Data mining analysis method applied to sandwich production control system

Abstract

The invention discloses a data mining analysis method applied to a sandwich production control system, and relates to the technical field of data mining. The method comprises the following steps: collecting multi-source data of a sandwich production control system; predicting the production state of the production control system by adopting a stream data processing technology; performing quality prediction on the sandwich by adopting a deep learning algorithm; performing defect detection on the sandwich by using a computer vision technology; performing fault prediction by adopting a deep learning algorithm and a correlation analysis algorithm; optimizing the production scheduling of each production link by adopting an optimization algorithm; and product data of the sandwich are collected and product traceability is carried out. On the basis of data mining, various technologies are adopted to process and analyze the mined data, so that the production efficiency and quality are improved, and the production cost is reduced; by predicting the production quality and carrying out defect detection and fault detection, the production management level is improved, and the optimization of the production process, the accurate control of the quality and the effective management of the cost are realized.

Description

Technical Field

[0001] The invention belongs to the technical field of data mining, and in particular relates to a data mining analysis method applied to a sandwich production control system. Background Art

[0002] As a widely popular food, sandwiches have high production requirements. The production process usually involves multiple links, such as raw material procurement, processing, assembly, packaging, quality inspection and inventory management. In order to ensure product quality and production efficiency, sandwich production requires precise control of each link, including the baking temperature of bread, the amount of sauce used, the combination of ingredients and the packaging of the final product. With the continuous improvement of market requirements for product quality and production speed, traditional manual operation and experience judgment can no longer meet the needs of modern production.

[0003] As a technology for extracting valuable information from big data, data mining technology has achieved remarkable results in many industries. In the manufacturing industry, data mining technology can effectively identify potential patterns and trends by analyzing data in the production process, thereby optimizing the production process.

[0004] At present, data mining technology has been widely used in the production control field of various manufacturing industries. For example, in food production, there have been studies using data mining methods to optimize production scheduling, reduce resource waste, and improve the quality control level of production lines. In the field of sandwich production, although there are relatively few related studies, with the improvement of automation and the development of data acquisition technology, the application potential of data mining technology has gradually emerged. Therefore, how to effectively apply data mining technology to the production control system of sandwiches has become an urgent problem to be solved. Summary of the invention

[0005] The object of the present invention is to provide a data mining and analysis method applied to a sandwich production control system, which can be achieved by the following technical solutions: The present application embodiment provides a data mining analysis method applied to a sandwich production control system, comprising the following steps: S1, collects historical production data, historical quality data and historical equipment operation data of the sandwich production control system; S2, using stream data processing technology to monitor the production status of the sandwich production control system in real time to obtain multi-source real-time production data, and combining the historical production data to predict the future production status of the sandwich production control system; S3, predicting the final quality of the sandwich using a deep learning algorithm based on the historical quality data; S4, during the production process of the sandwiches, using computer vision technology combined with cameras and image recognition algorithms to perform defect detection on the sandwiches to obtain defect data, and performing product screening and production line adjustment based on the defect data; S5, collecting real-time equipment operation data of the sandwich production equipment by monitoring the equipment status, and performing fault prediction by combining the historical equipment operation data and using a deep learning algorithm and an association analysis algorithm; S6, in the sandwich production process, using an optimization algorithm to optimize the production scheduling of each production link in the sandwich production control system; S7, collecting product data of the sandwiches, and tracing the sandwiches with quality problems in combination with the defect data.

[0006] Preferably, step S2 specifically includes: Collect data from multiple data sources from the production control system of Sandwich to form multi-source real-time production data; Preprocessing the multi-source real-time production data using stream data processing technology, wherein the preprocessing includes data cleaning and data aggregation; Based on the historical production data, a time series analysis method is used to conduct a comprehensive analysis to obtain the historical laws of production status; Based on the preprocessed multi-source real-time production data and combined with the historical laws of the production status, a deep learning model is used to predict the future production status, specifically including: Production volume forecast: predict the number of sandwiches produced in the future; Resource consumption forecasting: forecasting the consumption of raw materials in the sandwich production process; Wherein, the deep learning model includes a convolutional neural network or a recurrent neural network; The time series analysis method includes an ARIMA algorithm or a LSTM algorithm. Preferably, step S3 specifically includes: Acquire sample data from the historical quality data, and perform data preprocessing on the sample data; Determining at least one key process in the whole process of the sandwich production control system; Creating a corresponding key process prediction model for the at least one key process; Using sample data of key processes to train and calibrate the key process prediction model; Using the key process prediction model to predict the key process, and obtain the prediction result of the key process; creating a full process prediction model for the entire process based on the prediction results of the key process; The whole process prediction model of the whole process is created according to the prediction results of the key process, including: Create a feedback neural network model RNNS as a full-process prediction model; The minimum value of the prediction accuracy in the prediction results of the key process is used as the feedback value of the whole process prediction model and introduced into the feedback neural network model RNNS; Using sample data of the whole process to train and calibrate the whole process prediction model; The whole process prediction model is used to predict product quality.

