A control method and system for intelligent identification of mechanical equipment

Abstract

The present application relates to the field of intelligent control technology, and provides a control method and system for intelligent identification of mechanical equipment, including: determining the quality standard of textile production; conducting textile production process analysis to generate textile process node parameter information; laying out a node acquisition and identification network set; obtaining production node textiles, and obtaining a production node detection data stream through a node acquisition and identification network set; determining target production quality feature data; performing loss analysis on the production node detection data stream and target production quality feature data based on a digital twin network model; optimizing textile process node parameter information based on textile loss feature information to generate textile process node optimization parameter information. The method can solve the technical problem of poor textile production quality due to low precision and accuracy of textile equipment control in the textile production process, and can improve the precision and accuracy of textile equipment control, thereby improving the production quality of textiles.

Description

Technical Field

[0001] The present application relates to the field of intelligent control technology, and in particular to a control method and system for intelligent identification of mechanical equipment. Background Art

[0002] Textiles refer to products that have undergone textile processing, including yarn, woven fabrics, knitted fabrics, braided fabrics, etc. The production process of textiles is complex and involves many steps, including cotton cleaning, roving, warping, sizing, weaving and other steps. Many of these steps are closely linked, and the production quality of the previous step will directly affect the subsequent steps and even the final production quality of the textiles.

[0003] In the existing textile production process, fixed operating parameters are usually set to control textile equipment, and multiple process links of textile processing are not individually analyzed and controlled, resulting in low control accuracy of textile equipment, which affects the production quality of textile products.

[0004] In summary, the prior art has a technical problem in which the textile production quality is poor due to the low control precision and accuracy of textile equipment during the textile production process. Summary of the invention

[0005] Based on this, it is necessary to provide a control method and system for intelligent identification of mechanical equipment to address the above technical issues.

[0006] A control method for intelligent identification of mechanical equipment, comprising: obtaining a textile production design drawing, quantifying the quality standard of the textile production design drawing, and determining the textile production quality standard; analyzing the textile production process based on the textile production quality standard to generate textile process node parameter information; laying out a node acquisition and identification network set according to the textile process node parameter information; performing textile production control according to the textile process node parameter information to obtain production node textiles, and obtaining the production node detection data stream of the production node textile through the node acquisition and identification network set; determining the target production quality characteristic data of the production node textile according to the textile production quality standard and the textile process node parameter information; performing identification loss analysis on the production node detection data stream and the target production quality characteristic data based on a digital twin network model to obtain textile loss characteristic information; optimizing the textile process node parameter information based on the textile loss characteristic information to generate textile process node optimization parameter information, and performing linkage production control according to the textile process node optimization parameter information.

[0007] A control system for intelligent identification of mechanical equipment, comprising:

[0008] A textile production quality standard determination module, the textile production quality standard determination module is used to obtain textile production design drawings, quantify the quality standards of the textile production design drawings, and determine the textile production quality standards;

[0009] A textile process node parameter information generation module, the textile process node parameter information generation module is used to analyze the textile production process flow based on the textile production quality standard and generate textile process node parameter information;

[0010] A node collection and identification network set acquisition module, wherein the node collection and identification network set acquisition module is used to arrange and obtain a node collection and identification network set according to the textile process node parameter information;

[0011] A production node detection data stream obtaining module, the production node detection data stream obtaining module is used to control textile production according to the textile process node parameter information, obtain production node textiles, and obtain the production node detection data stream of the production node textiles through the node acquisition and identification network set;

[0012] A target production quality characteristic data determination module, the target production quality characteristic data determination module is used to determine the target production quality characteristic data of the textile at the production node according to the textile production quality standard and the textile process node parameter information;

[0013] An identification loss analysis module, wherein the identification loss analysis module is used to perform identification loss analysis on the production node detection data stream and the target production quality feature data based on a digital twin network model to obtain textile loss feature information;

[0014] A textile process node optimization parameter information generation module, wherein the textile process node optimization parameter information generation module is used to optimize the textile process node parameter information based on the textile loss characteristic information, generate the textile process node optimization parameter information, and perform linkage production control according to the textile process node optimization parameter information.

