Hole roughness monitoring method in deep hole drilling and related device

By combining a multivariate synchronous compression algorithm and an AlexNet deep transfer learning model with an eddy current displacement sensor, real-time and non-destructive monitoring of deep hole roughness was achieved, solving the problem of inaccurate data from manual monitoring and improving the processing efficiency of deep hole parts.

CN119022870APending Publication Date: 2024-11-26XIAN UNIV OF TECH
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411145833.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

In the existing technology, deep hole roughness monitoring mainly relies on manual methods, which leads to incomplete and inaccurate monitoring data, affecting the processing efficiency of deep hole parts.

Method used

A multivariate synchronous compression algorithm and a deep transfer learning model based on the AlexNet network are used. Combined with an eddy current displacement sensor, the transverse and longitudinal vibration signals during the deep hole drilling process are collected in real time. Through time-frequency synchronous compression feature fusion and deep transfer learning model training, the roughness of the deep hole can be monitored in real time.

Benefits of technology

It improves the accuracy and comprehensiveness of monitoring, enabling accurate monitoring of hole roughness under different working conditions, real-time adjustment of processing parameters, avoidance of defective products, and improvement of processing efficiency for deep hole parts.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119022870A_ABST
    Figure CN119022870A_ABST
Patent Text Reader

Abstract

The invention discloses a hole roughness monitoring method in deep hole drilling and a related device. The method comprises the following steps: acquiring a sample signal and a deep hole surface roughness value; performing time-frequency synchronous compression feature fusion on the sample signal through a multivariable synchronous compression algorithm to obtain a sample graph; constructing a deep transfer learning model based on an AlexNet network, and training and testing the deep transfer learning model by using the sample graph; and monitoring the hole roughness in deep hole drilling by using the tested depth transfer learning model. According to the monitoring method, sample signals from the deep hole drilling process can be received and processed in real time, the hole surface roughness condition can be rapidly recognized, machining parameters can be found and adjusted in time, unqualified products are avoided, and the problems that in the prior art, monitoring data are not comprehensive and not accurate due to manual monitoring, and the machining quality is poor are solved. The roughness of the deep hole cannot be effectively monitored, and the machining efficiency of a deep hole workpiece is affected.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of drilling hole roughness monitoring, and in particular relates to a method for monitoring hole roughness in deep hole drilling and a related device. Background Art

[0002] In mechanical processing, deep holes usually refer to holes with an aspect ratio greater than 5. At present, deep hole parts have been widely used in various high-tech industries. During the use of deep hole parts, the deep hole roughness error of deep hole parts will directly affect the service performance and service life of related mechanical equipment. Effective deep hole roughness error monitoring can improve the quality of deep hole processing and enhance the service performance and service life of mechanical equipment.

[0003] The problem of deep hole roughness error monitoring is currently mainly monitored manually, specifically by operators observing or measuring the quality of the processed holes through frequent graded feeding or retracting. This monitoring method is prone to cause secondary damage to the processed surface, and can only monitor the deep hole roughness error at a local position, resulting in incomplete and inaccurate monitoring data, and unable to effectively monitor the deep hole roughness, affecting the processing efficiency of deep hole parts. Summary of the invention

[0004] The present invention provides a method and related device for monitoring the roughness of holes in deep hole drilling, so as to solve the technical defects in the prior art that manual monitoring leads to incomplete and inaccurate monitoring data, cannot effectively monitor the roughness of deep holes, and affects the processing efficiency of deep hole parts.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, a method for monitoring hole roughness in deep hole drilling is provided, comprising:

[0007] Obtain sample signals and deep hole surface roughness values;

[0008] By using a multivariate synchronous compression algorithm, the sample signal is subjected to time-frequency synchronous compression feature fusion to obtain a sample graph;

[0009] Constructing a deep transfer learning model based on the AlexNet network, and using the sample graph to train and test the deep transfer learning model;

[0010] The deep transfer learning model that passes the test is used to monitor the hole roughness during deep hole drilling.