[0007] Preferably, step S4 specifically includes: Obtain the image of the target sandwich to be detected; Using the trained image segmentation network, identifying the target overall image corresponding to the target sandwich from the image to be detected, and identifying the defective portion image in the first shape; Determine a first sub-image according to the target overall image and the defective portion image, and determine a second sub-image according to the defective portion image; Performing defect detection on the first sub-image and the second sub-image respectively, and if a defect is detected in at least one sub-image, determining that the target sandwich has a defect; Analyzing defects of the target sandwich to obtain defect data, and performing product screening and production line adjustment based on the defect data; The determining of the first sub-image according to the target overall image and the defective partial image comprises: selecting an area in the target overall image whose position matches the defective partial image but whose size is smaller than the defective partial image as the area corresponding to the defective partial image, and filling the area in the target overall image corresponding to the defective partial image with a specified color to obtain the first sub-image; Determining the second sub-image according to the defective portion image includes: extracting a region whose position matches the defective portion image but whose size is larger than the defective portion image from the image to be detected, to obtain the second sub-image. Preferably, the step S4 further includes: In case a defect is detected, determining sub-image pixel coordinates of the defect in the corresponding sub-image; According to the position of the target overall image and / or the defective partial image in the image to be detected, the sub-image pixel coordinates of the defect are transformed to obtain the original image pixel coordinates of the defect in the image to be detected, so as to determine the position of the defect in the target based on the original image pixel coordinates of the defect in the image to be detected.

[0008] Preferably, in step S5, a deep learning algorithm is used to perform fault prediction, which specifically includes: Real-time equipment operation data is collected in real time through sensors installed on sandwich production equipment; Performing data preprocessing on the historical equipment operation data and the real-time equipment operation data, wherein the data preprocessing includes missing value processing, noise removal and normalization processing; Selecting equipment operation features that affect equipment failure from the historical equipment operation data, and using a statistical analysis method to screen out features that are highly correlated with equipment failure as historical equipment failure features; Filtering out real-time equipment failure characteristics from the real-time equipment operation data based on the historical equipment failure characteristics; Performing feature construction based on the real-time device fault feature to generate a real-time device data stream feature includes: By using sliding window technology, time series features are constructed from the real-time equipment fault features to capture the dynamic changes in the state of the production equipment; Calculate statistical characteristics of production equipment, including mean, maximum, standard deviation, and crest factor; Use Fourier transform to extract frequency domain features to identify abnormal patterns in production equipment; Aggregating the time series features, the statistical features and the frequency domain features into real-time device data stream features; Based on the historical equipment failure characteristics, a long short-term memory network and a convolutional neural network are combined to perform model training to generate an equipment failure prediction model; Inputting the real-time device data stream characteristics into the device fault prediction model to perform fault prediction; The real-time equipment operation data includes temperature, vibration, pressure, current, voltage, speed, working time and load. Preferably, in step S5, a correlation analysis algorithm is used to perform fault prediction, which specifically includes: Acquire real-time equipment operation data of the sandwich production equipment to be tested at multiple predetermined time points within a predetermined time period; Arranging the real-time equipment operation data at the plurality of predetermined time points into an operation parameter time series input matrix according to the time dimension and the sample dimension; Performing inter-parameter timing correlation feature analysis on the operating parameter timing input matrix to obtain global timing features of the operating parameters; Based on the global time series characteristics of the operating parameters, determining whether to generate a fault warning prompt; The step of performing a timing correlation feature analysis between parameters on the operating parameter timing input matrix to obtain a global timing feature of the operating parameters includes: The operating parameter time series input matrix is ​​subjected to matrix block processing and then passed through a ViT model including an embedding layer to obtain a sequence of local time series semantic feature vectors of the operating parameters; Performing consistency association analysis on the sequence of the local temporal semantic feature vectors of the operating parameters to obtain a consistency topological feature matrix of the operating parameters; Performing graph-structure-based associative coding on the sequence of the local temporal semantic feature vectors of the operating parameters and the consistent topological feature matrix of the operating parameters to obtain a consistent topological operating parameter global temporal feature matrix as the global temporal feature of the operating parameters; Wherein, determining whether to generate a fault warning prompt based on the global time series characteristics of the operating parameters includes: Expanding the consistency topology operation parameter global timing feature matrix into a consistency topology operation parameter global timing feature vector; Optimizing each eigenvalue of the consistent topology operation parameter global timing eigenvector to obtain an optimized consistent topology operation parameter global timing eigenvector; Passing the global time series feature vector of the optimized consistency topology operation parameter through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the operation state of the sandwich production equipment to be detected is abnormal; Based on the classification result, a fault warning prompt is generated.

