CN117961643A - Numerical control machine tool online fault monitoring method based on multi-axis current signals - Google Patents
Numerical control machine tool online fault monitoring method based on multi-axis current signals Download PDFInfo
- Publication number
- CN117961643A CN117961643A CN202410284293.4A CN202410284293A CN117961643A CN 117961643 A CN117961643 A CN 117961643A CN 202410284293 A CN202410284293 A CN 202410284293A CN 117961643 A CN117961643 A CN 117961643A
- Authority
- CN
- China
- Prior art keywords
- machine tool
- data
- axis
- numerical control
- vibration
- 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.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 19
- 238000005070 sampling Methods 0.000 claims description 16
- 230000002159 abnormal effect Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 9
- 238000012417 linear regression Methods 0.000 claims description 7
- 238000001228 spectrum Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000003066 decision tree Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000007637 random forest analysis Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000003754 machining Methods 0.000 abstract description 6
- 238000003745 diagnosis Methods 0.000 abstract description 4
- 238000013210 evaluation model Methods 0.000 abstract 1
- 238000012216 screening Methods 0.000 abstract 1
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a numerical control machine tool online fault monitoring method based on a multi-axis current signal, which comprises the following steps: the method comprises the following steps: collecting current and vibration data of a multi-shaft driving motor of a machine tool; extracting current and vibration signal characteristics, and screening the characteristics; and establishing a plurality of machine tool motor fault diagnosis models according to the multi-axis data characteristics, integrating the plurality of diagnosis models into a comprehensive evaluation model with higher robustness, and realizing online monitoring of the machine tool through the model obtained by the historical data. The invention fully considers the problem of frequent working condition change and multiaxial linkage machining during machine tool machining. According to actual demands, the problem of insufficient robustness of a single vibration signal under a variable working condition is solved by introducing current signals and multi-axis signal regression, and a model can be quickly retrained in an actual use process, so that the expansibility and the accuracy of machine tool fault diagnosis are continuously improved.
Description
Technical Field
The invention relates to the field of on-line monitoring of mechanical faults, in particular to the on-line monitoring of a machine tool accurately by introducing multiaxial motion information and current signals related to working condition changes aiming at the problems of quick change of processing working conditions, complex structure and operation of a numerical control machine tool and the like.
Background
The on-line monitoring technology can track the running state of the machine tool in real time, identify potential problems and take measures in time so as to avoid downtime. Through effective monitoring, faults and errors in production can be reduced, the utilization rate of the machine tool is improved, and finally the overall production efficiency is improved.
Vibration signals are particularly suitable for rotating machines such as motors, fans, pumps, etc. These devices often present mechanical problems through vibration. And the vibration signal exhibits a high sensitivity to various faults in the mechanical system, such as bearing faults, gear faults, unbalance, etc. Abnormal vibration patterns may generally indicate a particular type of fault, and on-line monitoring systems often employ vibration signals.
Due to the complex coupling relation among the mechanical structure of the machine tool, the numerical control system and the control components, the machine tool involves multi-axis linkage and frequent change of the rotating speed during machining. This may interfere with accurate monitoring of the vibration signal of the machine system itself, making it difficult to monitor the condition of the machine tool with a conventional single vibration signal.
The effective value of the current signal can reflect the change of the load during processing to a certain extent, and the frequency conversion can reflect the change of the rotating speed. Under the condition of lack of processing working conditions, the reaction working conditions can be approximately changed through current signals, and the method is very suitable for the object with larger working condition change, such as a machine tool. Meanwhile, the coupling relation of the machine tool is complex, and the machining process possibly comprises multi-axis combined motion and single-axis motion, so that the operation states of a plurality of axes are considered for combined analysis, and higher monitoring accuracy can be obtained.
Disclosure of Invention
The invention solves the problems that: in view of the above-mentioned problems, the invention combines the multi-axis current signals to monitor the machine tool on line in real time, and can realize more accurate monitoring results.
The technical scheme adopted by the invention is as follows: an online fault monitoring method of a numerical control machine based on multi-axis current signals, the method comprises the following steps:
(1) Using a vertical four-axis numerical control milling center as an analysis object to collect vibration and current signals of each motor;
(2) Obtaining the frequency spectrum of the current signal through Fast Fourier Transform (FFT), extracting the frequency spectrum characteristic of the current signal, and simultaneously extracting the time domain characteristics of the current signal and the vibration signal;
(3) Judging whether the data is in a shutdown state or not, and deleting the data in the shutdown state;
(4) According to the current characteristics of all the monitoring motors in the historical data, taking the current characteristics of all the monitoring motors as the input of a plurality of models, and taking the vibration signal characteristics of one shaft as the output of the models to solve all the models;
(5) Respectively obtaining proper abnormal discrimination thresholds of all models by using data of the training set;
(6) After the real-time data is characterized by being obtained in the step (2), inputting the real-time data into the model in the step (4) to obtain a predicted value;
(7) Comparing errors of the predicted vibration values and the actual vibration values in all the models, and alarming when all the models exceed the threshold value in the step (5), so that the online monitoring of the numerical control machine tool with high robustness is realized;
(8) And updating the model when the abnormal data is detected and the abnormal data is in a normal state.