[0011] Furthermore, the sample signal and the deep hole surface roughness value are obtained, specifically including:

[0012] An eddy current displacement sensor is used to collect the lateral vibration signal and the longitudinal vibration signal of the tool rod closest to the workpiece in real time during deep hole drilling at a sampling frequency of 3200 Hz, and the lateral vibration signal and the longitudinal vibration signal are used as sample signals;

[0013] Destructive testing is performed on the deep hole after drilling to obtain the surface roughness value of the deep hole, and the obtained surface roughness value of the deep hole is used to calibrate the obtained sample signal.

[0014] Furthermore, the sample signal is subjected to time-frequency synchronous compression feature fusion through a multivariate synchronous compression algorithm to obtain a sample graph, specifically including:

[0015] Through the multivariate synchronous compression algorithm, the time-frequency synchronous compression features of the sample signal are fused to obtain a time-frequency spectrum diagram of the fused signal, and the time-frequency spectrum diagram is used as a labeled sample diagram.

[0016] Furthermore, the time-frequency spectrum diagram includes a time-frequency spectrum T h (ω,b) and time-frequency spectrum T v (ω,b);

[0017] Where ω is frequency, b is time, and the time-frequency spectrum T is calculated respectively. h (ω, b) and the time-frequency spectrum T v (ω, b) The instantaneous frequency and instantaneous amplitude in each subband k.

[0018] Further, it also includes: calculating the multivariate joint instantaneous frequency and instantaneous amplitude of each sub-band K;

[0019] Calculate the time-frequency spectrum of the fused feature signal after multivariate synchronous compression transformation.

[0020] Furthermore, a deep transfer learning model based on the AlexNet network is constructed, and the deep transfer learning model is trained and tested using the sample graph, including:

[0021] Build a deep transfer learning model based on the AlexNet network and divide the sample graphs into training sample graph sets and test sample graph sets;

[0022] The deep transfer learning model is trained using the training sample atlas, and the trained deep transfer learning model is tested using the test sample atlas.

[0023] Furthermore, the deep transfer learning model includes a transfer sub-model and a fine-tuning sub-model.

[0024] In a second aspect, a system for monitoring the roughness of a hole during deep hole drilling is provided, comprising:

[0025] An acquisition module, used for acquiring sample signals and deep hole surface roughness values;

[0026] A feature fusion module is used to perform time-frequency synchronous compression feature fusion on sample signals;

[0027] Model building module, used to build a deep transfer learning model based on the AlexNet network;

[0028] Monitoring module, used for monitoring hole roughness during deep hole drilling.

[0029] According to a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the method for monitoring hole roughness in deep hole drilling as described above when executing the computer program.

[0030] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for monitoring hole roughness in deep hole drilling as described above is implemented.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. By using a multivariate synchronous compression algorithm to perform time-frequency synchronous compression feature fusion on the sample signal, feature information directly related to hole roughness can be extracted more effectively, and important time-frequency characteristics in the signal can be retained while reducing redundant information, thereby improving the accuracy of subsequent model analysis. The deep transfer learning model is constructed using the AlexNet network, and is trained and tested based on sample graphs, so that the model can accurately monitor hole roughness under different working conditions and conditions, and can receive and process sample signals from deep hole drilling in real time, quickly identify the surface roughness of the hole, and can promptly discover and adjust processing parameters to avoid the production of unqualified products. This solves the technical defects of the existing technology that manual monitoring leads to incomplete and inaccurate monitoring data, and cannot effectively monitor deep hole roughness, affecting the processing efficiency of deep hole parts.