[0015] The above-mentioned control method and system for intelligent identification of mechanical equipment can solve the technical problem of poor textile production quality due to low control precision and accuracy of textile equipment during the textile production process. First, obtain the textile production design drawings, quantify the quality standards of the textile production design drawings, and determine the textile production quality standards; then analyze the textile production process according to the textile production quality standards to generate textile process node parameter information; according to the textile process node parameter information, arrange the sensor network of each process node to obtain the node acquisition identification network set; perform textile production control according to the textile process node parameter information to obtain the production node textile, and collect the node control parameters of each production node through the node acquisition identification network set to obtain the production node detection data stream of the production node textile; according to the textile production quality standards and the textile process node parameter information, determine the target production quality characteristic data of the production node textile; based on the digital twin network model, perform identification loss analysis on the production node detection data stream and the target production quality characteristic data to obtain textile loss characteristic information; finally, optimize the textile process node parameter information according to the textile loss characteristic information to generate textile process node optimization parameter information, and perform linkage production control according to the textile process node optimization parameter information. The above method can improve the precision and accuracy of textile equipment management and control, thereby improving the production quality of textiles.

[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flow chart of a control method for intelligent identification of mechanical equipment is provided for this application;

[0018] Figure 2 A schematic diagram of a process for obtaining a production node detection data stream in a control method for intelligent identification of mechanical equipment is provided for this application;

[0019] Figure 3 The present application provides a flow chart of obtaining textile loss characteristic information in a control method for intelligent identification of mechanical equipment;

[0020] Figure 4 A structural schematic diagram of a control system for intelligent identification of mechanical equipment is provided for this application.

[0021] Explanation of the accompanying drawings: textile production quality standard determination module 1, textile process node parameter information generation module 2, node collection and identification network set acquisition module 3, production node detection data flow acquisition module 4, target production quality characteristic data determination module 5, identification loss analysis module 6, textile process node optimization parameter information generation module 7. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0023] like Figure 1 As shown, the present application provides a control method for intelligent identification of mechanical equipment, including:

[0024] Step S100: obtaining a textile production design drawing, quantifying the quality standard of the textile production design drawing, and determining the textile production quality standard;

[0025] Specifically, the method provided in the present application is used to intelligently identify the production quality of multiple process nodes in the production process of textile equipment, and optimize the equipment control parameters of the process nodes based on the identification results, thereby realizing intelligent control of textile equipment and improving the production quality of textiles, wherein textile equipment refers to textile equipment with automated production technology, such as: fully automatic textile all-in-one machine, etc. The method provided in the present application is specifically implemented in a control system for intelligent identification of mechanical equipment.

[0026] First, obtain the textile production design drawings, which are documents composed of graphics, symbols, text, and numbers, etc., used to express the textile design intention and manufacturing requirements, including basic information of textile production design, such as textile size, color, pattern shape, elasticity, air permeability, shrinkage, etc. Quantify the quality standards of textiles according to the textile production design drawings. The quality standard quantification refers to setting specific values ​​of multiple attribute indicators that meet the quality requirements according to the textile production design drawings to determine the textile production quality standards. The textile production quality standards include multiple attribute indicators of textiles, such as: the air permeability of textiles is 27000g / ㎡ / 24h of water vapor permeability, the elasticity of textiles is 10%, etc. Obtaining the textile production quality standards provides a basis for determining the parameter information of textile process nodes in the next step.

[0027] Step S200: performing a textile production process analysis based on the textile production quality standard to generate textile process node parameter information;

[0028] Specifically, the textile production process flow of textile equipment is obtained, wherein the textile production process flow includes two parts: spinning process and weaving process. The spinning process flow mainly includes process nodes such as cotton cleaning, cotton carding, combing, drawing, coarse yarn, fine yarn, winding, twisting, and reeling; the weaving process flow mainly includes process nodes such as warping, sizing, drawing in, weaving, and finishing.

[0029] According to the textile production quality standards, the production quality analysis of each process node in the textile production process is performed, the production quality standards of each process node in the textile production process are determined, and the operation control parameters of the node equipment are set according to the production quality standards of each process node to obtain the textile process node parameter information. For example: in the roving process node, the textile process node parameters include multiple operation control parameters such as quantitative, spindle speed, draft multiple, roller pressure, twist coefficient, etc. By obtaining the textile process node parameter information, support is provided for the next step of sensor layout of each process node.

[0030] Step S300: Arranging and obtaining a node collection and identification network set according to the textile process node parameter information;

[0031] Specifically, the sensor type and node equipment characteristics of each process node are obtained according to the textile process node parameter information, and sensors are arranged for each process node according to the sensor type and node equipment characteristics of each process node, and the sensors of each process node are all connected in communication with the control system of intelligent identification of mechanical equipment, and the communication connection means that the sensors of each process node input the acquired real-time sensor information into the control system of intelligent identification of mechanical equipment. For example: when the process node is sizing, the sizing node parameter information includes parameters such as drying temperature and sizing speed, then the types of sensors arranged are image acquisition devices, temperature sensors, speed sensors, etc., and the arrangement position of the sensors is selected according to the characteristics of the sizing equipment. A node acquisition and identification network set is obtained, and the node acquisition and identification network set includes multiple sensors of multiple process nodes.