Further, in the step (1), vibration data acquisition parameters are as follows: the sampling unit is m/s 2, and the sampling rate is 25.6Hz. The current signal acquisition parameters are as follows: the sampling unit is A and the sampling rate is 25.6Hz.
Further, the current signal extracted in the step (2) has a frequency corresponding to a maximum peak value in the frequency spectrum, and the time domain has an effective value (RMS). The time domain characteristic of the vibration signal is the effective value (RMS).
Further, in step (3), in order to reduce the data redundancy, all the data with the effective value (RMS) of the shaft current less than 0.005, namely considered as the shutdown state, are deleted.
Further, in the step (4), a random forest, a decision tree, a K nearest neighbor algorithm and a linear regression model are used for solving the relation between the current and the vibration signals, and through historical data, the current signal characteristics of a plurality of axes are mapped to vibration signal characteristics including a main axis, an X axis, a Y axis and a Z axis respectively to obtain a model of a corresponding axis.
Further, the error between the predicted effective value and the true effective value obtained by calculating the training set in the step (5) is calculated through a relative error. And determining abnormal judgment thresholds of the models according to errors of the training set, wherein the thresholds which are not exceeded by continuous 5 points are met in the errors of the training set and serve as hard thresholds. To increase the robustness, the threshold is increased by 20% to be determined as the final abnormality judgment threshold.
Further, in the step (6), the characteristics of the real-time data are extracted, the real-time data are judged, if the data are in a processing state, the predicted value of the vibration effective value is obtained through the solved multiple models, and otherwise, the calculation is not performed.
Further, in the step (7), the alarm is given by exceeding the alarm threshold value for five continuous times through the real-time data error.
Further, in the step (8), if the model gives an alarm, the machine needs to be stopped for checking the state of the machine tool, and if no fault is found, the data of the five time periods are stored in a historical database, and the model is retrained and updated.
The beneficial effects of the invention are as follows: the invention fully considers the problem of frequent working condition change and multiaxial linkage machining during machine tool machining. According to actual demands, the problem of insufficient robustness of a single vibration signal under a variable working condition is solved by introducing current signals and multi-axis signal regression, and a model can be quickly retrained in an actual use process, so that the expansibility and the accuracy of machine tool fault diagnosis are continuously improved.
Drawings
FIG. 1 is a general flow chart of the present invention;
FIG. 2 is a flow chart of the training of the present discovery on-line monitoring model;
FIG. 3 is a flow chart of the present discovery online monitoring system;
FIG. 4 is an extracted feature of the present invention;
FIG. 5 is a flow chart of the anomaly threshold value determination according to the present invention;
FIG. 6 is a regression result of the random forest model of the present invention;
FIG. 7 is a regression result of the decision tree model of the present invention;
FIG. 8 is a regression result of the K nearest neighbor algorithm model of the present invention;
FIG. 9 is a regression result of a linear regression model of the present invention;
Detailed Description
The invention will be further described with reference to the following examples of embodiments using a spindle as an example:
1. The example adopts a vertical four-axis numerical control milling center as an analysis object, and collects vibration and current signals of each motor, wherein the sampling mode is interval sampling, the time of each sample is 1s, and the sampling interval is 1s. The vibration data acquisition parameters are as follows: the sampling unit is m/s 2, and the sampling rate is 25.6Hz. The current signal acquisition parameters are as follows: the sampling unit is A and the sampling rate is 25.6Hz.
2. The whole flow of the method is shown in figure 1, and the specific implementation steps are as follows:
Step 1: extracting characteristics of vibration and current data acquired by each motor, wherein fig. 4 is the characteristics extracted by the invention;
The feature extraction process is as follows:
where x is the signal, rms and maxfrequency are the extracted features, fs is the sampling frequency, and n is the signal length, respectively.
Step 2: and deleting all data with the effective value of the shaft current less than 0.005 to obtain the screened characteristics.
Step 3: and solving the relation between the current and the vibration signals by using a random forest, a decision tree, a K nearest neighbor algorithm and a linear regression model, and mapping the current signal characteristics of a plurality of axes into vibration signal characteristics including a main axis, an X axis, a Y axis and a Z axis respectively through historical data to obtain a model of the corresponding axis.