[0033] 2. The use of a sampling frequency of 3200Hz can capture subtle changes in the vibration of the tool bar during deep hole drilling, which helps to more accurately reflect the dynamic characteristics of the drilling process. The high sampling rate ensures the comprehensiveness and accuracy of the data, providing a solid foundation for subsequent time-frequency analysis and feature extraction. Secondly, the simultaneous collection of transverse vibration signals and longitudinal vibration signals as sample signals can fully reflect the movement state of the tool bar during the drilling process. The vibration signals of these two dimensions complement each other, providing richer information and helping to more accurately identify the key factors affecting hole roughness.

[0034] 3. The synchronous compression algorithm improves the energy concentration of the time-frequency spectrum by concentrating the energy in the time-frequency distribution near the instantaneous frequency, making the time-frequency characteristics of the signal more prominent, helping to reduce the ambiguity on the time-frequency plane and improving the time-frequency resolution.

[0035] 4. Dividing the time-frequency spectrum into multiple sub-bands and calculating the instantaneous frequency and instantaneous amplitude of each sub-band can reveal the dynamic characteristics of the signal in different frequency ranges in more detail and more accurately identify the specific frequency components that affect the hole roughness.

[0036] 5. By calculating the multivariate joint instantaneous frequency and instantaneous amplitude of each sub-band K, the mutual correlation and mutual influence of different vibration signals in the same frequency range can be captured, thereby improving the accuracy and comprehensiveness of feature extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 The present invention provides a flow chart of a method for monitoring hole roughness in deep hole drilling;

[0039] Figure 2 The present invention provides a schematic diagram of a method for monitoring hole roughness in deep hole drilling;

[0040] Figure 3 Schematic diagram of the BTA deep hole drilling experimental device;

[0041] Figure 4 for Figure 3 Schematic diagram of the lateral vibration signal of the middle tool bar;

[0042] Figure 5 for Figure 3 Schematic diagram of the longitudinal vibration signal of the middle tool bar;

[0043] Figure 6 This is a schematic diagram of the measurement results of the surface roughness of deep hole No. 1;

[0044] Figure 7 This is a schematic diagram of the surface roughness measurement results of deep hole No. 6;

[0045] Figure 8 This is a schematic diagram of the surface roughness measurement results of deep hole No. 16;

[0046] Fig. 9It is a schematic diagram of the lateral vibration sample signal;

[0047] Fig.10 It is a schematic diagram of longitudinal vibration sample signal;

[0048] Fig.11 It is a time spectrum diagram;

[0049] Fig.12 Training loss graph for deep transfer learning model;

[0050] Fig.13 Schematic diagram of deep transfer learning model;

[0051] Fig.14 This is a schematic diagram of a hole roughness monitoring system for deep hole drilling.

[0052] In the figure: 1. spindle motor; 2. feed motor; 3. drill rod; 4. auxiliary support; 5. oiler; 6. workpiece; 7. displacement sensor; 8. data acquisition system; 9. computer. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0054] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0056] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0057] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0058] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0059] In mechanical processing, deep holes usually refer to holes with an aspect ratio greater than 5. At present, deep hole parts have been widely used in various high-tech industries. During the use of deep hole parts, the deep hole roughness error of deep hole parts will directly affect the service performance and service life of related mechanical equipment. Effective deep hole roughness error monitoring can improve the quality of deep hole processing and enhance the service performance and service life of mechanical equipment.

[0060] The problem of deep hole roughness error monitoring is currently mainly monitored manually, specifically by operators observing or measuring the quality of the processed holes through frequent graded feeding or retracting. This monitoring method is prone to cause secondary damage to the processed surface, and can only monitor the deep hole roughness error at a local position, resulting in incomplete and inaccurate monitoring data, and unable to effectively monitor the deep hole roughness, affecting the processing efficiency of deep hole parts.

[0061] In order to solve the above technical defects, the inventor provides a method and related device for monitoring hole roughness in deep hole drilling.