[0032] By constructing a node collection and identification network set, it provides support for the detection of operation control parameters of textile production nodes, and at the same time can improve the precision of operation control data acquisition in the production process of textile equipment.

[0033] Step S400: Controlling textile production according to the textile process node parameter information, obtaining production node textiles, and obtaining production node detection data streams of the production node textiles through the node acquisition and identification network set;

[0034] like Figure 2 As shown, in one embodiment, step S400 of the present application further includes:

[0035] Step S410: setting a structured data recognition channel and an image data recognition channel according to the node collection recognition network set;

[0036] Step S420: respectively collecting and acquiring the structured node detection data stream and the image node detection data stream of the production node textile based on the structured data recognition channel and the image data recognition channel;

[0037] Step S430: normalizing the structured node detection data stream, and performing denoising preprocessing on the image node detection data stream to obtain a standard structured node detection data stream and a standard image node detection data stream;

[0038] Step S440: Generate the production node detection data stream based on the standard structured node detection data stream and the standard image node detection data stream.

[0039] Specifically, the textile equipment is controlled to produce textiles according to the textile process node parameter information to obtain production node textiles, wherein the production node textiles refer to textiles completed at each process node in the textile production process, for example, in the sizing process node, the obtained production node textiles are textiles that have completed sizing. Then, the sensors of each process node in the node acquisition and identification network set are used to detect the operation control parameters of each process node to obtain the detection data of each process node.

[0040] First, according to the node collection and identification network set, a structured data identification channel and an image data identification channel are set, the structured data identification channel is used to collect and identify the structured data of each process node, wherein the structured data includes the size, elasticity, strength, etc. of the textile, and the image data identification channel is used to collect and identify the image data of each process node, such as: textile fabric color, pattern shape, pattern size, etc. The structured node detection data stream of the textile of the production node is acquired according to the structured data identification channel, and the image node detection data stream of the textile of the production node is acquired according to the image data identification channel.

[0041] Then the structured node detection data stream is normalized, and the normalization process refers to converting the structured node detection data into standardized data within the range of 0 to 1, wherein the data conversion process can be realized by the existing normalization processing formula, which will not be described in detail here, to obtain a standard structured node detection data stream. The image node detection data stream is subjected to denoising preprocessing, and the denoising preprocessing refers to reducing the influence of noise data in the image on the image detection data and improving the accuracy of image detection data recognition, wherein the image denoising preprocessing method can be customized based on the actual situation of the image, for example: mean filter denoising, median filter denoising, Gaussian filter denoising and other methods, to obtain a standard image node detection data stream. The standard structured node detection data stream and the standard image node detection data stream are combined to obtain a production node detection data stream. By obtaining the production node detection data stream, the node detection data of each process node textile can be obtained in real time, providing data support for the quality analysis of textiles at each process node.

[0042] Step S500: determining target production quality characteristic data of the textile at the production node according to the textile production quality standard and the textile process node parameter information;

[0043] Specifically, according to the textile production quality standard and the textile process node parameter information, the target production quality characteristic data of the textiles at each production node is determined, and the target production quality characteristic data refers to the standard quality characteristic data of the textiles at the production node, including characteristic data such as standard image, standard elasticity, standard water absorption, and standard air permeability, wherein the standard quality characteristic data of each production node is different. By obtaining the target production quality characteristic data, it provides support for the next step of comparing and analyzing the production node detection data flow and obtaining loss characteristic information.

[0044] Step S600: performing identification loss analysis on the production node detection data stream and the target production quality feature data based on the digital twin network model to obtain textile loss feature information;

[0045] like Figure 3 As shown, in one embodiment, step S600 of the present application further includes:

[0046] Step S610: Building the digital twin network model, wherein the digital twin network model includes a node twin network sub-model and a target twin network sub-model;

[0047] Specifically, the basic information of textile equipment is obtained, wherein the basic information includes the specifications of textile equipment, the physical structure of textile equipment, and the operation control parameters of each process node of textile production. A digital twin network model is constructed based on digital twin technology and the basic information of textile equipment, wherein the digital twin network model includes a node twin network sub-model and a target twin network sub-model, the input data of the node twin network sub-model is the production node detection data stream, the input data of the target twin network sub-model is the target production quality feature data, and the data processing method of the node twin network sub-model and the target twin network sub-model is the same. By building a digital twin model based on digital twin technology, the efficiency and accuracy of loss analysis in production node detection data stream identification can be improved.