[0009] Preferably, step S6 specifically includes: Preprocessing the historical production data and historical equipment operation data of the sandwich production control system, including denoising, filtering, extracting, cleaning, reducing and transforming; Perform data mining analysis on the pre-processed sandwich production data to obtain the correlation between the process parameters and quality indicators of each process of sandwich production, and establish a corresponding mathematical model and form visual knowledge. The data mining analysis includes correlation analysis, cluster analysis, regression analysis and prediction analysis; Based on the visualized knowledge, the entire process of sandwich production is intelligently feedback-controlled, and the production scheduling of each production link in the sandwich production control system is optimized. Preferably, the data mining analysis of the pre-processed sandwich production data specifically includes: Data association: Apriori algorithm is used to perform data association analysis on the preprocessed sandwich production data. According to the mined association relationship, the information of one attribute is inferred from the information of another attribute. When the confidence reaches a certain threshold, the association rule is considered to be established. Data clustering: Perform data clustering analysis on the preprocessed sandwich production data. According to the intrinsic properties of the sandwich production data, the data is divided into some clusters. The elements in each cluster have the same characteristics. By observing the characteristics of each cluster, further analysis is conducted on specific clusters to obtain the required information. Data regression: Perform data regression analysis on the pre-processed sandwich production data to determine the quantitative relationship between two or more variables, establish a corresponding mathematical model, fit a series of influencing factors and results to fit an equation, and then apply this equation to other similar events to form visual knowledge; Data prediction: Based on the formed visualization knowledge, data prediction analysis is performed on the pre-processed sandwich production data to form a prediction result, wherein the data prediction analysis method is a causal relationship prediction method; Among them, the production scheduling of each production link in the sandwich production control system is optimized, specifically including production resource optimization, human resource optimization, raw material and material supply chain optimization, production process optimization, energy consumption and ingredient optimization, logistics and transportation optimization and production scheduling optimization.

[0010] Preferably, step S7 specifically includes: Collect product data in the sandwich production control system; the product data includes raw material data, production process data, and logistics data; Using the product data as traceability data and storing it in a product database; The sandwiches with quality problems are traced back as follows: raw material data, production process data and logistics data are processed to obtain raw material score values, production process score values ​​and logistics transportation score values, which are then compared with their corresponding system score thresholds. Based on the comparison results, it is determined whether there are abnormal items. If so, it is determined to be the source of the product problem. Otherwise, it is determined that the traceability has failed. Based on big data, the stored traceability data and the defect data are processed and analyzed to obtain a quality trend coefficient, the calculated quality trend coefficient is compared with the quality trend interval preset by the sandwich production control system, and the authenticity of the traceability data is judged based on the comparison result. The beneficial effects of the present invention are as follows: the present invention adopts a variety of technologies to process and analyze the mined data on the basis of data mining, so as to improve production efficiency and quality and reduce production costs; by monitoring production status, predicting production quality, performing defect detection and fault detection, as well as performing production scheduling optimization and product traceability, the problems of low efficiency, quality fluctuations and high costs existing in traditional production are solved, and the level of intelligent production management can be improved; by utilizing massive data in the production process and applying advanced data analysis methods, the production process can be optimized, quality can be accurately controlled and cost can be effectively managed, providing sandwich production companies with stronger competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] For better understanding and implementation, the technical solution of the present application is described in detail below with reference to the accompanying drawings.

[0012] Figure 1 A flowchart of the steps of a data mining and analysis method applied to a sandwich production control system provided in an embodiment of the present application; Figure 2 A flowchart of the steps for production status prediction provided in an embodiment of the present application; Figure 3 A flowchart of the steps of quality prediction provided in the embodiment of the present application; Figure 4 A flowchart of the steps of fault prediction provided in an embodiment of the present application; Figure 5 A structural schematic diagram of a data mining and analysis system applied to a sandwich production control system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0013] In order to further explain the technical means and effects taken by the present invention to achieve the predetermined invention purpose, exemplary embodiments will be described in detail here, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of methods and systems consistent with some aspects of the present application as detailed in the attached claims.

[0014] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in this article refers to any or all possible combinations of one or more associated listed items.

[0015] The specific implementation methods, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0016] Example 1 See also Figure 1 The present application embodiment provides a data mining and analysis method applied to a sandwich production control system, comprising the following steps: S1, collects historical production data, historical quality data and historical equipment operation data of the sandwich production control system; S2, using stream data processing technology to monitor the production status of the sandwich production control system in real time to obtain multi-source real-time production data, and combining the historical production data to predict the future production status of the sandwich production control system; S3, predicting the final quality of the sandwich using a deep learning algorithm based on the historical quality data; S4, during the production process of the sandwiches, using computer vision technology combined with cameras and image recognition algorithms to perform defect detection on the sandwiches to obtain defect data, and performing product screening and production line adjustment based on the defect data; S5, collecting real-time equipment operation data of the sandwich production equipment by monitoring the equipment status, and performing fault prediction by combining the historical equipment operation data and using a deep learning algorithm and an association analysis algorithm; S6, in the sandwich production process, using an optimization algorithm to optimize the production scheduling of each production link in the sandwich production control system; S7, collecting product data of the sandwiches, and tracing the sandwiches with quality problems in combination with the defect data.

[0017] Specifically, since the existing sandwich production control system may have problems in the production process such as: differences in production efficiency, uneven production scheduling and resource allocation, fluctuations in the quality of different production links, and increases in production costs due to fluctuations in raw material costs, in order to solve the above problems, this application uses a variety of technologies on the basis of data mining to process and analyze the mined data, so as to improve production efficiency and quality and reduce production costs. Specifically, the following contents are included: first, historical production data, historical quality data and historical equipment operation data of the sandwich production control system are collected; then, stream data processing technology is used to monitor the production status of the production control system in real time to obtain multi-source real-time production data, and the subsequent production of the production control system is processed in combination with the historical production data. The production status is predicted; then, based on the historical quality data, a deep learning algorithm is used to predict the final quality of the sandwich; then, in the production process of the sandwich, computer vision technology is used in combination with a camera and an image recognition algorithm to detect defects in the sandwich to obtain defect data, and product screening and production line adjustment are performed based on the defect data; secondly, the real-time equipment operation data of the sandwich production equipment is collected through equipment status monitoring, and fault prediction is performed in combination with the historical equipment operation data and a deep learning algorithm; then, in the production process of the sandwich, an optimization algorithm is used to optimize the production scheduling of each production link in the production control system; finally, the product data of the sandwich is collected, and the sandwich with quality problems is traced in combination with the defect data. Through the above content, this application solves the problems of low efficiency, quality fluctuations and high costs existing in traditional production, and can also improve the level of intelligent production management; by utilizing massive data in the production process and applying advanced data analysis methods, the production process can be optimized, quality can be accurately controlled, and costs can be effectively managed, providing sandwich production companies with stronger competitiveness.

