Intelligent manufacturing process control chart mode identification method and system for metal sawing band sawing machine

Abstract

The invention belongs to the technical field of metal sawing band sawing machine manufacturing process quality management control, and discloses a metal sawing band sawing machine intelligent manufacturing process control chart mode identification method, which comprises the following steps: acquiring high-frequency, intermediate-frequency and low-frequency signals; then, taking the high-frequency signal, the intermediate-frequency signal and the low-frequency signal with the stationary characteristics as the input of a linear support vector machine and taking the high-frequency signal, the intermediate-frequency signal and the low-frequency signal with the non-stationary characteristics as the input of a multi-kernel nonlinear support vector machine to carry out period, trend and step mode identification; and finally, according to the obtained corresponding category, giving out corresponding abnormity judgment and prediction of the change trend of the next stage through a production process control method, and drawing a control chart at the same time to realize intelligent mode identification of the metal sawing band sawing machine manufacturing process control chart. According to the method, quality management in the manufacturing and production process of the metal sawing band sawing machine is effectively achieved, and production efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of quality management and control of the manufacturing process of a metal sawing band saw machine, and in particular relates to a control chart pattern recognition method and system for an intelligent manufacturing process of a metal sawing band saw machine. Background Art

[0002] With the rapid development of intelligent manufacturing, metal sawing machine tools, as the "mother machine" of industrial manufacturing equipment, have a direct impact on the efficiency, stability and intelligence of manufacturing equipment processing and enterprise competitiveness. As one of the important quality management and control technologies, the control chart pattern recognition method can accurately identify and promptly handle abnormal fluctuations in the production process through feature extraction, feature screening, pattern recognition algorithms, and result verification and optimization. Most of the existing manufacturing companies' applications of quality management and control use methods such as statistical process control (SPC) to collect and analyze production data, find out the root causes of quality problems, and take corresponding improvement measures.

[0003] However, there are few reports on quality management and control of the manufacturing process of metal sawing band saws, and how to achieve efficient quality management of the manufacturing process of metal sawing band saws has not been well solved. Therefore, the development of a pattern recognition method for intelligent manufacturing process control charts of metal sawing band saws has important practical significance and application value.

[0004] Through the above analysis, the problems and defects of the prior art are as follows:

[0005] (1) In terms of manufacturing process analysis and data collection preprocessing, it is necessary to give the key quality parameters of the control points of the metal sawing band saw machine manufacturing process, classify the stationary signals and non-stationary signals respectively, and preprocess the collected raw data to deal with various types of noise and other irrelevant information and missing information mixed in the raw data during the manufacturing process, so as to reduce its impact on the efficient management of the quality of the metal sawing band saw machine manufacturing production process.

[0006] (2) In terms of feature extraction and identification of quality parameter signals involved in the manufacturing process of metal sawing band saws, it is necessary to decompose the data of quality parameters of the manufacturing process of metal sawing band saws and perform feature extraction, and perform various pattern recognitions based on classification, so as to ensure the realization of efficient quality management of the manufacturing process of metal sawing band saws. Summary of the invention

[0007] In view of the problems existing in the prior art, the present invention provides a method for pattern recognition of a control chart of an intelligent manufacturing process of a metal sawing band saw.

[0008] The present invention is implemented in this way: a metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method comprises:

[0009] Step 1: Analyze the manufacturing process of metal sawing band saws, give the key quality parameters of the manufacturing process control points, classify the stationary signals and non-stationary signals, and collect data;

[0010] Step 2: pre-process the raw data to deal with all kinds of irrelevant information and missing information such as noise mixed in the raw data during the manufacturing process;

[0011] Step 3, using Fourier transform method to process stationary signals and using continuous wavelet analysis method with high and low filters to process non-stationary signals, respectively, by decomposing the data of quality parameters of the metal sawing band saw manufacturing process and extracting features, high-frequency, medium-frequency and low-frequency signals are obtained;

[0012] Step 4, respectively taking the high frequency, medium frequency and low frequency signals with stable characteristics as the input of the linear support vector machine SVM and taking the high frequency, medium frequency and low frequency signals with non-stationary characteristics as the input of the multi-core nonlinear support vector machine SVM to identify the period, trend and step mode;

[0013] Step 5, according to the corresponding category, the corresponding abnormal judgment and the prediction of the change trend in the next stage are given through the production process control method, and the control chart is drawn at the same time to realize the intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

[0014] Furthermore, in step 1, according to the manufacturing process of the metal sawing band saw, it is analyzed from its operating mechanism that the key quality parameters involving the feeding module, the holding module and the supporting module are steady signals, and the key quality parameters involving the sawing module, the tensioning module and the transmission module are non-steady signals. On the basis of classification, the steady signal and non-steady signal data are collected separately.