[0062] In a first aspect, an embodiment of the present invention provides a method for monitoring hole roughness in deep hole drilling, such as Figure 1 and Figure 2 As shown, including:

[0063] S101, obtaining sample signals and deep hole surface roughness values; illustratively, as Figure 3 As shown, a BTA deep hole drilling test bench is first built. A spindle motor 1 and a feed motor 2 are installed on the test bench. A drill rod 3 is installed on the spindle motor 1. The bottom of the drill rod 3 is supported by an auxiliary support 4. The end of the drill rod 3 is connected to the workpiece 6 through an oiler 5. Through the cooperation of the spindle motor 1 and the feed motor 2, 24 deep holes with a diameter of 17.5 mm and a depth of 90 mm are drilled on the 210*200*70 45 steel specimen. The drilling parameters are shown in Table 1 below.

[0064] Table 1 Deep hole machining parameters in the experiment

[0065]

[0066] Then, the eddy current displacement sensor 7 is used to collect the lateral vibration signal h(t) and longitudinal vibration signal v(t) of the drill rod 3 (tool rod) closest to the workpiece in deep hole drilling in real time at a sampling frequency of 3200 Hz, and the lateral vibration signal h(t) and longitudinal vibration signal v(t) are used as sample signals; Figure 4 and Figure 5 As shown, the lateral vibration signal and longitudinal vibration signal collected during a deep hole drilling process. Secondly, the deep hole after drilling is destructively tested to obtain the surface roughness value of the deep hole, and the obtained deep hole surface roughness value is used to calibrate the obtained sample signal.

[0067] S102, by using a multivariable synchronous compression algorithm, the sample signal is subjected to time-frequency synchronous compression feature fusion to obtain a sample graph; illustratively, by using a multivariable synchronous compression algorithm, the sample signal is subjected to time-frequency synchronous compression feature fusion to obtain a time-frequency spectrum of the fused signal, and the time-frequency spectrum is used as a labeled sample graph; secondly, the collected transverse vibration signal h(t) and longitudinal vibration signal v(t) are subjected to synchronous compression transformation respectively to obtain time-frequency spectrum T of the rearranged time-frequency signal respectively. h (ω,b) and T v (ω,b), where ω is frequency and b is time, and then calculate the time spectrum T h (ω,b) and T v The instantaneous frequency and instantaneous amplitude of (ω,b) in each sub-band k are shown as follows:

[0068]

[0069] in: and T h The instantaneous frequency and instantaneous amplitude of (ω, b); Sum and are respectively the instantaneous frequency and the instantaneous amplitude of T v (ω, b); ω is the frequency, b is the time, {ω k} k=1,...,K is the sub - band width.

[0070] Calculate the multivariate joint instantaneous frequency and the instantaneous amplitude of each sub - band k. The calculation process is as follows:

[0071]

[0072] Where: Sum and are respectively the joint instantaneous frequency and the instantaneous amplitude of T h (ω, b) and T v (ω, b).

[0073] Calculate the time - frequency spectrum of the fused feature signal after the multivariate synchrosqueezing transform, as shown in the following formula:

[0074]

[0075] Where: is the time - frequency spectrum after the fusion of the multivariate synchrosqueezing transform of T h (ω, b) and T v (ω, b), and δ(g) is the Dirac function.

[0076] Divide the sample atlas into training samples and test samples. As Figure 6-Figure 8 shown,刨开 the 24 deep holes after drilling along the axial direction, measure the surface roughness of the hole inner surface. If the measured value of the surface roughness Ra ≤ 1μm and the surface finish of the hole wall is good, the vibration signal corresponding to the hole wall is marked as a stable vibration sample; if the measured value of the surface roughness 1μm < Ra ≤ 3μm and there are slight marks on the hole wall, the corresponding section of the hole wall is marked as a slightly abnormal vibration sample; if the measured value of the surface roughness Ra > 3μm and there are serious marks on the hole wall, the corresponding section of the hole wall is marked as a seriously abnormal vibration sample.