[0018] like Figure 2 As shown, in an embodiment provided in the present application, step S2 specifically includes: Collect data from multiple data sources from the production control system of Sandwich to form multi-source real-time production data; Preprocessing the multi-source real-time production data using stream data processing technology, wherein the preprocessing includes data cleaning and data aggregation; Based on the historical production data, a time series analysis method is used to conduct a comprehensive analysis to obtain the historical laws of production status; Based on the preprocessed multi-source real-time production data and combined with the historical laws of the production status, a deep learning model is used to predict the future production status, specifically including: Production volume forecast: predict the number of sandwiches produced in the future; Resource consumption forecasting: forecasting the consumption of raw materials in the sandwich production process; Wherein, the deep learning model includes a convolutional neural network or a recurrent neural network; The time series analysis method includes an ARIMA algorithm or a LSTM algorithm.

[0019] Specifically, this embodiment uses stream data processing technology to monitor the status of the production control system in real time and make predictions based on historical data, which can effectively improve the intelligence level of the production process, timely discover potential problems, optimize resource allocation, improve production efficiency, and achieve accurate production scheduling and quality control. It should be noted that the ARIMA (AutoRegressive Integrated Moving Average) algorithm is a classic time series analysis method, mainly used to predict numerical data with time dependence; the LSTM (Long Short-Term Memory) algorithm is a special type of recurrent neural network, specially designed to solve the "gradient disappearance" problem faced by traditional recurrent neural networks when processing long sequences. LSTM can memorize and process long-term dependencies in time series data for a long time by introducing a structure called memory cell.

[0020] like Figure 3 As shown, in an embodiment provided in the present application, step S3 specifically includes: Acquire sample data from the historical quality data, and perform data preprocessing on the sample data; Determining at least one key process in the whole process of the sandwich production control system; Creating a corresponding key process prediction model for the at least one key process; Using sample data of key processes to train and calibrate the key process prediction model; Using the key process prediction model to predict the key process, and obtain the prediction result of the key process; creating a full process prediction model for the entire process based on the prediction results of the key process; The whole process prediction model of the whole process is created according to the prediction results of the key process, including: Create a feedback neural network model RNNS as a full-process prediction model; The minimum value of the prediction accuracy in the prediction results of the key process is used as the feedback value of the whole process prediction model and introduced into the feedback neural network model RNNS; Using sample data of the whole process to train and calibrate the whole process prediction model; The whole process prediction model is used to predict product quality.

[0021] Specifically, this embodiment predicts the key processes in the sandwich production process, and uses the prediction results of the key processes as an input for the overall product quality prediction, thereby improving the accuracy of the sandwich quality prediction and making the process of the sandwich production control system controllable; and facilitating the tracing of the root causes of product problems, thereby improving the process flow.

[0022] In an embodiment provided in the present application, step S4 specifically includes: Obtain the image of the target sandwich to be detected; Using the trained image segmentation network, identifying the target overall image corresponding to the target sandwich from the image to be detected, and identifying the defective portion image in the first shape; Determine a first sub-image according to the target overall image and the defective portion image, and determine a second sub-image according to the defective portion image; Performing defect detection on the first sub-image and the second sub-image respectively, and if a defect is detected in at least one sub-image, determining that the target sandwich has a defect; Analyzing defects of the target sandwich to obtain defect data, and performing product screening and production line adjustment based on the defect data; The determining of the first sub-image according to the target overall image and the defective partial image comprises: selecting an area in the target overall image whose position matches the defective partial image but whose size is smaller than the defective partial image as the area corresponding to the defective partial image, and filling the area in the target overall image corresponding to the defective partial image with a specified color to obtain the first sub-image; Determining the second sub-image according to the defective portion image includes: extracting a region whose position matches the defective portion image but whose size is larger than the defective portion image from the image to be detected, to obtain the second sub-image.

[0023] In an embodiment provided in the present application, the step S4 further includes: In case a defect is detected, determining sub-image pixel coordinates of the defect in the corresponding sub-image; According to the position of the target overall image and / or the defective partial image in the image to be detected, the sub-image pixel coordinates of the defect are transformed to obtain the original image pixel coordinates of the defect in the image to be detected, so as to determine the position of the defect in the target based on the original image pixel coordinates of the defect in the image to be detected.