[0015] Furthermore, in step 2, the raw data is preprocessed to process irrelevant information and missing information such as various noises mixed in the raw data during the manufacturing process:

[0016] Step 2.1) For a set of data k containing stationary signals mixed with various types of noise and other irrelevant information and missing information i , i=1,2,…,i,…,n, n>1 and is an odd number, first take out a set of sample data p i , i=1,2…,m,1<m<n, m is an odd number, when data p i When missing data occurs, the corresponding data is automatically filled in according to the time series, so that the mean value can be obtained. and standard deviation δ p for

[0017]

[0018] when Then p i is mixed noise, and is replaced by When the data k i Missing, That is, when data is missing, use sample data p i The mean value p is filled to avoid abnormality and inability to perform data analysis and solution;

[0019] Step 2.2) For a set of data h containing non-stationary signals mixed with various types of noise and other irrelevant information and missing information i , i=1,2,…,i,…,n, n>1 and is an odd number, when the data h i When missing data occurs, according to the time series, the current data That is, when data is missing, the corresponding data mean before and after the current moment is used to fill it, so that the mean value can be obtained. and standard deviation δ l for

[0020]

[0021] when and is an integer, which is adjusted in real time according to the noise size of the non-stationary signal, then h i is mixed noise, and is replaced by Avoid abnormalities that make it impossible to perform data analysis and solutions.

[0022] Further, in step 3, the Fourier transform method is used to process the stationary signal and the wavelet analysis method with high and low filters is used to process the non-stationary signal. By decomposing the data of the quality parameters of the metal sawing band saw manufacturing process and performing feature extraction, three signals of high frequency, medium frequency and low frequency are obtained, which contain the information of different process modules of the metal sawing band saw manufacturing process. The low-frequency part contains the transmission accuracy, support force, tension force, coolant flow rate and sawing workpiece temperature of the sawing module, tensioning module and transmission module, and the trend and step signals are the physical quantities that change slowly. The medium-frequency part contains motion signals such as sawing speed, driving torque and scraper speed. These signals will affect the sawing quality of the metal sawing band saw and are the physical quantities that change quickly. The high-frequency part contains non-periodic signals with sudden changes such as worker operation errors and equipment failures and noise signals that have not been preprocessed. For the above-mentioned types of stationary signals, the Fourier transform method is used as follows:

[0023]

[0024] w kis the frequency variable, j represents the imaginary unit; for the above-mentioned non-stationary signals, high and low filters are first used to process them respectively, and then the M-level wavelet transform is used as follows:

[0025]

[0026] where w aM and w bm is the wavelet coefficient, A M (n) is the Haar wavelet basis function, B m (n) is a Morlet wavelet basis function, which decomposes a signal related to the key quality parameter into multiple sub-signals, wherein each sub-signal can represent a different frequency range.

[0027] Furthermore, in step 4, the high-frequency, medium-frequency, and low-frequency signals with stationary characteristics are used as inputs of a linear support vector machine, and the high-frequency, medium-frequency, and low-frequency signals with non-stationary characteristics are used as inputs of a multi-core nonlinear support vector machine to identify periods, trends, and step patterns. During training, samples of a certain quality parameter are classified into one category in turn, so that a corresponding number of samples of key quality parameters analyzed by design are constructed.

[0028] Furthermore, in step 5, the corresponding abnormality judgment and the prediction of the change trend in the next stage are given according to the corresponding category through the production process control method, and the control chart is drawn at the same time to realize the intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

[0029] Another object of the present invention is to provide a metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system comprising:

[0030] The analysis module is used to analyze the manufacturing process of metal sawing band saws, provide key quality parameters of the manufacturing process control points, classify stationary signals and non-stationary signals, and perform data collection;

[0031] The preprocessing module is used to preprocess the raw data and process all kinds of irrelevant information and missing information such as noise mixed in the raw data during the manufacturing process;

[0032] The feature extraction module is used to process stationary signals by Fourier transform method and non-stationary signals by continuous wavelet analysis method with high and low filters. The high-frequency, medium-frequency and low-frequency signals are obtained by decomposing the data of quality parameters of the metal sawing band saw manufacturing process and extracting features.