[0048] Step S620: inputting the production node detection data stream and the target production quality feature data into the node twin network sub-model and the target twin network sub-model respectively, and outputting a node quality detection feature set and a target quality detection feature set;

[0049] In one embodiment, step S620 of the present application further includes:

[0050] Step S621: acquiring a structured data processing layer and an image data processing layer according to the digital twin network model;

[0051] Step S622: Based on the structured data processing layer, attribute classification and division are performed on the standard structured node detection data stream to obtain a node attribute point detection data stream;

[0052] Specifically, the production node detection data stream is input into the node twin network sub-model, and the target production quality feature data is input into the target twin network sub-model. According to the digital twin network model, a structured data processing layer and an image data processing layer are obtained, wherein the structured data processing layer is used to perform data processing on the standard structured node detection data stream, and the image data processing layer is used to perform data processing on the standard image node detection data stream.

[0053] The standard structured node detection data stream is input into the structured data processing layer, and the standard structured node detection data stream is attribute-classified and divided through the structured data processing layer, wherein attribute classification refers to classifying the quality detection types in the standard structured node detection data stream, and the quality detection types include textile strength, water absorption, air permeability, etc.; attribute division refers to dividing the detection quality grade of the attribute type according to the degree of influence of the attribute type on the quality of the textile, wherein the greater the influence of the attribute type on the quality of the textile, the higher the detection quality grade, and vice versa, the smaller the influence on the quality of the textile, the smaller the detection quality grade. The node attribute point detection data stream is obtained according to the attribute classification result and the attribute division result. By performing attribute classification and division on the standard structured node detection data stream, the precision of the comparison of the standard structured node detection data stream can be improved, thereby producing the accuracy of node textile quality identification. Step S623: Feature extraction is performed on the standard image node detection data stream through the image data processing layer to obtain a node image feature data stream;

[0054] In one embodiment, step S623 of the present application further includes:

[0055] Step S6231: obtaining a textile semantic segmentation model according to the image data processing layer, and performing semantic segmentation on the standard image node detection data stream based on the textile semantic segmentation model to obtain a textile semantic segmentation result;

[0056] Step S6232: Acquire textile image detection factors, where the textile image detection factors include size structure, printing and dyeing difference, and surface defects;

[0057] Step S6233: determining a preset convolution feature set of textiles according to the textile image detection factors;

[0058] Step S6234: performing traversal convolution calculation on the textile semantic segmentation result according to the preset convolution feature set of the textile, and obtaining the node image feature data stream that meets the predetermined convolution value range.

[0059] Specifically, the image data processing layer performs feature extraction on the standard image node detection data stream. First, a textile semantic segmentation model is obtained according to the image data processing layer. The textile semantic segmentation model refers to an intelligent model for semantic segmentation of textile node images based on computer vision technology and deep learning. Then, the standard image node detection data stream is semantically segmented according to the textile semantic segmentation model. The semantic segmentation refers to automatically identifying the image content and image contour in the standard image node detection data stream through the textile semantic segmentation model, and segmenting the image in the standard image node detection data stream according to the image content and image contour, dividing the image into multiple sub-blocks, and the image content includes different types of image types such as patterns and figures, wherein different image contents have different pixels, and the image contour refers to closed lines in different pixel areas, and the textile semantic segmentation result is obtained. By performing semantic segmentation on the standard image node detection data stream, the accuracy of image feature recognition in the standard image node detection data stream can be improved.

[0060] The textile image detection factors are obtained, and the textile image detection factors include size structure, printing and dyeing difference, and surface defects, and the surface defects refer to spots and flaws on the surface of the textile. Then, according to the textile image detection factors, a set of preset convolutional features of textiles is determined, and the preset convolutional features refer to the parts of the textile image that need to be detected and identified, such as the overall size of the textile, the size of the pattern in the textile, the shape of the pattern, etc.

[0061] According to the preset convolution feature set of the textile, multiple image feature convolution kernels are determined, and then the textile semantic segmentation result is traversed and convolved according to the multiple image feature convolution kernels to obtain a node image feature data stream that meets the preset convolution value range, wherein the preset convolution value range refers to the similarity threshold of the image convolution feature, wherein the specific similarity value of the preset convolution value range can be set by a person skilled in the art based on the actual situation, for example: the preset convolution value range is set to 95% similarity, and the number of the preset convolution value range is the same as the number of preset convolution features in the preset convolution feature set, and has a corresponding relationship. When the node image feature after traversal convolution satisfies the corresponding preset convolution value range, the node image feature is characterized as an image feature that meets the preset convolution feature set; when the node image feature does not meet the corresponding preset convolution value range, the image feature is discarded. Obtain a node image feature data stream that meets the preset convolution value range. By obtaining the node image feature data stream, the image feature data of each node can be accurately obtained, and the accuracy of the node textile image quality comparison can be improved.