[0024] Specifically, this embodiment uses computer vision technology combined with cameras and image recognition algorithms to detect defects in sandwiches, monitors the appearance and quality of finished products of sandwiches in real time during the production process, such as the color of bread baking, the uniformity of sandwich fillings, the integrity of packaging, etc., and automatically detects and identifies unqualified (defective) products, and removes or adjusts the production process, thereby improving the automation of the production line, reducing the cost and error rate of manual quality inspection, and ensuring that each product meets the standards. On the other hand, this embodiment uses a trained image segmentation network to identify the target overall image and the defective partial image in the first shape corresponding to the target sandwich from the image to be inspected through one recognition; then, by generating two sub-graphs, defect detection is performed on the sub-graphs respectively. If at least one sub-graph can detect defects, it is determined that the target has defects. In this way, the integrity of the target sandwich is considered, and the focus is on the local area, which improves the accuracy of sandwich defect detection, and the process is simple, which is particularly suitable for product defect detection on sandwich production lines.

[0025] like Figure 4 As shown, in one embodiment provided in the present application, a deep learning algorithm is used to perform fault prediction in step S5, specifically including: Real-time equipment operation data is collected in real time through sensors installed on sandwich production equipment; Performing data preprocessing on the historical equipment operation data and the real-time equipment operation data, wherein the data preprocessing includes missing value processing, noise removal and normalization processing; Selecting equipment operation features that affect equipment failure from the historical equipment operation data, and using a statistical analysis method to screen out features that are highly correlated with equipment failure as historical equipment failure features; Filtering out real-time equipment failure characteristics from the real-time equipment operation data based on the historical equipment failure characteristics; Performing feature construction based on the real-time device fault feature to generate a real-time device data stream feature includes: By using sliding window technology, time series features are constructed from the real-time equipment fault features to capture the dynamic changes in the state of the production equipment; Calculate statistical characteristics of production equipment, including mean, maximum, standard deviation, and crest factor; Use Fourier transform to extract frequency domain features to identify abnormal patterns in production equipment; Aggregating the time series features, the statistical features and the frequency domain features into real-time device data stream features; Based on the historical equipment failure characteristics, a long short-term memory network and a convolutional neural network are combined to perform model training to generate an equipment failure prediction model; Inputting the real-time device data stream characteristics into the device fault prediction model to perform fault prediction; The real-time equipment operation data includes temperature, vibration, pressure, current, voltage, speed, working time and load.

[0026] Specifically, this embodiment can achieve accurate fault prediction of sandwich production equipment by combining real-time data of equipment status monitoring, historical equipment data and deep learning algorithms. The specific steps include data collection and preprocessing, feature engineering, deep learning model training, fault prediction and identification. In this way, not only can the reliability of equipment operation be improved and the failure rate be reduced, but also production scheduling and resource allocation can be optimized, and production efficiency and economic benefits can be improved.

[0027] In an embodiment provided in the present application, in step S5, a correlation analysis algorithm is used to perform fault prediction, specifically including: Acquire real-time equipment operation data of the sandwich production equipment to be tested at multiple predetermined time points within a predetermined time period; Arranging the real-time equipment operation data at the plurality of predetermined time points into an operation parameter time series input matrix according to the time dimension and the sample dimension; Performing inter-parameter timing correlation feature analysis on the operating parameter timing input matrix to obtain global timing features of the operating parameters; Based on the global time series characteristics of the operating parameters, determining whether to generate a fault warning prompt; The step of performing a timing correlation feature analysis between parameters on the operating parameter timing input matrix to obtain a global timing feature of the operating parameters includes: The operating parameter time series input matrix is ​​subjected to matrix block processing and then passed through a ViT model including an embedding layer to obtain a sequence of local time series semantic feature vectors of the operating parameters; Performing consistency association analysis on the sequence of the local temporal semantic feature vectors of the operating parameters to obtain a consistency topological feature matrix of the operating parameters; Performing graph-structure-based associative coding on the sequence of the local temporal semantic feature vectors of the operating parameters and the consistent topological feature matrix of the operating parameters to obtain a consistent topological operating parameter global temporal feature matrix as the global temporal feature of the operating parameters; Wherein, determining whether to generate a fault warning prompt based on the global time series characteristics of the operating parameters includes: Expanding the consistency topology operation parameter global timing feature matrix into a consistency topology operation parameter global timing feature vector; Optimizing each eigenvalue of the consistent topology operation parameter global timing eigenvector to obtain an optimized consistent topology operation parameter global timing eigenvector; Passing the global time series feature vector of the optimized consistency topology operation parameter through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the operation state of the sandwich production equipment to be detected is abnormal; Based on the classification result, a fault warning prompt is generated.

[0028] Specifically, this embodiment uses a correlation analysis algorithm to predict the faults of the sandwich production equipment. By collecting multiple operating data of the sandwich production equipment and introducing data processing and analysis algorithms at the back end to perform time series correlation analysis of the operating parameters of the sandwich production equipment, it is possible to determine whether there is an abnormality in the operating status of the sandwich production equipment, and issue an early warning prompt when an abnormality is detected. In this way, automatic detection and early warning of the working status of the sandwich production equipment can be achieved, thereby improving the reliability and stability of the sandwich production equipment, reducing downtime and production losses caused by failures, and improving production efficiency and product quality.

