CN108723895B - Signal segmentation method for real-time monitoring of drilling machining state - Google Patents
Signal segmentation method for real-time monitoring of drilling machining state Download PDFInfo
- Publication number
- CN108723895B CN108723895B CN201810516646.3A CN201810516646A CN108723895B CN 108723895 B CN108723895 B CN 108723895B CN 201810516646 A CN201810516646 A CN 201810516646A CN 108723895 B CN108723895 B CN 108723895B
- Authority
- CN
- China
- Prior art keywords
- signal
- drilling
- envelope
- real
- shannon
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23Q—DETAILS, COMPONENTS, OR ACCESSORIES FOR MACHINE TOOLS, e.g. ARRANGEMENTS FOR COPYING OR CONTROLLING; MACHINE TOOLS IN GENERAL CHARACTERISED BY THE CONSTRUCTION OF PARTICULAR DETAILS OR COMPONENTS; COMBINATIONS OR ASSOCIATIONS OF METAL-WORKING MACHINES, NOT DIRECTED TO A PARTICULAR RESULT
- B23Q17/00—Arrangements for observing, indicating or measuring on machine tools
- B23Q17/09—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool
- B23Q17/0952—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool during machining
Abstract
A signal segmentation method for real-time monitoring of drilling machining state mainly realizes full-automatic detection of actual cutting (elimination of idle cutting) to select a signal segment representing machining state stability and solves the problem of excessive data in the signal processing process. The key points of the technical scheme are as follows: firstly, performing wavelet packet decomposition on drilling machining monitoring signals, and respectively calculating normalized shannon energy of each layer of signals; secondly, reconstructing a Shannon envelope of a Shannon energy maximum layer signal as a drilling monitoring signal envelope; smoothing signal envelope by using a moving average algorithm, and realizing double-threshold self-adaptive signal detection and segmentation by continuously updating noise signal estimation in real time; and finally, calculating the theoretical length of the cutting signal section according to the drilling processing parameters, and introducing the theoretical length into detection to serve as an additional constraint condition for judgment to correct the result, so that the actual cutting signal and the empty cutting signal are accurately separated in the drilling process. The method can be widely applied to a real-time online monitoring system of the machining state.
Description
Technical Field
The invention relates to a signal processing method for monitoring drilling machining state, in particular to the field of on-line real-time state monitoring needing to be continuously carried out for a long time.
Background
With the introduction of the strong national strategy of "china manufacturing 2025", governments, enterprises, universities and the like are actively developing the integration innovation and engineering application of the integration of new generation information technology and manufacturing equipment, closely surrounding key links in the key manufacturing field. The machining state monitoring is an important component of predictive maintenance, plays an important role in ensuring the processing quality of products and reducing enterprise loss, and meanwhile, the real-time performance of the state monitoring is an important evaluation index for measuring the performance of a state monitoring system.
In consideration of cost and other reasons, most of the detection on the equipment state in actual production utilizes a handheld data acquisition analyzer to analyze acquired data to evaluate the current operating condition of the equipment. However, in some production with great potential safety hazard or great loss caused by fault shutdown, continuous monitoring for a long time is required instead of relying on regular handheld detection and analysis.
With the rapid development of computer and sensor technologies, sensors are generally used to monitor the entire processing system in the high precision processing field such as aerospace to ensure the quality of parts, and in order to make the monitoring signals fully fit the actual processing state to meet the detection precision, a mode of increasing the sampling frequency is generally adopted. However, as the sampling frequency is continuously increased, the data volume of the monitoring signal is also very large, which puts higher requirements on the memory and the running speed of the calculation, and also restricts the real-time performance and the high efficiency of the state monitoring.
Therefore, an effective way to solve the above problems is to make the computer automatically remove the signal data (such as tool changing stage signal, feeding stage signal and idle signal) irrelevant to the processing state on the premise of ensuring that the monitoring signal sample can fully represent the system state, and avoid the irrelevant signals from being analyzed by the monitoring system, thereby greatly reducing the load of the signal processing stage of the monitoring system, shortening the analysis processing time of the state monitoring system to a certain extent, providing possibility for realizing faster and faster online real-time monitoring, and also reducing the production cost of enterprises. In addition, in the state monitoring process, the relevant signals representing the operation stage of the equipment are subjected to processing such as feature extraction, and the whole signals are not analyzed, so that the first step of state detection is how to select the representative part of the monitoring signals.