[0077] For the vibration signals collected from drilling 24 deep holes, the signals collected every 2 seconds are defined as one sample. Among them, a certain sample signal is as Fig. 9 Sum and Fig.10 shown. A total of 1187 normal vibration samples, 1185 slightly abnormal vibration samples, and 1160 seriously abnormal vibration samples are marked. Among the samples for learning, 70% of each type of sample is set as the training sample, and 30% is set as the test sample. Perform the multivariate synchrosqueezing transform on the lateral vibration and longitudinal vibration sample signals to obtain the time - frequency spectrum of the fused vibration signal as Fig.11 shown.

[0078] It should be noted that the Chinese character "刨开" in the original text seems a bit unclear in this context. I've tried to translate it as best as possible according to the overall meaning, but it might need further clarification in the original technical content.S103, constructing a deep transfer learning model based on the AlexNet network, and using the sample graph to train and test the deep transfer learning model; illustratively, Fig.13 As shown in Figure 1, the deep transfer learning model includes a transfer sub-model and a fine-tuning sub-model. The parameters of the transfer sub-model are directly called and blocked, and the parameters of the fine-tuning sub-model are slightly adjusted through training. A deep transfer learning model based on the AlexNet network is constructed, and the sample graphs are divided into a training sample graph set and a test sample graph set. The deep transfer learning model is trained using the training sample graph set, and the time-frequency spectrum graph set is input into the AlexNet network deep transfer learning model. The model is trained and tested. The loss curve of the training process is shown in Figure 1. Fig.12 As shown, the trained deep transfer learning model is tested using the test sample atlas.

[0079] S104. Using the deep transfer learning model that has passed the test, the roughness of the hole in deep hole drilling is monitored; illustratively, the trained AlexNet network deep transfer learning model is used as a feature extractor to extract feature information from the fused vibration signal time spectrum, and the extracted feature information is input into the softmax classifier to identify the surface roughness of the deep hole. The recognition accuracy of the training samples and the test samples is shown in Table 2:

[0080] Table 2. Recognition accuracy of the proposed method for training samples and test samples

[0081]

[0082] When monitoring the hole roughness in deep hole drilling through the deep transfer learning model, a multivariable synchronous extrusion transformation method is introduced to fuse multi-sensor signals and enhance feature information. Since the vibration frequency of the slender drill rod 10 changes with the increase of drilling depth (even if the cutting parameters remain unchanged), and the available marker signals are limited, the deep transfer learning model based on AlexNet is used to extract the vibration feature information related to the processing quality from the entire frequency band of the fused signal, and the vibration signals collected from the vertical and horizontal directions of the drill rod 10 are fused through the multivariable synchronous extrusion transformation method. Then, the deep transfer learning method based on AlexNet is introduced to extract the hidden layer features related to the processing quality from the time-frequency spectrum image of the fused signal. Finally, the extracted hidden layer features are input into softmax to identify the surface roughness of the deep hole. The method has the characteristics of simplicity, reliability, ease of operation, low cost, and strong real-time performance during operation. It is suitable for real-time monitoring of deep hole roughness in engineering practice, and realizes real-time, online, and non-destructive detection of hole roughness in deep hole drilling, which has important engineering practical value. In addition, the monitoring method uses the AlexNet network to build a deep transfer learning model, and trains and tests based on sample graphs, so that the model can accurately monitor the hole roughness under different working conditions and can receive and process sample signals from the deep hole drilling process in real time, quickly identify the surface roughness of the hole, and can promptly discover and adjust processing parameters to avoid the production of unqualified products. It solves the technical defects of the existing technology that manual monitoring leads to incomplete and inaccurate monitoring data, and cannot effectively monitor the roughness of deep holes, thereby affecting the processing efficiency of deep hole parts.

[0083] Second, as Fig.14 As shown, an embodiment of the present invention provides a system for monitoring hole roughness in deep hole drilling, comprising:

[0084] An acquisition module, used for acquiring sample signals and deep hole surface roughness values;

[0085] A feature fusion module is used to perform time-frequency synchronous compression feature fusion on sample signals;

[0086] Model building module, used to build a deep transfer learning model based on the AlexNet network;

[0087] Monitoring module, used for monitoring hole roughness during deep hole drilling.