[0029] In an embodiment provided in the present application, step S6 specifically includes: Preprocessing the historical production data and historical equipment operation data of the sandwich production control system, including denoising, filtering, extracting, cleaning, reducing and transforming; Perform data mining analysis on the pre-processed sandwich production data to obtain the correlation between the process parameters and quality indicators of each process of sandwich production, and establish a corresponding mathematical model and form visual knowledge. The data mining analysis includes correlation analysis, cluster analysis, regression analysis and prediction analysis; Based on the visualized knowledge, the entire process of sandwich production is intelligently feedback-controlled, and the production scheduling of each production link in the sandwich production control system is optimized; The data mining and analysis of the pre-processed sandwich production data specifically includes: Data association: Apriori algorithm is used to perform data association analysis on the preprocessed sandwich production data. According to the mined association relationship, the information of one attribute is inferred from the information of another attribute. When the confidence reaches a certain threshold, the association rule is considered to be established. Data clustering: Perform data clustering analysis on the preprocessed sandwich production data. According to the intrinsic properties of the sandwich production data, the data is divided into some clusters. The elements in each cluster have the same characteristics. By observing the characteristics of each cluster, further analysis is conducted on specific clusters to obtain the required information. Data regression: Perform data regression analysis on the pre-processed sandwich production data to determine the quantitative relationship between two or more variables, establish a corresponding mathematical model, fit a series of influencing factors and results to fit an equation, and then apply this equation to other similar events to form visual knowledge; Data prediction: Based on the formed visualization knowledge, data prediction analysis is performed on the pre-processed sandwich production data to form a prediction result, wherein the data prediction analysis method is a causal relationship prediction method; Among them, the production scheduling of each production link in the sandwich production control system is optimized, specifically including production resource optimization, human resource optimization, raw material and material supply chain optimization, production process optimization, energy consumption and ingredient optimization, logistics and transportation optimization and production scheduling optimization.

[0030] Specifically, this embodiment uses association analysis, cluster analysis, regression analysis and predictive analysis to study the association between the process parameters and quality indicators of each process of sandwich production, and establishes corresponding mathematical models to form visual knowledge, which provides a basis and guidance for the improvement of sandwich production technology and quality, and improves the efficiency and quality of each process of sandwich production. The optimization algorithm is also used to schedule and optimize each production link in the sandwich production control system, including production plan optimization, resource scheduling (equipment and personnel), raw material supply chain optimization, production process scheduling, energy consumption and cost optimization, logistics and distribution scheduling, etc. These optimization measures can improve production efficiency, reduce costs, ensure product quality, and ensure the smooth operation of the production process and on-time delivery. It should be noted that the above-mentioned optimization algorithms include linear programming algorithms, integer programming algorithms, deep learning, etc.

[0031] In an embodiment provided in the present application, step S7 specifically includes: Collect product data in the sandwich production control system; the product data includes raw material data, production process data, and logistics data; Using the product data as traceability data and storing it in a product database; The sandwiches with quality problems are traced back as follows: raw material data, production process data and logistics data are processed to obtain raw material score values, production process score values ​​and logistics transportation score values, which are then compared with their corresponding system score thresholds. Based on the comparison results, it is determined whether there are abnormal items. If so, it is determined to be the source of the product problem. Otherwise, it is determined that the traceability has failed. Based on big data, the stored traceability data and the defect data are processed and analyzed to obtain a quality trend coefficient, the calculated quality trend coefficient is compared with the quality trend interval preset by the sandwich production control system, and the authenticity of the traceability data is judged based on the comparison result.

[0032] Specifically, this embodiment can accurately determine the source of the problem during the traceability process through the above process to avoid misjudgment and traceability failure, and can discover potential traceability data falsification problems. By collecting and analyzing the product data, defect data and detailed information on the production process of the sandwiches, combined with traceability technology, it is possible to quickly trace sandwiches with quality problems. By accurately locating the root cause of the problem and taking effective corrective measures, product quality and the stability of the production process can be ensured, and consumer trust can be improved. In addition, the use of technical means such as the Internet of Things, blockchain and artificial intelligence can further enhance the real-time, accuracy and credibility of product traceability.

[0033] Example 2 See also Figure 5 The present application embodiment provides a data mining and analysis system applied to a sandwich production control system, comprising Data acquisition module: collects historical production data, historical quality data and historical equipment operation data of the sandwich production control system; Monitoring and prediction module: using stream data processing technology to monitor the production status of the sandwich production control system in real time to obtain multi-source real-time production data, and combining the historical production data to predict the future production status of the sandwich production control system; Quality prediction module: based on the historical quality data, using a deep learning algorithm to predict the final quality of the sandwich; Defect detection module: During the production process of sandwiches, computer vision technology combined with cameras and image recognition algorithms are used to detect defects in sandwiches to obtain defect data, and product screening and production line adjustments are performed based on the defect data; Fault prediction module: collects real-time equipment operation data of sandwich production equipment by monitoring equipment status, and performs fault prediction by combining the historical equipment operation data and using deep learning algorithm and association analysis algorithm; Production scheduling module: in the production process of sandwiches, an optimization algorithm is used to optimize the production scheduling of each production link in the sandwich production control system; Product traceability module: collects product data of sandwiches, and combines the defect data to trace the sandwiches with quality problems.