Taking the monitoring of the drilling state as an example, if the monitoring system is required to perform long-time continuous monitoring in actual production, the sampling frequency is increased to ensure the accuracy of the monitoring signal. Then, the signal data acquired in the state monitoring process will be massive, and if all the signal data are processed, the real-time performance of monitoring will be affected undoubtedly, and even detection fails. The hypothetical drilling process state monitoring system automatically selects a signal representative of the cutting process to analyze and process, and automatically removes signals not related to the actual cutting, such as signals during the time the drill changes tools and processes the tool between the completion of one hole and the next. This results in a significantly reduced signal data for drilling conditions, while the processing of only cutting condition related signals makes the condition monitoring system more reliable.
Disclosure of Invention
In order to improve the real-time performance and reliability of drilling machining state monitoring, the invention aims to provide a method for eliminating irrelevant signals by fully automatically detecting corresponding actual cutting signals in drilling, and provide reference and reference for other state monitoring real-time performance.
The technical scheme adopted by the invention for solving the technical problems is as follows: firstly, performing wavelet packet decomposition on drilling machining monitoring signals, and respectively calculating normalized shannon energy of each layer of signals; then, reconstructing a Shannon envelope of a Shannon energy maximum layer signal to serve as a drilling monitoring signal envelope; then, smoothing signal envelope by using a moving average algorithm, and realizing double-threshold self-adaptive signal detection and segmentation by continuously updating noise signal estimation in real time; and finally, calculating the theoretical length of the cutting signal section according to the drilling processing parameters, and introducing the theoretical length into detection to serve as an additional constraint condition for judgment to correct the result, so that the actual cutting signal and the empty cutting signal are accurately separated in the drilling process.
The wavelet packet decomposition is to decompose drilling monitoring signals X (t) into 8 different frequency bands by adopting db5 wavelet 3-layer decomposition: s130, s131, …, s137, wavelet basis functions and the number of decomposition levels are determined experimentally comparing the signal segmentation effects.
The normalized shannon energy of each layer of signals after decomposition is calculated, and the calculation process is as follows: first, s130, s131, …, s137 are normalized by x (i) s13inormS13i/max (| s13i |); then, the shannon energy e (i) ═ -x of the normalized signal is calculated2(i)log[x2(i)](ii) a Finally, a shannon envelope p (i) ═ e (i) — M (e (i))]S (E (i)), M (E (i)) is the Shannon energy mean, S (E (i)) the Shannon energy variance. After the shannon envelope of the signal with the maximum shannon energy layer is reconstructed to be used as the drilling monitoring signal envelope, the envelope of the signal with the maximum shannon energy layer is obtained after the shannon energy E (0), E (1),. and E (7) of the decomposed signals s130, s131, … and s137 of the wavelet packet are respectively obtained, and finally the envelope is reconstructed to obtain the drilling monitoring signal envelope.
Smoothing the signal envelope by using a moving average algorithm, finding that more burrs exist after the Shannon envelope p (i) of the drilling signal is obtained to influence the subsequent processing, therefore, smoothing by using the moving average algorithm, namely, continuously taking m adjacent data of the N data of the Shannon envelope p (i) for weighted average to smooth,wherein wiIs a weight coefficient, andp, q are positive integers less than m, and p + q is 1.
The method for realizing the double-threshold self-adaptive signal detection and segmentation by continuously updating the noise signal estimation in real time comprises the following specific steps: firstly, double threshold assignment th1 and th2 are initialized, a large threshold th2 is used for filtering noise, and a small threshold th1 is used for accurate segmentation; then, setting a fixed length l as a × f, and a window shift w as b × f, and starting detection by a sliding window from a signal starting point, wherein a and b are natural numbers, and f is a signal sampling frequency; in the detection process, th1 and th2 are updated according to the following update rules: th1 ═ c × mean (X (i: i + l)), th2 ═ nois + d × | th1 — nois |, nois being real-time noise estimate nois ═ m × mean (X (i: i + l)), mean (X) being the average of signal X (t), and satisfying 0 < m < c < d < 1; and finally, the signal meeting the condition that the X is more than th1 is the actual cutting signal in the drilling process, and the rest signals are non-cutting signals.