[0033] The recognition module is used to use the high-frequency, medium-frequency and low-frequency signals with stable characteristics as the input of the linear support vector machine SVM and use the high-frequency, medium-frequency and low-frequency signals with non-stationary characteristics as the input of the multi-core nonlinear support vector machine SVM to identify the period, trend and step mode;

[0034] The prediction module is used to give corresponding abnormal judgments and predictions of the next stage of change trends based on the corresponding categories through the production process control method, and draw control charts at the same time to realize intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

[0035] Another object of the present invention is to provide a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method.

[0036] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method.

[0037] Another object of the present invention is to provide an information data processing terminal, which is used to implement the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system.

[0038] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0039] First, the existing process control chart pattern recognition is difficult to handle the variable working conditions, multiple processes and a large amount of process data with "noise" in the production process of metal sawing band saws. Therefore, it is necessary to reasonably analyze the key quality control parameters of its production process, and use process data to achieve accurate and intelligent management and control of the production process and effective diagnosis of the operating status of the manufacturing process through data preprocessing, intelligent algorithms and other technologies to ensure a smooth production process and good product quality. In response to the above problems, the present invention provides a pattern recognition method for intelligently managing and controlling key quality parameters of the production process of metal sawing band saws. This patent proposes a method for analyzing the manufacturing process of metal sawing band saws, giving key quality parameters of the manufacturing process control points, and classifying stationary signals and non-stationary signals respectively. It is effective and efficient to process irrelevant information and missing information such as various types of noise mixed in the original data during the manufacturing process. It adopts the Fourier transform method to process stationary signals and the continuous wavelet analysis method with high and low filters to process non-stationary signals. By decomposing the data of the quality parameters of the metal sawing band saw manufacturing process and extracting features, high-frequency, medium-frequency and low-frequency signals are obtained, and they are effectively classified according to their production processes. Then, the high-frequency, medium-frequency and low-frequency signals with stationary characteristics are used as the input of the linear support vector machine and the high-frequency, medium-frequency and low-frequency signals with non-stationary characteristics are used as the input of the multi-core nonlinear support vector machine to identify the period, trend and step mode, which well solves the problem of pattern recognition of the control chart of the production process of metal sawing band saws. Compared with the process control chart pattern recognition method of the traditional analysis method, the present invention effectively gives the key quality parameter classification of the manufacturing process control points based on the analysis of the operation mechanism of the metal sawing band saw manufacturing process. According to their different types, the Fourier transform method, wavelet analysis method, linear support vector machine and multi-core nonlinear support vector machine are used respectively to design an intelligent metal sawing band saw production process control chart pattern recognition method, which improves the recognition accuracy and ability of its production process control chart.

[0040] Second, the technical solution of the present invention uses advanced pattern recognition technology to intelligently analyze and identify the control diagram in the production process of metal sawing band saws. This method aims to achieve early warning and accurate identification of abnormal fluctuations in the production process, thereby improving production efficiency and product quality.

[0041] Intelligent feature extraction: Automatically extract key feature information from control charts, which can accurately reflect the stability and abnormal fluctuations of the production process. The feature extraction process has a high level of automation and intelligence, reducing the impact of manual intervention and subjective judgment.

[0042] Accurate pattern recognition: Use advanced pattern recognition algorithms, such as neural networks and support vector machines, to learn and classify the extracted features.

[0043] The recognition algorithm has high recognition accuracy and generalization ability, and can accurately identify different types of abnormal patterns.

[0044] Real-time monitoring and early warning: Realize real-time monitoring of the production process and capture abnormal fluctuation signals in time. Provide early warning function, when abnormalities occur in the production process, it can quickly issue an alarm and give corresponding treatment suggestions.

[0045] System integration and visualization: Integrate the technical solution with the existing production management system to achieve seamless connection and sharing of data. Provide a visual interface and operating platform to facilitate users to view, analyze and process data.