The linear regression model using the principal axis as an example is shown as follows:
VRm＝A*ERm+B*ERFm+C*(ERm*ERFm)+D*ERx+E*ERFx+F
*(ERx*ERFx)+G*ERy+H*ERFy+I*(ERy*ERFy)+J*ERz+K
*ERFzz+L*(ERz*ERFz)
Wherein VR m represents the vibration effective value of the principal axis after regression; ER m represents the effective value of the current of the shaft; ERF m represents the current slew frequency of the spindle; the subscripts denote the respective axes.
Step 4: and obtaining a threshold value for judging the abnormal data through the calculated model. FIG. 5 is a flow chart of the anomaly threshold value determination according to the present invention;
(1) And calculating the relative error of the predicted effective value and the real effective value. The relative error calculation formula is:
Where E i is the error of the ith model, RMS pi is the predicted value of the real-time data in the ith model, and RMS r is the true effective value of the real-time data. i=1, 2,3,4, representing a random forest, a decision tree, a K nearest neighbor algorithm, and a linear regression model, respectively. FIGS. 6-9 are regression results of a random forest, decision tree, K nearest neighbor algorithm, and linear regression model, respectively;
(2) According to the calculation step of fig. 5 and the regression error results of the 4 models, the threshold values of the respective models are calculated respectively, and the threshold value is increased by 20% as the judgment threshold value of the real-time data.
Step 5: and extracting the characteristics of the real-time data, judging the real-time data, and obtaining a predicted value of the vibration effective value through a plurality of solved models if the data is in a processing state, otherwise, not calculating.
Step 6: and alarming through errors obtained by the real-time data, if the 4 models exceed the alarm threshold for five continuous times, the models alarm, and if the 4 models do not exceed the threshold at the same time, the models do not alarm.
Step 7: if the model gives an alarm, the machine tool needs to be stopped for checking the state of the machine tool, and if no fault is found, the data of five time periods are stored in a historical database, and the model is retrained and updated.
3. Finally, it should be noted that: the above embodiments are only for illustrating the technical aspects of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the above embodiments, it should be understood by those of ordinary skill in the art that: modifications and equivalents may be made to the specific embodiments of the invention without departing from the spirit and scope of the invention, which is intended to be covered by the claims.
Claims (9)
1. The online fault monitoring method of the numerical control machine tool based on the multi-axis current signal is characterized by comprising the following steps of:
(1) Using a vertical four-axis numerical control milling center as an analysis object to collect vibration and current signals of each motor;
(2) Obtaining the frequency spectrum of the current signal through Fast Fourier Transform (FFT), extracting the frequency spectrum characteristic of the current signal, and simultaneously extracting the time domain characteristics of the current signal and the vibration signal;
(3) Judging whether the data is in a shutdown state or not, and deleting the data in the shutdown state;
(4) According to the current characteristics of all the monitoring motors in the historical data, taking the current characteristics of all the monitoring motors as the input of a plurality of models, and taking the vibration signal characteristics of one shaft as the output of the models to solve all the models;
(5) Respectively obtaining proper abnormal discrimination thresholds of all models by using data of the training set;
(6) After the real-time data is characterized by being obtained in the step (2), inputting the real-time data into the model in the step (4) to obtain a predicted value;
(7) Comparing errors of the predicted vibration values and the actual vibration values in all the models, and alarming when all the models exceed the threshold value in the step (5), so that the online monitoring of the numerical control machine tool with high robustness is realized;
(8) And updating the model when the abnormal data is detected and the abnormal data is in a normal state.
2. The on-line fault monitoring method for a numerical control machine tool based on multi-axis current signals according to claim 1, wherein vibration data acquisition parameters in the step (1) are as follows: the sampling unit is m/s 2, the sampling rate is 25.6Hz, and the current signal acquisition parameters are as follows: the sampling unit is A and the sampling rate is 25.6Hz.
3. The method for online fault monitoring of a numerical control machine tool based on multi-axis current signals according to claim 1, wherein the current signal spectrum extracted in the step (2) is characterized by a frequency corresponding to a maximum peak value in a spectrum, the time domain is characterized by an effective value (RMS), and the time domain is characterized by an effective value (RMS).
4. The method for on-line fault monitoring of a numerically controlled machine tool based on multi-axis current signals according to claim 1, wherein in step (3), in order to reduce data redundancy, all the data whose axis current effective value (RMS) is less than 0.005, namely considered as a shutdown state, are deleted.