The theoretical length of the cutting signal section is calculated according to the drilling machining parameters, the theoretical length is introduced into detection to serve as an extra constraint condition for judgment to correct the result, the theoretical length of the signal section is calculated according to the fact that information about feeding quantity F, drilling length Q, sampling frequency F and the like can be generally obtained in a numerical control machining program in the drilling machining and signal acquisition processes, therefore, the theoretical length L of the cutting signal can be determined to be Qf/F, L serves as an extra judgment condition outside a threshold value in the signal detection process, namely in the judgment process, when the cutting signal section length len obtained after threshold value division is far smaller than the theoretical length L of the cutting signal section, the program refuses to execute division, and the judgment is considered to be misjudgment caused by noise jitter.
The invention has the beneficial effects that: a signal dividing method for real-time monitoring of drilling state features that the actual cutting monitor signal in drilling can be automatically selected and the non-cutting signal can be eliminated. In some current commercial tool monitoring systems or laboratories, the choice of analysis signal is mostly a manual selection by the user, which is difficult due to cutting system variations or human error. Meanwhile, in the monitoring system, the monitoring signals in the processing stage are analyzed, and not all the acquired signals are taken as processing objects, so that the reliability of the monitoring system can be improved by automatically selecting the useful section signals. On the other hand, the automatic selection of the cutting stage signal and the elimination of the non-cutting signal for processing can greatly shorten the analysis processing time of the monitoring system, thereby improving the on-line monitoring real-time performance. Therefore, the invention has the direct benefit of improving the real-time performance and reliability of the online monitoring system.
Drawings
FIG. 1 is a schematic flow diagram of the present invention.
Figure 2 drill process vibration signal.
FIG. 3 shows the wavelet packet decomposition result of the drilling vibration signal.
Fig. 4 shannon energy values of each layer of signal after wavelet packet decomposition.
Fig. 5 shows the envelope of the vibration signal after the reconstruction of the shannon energy maximum layer signal.
FIG. 6 shows the smoothed envelope of the drilling vibration signal by the moving average algorithm.
Fig. 7 detection of vibration signals during actual drilling stages.
Fig. 8 shows the signal detection correction result after the length control is added.
Detailed Description
For the purpose of illustrating the technical solutions and technical objects of the present invention, the present invention will be described in further detail with reference to the accompanying drawings and examples.
Embodiment 1, it mainly realizes the full automatic detection of actual cutting (eliminating empty cutting) to select the stable signal segment representing the processing state and solve the problem of excessive data in the signal processing process. The key points of the technical scheme are as follows: firstly, performing wavelet packet decomposition on drilling machining monitoring signals, and respectively calculating normalized shannon energy of each layer of signals; then, reconstructing a Shannon envelope of a Shannon energy maximum layer signal to serve as a drilling monitoring signal envelope; then, smoothing signal envelope by using a moving average algorithm, and realizing double-threshold self-adaptive signal detection and segmentation by continuously updating noise signal estimation in real time; and finally, calculating the theoretical length of the cutting signal section according to the drilling processing parameters, and introducing the theoretical length into detection to serve as an additional constraint condition for judgment to correct the result, so that the actual cutting signal and the empty cutting signal are accurately separated in the drilling process. See fig. 1-8.
Example 2, the wavelet packet decomposition is to decompose the drilling monitoring signal x (t) into 8 different frequency bands by adopting db5 wavelet 3-layer decomposition: s130, s131, …, s137, wavelet basis functions and the number of decomposition levels are determined experimentally comparing the signal segmentation effects. Referring to fig. 1 to 8, the rest is the same as embodiment 1.
Embodiment 3, the normalized shannon energy of each layer of signals after decomposition is calculated, and the shannon envelope of the signal of the layer with the maximum shannon energy is reconstructed and then used as the envelope of the drilling monitoring signal, and the calculation process is as follows: first, s130, s131, …, s137 are normalized by x (i) s13inormS13i/max (| s13i |); then, the shannon energy e (i) ═ -x of the normalized signal is calculated2(i)log[x2(i)](ii) a Finally, a shannon envelope p (i) ═ e (i) — M (e (i))]S (E (i)), M (E (i)) is the Shannon energy mean, S (E (i)) the Shannon energy variance. Referring to fig. 1 to 8, the rest of the embodiment is the same as the above embodiment.