[0046] Application effect of technical solutions:

[0047] Improve production efficiency: Through real-time monitoring and early warning functions, abnormal fluctuations in the production process can be discovered and handled in a timely manner to avoid production interruptions and delays. Optimize production processes and improve production efficiency and equipment utilization.

[0048] Improve product quality: Accurately identify quality problems in the production process and take timely measures to improve and optimize. Improve product stability and consistency, and reduce defective product rates and customer complaints.

[0049] Reduce production costs: Reduce production costs and waste by reducing production interruptions and defective product rates. Improve production efficiency and equipment utilization to further reduce production costs.

[0050] 4. Comparison of technical solutions at home and abroad

[0051] Compared with existing manufacturing process control technologies at home and abroad, this technical solution has the following advantages:

[0052] Higher level of intelligence: It uses advanced pattern recognition technology and algorithms to achieve intelligent analysis and recognition of control charts. Compared with traditional manual monitoring and judgment methods, it has a higher level of intelligence and automation.

[0053] Stronger recognition accuracy and generalization ability: The recognition algorithm has high recognition accuracy and generalization ability, and can accurately identify different types of abnormal patterns. Compared with other similar recognition methods, it has stronger robustness and adaptability.

[0054] More comprehensive functions and application scenarios: A comprehensive solution that integrates real-time monitoring, early warning, data analysis, visualization and other functions. It is suitable for the production and manufacturing process of different types of metal sawing band saws and has broad application prospects and market potential.

[0055] To sum up, the technical solution of the pattern recognition method of the control chart in the production and manufacturing process of metal sawing band saws has obvious advantages in terms of intelligence level, recognition accuracy and generalization ability, functions and application scenarios, filling the technical gap in the industry at home and abroad, and providing strong support for technological innovation and industrial upgrading in the field of metal sawing band saw production and manufacturing.

[0056] Third, in the manufacturing process of metal sawing band saws, due to the complex process, diverse parameters and diverse signals (coexistence of stable and non-stationary signals), traditional process control methods are difficult to monitor changes in key quality parameters in real time, and are even more unable to accurately identify and predict abnormal situations. Existing technologies usually rely on manual experience or simple single variable analysis, resulting in insufficient control accuracy and slow response speed, which in turn affects product consistency and manufacturing efficiency. In addition, noise interference and data loss further increase the difficulty of monitoring manufacturing process parameters, limiting the promotion and application of intelligent production.

[0057] Currently in industrial applications, many sawing band saw manufacturers lack effective data processing and pattern recognition methods, especially when faced with complex signals. It is difficult to quickly extract key features, resulting in the inability to accurately identify cycles, trends or sudden abnormal patterns in the production process. This deficiency directly affects the stability of the manufacturing process, while increasing the scrap rate and production costs. In addition, the existing control chart drawing methods are mostly static processing, which makes it difficult to achieve real-time response to dynamic changes, is not conducive to rapid adjustment of production processes, and reduces industrial competitiveness.

[0058] The present invention solves the problem of difficulty in decomposing and extracting features of stationary and non-stationary signals in the manufacturing process of metal sawing band saws by introducing an intelligent pattern recognition method based on signal processing and support vector machines, combined with Fourier transform and continuous wavelet analysis technology. At the same time, the linear and nonlinear support vector machine (SVM) models are used to achieve accurate classification of cycles, trends and step patterns, and through intelligent drawing of control charts and trend prediction, it provides efficient technical means for abnormal judgment and process optimization of the manufacturing process, greatly improving the intelligent level of the manufacturing process.

[0059] The present invention significantly improves the quality control efficiency of the metal sawing band saw manufacturing process and reduces production losses caused by failure to detect abnormalities in a timely manner. At the same time, through intelligent pattern recognition and trend prediction technology, the stability of the manufacturing process and the consistency of the product can be effectively improved to meet the needs of high-precision and high-reliability industrial production. In addition, combined with the real-time drawing of intelligent control charts and multi-dimensional data analysis methods, the present invention has achieved a technological leap from traditional static monitoring to dynamic intelligent control, providing an innovative solution with broad application value for the field of sawing band saw manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of a method for pattern recognition of a metal sawing band saw intelligent manufacturing process control diagram provided by an embodiment of the present invention.

