.

[0080] Data processing and database construction:

[0081] The data intelligent storage module uses an advanced distributed platform to connect to the enterprise big data center and applies clustering analysis technology to establish a fabric feature database for different fabrics based on fabric feature data.

[0082] Group the data of the setting process of different fabrics, extract feature analysis similarities, divide them into different clusters, and provide an analysis basis for the network optimization module.

[0083] Based on the results of the network optimization module, establish a specific hot air setting machine process database for different fabrics and store it in the cloud platform center.

[0084] Genetic algorithm and neural network training:

[0085] Preprocess the collected data and divide it into a training set and a test set.

[0086] Initialize the genetic algorithm parameters, including DNA length, population size, crossover rate, mutation rate, number of generations of evolution, and hyperparameter bounds.

[0087] Input the independent variables (fabric feature data and equipment operation data) in the training set data into the genetic algorithm model, process them into an independent variable matrix, and then perform binary coding to generate an initial population. The independent variable matrix contains the initial weights and thresholds of the neural network.

[0088] Perform crossover and mutation operations on the population to generate a new generation of population.

[0089] Define the fitness function (F(j) = 1 / E(j), where E(j) = (1 / 2)*Sum[(pj(w,x) - yj)^2]), and calculate the fitness of each individual in the new population.

[0090] Select new individuals according to fitness and the "roulette principle" to generate a new population. Repeat the above steps until the fitness meets the requirements, and output the optimal weights and thresholds of the neural network.

[0091] Apply the optimal weights and thresholds to the neural network, perform forward propagation and error backpropagation, and perform iterative calculations. When the dual objectives of energy consumption and efficiency are met, the predicted setting energy consumption, predicted setting time, and predicted fabric grammage are obtained.

[0092] Process parameter matching and verification:

[0093] Select a new fabric sample whose characteristic data conforms to a certain type of fabric in the fabric characteristic database. The enterprise big data center sends an application to the cloud platform center.

[0094] The intelligent decision-making module accepts the application, retrieves the process database, matches the optimal process parameters for the current hot air setting machine model, fabric, and production requirements, and returns the results to the enterprise to adjust the equipment operation parameters in real time.

[0095] Use the hot air setting machine to set the fabric sample according to the optimized process parameters, and record the actual setting energy consumption, setting time, and fabric grammage after setting.

[0096] Experimental results

[0097] We selected 5 different types of fabrics for experiments, and each fabric was subjected to 5 repeated experiments. The process parameters and results before and after optimization were compared as shown in the following table:

[0098]

[0099]

[0100] Experimental conclusions

[0101] Energy consumption reduction: It can be seen from the experimental data that after optimization by the method of the present invention, the setting energy consumption of different types of fabrics has been significantly reduced, and the average reduction rate reaches about 15%. This shows that the present invention can effectively reduce the energy consumption of the setting process.

[0102] Time shortening: The setting time has also been significantly shortened, with an average shortening of about 20%. This shows that the optimized process parameters improve the efficiency of the setting process.

[0103] Accurate control of grammage: After optimization, the fabric grammage is closer to the target value and the deviation is smaller, indicating that the present invention can more accurately control the quality of the fabric after setting.

[0104] Verification of beneficial effects: The experimental results verified the beneficial effects of the present invention, achieved the advance prediction and optimization of the energy consumption, time and gram weight of fabric setting, improved the prediction accuracy through the neural network improved by the genetic algorithm, matched more optimal operating parameters for enterprises based on fabric characteristics and process databases, and realized the iterative update of the process databases.