[0034] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0035] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0036] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0037] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A data mining analysis method applied to a sandwich production control system, characterized by: The steps include: S1, collects historical production data, historical quality data and historical equipment operation data of the sandwich production control system; S2, using stream data processing technology to monitor the production status of the sandwich production control system in real time to obtain multi-source real-time production data, and combining the historical production data to predict the future production status of the sandwich production control system; S3, predicting the final quality of the sandwich using a deep learning algorithm based on the historical quality data; S4, during the production process of the sandwiches, using computer vision technology combined with cameras and image recognition algorithms to perform defect detection on the sandwiches to obtain defect data, and performing product screening and production line adjustment based on the defect data; S5, collecting real-time equipment operation data of the sandwich production equipment by monitoring the equipment status, and performing fault prediction by combining the historical equipment operation data and using a deep learning algorithm and an association analysis algorithm; S6, in the sandwich production process, using an optimization algorithm to optimize the production scheduling of each production link in the sandwich production control system; S7, collecting product data of the sandwiches, and tracing the sandwiches with quality problems in combination with the defect data.

2. The data mining and analysis method for sandwich production control system according to claim 1 is characterized in that: Step S2 specifically includes: Collect data from multiple data sources from the production control system of Sandwich to form multi-source real-time production data; Preprocessing the multi-source real-time production data using stream data processing technology, wherein the preprocessing includes data cleaning and data aggregation; Based on the historical production data, a time series analysis method is used to conduct a comprehensive analysis to obtain the historical laws of production status; Based on the preprocessed multi-source real-time production data and combined with the historical laws of the production status, a deep learning model is used to predict the future production status, specifically including: Production volume forecast: predict the number of sandwiches produced in the future; Resource consumption forecasting: forecasting the consumption of raw materials in the sandwich production process; Wherein, the deep learning model includes a convolutional neural network or a recurrent neural network; The time series analysis method includes an ARIMA algorithm or a LSTM algorithm.

3. The data mining and analysis method for sandwich production control system according to claim 1 is characterized in that: Step S3 specifically includes: Acquire sample data from the historical quality data, and perform data preprocessing on the sample data; Determining at least one key process in the entire process of the sandwich production control system; Creating a corresponding key process prediction model for the at least one key process; Using sample data of key processes to train and calibrate the key process prediction model; Using the key process prediction model to predict the key process, and obtain the prediction result of the key process; creating a full process prediction model for the entire process based on the prediction results of the key processes; The whole process prediction model of the whole process is created according to the prediction results of the key process, including: Create a feedback neural network model RNNS as a full-process prediction model; The minimum value of the prediction accuracy in the prediction results of the key process is used as the feedback value of the whole process prediction model and introduced into the feedback neural network model RNNS; Using sample data of the whole process to train and calibrate the whole process prediction model; The whole process prediction model is used to predict product quality.

4. The data mining and analysis method for sandwich production control system according to claim 1 is characterized in that: Step S4 specifically includes: Obtain the image of the target sandwich to be detected; Using the trained image segmentation network, identifying the target overall image corresponding to the target sandwich from the image to be detected, and identifying the defective portion image in the first shape; Determine a first sub-image according to the target overall image and the defective portion image, and determine a second sub-image according to the defective portion image; Performing defect detection on the first sub-image and the second sub-image respectively, and if a defect is detected in at least one sub-image, determining that the target sandwich has a defect; Analyzing defects of the target sandwich to obtain defect data, and performing product screening and production line adjustment based on the defect data; The determining of the first sub-image according to the target overall image and the defective partial image comprises: selecting an area in the target overall image whose position matches the defective partial image but whose size is smaller than the defective partial image as the area corresponding to the defective partial image, and filling the area in the target overall image corresponding to the defective partial image with a specified color to obtain the first sub-image; Determining the second sub-image according to the defective portion image includes: extracting a region whose position matches the defective portion image but whose size is larger than the defective portion image from the image to be detected, to obtain the second sub-image.

5. The data mining and analysis method for sandwich production control system according to claim 4 is characterized in that: The step S4 further comprises: In case a defect is detected, determining sub-image pixel coordinates of the defect in the corresponding sub-image; According to the position of the target overall image and / or the defective partial image in the image to be detected, the sub-image pixel coordinates of the defect are transformed to obtain the original image pixel coordinates of the defect in the image to be detected, so as to determine the position of the defect in the target based on the original image pixel coordinates of the defect in the image to be detected.

6. The data mining and analysis method for sandwich production control system according to claim 1, characterized in that: In step S5, a deep learning algorithm is used to predict faults, which specifically includes: Real-time equipment operation data is collected in real time through sensors installed on sandwich production equipment; Performing data preprocessing on the historical equipment operation data and the real-time equipment operation data, wherein the data preprocessing includes missing value processing, noise removal and normalization processing; Selecting equipment operation features that affect equipment failure from the historical equipment operation data, and using a statistical analysis method to screen out features that are highly correlated with equipment failure as historical equipment failure features; Filtering out real-time equipment failure characteristics from the real-time equipment operation data based on the historical equipment failure characteristics; Performing feature construction based on the real-time device fault feature to generate a real-time device data stream feature includes: By using sliding window technology, time series features are constructed from the real-time equipment fault features to capture the dynamic changes in the state of the production equipment; Calculate statistical characteristics of production equipment, including mean, maximum, standard deviation, and crest factor; Use Fourier transform to extract frequency domain features to identify abnormal patterns in production equipment; Aggregating the time series features, the statistical features and the frequency domain features into real-time device data stream features; Based on the historical equipment failure characteristics, a long short-term memory network and a convolutional neural network are combined to perform model training to generate an equipment failure prediction model; Inputting the real-time device data stream characteristics into the device fault prediction model to perform fault prediction; The real-time equipment operation data includes temperature, vibration, pressure, current, voltage, speed, working time and load.