[0061] Figure 2 It is a structural block diagram of a pattern recognition system for a metal sawing band saw intelligent manufacturing process control diagram provided by an embodiment of the present invention.

[0062] Figure 3 It is a structural diagram of a method for pattern recognition of a metal sawing band saw intelligent manufacturing process control diagram provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0064] like Figure 1 , 3 As shown, a method for pattern recognition of a metal sawing band saw intelligent manufacturing process control diagram provided by an embodiment of the present invention comprises the following steps:

[0065] S101, analyzing the manufacturing process of metal sawing band saws, providing key quality parameters of the manufacturing process control points, classifying stationary signals and non-stationary signals, and performing data collection;

[0066] S102, pre-processing the raw data to process irrelevant information and missing information such as various noises mixed in the raw data during the manufacturing process;

[0067] S103, using Fourier transform method to process stationary signals and using continuous wavelet analysis method with high and low filters to process non-stationary signals, respectively, by decomposing the data of quality parameters of the metal sawing band saw manufacturing process and extracting features, obtaining three types of signals: high frequency, medium frequency and low frequency;

[0068] S104, respectively using the high frequency, medium frequency and low frequency signals with stable characteristics as inputs of a linear support vector machine (SVM) and using the high frequency, medium frequency and low frequency signals with non-stationary characteristics as inputs of a multi-core nonlinear support vector machine (SVM) to identify cycles, trends and step patterns;

[0069] S105, according to the corresponding category, the corresponding abnormal judgment and the prediction of the change trend in the next stage are given through the production process control method, and the control chart is drawn at the same time to realize the intelligent pattern recognition of the control chart of the metal sawing band saw manufacturing process.

[0070] The manufacturing process of metal sawing band saws includes multiple steps, and the key to its quality control lies in the effective monitoring and identification of process parameters. By analyzing the key quality parameters of each control point in the manufacturing process, such as sawing speed, blade vibration amplitude, cutting flatness, etc., the difference between stable signals (such as uniform sawing speed) and non-stationary signals (such as vibration amplitude changes) is clarified, and sensors are used to collect real-time data of these signals, providing a basis for subsequent data processing and pattern recognition.

[0071] The collected raw data often contains a variety of noise and interference, such as equipment vibration, environmental noise, sensor error, etc. Through preprocessing methods such as filtering, interpolation and normalization, irrelevant and missing information in the data is removed, so as to extract effective signals related to the sawing machine manufacturing process and ensure the accuracy of subsequent analysis. This step lays the data foundation for the intelligent recognition of control chart patterns.

[0072] According to the stationary characteristics of the signal, Fourier transform and continuous wavelet analysis methods are used to process stationary signals and non-stationary signals respectively. Fourier transform can decompose stationary signals into different frequency components and capture their stable periodic characteristics; while wavelet analysis uses high and low filters to perform multi-resolution decomposition on non-stationary signals and extract the high-frequency, medium-frequency and low-frequency features of the signal, thereby retaining the time and frequency information of the signal and providing multi-dimensional feature data for pattern recognition.

[0073] The extracted high-frequency, medium-frequency and low-frequency signal features are respectively input into the support vector machine (SVM) model for pattern recognition. For stationary signals, linear SVM is used to identify patterns such as cycles, trends and steps by building linear decision boundaries; for non-stationary signals, multi-core nonlinear SVM is used to identify complex nonlinear patterns through the mapping ability of multidimensional kernel functions. This process accurately matches the key signals in the manufacturing process with the pattern types of the control chart.

[0074] After pattern recognition, the cycle, trend and step patterns generated by the control chart are combined with production process control theory to determine the abnormalities in the manufacturing process. For example, when the step change pattern of a parameter is identified, it can be determined that there may be sudden abnormalities in the manufacturing process; trend changes indicate possible process drift or equipment problems in the future. Based on the pattern recognition results, the system predicts the change trend of the next stage of the manufacturing process, thereby optimizing manufacturing decisions.

[0075] Finally, the results of pattern recognition and trend prediction are visualized as control charts, which show the changes of key parameters in the manufacturing process in real time. This intelligent control chart can not only reflect the current process status, but also provide a reference for abnormal monitoring and process optimization, helping users to adjust production strategies in time and ensure the stability of the metal sawing band saw manufacturing process and product quality.