[0062] Step S624: Determine the node quality detection feature set based on the node attribute point detection data stream and the node image feature data stream.

[0063] Specifically, a node quality detection feature set is obtained, and the node quality detection feature set includes the node attribute point detection data stream and the node image feature data stream. By obtaining the node quality detection feature set, data support is provided for the next step of textile node quality comparison.

[0064] Step S630: introducing a feature loss function to perform loss analysis on the node quality detection feature set and the target quality detection feature set, and outputting node quality loss data;

[0065] In one embodiment, step S630 of the present application further includes:

[0066] Step S631: The feature loss function is specifically:

[0067] ;

[0068] Step S632: wherein: is the loss function of the node quality detection feature set and the target quality detection feature set, is the feature weight assignment information for the data sample, i is the feature order of the data sample, p is the total number of features of the data sample, is the current feature similarity between the node quality detection feature set and the target quality detection feature set, It is the preset feature similarity between the node quality detection feature set and the target quality detection feature set.

[0069] Specifically, a feature loss function is constructed, and the expression of the feature loss function is: ;

[0070] in, is a loss function of the node quality detection feature set and the target quality detection feature set; The feature weight allocation information is for the data sample, wherein the data sample includes multiple textile feature types, such as strength, water absorption, air permeability, etc. The feature weight of the data sample can be set based on the degree of influence of the data sample type on the node textile quality, wherein the higher the degree of influence of the data sample type on the textile quality, the higher the feature weight of the data sample; the lower the degree of influence of the data sample type on the textile quality, the lower the feature weight of the data sample; wherein the specific setting method of the feature weight can be implemented by the existing coefficient of variation method, and the coefficient of variation method is a weighting method commonly used by technicians in this field, which will not be described in detail here.

[0071] i is the feature order of the data sample, which refers to the order of each feature shown in each process node of the textile, and the similarity calculation of the feature is performed in sequence according to the feature order; p is the total number of features of the data sample; is the current feature similarity between the node quality detection feature set and the target quality detection feature set; The preset feature similarity of the node quality detection feature set and the target quality detection feature set is a standard for measuring the similarity of each feature, wherein the preset feature similarity of each feature is different, and the preset feature similarity can be customized by a person skilled in the art based on the feature.

[0072] Finally, the feature loss of the node quality detection feature set and the target quality detection feature set is calculated according to the feature loss function to obtain the node quality loss data.

[0073] Step S640: performing hierarchical feature marking on the node mass loss data to obtain the textile loss feature information.

[0074] Specifically, the node quality is marked with hierarchical features according to the value of the node quality loss data, wherein the larger the value of the node quality loss data is, the larger the loss level of the corresponding node is, wherein the specific node quality level marking method can be set according to the actual situation of the node quality loss data, for example: when the node quality loss data is 5%, the loss level is set to level one; when the node quality loss data is 9%, the loss level is set to level two; and the textile loss feature information is obtained, and the textile loss feature information includes the hierarchical feature marking result of the node quality loss data. By obtaining the textile loss feature information, support is provided for the next step of parameter optimization of the textile process node.

[0075] Step S700: Optimizing the textile process node parameter information based on the textile loss characteristic information, generating textile process node optimization parameter information, and performing linkage production control according to the textile process node optimization parameter information.

[0076] Specifically, the quality standard loss threshold of the textile process node is obtained. The quality standard loss threshold can be set based on the actual situation, wherein the higher the quality requirement of the textile process node, the larger the value of the quality standard loss threshold. The textile loss characteristic information is judged according to the quality standard loss threshold. When the textile loss characteristic information is less than the quality standard loss threshold, the loss of the textile quality at the process node is obtained according to the textile loss characteristic information, and the textile process node parameter information is optimized according to the loss of the textile quality at the process node. For example, at the sizing process node, if the air permeability of the textile at the process node does not meet the standard, the sizing density in the sizing process can be appropriately reduced to improve the air permeability of the textile. Generate the optimized parameter information of the textile process node, and perform linkage production control on the textile equipment of each process node according to the optimized parameter information of the textile process node.

[0077] By optimizing the textile process node parameter information according to the textile loss characteristic information, the production control parameters of each process node of the textile equipment can be adjusted in time, the precision and accuracy of the textile equipment management and control can be improved, thereby improving the production quality of the textile. In one embodiment, step S700 of the present application also includes:

[0078] Step S710: constructing a textile production quality feature label library;

[0079] Step S720: classifying the textile loss feature information based on the textile production quality feature label library to obtain textile production quality label information;

[0080] Step S730: When the textile production quality label information does not reach the quality benchmark value, the production node textile is marked to obtain a textile pre-inspection mark;

[0081] Step S740: performing early warning tracking on the textile at the production node based on the textile pre-inspection mark.