7. The data mining and analysis method for sandwich production control system according to claim 6 is characterized in that: In step S5, a correlation analysis algorithm is used to perform fault prediction, which specifically includes: Acquire real-time equipment operation data of the sandwich production equipment to be tested at multiple predetermined time points within a predetermined time period; Arranging the real-time equipment operation data at the plurality of predetermined time points into an operation parameter time series input matrix according to the time dimension and the sample dimension; Performing inter-parameter timing correlation feature analysis on the operating parameter timing input matrix to obtain global timing features of the operating parameters; Based on the global time series characteristics of the operating parameters, determining whether to generate a fault warning prompt; The step of performing a timing correlation feature analysis between parameters on the operating parameter timing input matrix to obtain a global timing feature of the operating parameters includes: The operating parameter time series input matrix is ​​subjected to matrix block processing and then passed through a ViT model including an embedding layer to obtain a sequence of local time series semantic feature vectors of the operating parameters; Performing consistency association analysis on the sequence of the local temporal semantic feature vectors of the operating parameters to obtain a consistency topological feature matrix of the operating parameters; Performing graph-structure-based associative coding on the sequence of the local temporal semantic feature vectors of the operating parameters and the consistent topological feature matrix of the operating parameters to obtain a consistent topological operating parameter global temporal feature matrix as the global temporal feature of the operating parameters; Wherein, determining whether to generate a fault warning prompt based on the global time series characteristics of the operating parameters includes: Expanding the consistency topology operation parameter global timing feature matrix into a consistency topology operation parameter global timing feature vector; Optimizing each eigenvalue of the consistent topology operation parameter global timing eigenvector to obtain an optimized consistent topology operation parameter global timing eigenvector; Passing the global time series feature vector of the optimized consistency topology operation parameter through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the operation state of the sandwich production equipment to be detected is abnormal; Based on the classification result, a fault warning prompt is generated.

8. The data mining and analysis method for sandwich production control system according to claim 1, characterized in that: Step S6 specifically includes: Preprocessing the historical production data and historical equipment operation data of the sandwich production control system, including denoising, filtering, extracting, cleaning, reducing and transforming; Perform data mining analysis on the pre-processed sandwich production data to obtain the correlation between the process parameters and quality indicators of each process of sandwich production, and establish a corresponding mathematical model and form visual knowledge. The data mining analysis includes correlation analysis, cluster analysis, regression analysis and prediction analysis; Based on the visualized knowledge, the entire process of sandwich production is intelligently feedback-controlled, and the production scheduling of each production link in the sandwich production control system is optimized.

9. The data mining and analysis method for sandwich production control system according to claim 8, characterized in that: The data mining analysis of the pre-processed sandwich production data specifically includes: Data association: Apriori algorithm is used to perform data association analysis on the preprocessed sandwich production data. According to the mined association relationship, the information of one attribute is inferred from the information of another attribute. When the confidence reaches a certain threshold, the association rule is considered to be established. Data clustering: Perform data clustering analysis on the preprocessed sandwich production data. According to the intrinsic properties of the sandwich production data, the data is divided into some clusters. The elements in each cluster have the same characteristics. By observing the characteristics of each cluster, further analysis is conducted on specific clusters to obtain the required information. Data regression: Perform data regression analysis on the pre-processed sandwich production data to determine the quantitative relationship between two or more variables, establish a corresponding mathematical model, fit a series of influencing factors and results to fit an equation, and then apply this equation to other similar events to form visual knowledge; Data prediction: Based on the formed visualization knowledge, data prediction analysis is performed on the pre-processed sandwich production data to form a prediction result, wherein the data prediction analysis method is a causal relationship prediction method; Among them, the production scheduling of each production link in the sandwich production control system is optimized, specifically including production resource optimization, human resource optimization, raw material and material supply chain optimization, production process optimization, energy consumption and ingredient optimization, logistics and transportation optimization and production scheduling optimization.

10. The data mining and analysis method for sandwich production control system according to claim 1, characterized in that: Step S7 specifically includes: Collect product data in the sandwich production control system; the product data includes raw material data, production process data, and logistics data; Using the product data as traceability data and storing it in a product database; The sandwiches with quality problems are traced back as follows: raw material data, production process data and logistics data are processed to obtain raw material score values, production process score values ​​and logistics transportation score values, which are then compared with their corresponding system score thresholds. Based on the comparison results, it is determined whether there are abnormal items. If so, it is determined to be the source of the product problem. Otherwise, it is determined that the traceability has failed. Based on big data, the stored traceability data and the defect data are processed and analyzed to obtain a quality trend coefficient, the calculated quality trend coefficient is compared with the quality trend interval preset by the sandwich production control system, and the authenticity of the traceability data is judged based on the comparison result.

Images