[0076] In S101 provided in the embodiment of the present invention, according to the manufacturing process of the metal sawing band saw machine, it is analyzed from its operating mechanism that the key quality parameters involving the feeding module, the holding module and the supporting module are stable signals, and the key quality parameters involving the sawing module, the tensioning module and the transmission module are non-stationary signals. On the basis of classification, the steady signal and non-stationary signal data are collected respectively.

[0077] In S102 provided in the embodiment of the present invention, the original data is preprocessed to process irrelevant information and missing information such as various types of noise mixed in the original data during the manufacturing process:

[0078] Step 2.1) For a set of data k containing stationary signals mixed with various types of noise and other irrelevant information and missing information i , i=1,2,…,i,…,n, n>1 and is an odd number, first take out a set of sample data p i , i=1,2…,m,1<m<n, m is an odd number, when data p i When missing data occurs, the corresponding data is automatically filled in according to the time series, so that the mean value can be obtained. and standard deviation δ p for

[0079]

[0080] when Then p i is mixed noise, and is replaced by When the data k i Missing, That is, when data is missing, use sample data p i Mean Fill it in to avoid abnormality and inability to perform data analysis and solution;

[0081] Step 2.2) For a set of data h containing non-stationary signals mixed with various types of noise and other irrelevant information and missing information i , i=1,2,…,i,…,n, n>1 and is an odd number, when the data h i When missing data occurs, according to the time series, the current data That is, when data is missing, the corresponding data mean before and after the current moment is used to fill it, so that the mean value can be obtained. and standard deviation δ l for

[0082]

[0083] when and is an integer, which is adjusted in real time according to the noise size of the non-stationary signal, then h i is mixed noise, and is replaced by Avoid abnormalities that make it impossible to perform data analysis and solutions.

[0084] In S103 provided by the embodiment of the present invention, the Fourier transform method is used to process the stationary signal and the wavelet analysis method with high and low filters is used to process the non-stationary signal. By decomposing the data of the quality parameters of the metal sawing band saw manufacturing process and performing feature extraction, three signals of high frequency, medium frequency and low frequency are obtained, which contain information of different process modules of the metal sawing band saw manufacturing process. The low frequency part contains the transmission accuracy, supporting force, tensioning force, coolant flow rate and sawing workpiece temperature of the sawing module, tensioning module and transmission module, and the trend and step signals are physical quantities that change slowly. The medium frequency part contains motion signals such as sawing speed, driving torque, scraper speed, etc. These signals will affect the sawing quality of the metal sawing band saw machine and are physical quantities that change quickly. The high frequency part contains non-periodic signals with sudden changes such as worker operation errors and equipment failures and noise signals that have not been preprocessed. For the above-mentioned types of stationary signals, the Fourier transform method is used as follows:

[0085]

[0086] w k is the frequency variable, j represents the imaginary unit; for the above-mentioned non-stationary signals, high and low filters are first used to process them respectively, and then the M-level wavelet transform is used as follows:

[0087]

[0088] where w aM and w bm is the wavelet coefficient, A M (n) is the Haar wavelet basis function, B m(n) is a Morlet wavelet basis function, which decomposes a signal related to the key quality parameter into multiple sub-signals, wherein each sub-signal can represent a different frequency range.

[0089] In S104 provided in the embodiment of the present invention, the high-frequency, medium-frequency and low-frequency signals with stable characteristics are respectively used as inputs of a linear support vector machine, and the high-frequency, medium-frequency and low-frequency signals with non-stationary characteristics are used as inputs of a multi-core nonlinear support vector machine to identify periods, trends and step patterns. During training, samples of a certain quality parameter are classified into one category in turn, so that the corresponding number of samples of the key quality parameters analyzed by the design are constructed.

[0090] In S105 provided by the embodiment of the present invention, the corresponding abnormality judgment and the prediction of the change trend in the next stage are given according to the corresponding category through the production process control method, and the control chart is drawn at the same time to realize the intelligent pattern recognition of the control chart of the metal sawing band saw manufacturing process.