Embodiment 4, smoothing the signal envelope by using the moving average algorithm, and finding that more burrs exist after the envelope p (i) of the drilling signal is obtained, the subsequent processing is affected, so the smoothing processing is performed by using the moving average algorithm, that is, m adjacent data are continuously taken from N data of the envelope p (i) to perform weighted average for smoothing,wherein wi is a weight coefficient, andp, q are positive integers less than m, and p + q is 1. Referring to fig. 1 to 8, the rest of the embodiment is the same as the above embodiment.
Embodiment 5, the method for implementing dual-threshold adaptive signal detection and segmentation by continuously updating the noise signal estimation in real time includes: first, the dual threshold assignments th1, th2 are initialized, the large threshold th2 for noise filtering and the small threshold th1 for accurate segmentation. Then, a sliding window with a fixed length l being a × f and a window shift w being b × f (a, b being natural numbers, f being a signal sampling frequency) is set to be detected from the signal starting point. In the detection process, th1 and th2 are updated according to the following update rules: th1 ═ c × mean (X (i: i + l)), th2 ═ nois + d × | th1 — nois |, nois being the real-time noise estimate nois ═ m × mean (X (i: i + l)), mean (X) being the average of the signal X (t), and satisfying 0 < m < c < d < 1. And finally, the signal meeting the condition that the X is more than th1 is the actual cutting signal in the drilling process, and the rest signals are non-cutting signals. Referring to fig. 1 to 8, the rest of the embodiment is the same as the above embodiment.
Example 6, the theoretical length of the cutting signal segment is calculated according to the drilling processing parameters, and the theoretical length is introduced into detection as an additional constraint condition for judgment to correct the result. The theoretical length of the signal section is calculated according to the fact that information about the feeding amount F, the drilling length Q, the sampling frequency F and the like can be generally obtained in a numerical control machining program in the process of drilling machining and signal acquisition, and therefore the theoretical length L of the cutting signal can be determined to be Qf/F. When signal detection is carried out, L is taken as an extra judgment condition outside a threshold value, namely in the judgment process, when the length len of a cut signal segment obtained after threshold value division is far smaller than the theoretical length L of the cut signal segment, the program refuses to execute division, and the judgment is considered to be misjudgment caused by noise jitter. Referring to fig. 1 to 8, the rest of the embodiment is the same as the above embodiment.
Example 7, the specific procedure was as follows:
as shown in fig. 1, a general condition monitoring system mainly includes a data acquisition system, a signal processing system and a condition diagnosis system, but in practice, signals often need to be preprocessed before data acquisition is completed and sent to the signal processing system. For example, in a drilling process tool monitoring system, the object of signal processing should be the signal when the tool is cutting stably, and not the entire signal is suitable for tool monitoring. In fact, the acquired signals include signals of tool changing, tool feeding, idling and the like of the machine tool, and as shown in fig. 2, the workload of processing by the signal processing system can be greatly reduced by detecting and dividing the actual cutting signals. Therefore, the invention provides an algorithm for automatically detecting actual cutting, which mainly comprises the steps of envelope extraction and self-adaptive double-threshold detection.
Firstly, a db5 wavelet 3-layer decomposition is adopted for the acquired drilling vibration signal, the main purpose is to carry out denoising processing on the signal, and the result after the drilling vibration signal decomposition is shown in fig. 3. Regarding signal envelope extraction, the general envelope extraction method comprises Hilbert-Huang transform, mathematical morphology, phase shift wavelet and normalized Shannon energy method, and the first three methods have the main defects of sensitivity to noise, large signal fluctuation caused by poor anti-interference capability and inconvenience for subsequent analysis. Therefore, a method of combining wavelet and normalized shannon energy is adopted to extract the envelope of the drilling vibration signal.
As shown in fig. 4, normalized shannon energy values of 8 layers of signals after wavelet packet decomposition are respectively obtained, the obtained result is that E is ═ 26.32,47.85,26.22,46.11,14.78,30.67,15.79 and 30.78], a layer 2 signal with the largest energy is selected, and after an envelope is obtained, the envelope is reconstructed to generate a drilling vibration signal envelope, as shown in fig. 5. In practice, however, the extracted drilling vibration signal contains a lot of burrs due to strong noise in the drilling environment, and the burrs generated by the noise directly affect the detection of the drilling vibration signal in the following steps, as shown in the circles in the figure. Therefore, the reconstructed vibration signal envelope is smoothed by using a moving average algorithm, specifically, m — 5 adjacent data points of the envelope data point are continuously taken to be weighted average to smooth the vibration signal envelope, and the effect after smoothing is as shown in fig. 6.