[0105] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing process parameters of a hot air setting machine based on big data analysis, comprising a system, characterized in that: The system includes a data acquisition module, a data intelligent storage module, a network optimization module, and an intelligent decision-making module, which can collect, analyze, and optimize process parameters in real time to achieve energy saving and efficiency improvement of the heat setting process; The steps for optimizing the process parameters include: Step 1: Use the data acquisition module The yarn-level CT scanning equipment and intelligent sensors are used to obtain the characteristic data of the shaped fabric and the operation data of the hot air setting machine, and then transmitted to the cloud data center through the network; Step 2: Use the data storage module Rely on cluster analysis technology to realize classified storage of data, and establish corresponding optimization process database for specific fabrics based on the optimization results of network prediction module; Step 3: Use the network prediction module A neural network improved based on genetic algorithm is used. The neural network includes an input layer, a hidden layer and an output layer. The input layer is used to receive the characteristic data of the set fabric and the operation data of the hot air setting machine. The output layer is used to output the predicted setting energy consumption, predicted setting time and predicted fabric weight. The optimal weight and optimal threshold of the hidden layer come from the genetic algorithm. The optimization results that meet the dual requirements of energy consumption and efficiency and the corresponding fabric characteristic data and equipment operation parameters will be uploaded to the cloud platform center. Step 4: Use the intelligent decision-making module Rely on the suggestions of the data intelligence module and the results of the network prediction module to adjust the operating parameters of the equipment in real time.

2. The method for optimizing process parameters of a hot air setting machine based on big data analysis according to claim 1, characterized in that: The data acquisition module includes: a yarn-level CT scanning device and an intelligent sensor. The CT scanning device is a yarn-level CT scanning device, which is responsible for obtaining accurate fabric micro-slices and providing effective micro-data of the characteristics of the shaped fabric for the network prediction module described in Claim 1; the intelligent sensor is responsible for accurately and reliably obtaining the operating data of the hot air setting machine and sending it to the cloud platform center through wireless communication technology, providing effective hot air setting machine data for the network prediction module described in Claim 1.

3. The method for optimizing process parameters of a hot air setting machine based on big data analysis according to claim 1, characterized in that: The fabric characteristic data and the operation data of the hot air setting machine include: the fabric characteristic data include fabric micro data and fabric macro data, wherein the micro data include fabric fiber volume fraction, yarn cross-sectional area and porosity, and the macro data include thickness, gram weight and machine width of the set fabric; the operation data of the hot air setting machine mainly include vehicle speed, oven data and overfeed rate, wherein the oven data include oven temperature, oven humidity, hot air speed and steam valve opening.

4. The method for optimizing process parameters of a hot air setting machine based on big data analysis according to claim 1, characterized in that: The data intelligent storage module includes: using an advanced distributed platform, connecting to the enterprise's big data center, and adopting clustering analysis technology to first establish a fabric feature database for different fabrics, then grouping the data of different fabric shaping processes, and extracting features to analyze similarities, and then dividing them into different clusters, providing a reliable analysis basis for subsequent network prediction models; at the same time, based on the results of the network prediction model, specific hot air shaping machine process databases are established for different fabrics, and these libraries are stored in the cloud platform center.

5. The method for optimizing process parameters of a hot air setting machine based on big data analysis according to claim 1, characterized in that: The network optimization module comprises: firstly designing a neural network framework, wherein the input layer is used to receive characteristic data of the shaped fabric and operation data of the hot air setting machine, and the output layer is used to output predicted setting energy consumption, predicted setting time and predicted fabric weight; then initializing the weights and thresholds of the neural network, and performing binary encoding, taking the inverse of the square of the output error of the neural network as fitness, and then generating new weights and thresholds through selection, crossover and mutation until the termination condition is met, outputting the optimal weights and thresholds, and serving as the weights and thresholds of the neural network, and then calculating forward propagation and error back propagation, performing gradient update based on the principle of minimum error, and outputting predicted setting energy consumption, predicted setting time and predicted fabric weight when the dual requirements of energy consumption and efficiency are met; Among them, the optimal weights and optimal thresholds of the hidden layer of the neural network come from the genetic algorithm. The predicted values ​​of the neural network after optimization by the genetic algorithm will be uploaded to the cloud platform center and stored in the hot air setting machine process database of the corresponding fabric feature database.

6. The method for optimizing process parameters of a hot air setting machine based on big data analysis according to claim 1, characterized in that: The intelligent decision-making module includes: when the company's hot air setting machine heat sets the fabric that appears in the fabric feature database, the company's big data center will send an application to the cloud platform center, and the cloud platform center will accept the application, search the process database, match the optimal process parameters for the current model of hot air setting machine, fabric and production requirements, and return the results to the company to adjust the working status of the production equipment in real time.

Images