[0091] like Figure 2 As shown, a metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system provided by an embodiment of the present invention includes:

[0092] The analysis module is used to analyze the manufacturing process of metal sawing band saws, provide key quality parameters of the manufacturing process control points, classify stationary signals and non-stationary signals, and perform data collection;

[0093] The preprocessing module is used to preprocess the raw data and process all kinds of irrelevant information and missing information such as noise mixed in the raw data during the manufacturing process;

[0094] The feature extraction module is used to process stationary signals by Fourier transform method and non-stationary signals by continuous wavelet analysis method with high and low filters. The high-frequency, medium-frequency and low-frequency signals are obtained by decomposing the data of quality parameters of the metal sawing band saw manufacturing process and extracting features.

[0095] The recognition module is used to use the high-frequency, medium-frequency and low-frequency signals with stable characteristics as the input of the linear support vector machine SVM and use the high-frequency, medium-frequency and low-frequency signals with non-stationary characteristics as the input of the multi-core nonlinear support vector machine SVM to identify the period, trend and step mode;

[0096] The prediction module is used to give corresponding abnormal judgments and predictions of the next stage of change trends based on the corresponding categories through the production process control method, and draw control charts at the same time to realize intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

[0097] Another object of the present invention is to provide a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method.

[0098] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method.

[0099] Another object of the present invention is to provide an information data processing terminal, which is used to implement the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system.

[0100] Compared with existing production and manufacturing process control technologies at home and abroad, this technical solution adopts advanced pattern recognition technology and algorithms to realize intelligent analysis and recognition of control charts. Compared with traditional manual monitoring and judgment methods, it has a higher level of intelligence and automation. The recognition algorithm used can accurately identify different types of abnormal patterns. Compared with other similar recognition methods, it has higher recognition accuracy and generalization ability and stronger robustness and adaptability. It also realizes a comprehensive solution integrating real-time monitoring, early warning, data analysis, visualization and other functions, which is suitable for different types of metal sawing band saw production and manufacturing processes and has broad application prospects and market potential.

[0101] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0102] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for pattern recognition of control diagrams of intelligent manufacturing process of metal sawing band saws, characterized in that: The following steps are involved: Step 1: Analyze the manufacturing process of metal sawing band saws, give the key quality parameters of the manufacturing process control points, classify the stationary signals and non-stationary signals, and collect data; Step 2: preprocess the raw data to process all kinds of irrelevant information and missing information such as noise mixed in the raw data during the manufacturing process; Step 3, using Fourier transform method to process stationary signals and using continuous wavelet analysis method with high and low filters to process non-stationary signals, respectively, by decomposing the data of quality parameters of the metal sawing band saw manufacturing process and extracting features, high-frequency, medium-frequency and low-frequency signals are obtained; Step 4, respectively taking the high frequency, medium frequency and low frequency signals with stable characteristics as the input of the linear support vector machine SVM and taking the high frequency, medium frequency and low frequency signals with non-stationary characteristics as the input of the multi-core nonlinear support vector machine SVM to identify the period, trend and step mode; Step 5, according to the corresponding category, the corresponding abnormal judgment and the prediction of the change trend in the next stage are given through the production process control method, and the control chart is drawn at the same time to realize the intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

2. The method for pattern recognition of intelligent manufacturing process control diagram of metal sawing band saw according to claim 1, characterized in that: In the step 1, according to the manufacturing process of the metal sawing band saw machine, it is analyzed from its operating mechanism that the key quality parameters involving the feeding module, the holding module and the supporting module are stable signals, and the key quality parameters involving the sawing module, the tensioning module and the transmission module are non-stationary signals. On the basis of classification, the data of the stable signals and the non-stationary signals are collected respectively.

3. The metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method according to claim 1, characterized in that: In step 2, the raw data is preprocessed to process irrelevant information and missing information such as various noises mixed in the raw data during the manufacturing process: Step 2.1) For a set of data k containing stationary signals mixed with various types of noise and other irrelevant information and missing information i , i=1,2,…,i,…,n, n>1 and is an odd number, first take out a set of sample data p i , i=1,2…,m,1<m<n, m is an odd number, when data p i When missing data occurs, the corresponding data is automatically filled in according to the time series, so that the mean value can be obtained. and standard deviation δ p for when Then p i is mixed noise, and is replaced by When the data k i There is a missing That is, when data is missing, use sample data p i Mean Fill it in to avoid abnormality and inability to perform data analysis and solution; Step 2.2) For a set of data h containing non-stationary signals mixed with various types of noise and other irrelevant information and missing information i , i=1,2,…,i,…,n, n>1 and is an odd number, when the data h i When missing data occurs, according to the time series, the current data That is, when data is missing, the corresponding data mean before and after the current moment is used to fill it, so that the mean value can be obtained. and standard deviation δ l for when and is an integer, which is adjusted in real time according to the noise size of the non-stationary signal, then h i is mixed noise, and is replaced by Avoid abnormalities that make it impossible to perform data analysis and solutions.