[0082] Specifically, a textile production quality feature label library is constructed, in which textile loss feature information and corresponding feature labels are stored. For example, when the feature loss value in the textile loss feature information is less than 3%, the textile is a special grade product; when the feature loss value in the textile loss feature information is greater than or equal to 3% and less than 5%, the textile is a first grade product. The textile loss feature information is classified according to the textile production quality feature label library, and the classification results are marked with feature labels to obtain textile production quality label information. The quality reference value is obtained. The quality reference value can be set based on the actual situation of the textile quality inspection requirements. For example, second grade products and above are qualified.

[0083] The textile production quality label information is judged according to the quality reference value. When the textile production quality label information is less than the quality reference value, the textile at the production node is marked as substandard product quality to obtain a textile pre-inspection mark. Finally, the textile at the production node is tracked according to the textile pre-inspection mark. By pre-inspecting the textile at the production node, the textile early warning tracking time can be reduced and the efficiency and accuracy of textile early warning tracking can be improved.

[0084] In one embodiment, Figure 4 The invention provides a control system for intelligent identification of mechanical equipment, including: a textile production quality standard determination module 1, a textile process node parameter information generation module 2, a node acquisition and identification network set acquisition module 3, a production node detection data flow acquisition module 4, a target production quality feature data determination module 5, an identification loss analysis module 6, a textile process node optimization parameter information generation module 7, wherein:

[0085] A textile production quality standard determination module 1, the textile production quality standard determination module 1 is used to obtain textile production design drawings, quantify the quality standards of the textile production design drawings, and determine the textile production quality standards;

[0086] A textile process node parameter information generating module 2, wherein the textile process node parameter information generating module 2 is used to analyze the textile production process flow based on the textile production quality standard and generate textile process node parameter information;

[0087] A node collection and identification network set acquisition module 3, wherein the node collection and identification network set acquisition module 3 is used to arrange and obtain a node collection and identification network set according to the textile process node parameter information;

[0088] A production node detection data flow obtaining module 4, wherein the production node detection data flow obtaining module 4 is used to control textile production according to the textile process node parameter information, obtain production node textiles, and obtain the production node detection data flow of the production node textiles through the node acquisition and identification network set;

[0089] A target production quality characteristic data determination module 5, the target production quality characteristic data determination module 5 is used to determine the target production quality characteristic data of the textile at the production node according to the textile production quality standard and the textile process node parameter information;

[0090] An identification loss analysis module 6, wherein the identification loss analysis module 6 is used to perform identification loss analysis on the production node detection data stream and the target production quality feature data based on a digital twin network model to obtain textile loss feature information;

[0091] A textile process node optimization parameter information generation module 7, wherein the textile process node optimization parameter information generation module 7 is used to optimize the textile process node parameter information based on the textile loss characteristic information, generate the textile process node optimization parameter information, and perform linkage production control according to the textile process node optimization parameter information.

[0092] In one embodiment, the system further comprises:

[0093] An identification channel setting module, the identification channel setting module is used to set a structured data identification channel and an image data identification channel according to the node collection identification network set;

[0094] A node detection data stream acquisition module, the node detection data stream acquisition module is used to respectively acquire the structured node detection data stream and the image node detection data stream of the production node textile based on the structured data recognition channel and the image data recognition channel;

[0095] A standard node detection data stream obtaining module, wherein the standard node detection data stream obtaining module is used to perform normalization processing on the structured node detection data stream, perform denoising preprocessing on the image node detection data stream, and obtain a standard structured node detection data stream and a standard image node detection data stream;

[0096] A production node detection data stream generation module is used to generate the production node detection data stream based on the standard structured node detection data stream and the standard image node detection data stream.

[0097] In one embodiment, the system further comprises:

[0098] A digital twin network model building module, wherein the digital twin network model building module is used to build the digital twin network model, wherein the digital twin network model includes a node twin network sub-model and a target twin network sub-model;

[0099] A quality detection feature set output module, the quality detection feature set output module is used to input the production node detection data stream and the target production quality feature data into the node twin network sub-model and the target twin network sub-model respectively, and output a node quality detection feature set and a target quality detection feature set;

[0100] A loss analysis module, wherein the loss analysis module is used to introduce a characteristic loss function to perform loss analysis on a node quality detection feature set and a target quality detection feature set, and output node quality loss data;

[0101] A hierarchical feature marking module is used to perform hierarchical feature marking on the node mass loss data to obtain the textile loss feature information.