5. The online fault monitoring method of the numerical control machine tool based on the multi-axis current signals according to claim 1, wherein in the step (4), a random forest, a decision tree, a K nearest neighbor algorithm and a linear regression model are used for solving the relation between the current and the vibration signals, and the current signal characteristics of a plurality of axes are mapped into vibration signal characteristics including a main axis, an X axis, a Y axis and a Z axis respectively through historical data to obtain a model of the corresponding axis.
6. The online fault monitoring method of the numerical control machine based on the multi-axis current signals according to claim 1, wherein the error between the predicted effective value and the true effective value obtained by calculating the training set in the step (5) is calculated through a relative error; and determining the abnormal judgment threshold value of each model according to the error of the training set, wherein the threshold value which is not exceeded by continuous 5 points in the error of the training set is used as a hard threshold value, and the threshold value is increased by 20% to determine the final abnormal judgment threshold value for increasing the robustness.
7. The online fault monitoring method of the numerical control machine tool based on the multi-axis current signals according to claim 1, wherein the characteristic of real-time data is extracted in the step (6), the real-time data is judged, if the data is in a processing state, a predicted value of a vibration effective value is obtained through a plurality of solved models, and otherwise, calculation is not performed.
8. The online fault monitoring method of the numerical control machine tool based on the multi-axis current signal according to claim 1, wherein in the step (7), the alarm is given by exceeding an alarm threshold value for five consecutive times through real-time data errors.
9. The on-line fault monitoring method of a numerical control machine based on multi-axis current signals according to claim 1, wherein in step (8), if the model is alarmed, the machine state is checked by stopping the machine, if no fault is found, the data of five time periods are stored in a history database, and the model is retrained and updated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410284293.4A CN117961643A (en) | 2024-03-13 | 2024-03-13 | Numerical control machine tool online fault monitoring method based on multi-axis current signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410284293.4A CN117961643A (en) | 2024-03-13 | 2024-03-13 | Numerical control machine tool online fault monitoring method based on multi-axis current signals |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117961643A true CN117961643A (en) | 2024-05-03 |
Family
ID=90854867
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410284293.4A Pending CN117961643A (en) | 2024-03-13 | 2024-03-13 | Numerical control machine tool online fault monitoring method based on multi-axis current signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117961643A (en) |
-
2024
- 2024-03-13 CN CN202410284293.4A patent/CN117961643A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Luo et al. | Early fault detection of machine tools based on deep learning and dynamic identification | |
| CN111947928B (en) | Multi-source information fusion bearing fault prediction system and method | |
| CN104808585B (en) | A kind of quick inspection method of lathe health status | |
| US6587737B2 (en) | Method for the monitoring of a plant | |
| CN107291991B (en) | Early defect early warning method for wind turbine generator based on dynamic network sign | |
| CN116701984B (en) | Automatic fault diagnosis system of numerical control machine tool based on machine learning | |
| CN110907170B (en) | Wind turbine generator gearbox bearing temperature state monitoring and fault diagnosis method | |
| CN115935243B (en) | Fault analysis method based on data processing | |
| CN112149953B (en) | Electromechanical equipment operation safety assessment method based on multimode linkage and multistage cooperation | |
| CN114962239B (en) | Equipment fault detection method based on intelligent Internet of things | |
| CN114326593B (en) | Cutter life prediction system and method | |
| CN114800040A (en) | Cutter wear monitoring method and system based on process-state data correlation | |
| CN113237619B (en) | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration | |
| CN108363822A (en) | Cutter rigid predictions are used to inhibit the device and method of cutting-vibration | |
| CN117961643A (en) | Numerical control machine tool online fault monitoring method based on multi-axis current signals | |
| CN112621381B (en) | Intelligent health state evaluation method and device for machine tool feeding system | |
| JP3103193B2 (en) | Diagnostic equipment for rotating machinery | |
| CN114323664A (en) | Method for detecting abnormal gas vibration of gas turbine | |
| CN115326393A (en) | Wind turbine generator bearing pair fault diagnosis method based on temperature information | |
| CN112525337B (en) | Pretreatment method for vibration monitoring data of mechanical press | |
| CN117471308A (en) | Fan fault diagnosis method based on big data multi-feature technology mining | |
| CN115263644A (en) | Intelligent early warning method for faults of top cover of water turbine | |
| Lin et al. | Performance analysis of rotating machinery using enhanced cerebellar model articulation controller (E-CMAC) neural networks | |
| CN114239708A (en) | Combustion engine abnormity detection method based on quality control chart theory | |
| CN109102001A (en) | A kind of gene improve the rotor on-line fault diagnosis method of neural 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 |