4. The metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method according to claim 1, characterized in that: In the step 3, the Fourier transform method is used to process the stationary signal and the wavelet analysis method with high and low filters is used to process the non-stationary signal. By decomposing the data of the quality parameters of the metal sawing band saw machine manufacturing process and performing feature extraction, three signals of high frequency, medium frequency and low frequency are obtained, which contain the information of different process modules of the metal sawing band saw machine manufacturing process. The low-frequency part contains the transmission accuracy, support force, tension force, coolant flow rate and sawing workpiece temperature of the sawing module, tensioning module and transmission module, and the trend and step signals are the physical quantities that change slowly. The medium-frequency part contains the motion signals such as sawing speed, driving torque and scraper speed. These signals will affect the sawing quality of the metal sawing band saw machine and are the physical quantities that change quickly. The high-frequency part contains the non-periodic signals of sudden changes such as worker operation errors and equipment failures and the noise signals that have not been pre-processed. For the above-mentioned types of stationary signals, the Fourier transform method is used as follows: w k is the frequency variable, j represents the imaginary unit; for the above-mentioned non-stationary signals, high and low filters are first used to process them respectively, and then the M-level wavelet transform is used as follows: where w aM and w bm is the wavelet coefficient, A M (n) is the Haar wavelet basis function, B m (n) is a Morlet wavelet basis function, which decomposes a signal related to the key quality parameter into multiple sub-signals, wherein each sub-signal can represent a different frequency range.

5. The metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method according to claim 1, characterized in that: In step 4, the high-frequency, medium-frequency and low-frequency signals with stationary characteristics are respectively used as inputs of a linear support vector machine, and the high-frequency, medium-frequency and low-frequency signals with non-stationary characteristics are used as inputs of a multi-core nonlinear support vector machine to identify periods, trends and step patterns. During training, samples of a certain quality parameter are classified into one category in turn, so that the corresponding number of samples of the key quality parameters analyzed by the design are constructed.

6. The metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method according to claim 1, characterized in that: In step 5, the corresponding abnormality judgment and the prediction of the change trend in the next stage are given according to the corresponding category through the production process control method, and the control chart is drawn at the same time to realize the intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

7. A metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system implementing the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition method as claimed in any one of claims 1 to 6, characterized in that: The metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system comprises: The analysis module is used to analyze the manufacturing process of metal sawing band saws, provide key quality parameters of the manufacturing process control points, classify stationary signals and non-stationary signals, and perform data collection; The preprocessing module is used to preprocess the raw data and process all kinds of irrelevant information and missing information such as noise mixed in the raw data during the manufacturing process; The feature extraction module is used to process stationary signals by Fourier transform method and non-stationary signals by continuous wavelet analysis method with high and low filters. The high-frequency, medium-frequency and low-frequency signals are obtained by decomposing the data of quality parameters of the metal sawing band saw manufacturing process and extracting features. The recognition module is used to use the high-frequency, medium-frequency and low-frequency signals with stable characteristics as the input of the linear support vector machine SVM and use the high-frequency, medium-frequency and low-frequency signals with non-stationary characteristics as the input of the multi-core nonlinear support vector machine SVM to identify the period, trend and step mode; The prediction module is used to give corresponding abnormal judgments and predictions of the next stage of change trends based on the corresponding categories through the production process control method, and draw control charts at the same time to realize intelligent pattern recognition of the metal sawing band saw manufacturing process control chart.

8. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the metal sawing band saw intelligent manufacturing process control chart pattern recognition method as described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the method for pattern recognition of a metal sawing band saw intelligent manufacturing process control chart as described in any one of claims 1 to 6.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the metal sawing band saw machine intelligent manufacturing process control chart pattern recognition system as described in claim 7.

Images