[0102] In one embodiment, the system further comprises:

[0103] A data processing layer acquisition module, wherein the data processing layer acquisition module is used to acquire a structured data processing layer and an image data processing layer according to the digital twin network model;

[0104] A node attribute point detection data stream obtaining module, wherein the node attribute point detection data stream obtaining module is used to classify and divide the standard structured node detection data stream by attributes based on the structured data processing layer to obtain a node attribute point detection data stream;

[0105] A node image feature data stream acquisition module, the node image feature data stream acquisition module is used to extract features from the standard image node detection data stream through the image data processing layer to obtain a node image feature data stream;

[0106] A node quality detection feature set determination module is used to determine the node quality detection feature set based on the node attribute point detection data stream and the node image feature data stream.

[0107] In one embodiment, the system further comprises:

[0108] A semantic segmentation module, the semantic segmentation module is used to obtain a textile semantic segmentation model according to the image data processing layer, and perform semantic segmentation on the standard image node detection data stream based on the textile semantic segmentation model to obtain a textile semantic segmentation result;

[0109] A textile image detection factor acquisition module, wherein the textile image detection factor acquisition module is used to acquire textile image detection factors, wherein the textile image detection factors include size structure, printing and dyeing difference, and surface defects;

[0110] A textile preset convolution feature set determination module, the textile preset convolution feature set determination module is used to determine a textile preset convolution feature set according to the textile image detection factors;

[0111] A convolution calculation module is used to perform traversal convolution calculation on the textile semantic segmentation result according to the preset convolution feature set of the textile, so as to obtain the node image feature data stream that meets the predetermined convolution value range.

[0112] In one embodiment, the system further comprises:

[0113] Feature loss function module, the feature loss function module refers to the feature loss function specifically:

[0114] ;

[0115] A parameter summary module, wherein the parameter summary module refers to the module wherein: is the loss function of the node quality detection feature set and the target quality detection feature set, is the feature weight assignment information for the data sample, i is the feature order of the data sample, p is the total number of features of the data sample, is the current feature similarity between the node quality detection feature set and the target quality detection feature set, It is the preset feature similarity between the node quality detection feature set and the target quality detection feature set.

[0116] In one embodiment, the system further comprises:

[0117] A textile production quality feature label library construction module, wherein the textile production quality feature label library construction module is used to construct a textile production quality feature label library;

[0118] A loss characteristic information classification module, wherein the loss characteristic information classification module is used to classify the textile loss characteristic information based on the textile production quality characteristic label library to obtain textile production quality label information;

[0119] A textile pre-inspection mark obtaining module, wherein the textile pre-inspection mark obtaining module is used to mark the production node textiles when the textile production quality label information does not reach the quality reference value, and obtain the textile pre-inspection mark;

[0120] An early warning tracking module is used to perform early warning tracking on the production node textile based on the textile pre-inspection mark.

[0121] In summary, the present application provides a control method and system for intelligent identification of mechanical equipment, which has the following technical effects:

[0122] 1. The technical problem of poor textile production quality caused by low precision and accuracy of textile equipment control in the textile production process is solved. By optimizing the textile process node parameter information according to the textile loss characteristic information, the production control parameters of each process node of the textile equipment can be adjusted in time, and the precision and accuracy of textile equipment control can be improved, thereby improving the production quality of textiles.

[0123] 2. By constructing a node collection and identification network set, the precision of obtaining the operation control parameters of textile equipment can be improved. By classifying and dividing the attributes of the standard structured node detection data stream, the precision of the comparison of the standard structured node detection data stream can be improved, thereby improving the accuracy of the production node textile quality identification.

[0124] 3. By pre-inspecting and marking textiles at production nodes, the time for textile early warning tracking can be reduced and the efficiency and accuracy of textile early warning tracking can be improved.

[0125] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0126] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A control method for intelligent identification of mechanical equipment, characterized in that: The method comprises: Obtaining textile production design drawings, quantifying the quality standards of the textile production design drawings, and determining the textile production quality standards; Perform textile production process analysis based on the textile production quality standard to generate textile process node parameter information; Arrange and obtain a node collection and identification network set according to the textile process node parameter information; Controlling textile production according to the textile process node parameter information, obtaining production node textiles, and obtaining production node detection data streams of the production node textiles through the node acquisition and identification network set; Determining target production quality characteristic data of the textile at the production node according to the textile production quality standard and the textile process node parameter information; Based on the digital twin network model, the production node detection data stream and the target production quality feature data are subjected to identification loss analysis to obtain textile loss feature information; The textile process node parameter information is optimized based on the textile loss characteristic information to generate textile process node optimization parameter information, and linkage production control is performed according to the textile process node optimization parameter information.