The step after the extraction of the envelope of the vibration signal is finished is to detect the envelope by using a threshold value, which is also the core of the invention. In view of conventional burst signal detection, a single threshold segmentation method is generally adopted. The single threshold method is a traditional, simple signal segmentation method, which performs segmentation based on a selected threshold, and considers the part above the threshold as a useful signal and the part below the threshold as a noise signal, but the greatest disadvantage is that the interference resistance is poor, the segmentation accuracy depends on the threshold selection, and when the amplitude of the interference signal is greater than the selected threshold, the single threshold segmentation method treats the interference noise as a useful signal, causing segmentation errors.
The continued improvement of single threshold segmentation algorithms has resulted in the emergence of dual threshold segmentation methods. The method well solves the problem of strong noise, and is provided with two thresholds respectively, wherein one threshold is small, the other threshold is large, the large threshold is used for filtering interference noise, and the small threshold is used for accurately segmenting signals. However, the threshold is fixed in advance, the threshold is difficult to determine, the method also has certain limitation on the sudden strong noise situation, and the segmentation precision is difficult to guarantee. Therefore, the invention further improves the defect of double threshold, and the main improvements are reflected in two points: 1. according to the real-time estimation of the noise, the self-adaptive updating of the double thresholds is realized; 2. and the length control is added to correct the detection result so as to further avoid detection failure caused by strong burst noise.
The specific process of the dual-threshold adaptive update is as follows: the sliding window with the window length l being set to 512 and the window shift w being set to 128 detects the signal, and estimates noise as nois mean (X (i: i +512)) in real time, and the real-time updating rule of the threshold is as follows: th1 ═ 0.2 xmean (X (i: i +512)), th2 ═ nois +0.5 × | th1-nois |, and th1 < th 2. The threshold th2 is used to filter out noise, th1 is used for accurate segmentation.
As shown in fig. 7, if sudden strong noise occurs in the environment, the dual-threshold adaptive algorithm is more accurate in detecting the segments of the signal, but still causes segmentation failure. Therefore, further modification of the segmentation result by the length control is introduced. The length correction means that in an actual drilling process, related machining parameters, such as feed amount F, drilling length Q, sampling frequency F and the like, are generally given in a numerical control machining program, and theoretical lengths of cutting signals under different drilling lengths, namely L-Qf/F, can be obtained according to different parameter information.
In the signal detection process, if the detected drilling signal length len is far smaller than the calculated theoretical length L, as in the case of fig. 7, it is considered that the drilling signal length len is jitter caused by strong noise, and the program refuses to perform segmentation, so that the situation can be avoided, and the final segmentation result of the dual-threshold adaptive updating algorithm after combining length control is shown in fig. 8.
Claims (7)
1. A signal segmentation method for real-time monitoring of drilling machining state is characterized by comprising the following steps: firstly, performing wavelet packet decomposition on drilling machining monitoring signals, and respectively calculating normalized shannon energy of each layer of decomposed signals; then, reconstructing a Shannon envelope of a Shannon energy maximum layer signal to serve as a drilling monitoring signal envelope; then, smoothing signal envelope by using a moving average algorithm, and realizing double-threshold self-adaptive signal detection and segmentation by continuously updating noise signal estimation in real time; and finally, calculating the theoretical length of the cutting signal section according to the drilling processing parameters, and introducing the theoretical length into detection to serve as an additional constraint condition for judgment to correct the result, so that the actual cutting signal and the empty cutting signal are accurately separated in the drilling process.
2. The signal segmentation method for real-time monitoring of drilling machining states as claimed in claim 1, wherein: the wavelet packet decomposition is to decompose drilling monitoring signals X (t) into 8 different frequency bands by adopting db5 wavelet 3-layer decomposition: s130, s131, …, s137, wavelet basis functions and the number of decomposition levels are determined experimentally comparing the signal segmentation effects.