[0088] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the method for monitoring hole roughness in deep hole drilling as described above when executing the computer program.

[0089] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for monitoring hole roughness in deep hole drilling as described above is implemented.

[0090] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

[0091] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0092] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0093] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0094] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0095] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for monitoring hole roughness during deep hole drilling, characterized in that: include: Obtain sample signals and deep hole surface roughness values; By using a multivariate synchronous compression algorithm, the sample signal is subjected to time-frequency synchronous compression feature fusion to obtain a sample graph; Constructing a deep transfer learning model based on the AlexNet network, and using the sample graph to train and test the deep transfer learning model; The deep transfer learning model that passes the test is used to monitor the hole roughness during deep hole drilling.

2. The monitoring method according to claim 1, characterized in that: Obtain sample signals and deep hole surface roughness values, including: An eddy current displacement sensor is used to collect the lateral vibration signal and the longitudinal vibration signal of the tool rod closest to the workpiece in real time during deep hole drilling at a sampling frequency of 3200 Hz, and the lateral vibration signal and the longitudinal vibration signal are used as sample signals; Destructive testing is performed on the deep hole after drilling to obtain the surface roughness value of the deep hole, and the obtained surface roughness value of the deep hole is used to calibrate the obtained sample signal.

3. The monitoring method according to claim 1, characterized in that: By using a multivariate synchronous compression algorithm, the sample signal is subjected to time-frequency synchronous compression feature fusion to obtain a sample graph, which specifically includes: Through the multivariate synchronous compression algorithm, the time-frequency synchronous compression features of the sample signal are fused to obtain a time-frequency spectrum of the fused signal, and the time-frequency spectrum is used as a labeled sample graph.

4. The monitoring method according to claim 3, characterized in that: The time-frequency spectrum diagram includes a time-frequency spectrum T h (ω,b) and time-frequency spectrum T v (ω,b); Where ω is frequency, b is time, and the time-frequency spectrum T is calculated respectively. h (ω, b) and the time-frequency spectrum T v (ω, b) Instantaneous frequency and instantaneous amplitude in each subband k.

5. The monitoring method according to claim 4, characterized in that: Also includes: Calculate the multivariate joint instantaneous frequency and instantaneous amplitude for each sub-band K; Calculate the time-frequency spectrum of the fused feature signal after multivariate synchronous compression transformation.

6. The monitoring method according to claim 1, characterized in that: Constructing a deep transfer learning model based on the AlexNet network, and using the sample graph to train and test the deep transfer learning model, including: Build a deep transfer learning model based on the AlexNet network and divide the sample graphs into training sample graph sets and test sample graph sets; The deep transfer learning model is trained using the training sample atlas, and the trained deep transfer learning model is tested using the test sample atlas.

7. The monitoring method according to claim 6, characterized in that: The deep transfer learning model includes a transfer sub-model and a fine-tuning sub-model.

8. A system for monitoring the roughness of a hole during deep hole drilling, characterized in that: include: An acquisition module, used for acquiring sample signals and deep hole surface roughness values; A feature fusion module is used to perform time-frequency synchronous compression feature fusion on sample signals; Model building module, used to build a deep transfer learning model based on the AlexNet network; Monitoring module, used for monitoring hole roughness during deep hole drilling.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable in the processor. When the processor executes the computer program, the method realizes the steps of the method for monitoring hole roughness in deep hole drilling according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for monitoring hole roughness in deep hole drilling according to any one of claims 1 to 7 is implemented.

Citation Information

Cited By

  • Device for measuring surface roughness during drilling, especially of composites

    PL248884B1

  • Device for measuring surface roughness during drilling, especially of composites

    PL451653A1