2. The method according to claim 1, characterized in that The step of obtaining the production node detection data stream of the production node textile through the node collection and identification network set includes: According to the node collection and recognition network set, a structured data recognition channel and an image data recognition channel are set; Based on the structured data recognition channel and the image data recognition channel, respectively collect and obtain the structured node detection data stream and the image node detection data stream of the production node textile; Normalizing the structured node detection data stream, and performing denoising preprocessing on the image node detection data stream to obtain a standard structured node detection data stream and a standard image node detection data stream; The production node detection data stream is generated based on the standard structured node detection data stream and the standard image node detection data stream.

3. The method according to claim 2, characterized in that The obtaining of textile loss characteristic information comprises: Building the digital twin network model, wherein the digital twin network model includes a node twin network sub-model and a target twin network sub-model; Input the production node detection data stream and the target production quality feature data into the node twin network sub-model and the target twin network sub-model respectively, and output a node quality detection feature set and a target quality detection feature set; The feature loss function is introduced to perform loss analysis on the node quality detection feature set and the target quality detection feature set, and the node quality loss data is output; The node mass loss data is subjected to hierarchical feature marking to obtain the textile loss feature information.

4. The method according to claim 3, characterized in that The method comprises: According to the digital twin network model, a structured data processing layer and an image data processing layer are obtained; Based on the structured data processing layer, attribute classification and division are performed on the standard structured node detection data stream to obtain a node attribute point detection data stream; Performing feature extraction on the standard image node detection data stream through the image data processing layer to obtain a node image feature data stream; The node quality detection feature set is determined based on the node attribute point detection data stream and the node image feature data stream.

5. The method according to claim 4, characterized in that The step of obtaining the node image feature data stream comprises: Obtaining a textile semantic segmentation model according to the image data processing layer, and performing semantic segmentation on the standard image node detection data stream based on the textile semantic segmentation model to obtain a textile semantic segmentation result; Acquiring textile image detection factors, wherein the textile image detection factors include size structure, printing and dyeing difference, and surface defects; Determining a preset convolution feature set of textiles according to the textile image detection factors; The textile semantic segmentation result is subjected to traversal convolution calculation according to the preset convolution feature set of the textile to obtain the node image feature data stream that meets the predetermined convolution value range.

6. The method according to claim 3, characterized in that The feature loss function is specifically: ； in, is the loss function of the node quality detection feature set and the target quality detection feature set, is the feature weight assignment information of the data sample, i is the feature order of the data sample, p is the total number of features of the data sample, is the current feature similarity between the node quality detection feature set and the target quality detection feature set, It is the preset feature similarity between the node quality detection feature set and the target quality detection feature set.

7. The method according to claim 1, characterized in that The method comprises: Construct a label library of textile production quality characteristics; Classifying the textile loss feature information based on the textile production quality feature label library to obtain textile production quality label information; When the textile production quality label information does not reach the quality benchmark value, marking the production node textile to obtain a textile pre-inspection mark; Early warning tracking of the textiles at the production node is performed based on the textile pre-inspection mark.

8. A control system for intelligent identification of mechanical equipment, characterized in that: The system is used to perform any one of the control methods for intelligent identification of mechanical equipment described in claims 1-7, and comprises: A textile production quality standard determination module, the textile production quality standard determination module is used to obtain textile production design drawings, quantify the quality standards of the textile production design drawings, and determine the textile production quality standards; A textile process node parameter information generation module, the textile process node parameter information generation module is used to analyze the textile production process flow based on the textile production quality standard and generate textile process node parameter information; A node collection and identification network set acquisition module, wherein the node collection and identification network set acquisition module is used to arrange and obtain a node collection and identification network set according to the textile process node parameter information; A production node detection data stream obtaining module, the production node detection data stream obtaining module is used to control textile production according to the textile process node parameter information, obtain production node textiles, and obtain the production node detection data stream of the production node textiles through the node acquisition and identification network set; A target production quality characteristic data determination module, the target production quality characteristic data determination module is used to determine the target production quality characteristic data of the textile at the production node according to the textile production quality standard and the textile process node parameter information; An identification loss analysis module, wherein the identification loss analysis module is used to perform identification loss analysis on the production node detection data stream and the target production quality feature data based on a digital twin network model to obtain textile loss feature information; A textile process node optimization parameter information generation module, wherein the textile process node optimization parameter information generation module is used to optimize the textile process node parameter information based on the textile loss characteristic information, generate the textile process node optimization parameter information, and perform linkage production control according to the textile process node optimization parameter information.

Images