3. The signal segmentation method for real-time monitoring of drilling conditions according to claim 1 or 2, characterized in that: the normalized shannon energy of each layer of signals after decomposition is calculated, and the calculation process is as follows: first, s130, s131, …, s137 are normalized by x (i) s13inormS13i/max (| s13i |); then, the shannon energy e (i) ═ -x of the normalized signal is calculated2(i)log[x2(i)](ii) a Finally, a shannon envelope p (i) ═ e (i) — M (e (i))]S (E (i)), M (E (i)) is the Shannon energy mean, S (E (i)) the Shannon energy variance.
4. The signal segmentation method for real-time monitoring of drilling machining states as claimed in claim 3, wherein: after the shannon envelope of the signal with the maximum shannon energy layer is reconstructed to be used as the drilling monitoring signal envelope, the envelope of the signal with the maximum shannon energy layer is obtained after the shannon energy E (0), E (1),. and E (7) of the decomposed signals s130, s131, … and s137 of the wavelet packet are respectively obtained, and finally the envelope is reconstructed to obtain the drilling monitoring signal envelope.
5. The signal segmentation method for real-time monitoring of drilling machining states as claimed in claim 3, wherein: smoothing the signal envelope by using a moving average algorithm, finding that more burrs exist after the Shannon envelope p (i) of the drilling signal is obtained to influence the subsequent processing, therefore, smoothing by using the moving average algorithm, namely, continuously taking m adjacent data of the N data of the Shannon envelope p (i) for weighted average to smooth,wherein wiIs a weight coefficient, andp, q are positive integers less than m, and p + q is 1.
6. The signal segmentation method for real-time monitoring of drilling machining states as claimed in claim 1, wherein: the method for realizing the double-threshold self-adaptive signal detection and segmentation by continuously updating the noise signal estimation in real time comprises the following specific steps: firstly, double threshold assignment th1 and th2 are initialized, a large threshold th2 is used for filtering noise, and a small threshold th1 is used for accurate segmentation; then, setting a fixed length l as a × f, and a window shift w as b × f, and starting detection by a sliding window from a signal starting point, wherein a and b are natural numbers, and f is a signal sampling frequency; in the detection process, th1 and th2 are updated according to the following update rules: th1 ═ c × mean (X (i: i + l)), th2 ═ nois + d × | th1 — nois |, nois being real-time noise estimate nois ═ m × mean (X (i: i + l)), mean (X) being the average of signal X (t), and satisfying 0 < m < c < d < 1; and finally, the signal meeting the condition that the X is more than th1 is the actual cutting signal in the drilling process, and the rest signals are non-cutting signals.
7. The signal segmentation method for real-time monitoring of drilling machining states as claimed in claim 1, wherein: the theoretical length of the cutting signal segment is calculated according to the drilling processing parameters, the theoretical length is introduced into detection to serve as an extra constraint condition for judgment to correct the result, the theoretical length of the signal segment is calculated according to the fact that information about the feeding amount F, the drilling length Q and the sampling frequency F can be obtained in a numerical control processing program in the process of drilling processing and signal acquisition, therefore, the theoretical length L of the cutting signal can be determined to be Qf/F, when signal detection is carried out, L serves as an extra judgment condition outside a threshold value, namely in the judgment process, when the length len of the cutting signal segment obtained after threshold value division is far smaller than the theoretical length L of the cutting signal segment, the program refuses to carry out division, and the judgment is considered to be misjudgment caused by noise jitter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810516646.3A CN108723895B (en) | 2018-05-25 | 2018-05-25 | Signal segmentation method for real-time monitoring of drilling machining state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810516646.3A CN108723895B (en) | 2018-05-25 | 2018-05-25 | Signal segmentation method for real-time monitoring of drilling machining state |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108723895A CN108723895A (en) | 2018-11-02 |
| CN108723895B true CN108723895B (en) | 2019-12-31 |
Family
ID=63936120
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810516646.3A Active CN108723895B (en) | 2018-05-25 | 2018-05-25 | Signal segmentation method for real-time monitoring of drilling machining state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108723895B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934100B (en) * | 2019-01-24 | 2021-07-13 | 西安交通大学 | Amplitude-frequency time-varying process signal segmentation method based on sliding window |
| CN112958840B (en) * | 2021-02-10 | 2022-06-14 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Automatic segmentation method for cutting force signal in precision part machining |
| CN113040784B (en) * | 2021-04-21 | 2022-07-05 | 福州大学 | Myoelectric noise filtering method for electrocardiosignal |
| CN113579851B (en) * | 2021-07-26 | 2022-11-29 | 天津大学 | Non-stationary drilling process monitoring method based on adaptive segmented PCA |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11320201A (en) * | 1998-05-18 | 1999-11-24 | Mitsubishi Materials Corp | Grooving method and its device |
| CN101549468A (en) * | 2009-04-24 | 2009-10-07 | 北京邮电大学 | Image-based on-line detection and compensation system and method for cutting tools |
| CN102172849A (en) * | 2010-12-17 | 2011-09-07 | 西安交通大学 | Cutter damage adaptive alarm method based on wavelet packet and probability neural network |
| TWI472402B (en) * | 2012-02-10 | 2015-02-11 | 中原大學 | Tool flutter monitoring method |
| CN103745476B (en) * | 2014-01-22 | 2016-08-24 | 湘潭大学 | The mobile phone dividing plate sand grains detection method that local peaking analyzes is scanned based on line |
| CN104367344A (en) * | 2014-10-10 | 2015-02-25 | 广东工业大学 | Method and device for detecting instant heart rate of fetus on basis of Shanon envelope |
| CN106563973A (en) * | 2015-10-13 | 2017-04-19 | 桂林市君威机电科技有限公司 | Dynamic milling force measuring method based on tool vibration displacement |
| CN107350900B (en) * | 2017-07-06 | 2019-04-12 | 西安交通大学 | A kind of tool condition monitoring method extracted based on the chip breaking time |
-
2018
- 2018-05-25 CN CN201810516646.3A patent/CN108723895B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108723895A (en) | 2018-11-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108723895B (en) | Signal segmentation method for real-time monitoring of drilling machining state | |
| CN111275307B (en) | Quality control method for high-frequency continuous observation data of automatic online station for water quality | |
| CN108620950B (en) | Method and system for monitoring machining state of turning tool | |
| CN115829422B (en) | Industrial equipment operation abnormal state identification method based on big data | |
| CN110705181A (en) | Rolling bearing residual life prediction method based on convolution length-time memory cyclic neural network | |
| CN116237817B (en) | Five-axis linkage numerical control machine intelligent monitoring system based on Internet of things | |
| CN108629864B (en) | Vibration-based electric spindle radial precision characterization method and system | |
| CN113305645A (en) | Numerical control machine tool cutter residual life prediction method based on mixed neural model | |
| CN112207631B (en) | Method for generating tool detection model, method, system, device and medium for detecting tool detection model | |
| CN113146359A (en) | Numerical control machine tool control system for intelligently monitoring and diagnosing cutting state | |
| CN113752089A (en) | Cutter state monitoring method based on singularity Leersian index | |
| CN108684051B (en) | Wireless network performance optimization method based on causal diagnosis, electronic equipment and storage medium | |
| CN114611745A (en) | Machine tool evaluation method, machine tool evaluation system, and medium | |
| CN110991564A (en) | Variable working condition bearing fault diagnosis method based on multi-scale dispersion entropy deviation mean value and nonlinear mode decomposition | |
| CN117420967A (en) | Method and system for improving storage performance of software acquisition data | |
| CN114800040A (en) | Cutter wear monitoring method and system based on process-state data correlation | |
| CN113536042B (en) | Time series abnormity detection method, device and equipment | |
| CN111993158B (en) | Method for generating tool detection model, method, system, device and medium for detecting tool detection model | |
| CN114818815A (en) | Method and system for acquiring timed arrival time of blade tip for blade vibration measurement | |
| CN110907544B (en) | Identification method for abnormal step data of content of dissolved gas in transformer oil | |
| CN114357851A (en) | Numerical control milling surface roughness prediction method based on DAE-RNN | |
| CN106881630B (en) | High-speed milling flutter online recognition method based on adaptive-filtering Yu AR models | |
| CN116214263B (en) | Cutter residual life prediction method, system and computer | |
| CN115555920B (en) | Online chatter detection method and system based on adaptive variation modal decomposition | |
| Zou et al. | A Novel Machine RUL Prediction Method in Small Sample based on Fully Convolutional Variational Auto-Encoding Network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |