CN117272030B - Method for sampling and processing dynamic signal packet - Google Patents
Method for sampling and processing dynamic signal packet Download PDFInfo
- Publication number
- CN117272030B CN117272030B CN202311544677.7A CN202311544677A CN117272030B CN 117272030 B CN117272030 B CN 117272030B CN 202311544677 A CN202311544677 A CN 202311544677A CN 117272030 B CN117272030 B CN 117272030B
- Authority
- CN
- China
- Prior art keywords
- signal
- sampling
- signals
- frequency
- clock
- 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
- 238000005070 sampling Methods 0.000 title claims abstract description 118
- 238000012545 processing Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000010801 machine learning Methods 0.000 claims abstract description 10
- 238000013507 mapping Methods 0.000 claims abstract description 10
- 230000006870 function Effects 0.000 claims abstract description 8
- 238000012544 monitoring process Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 8
- 230000005236 sound signal Effects 0.000 claims description 8
- 238000013461 design Methods 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 238000012952 Resampling Methods 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 3
- 230000006872 improvement Effects 0.000 description 9
- 230000002159 abnormal effect Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 238000003672 processing method Methods 0.000 description 3
- 239000013589 supplement Substances 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to a method for sampling and processing dynamic signal packets, which realizes a scheme of sampling packets by using two clocks by using the technology of the method for sampling and processing the dynamic signal packets and adapts to signal sampling requirements in different frequency ranges by dynamically adjusting the highest frequency of the clocks. And a machine learning algorithm is adopted to establish a function mapping relation between the characteristics and the sampling signal frequency, so that a clock self-adaptive highest-frequency grouping sampling process is realized, and then a second clock is used for accurate sampling. The method can improve the efficiency and accuracy of signal acquisition, and is suitable for acquisition and processing of vibration signals of large-scale equipment in the fields of aerospace, rail transit, large-scale electric equipment, ship engineering and the like.
Description
Technical Field
The invention relates to a sampling and processing method, in particular to a dynamic signal grouping sampling and processing method based on machine learning, which is applied to the technical field of signal sampling.
Background
In the fields of aerospace, rail transit, large-scale electric power equipment, ship engineering and the like, the method is very important for the acquisition and processing of dynamic signals such as vibration, sound, rotating speed, temperature and the like of large-scale equipment. The frequency characteristics of these dynamic signals are different, and the conventional method generally needs to adopt multiple sets of clocks and proper sampling rates to perform time-division sampling on the signals in different frequency bands, which not only greatly increases the acquisition cost, but also results in lower accuracy of signal sampling.
Disclosure of Invention
Aiming at the prior art, the invention aims to solve the technical problems that the acquisition cost is greatly increased by adopting a mode of sampling signals by a plurality of groups of clocks, and the accuracy of signal sampling is lower.
In order to solve the above problems, the present invention provides a method for sampling and processing a dynamic signal packet, which includes the following steps:
s1, initializing a system:
setting two clocks in the system as time base of sampling; initializing a machine learning model, and preparing algorithms and parameters required by feature extraction and feature prediction;
s12, signal acquisition and grouping:
the method comprises the steps that dynamic signals are collected through a first clock arranged in a system, the signals comprise vibration signals, sound signals, rotating speed signals, temperature signals and other signals, and the collected signals are divided into a plurality of groups by means of grouping sampling;
s3, feature extraction and processing:
extracting the characteristics of each group of sampling signals; calculating the amplitude characteristic, the statistical characteristic and the time domain characteristic of each group of signals; the amplitude variation, statistical distribution and time characteristics of the signal are described by these features.
S4, mapping functions of the features and sampling signal frequencies:
establishing a function mapping relation between the features and the sampling signal frequency by using a machine learning algorithm, and learning a correlation rule between the features and the sampling signal frequency by using a training model to obtain a feature prediction model;
s5, dynamically adjusting the clock frequency:
predicting the characteristics of each group of sampling signals according to the characteristic prediction model; judging the frequency range of the signal according to the prediction result, and correspondingly adjusting the highest frequency of the first clock;
s6, sampling and processing:
resampling the signals according to the adjusted first clock frequency, and grouping and sampling by using the dynamically adjusted clock to ensure that the dynamic characteristics of all the signals can be captured; the second clock is dynamically adjusted according to the frequency of the signals acquired by the first clock, and then certain signals are independently sampled, wherein the signals sampled by the second clock are more accurate relative to the signals sampled by the first clock; the secondary acquired signals are further processed, analyzed or transmitted to meet specific application requirements.
In the method for sampling and processing the dynamic signal packet, two clocks are used, and the sampling of all signals is realized by dynamically adjusting the highest frequencies of the two clocks; according to the method, only two clocks are used for grouping sampling, the highest frequency of one clock is dynamically adjusted, the amplitude change, the statistical characteristic and the time domain characteristic of signals with different frequencies can be captured, the frequency of the other clock is adjusted according to the signals with different frequencies to independently sample different signals, and the accuracy of signal sampling can be effectively improved.
As a further improvement of the present application, in the step S5, a high-frequency signal and a low-frequency signal are taken as examples; when the prediction result shows that the signal belongs to the low-frequency signal, the clock frequency is reduced to adapt to the sampling requirement of the low-frequency signal; if the prediction result shows that the signal belongs to the high-frequency signal, the clock frequency is increased to meet the sampling requirement of the high-frequency signal.
As a further improvement of the present application, in the step S2, the signals in different frequency ranges are sampled in groups by performing the group sampling by the first clock, wherein in order to achieve the sampling of the dynamic signals, a suitable highest clock frequency is selected to cover the frequency ranges of all the signals; the signals are divided into different groups according to the frequency range of the signals, for example, vibration signals, sound signals, temperature signals can be divided into a low frequency group and a high frequency group, and the sampling period of each group is determined according to the highest frequency of the signals of the group.
As a further improvement of the present application, the feature extraction in the step S3 measures the sharpness of the signal peak using kurtosis features, that is, the peak degree of the signal distribution; the calculation of kurtosis is used for helping to identify nonlinear features in the signal, and has important significance for analysis of dynamic signals and evaluation of vibration characteristics.
As a further improvement of the present application, in the step S5, the frequency range is determined using the feature prediction result. Setting a threshold according to actual application requirements and system design; when the feature prediction result is smaller than or equal to a threshold value or belongs to a low frequency range, the sampling signal is judged to be a low frequency signal, and when the high frequency signal is sampled, the clock frequency is reduced to adapt to the sampling requirement of the low frequency signal; the sampling frequency can be reduced by reducing the highest frequency of the clock, so that the dynamic characteristics of the low-frequency signals can be fully captured; when the feature prediction result is larger than a threshold value or belongs to a high frequency range, the sampling signal is judged to be a high frequency signal, and when the high frequency signal is sampled, the clock frequency is increased to meet the sampling requirement of the high frequency signal; the sampling frequency can be increased by increasing the highest frequency of the clock, ensuring that the dynamics of the high frequency signal can be adequately captured.
As a further improvement of the present application, in said step S6, the signals are grouped according to the accurate sampling of the second clock at more accurate time intervals, the number of sampling points in each group depending on the sampling frequency and the sampling duration, so that the adaptive group length can be adapted better to the variation of the signals. The sampled data within each group is processed and then stored or transmitted as needed. The storage may use a suitable data format and storage medium for subsequent analysis and application.
As a further improvement of the present application, after the signal sampling in step S6, a step of performing monitoring and early warning on the waveform is added, wherein the monitoring and early warning on the waveform is performed by a signal waveform monitoring system;
the waveform monitoring system comprises a standard waveform database, a waveform comparison module, an image processing module and an alarm module, wherein the waveform comparison module and the image processing module are both in signal connection with the standard waveform database, and the alarm module is in signal connection with the waveform comparison module.
As a supplement to the further improvement of the application, the specific steps of the waveform for monitoring and early warning are as follows:
s61, firstly converting the sampled signals into wave bands, obtaining corresponding waveforms, corresponding the wave bands to a standard waveform database, and obtaining the shapes of the corresponding wave bands on the standard waveform database;
s62, comparing the shapes of the signal wave band and the corresponding wave band of the standard waveform through the image processing module, judging the amplitude of the target signal wave through the deviation between the signal wave band and the standard waveform, and when the larger deviation exists between the signal wave band and the standard wave band, indicating that the problem of overlarge or undersize amplitude exists in the target signal wave;
and S63, when the waveform amplitude has a problem, the alarm module alarms staff at the control center, and the staff can implement different plans according to the deviation condition of the waveform.
As a supplement to yet another improvement of the present application, the standard waveform database is a waveform of a target signal in the absence of an anomaly, the target signal also including a vibration signal, a sound signal, a rotational speed signal, a temperature signal, and other signals.
As a supplement to yet another improvement of the present application, the alarm module includes an in-field alarm mounted on the target signaling device and an out-of-line alarm disposed within the tower.
In summary, by the dynamic signal packet sampling and processing method of the present invention, a scheme of packet sampling using two clocks is realized, and the highest frequency of the clocks is dynamically adjusted to adapt to the signal sampling requirements of different frequency ranges. And a machine learning algorithm is adopted to establish a function mapping relation between the characteristics and the sampling signal frequency, so that a clock self-adaptive highest-frequency grouping sampling process is realized, and then a second clock is used for accurate sampling. The method can improve the efficiency and accuracy of signal acquisition, and is suitable for acquisition and processing of vibration signals of large-scale equipment in the fields of aerospace, rail transit, large-scale electric power equipment, ship engineering and the like.
Drawings
FIG. 1 is a main flow chart of a first embodiment of the present application;
FIG. 2 is a schematic diagram of monitoring and early warning waveforms according to a second embodiment of the present disclosure;
FIG. 3 is a graph comparing signal bands with standard waveforms in a second embodiment of the present application;
fig. 4 is a diagram showing a comparison between a signal band and a standard band corresponding to a standard waveform in the second embodiment of the present application.
Detailed Description
Two embodiments of the present application are described in detail below with reference to the accompanying drawings.
First embodiment:
fig. 1 shows a method of dynamic signal packet sampling and processing, comprising the steps of:
s1, initializing a system:
setting two clocks in the system as time base of sampling; initializing a machine learning model, and preparing algorithms and parameters required by feature extraction and feature prediction;
s12, signal acquisition and grouping:
the method comprises the steps that dynamic signals are collected through a first clock arranged in a system, the signals comprise vibration signals, sound signals, rotating speed signals, temperature signals and other signals, and the collected signals are divided into a plurality of groups by means of grouping sampling;
s3, feature extraction and processing:
extracting the characteristics of each group of sampling signals; calculating the amplitude characteristic, the statistical characteristic and the time domain characteristic of each group of signals; the amplitude variation, statistical distribution and time characteristics of the signal are described by these features.
S4, mapping functions of the features and sampling signal frequencies:
establishing a function mapping relation between the features and the sampling signal frequency by using a machine learning algorithm, and learning a correlation rule between the features and the sampling signal frequency by using a training model to obtain a feature prediction model;
s5, dynamically adjusting the clock frequency:
predicting the characteristics of each group of sampling signals according to the characteristic prediction model; judging the frequency range of the signal according to the prediction result, and correspondingly adjusting the highest frequency of the first clock;
take high frequency signal and low frequency signal as examples; when the prediction result shows that the signal belongs to the low-frequency signal, the clock frequency is reduced to adapt to the sampling requirement of the low-frequency signal; if the prediction result shows that the signal belongs to the high-frequency signal, the clock frequency is increased to meet the sampling requirement of the high-frequency signal.
S6, sampling and processing:
resampling the signals according to the adjusted first clock frequency, and grouping and sampling by using the dynamically adjusted clock to ensure that the dynamic characteristics of all the signals can be captured; the second clock is dynamically adjusted according to the frequency of the signals acquired by the first clock, and then certain signals are independently sampled, wherein the signals sampled by the second clock are more accurate relative to the signals sampled by the first clock; the secondary acquired signals are further processed, analyzed or transmitted to meet specific application requirements.
In the step S2, the signals in different frequency ranges are sampled in groups by the first clock, wherein in order to realize the sampling of the dynamic signals, a proper highest clock frequency is selected to cover the frequency ranges of all the signals; assuming that the highest clock frequency is selected to be：
（1）
Wherein,fmax is the dynamically adjusted clock maximum frequency,fbase is the reference clock frequency and,αis a frequency adjustment coefficient.
Sampling period:
（2）
the signals are divided into different groups according to the frequency range of the signals, for example, vibration signals, sound signals, temperature signals can be divided into a low frequency group and a high frequency group, and the sampling period of each group is determined according to the highest frequency of the signals of the group.
In the step S3, the feature extraction uses kurtosis features, that is, the peak degree of the signal distribution, and measures the sharpness of the signal peak in the following manner:
（3）
where x represents the data sequence of the sampled signal within a time window and N represents the number of signal samples. Mean represents the average value of the sampled signal data sequence, and standard deviation represents the standard deviation.
The calculation of kurtosis is used for helping to identify nonlinear features in the signal, and has important significance for analysis of dynamic signals and evaluation of vibration characteristics.
In step S4, we use a machine learning algorithm to build a functional mapping between features and sampling signal frequencies. And learning a correlation rule between the features and the sampling signal frequency through training the model to obtain a feature prediction model.
And extracting kurtosis characteristics and preparing a data set. The functional mapping of features to sampling signal frequencies can be expressed as the following equation:
（4）
where f_predicted represents the sampling signal frequency output by the feature prediction model, and Features is the feature vector input to the feature prediction model.
In the step S5, the frequency range is determined using the feature prediction result. Setting a threshold according to actual application requirements and system design;
when feature predictionf_predictedThe result is less than or equal to the threshold value or falls within a low frequency range, the sampled signal isDetermining a low-frequency signal, and reducing the clock frequency to adapt to the sampling requirement of the low-frequency signal when the high-frequency signal is sampled; the sampling frequency can be reduced by reducing the highest frequency of the clock, so that the dynamic characteristics of the low-frequency signals can be fully captured;
as a result of feature predictionf_predictedThe sampling signal is judged to be a high-frequency signal when the sampling signal is greater than a threshold value or belongs to a high-frequency range, and when the high-frequency signal is sampled, the clock frequency is increased to meet the sampling requirement of the high-frequency signal, and the sampling frequency can be increased by increasing the highest frequency of the clock, so that the dynamic characteristics of the high-frequency signal can be fully captured.
By dynamically adjusting the frequency of the first clock, the system can adaptively adapt to sampling requirements of signals with different frequencies according to real-time characteristic prediction results, and then the second clock adjusts the frequency according to the signals sampled by the first clock, so that signal acquisition can be more accurately performed on certain specific signals. Therefore, the data integrity and accuracy in the sampling process can be ensured to the greatest extent, and the efficiency and accuracy of signal processing are improved. Meanwhile, the formulation of the dynamic adjustment clock frequency is expressed as follows:
（5）
where α_i is an adjustment coefficient set according to the system design and actual requirements, and is used to determine a scaling factor for decreasing or increasing the clock frequency.
According to the characteristic prediction result and the frequency range judgment, the system can automatically adjust the highest frequency of the first clock so as to adapt to the sampling requirements of signals with different frequencies, and then the second clock carries out frequency adjustment according to the signals sampled by the first clock, so that signal acquisition can be carried out on certain specific signals more accurately. Therefore, the dynamic signals can be effectively collected and processed, and the accuracy and the integrity of the signals are ensured.
In the step S6, the signals are grouped according to the accurate sampling of the second clock at more accurate time intervals, and the number of sampling points in each group depends on the sampling frequency and the sampling duration, so that the adaptive group length can adapt to the change of the signals better. The sampled data within each group is processed and then stored or transmitted as needed. The storage may use a suitable data format and storage medium for subsequent analysis and application.
Second embodiment:
after the signal is sampled in the step S6, a step of monitoring and early warning the waveform is added, and the operation stability of the sampled equipment can be effectively monitored through the detection and early warning of the waveform, so that when the stability is abnormal, the personnel can conveniently and timely carry out corresponding emergency measures, and the operation stability of the equipment is effectively maintained.
The monitoring and early warning of the waveform is implemented through a signal waveform monitoring system, the waveform monitoring system comprises a standard waveform database, a waveform comparison module, an image processing module and an alarm module, the waveform comparison module and the image processing module are both in signal connection with the standard waveform database, the alarm module is in signal connection with the waveform comparison module, and the alarm module comprises an external alarm arranged in a tower and an on-site alarm arranged on target signal sending equipment.
When the method is applied to the aerospace field, the tower, namely the ground command control center, the target signal sending device, namely the aircraft, and when deviation between the signal waveform and the corresponding waveform in the standard waveform database is detected after sampling, the condition that the signal is abnormal is indicated on the aircraft, on one hand, the field alarm on the aircraft can be used for alarming in real time, so that workers can maintain and remove obstacles in time, the normal operation of the aircraft is effectively ensured, on the other hand, after the tower on the ground receives the alarm signal of the external alarm, the target aircraft can be contacted on the one hand, abnormal alarm or information interaction can be carried out on the target aircraft, and when the abnormality is larger, rescue measures can be timely made, thereby effectively improving safety precaution.
Fig. 2 shows the specific steps of the waveform for monitoring and early warning as follows:
s61, firstly converting the sampled signals into wave bands, obtaining corresponding waveforms, corresponding the wave bands to a standard waveform database, and obtaining the shapes of the corresponding wave bands on the standard waveform database;
s62, as shown in FIG. 3, comparing the shapes of the signal wave band and the corresponding wave band of the standard waveform by the image processing module, judging the amplitude of the target signal wave by the deviation between the signal wave band and the standard wave band, and when the larger deviation exists between the signal wave band and the standard wave band, indicating that the problem of overlarge or overlarge amplitude exists in the target signal wave;
and S63, when the waveform amplitude has a problem, the alarm module alarms staff at the control center, and the staff can implement different plans according to the deviation condition of the waveform.
The standard waveform database is the waveform of a target signal under the condition of no abnormality, and the target signal also comprises a vibration signal, a sound signal, a rotating speed signal, a temperature signal and other signals.
The present embodiment is based on the first embodiment, and the other parts have the same content as the first embodiment.
Specifically, after sampling in step S6, the sampled waveform is compared with waveforms in the standard waveform database, and whether the target signal is stable or not can be primarily judged through the waveforms, and whether an abnormal condition exists or not is judged.
In the step S62, the deviation in the minimum point is marked By using the signal band as an example, and the deviation between the signal band and the standard band in the step S62 is calculated as follows:
sa, obtaining coordinates of a comparison point:
the coordinates of the lowest point, the highest point and the middle point of the signal wave band are respectively marked as (X1, Y1), (X2, Y2), (X3 and Y3), under the condition that the coordinates of the X axis are the same, the coordinates of the Y axis three points of the corresponding wave band of the standard wave form are respectively marked as Y1, Y2 and Y3, and under the condition that the coordinates of the Y axis are the same, the coordinates of the X axis three points of the corresponding wave band of the standard wave form are respectively marked as X1, X2 and X3;
sb, Y-axis deviation calculation:
under the condition that the X-axis coordinates are the same, the deviation of the Y-axis between the corresponding three points of the two wave bands is (+/-) (Y1-Y1), + - (Y2-Y2) and (+/-) (Y3-Y3), three deviation values are compared, the largest deviation value is selected to be compared with the corresponding threshold value By, when the largest deviation value is larger than the threshold value By, the problem that the amplitude of the Y-axis is abnormal is solved, wherein when the corresponding point of the signal wave band is located above the corresponding point of the standard wave band, the condition that the amplitude of the signal wave band is larger is solved, and otherwise, the amplitude is smaller;
sc and X axis deviation calculation:
under the condition that the Y-axis coordinates are the same, the deviation of the X-axis between the corresponding three points of the two wave bands is (+/-) (X1-X1), + - (X2-X2) and (+/-) (X3-X3), the three deviation values are compared, the largest deviation value is selected to be compared with the corresponding threshold value Ax, when the largest deviation value is larger than the threshold value Ax, the problem that the amplitude of the X-axis is abnormal is solved, wherein when the corresponding point of the signal wave band is positioned at the left side of the corresponding point of the standard wave band, the condition that the amplitude of the signal wave band is larger is solved, and otherwise, the amplitude is smaller.
Notably, are: the "±" symbol is used as compensation, so that the calculated deviation value of the X axis or the Y axis is always positive, and the thresholds Ax and By are also positive.
The scope of protection of the above-described embodiments employed in the present application is not limited to the above-described embodiments, and various changes made by those skilled in the art without departing from the spirit of the present application are still within the scope of protection of the present invention.
Claims (10)
1. A method for dynamic signal packet sampling and processing, characterized by: the method comprises the following steps:
s1, initializing a system:
setting two clocks in the system as time base of sampling; initializing a machine learning model, and preparing algorithms and parameters required by feature extraction and feature prediction;
s12, signal acquisition and grouping:
the method comprises the steps that dynamic signals are collected through a first clock arranged in a system, the signals comprise vibration signals, sound signals, rotating speed signals, temperature signals and other signals, and the collected signals are divided into a plurality of groups by means of grouping sampling;
s3, feature extraction and processing:
extracting the characteristics of each group of sampling signals; calculating the amplitude characteristic, the statistical characteristic and the time domain characteristic of each group of signals; describing the amplitude variation, statistical distribution and time characteristics of the signal by these features;
s4, mapping functions of the features and sampling signal frequencies:
establishing a function mapping relation between the features and the sampling signal frequency by using a machine learning algorithm, and learning a correlation rule between the features and the sampling signal frequency by using a training model to obtain a feature prediction model;
s5, dynamically adjusting the clock frequency:
predicting the characteristics of each group of sampling signals according to the characteristic prediction model; judging the frequency range of the signal according to the prediction result, and correspondingly adjusting the highest frequency of the first clock;
s6, sampling and processing:
resampling the signals according to the adjusted first clock frequency, and grouping and sampling by using the dynamically adjusted clock to ensure that the dynamic characteristics of all the signals can be captured; the second clock is dynamically adjusted according to the frequency of the signals acquired by the first clock, and then certain signals are independently sampled, wherein the signals sampled by the second clock are more accurate relative to the signals sampled by the first clock; the secondary acquired signals are further processed, analyzed or transmitted to meet specific application requirements.
2. A method of dynamic signal packet sampling and processing as claimed in claim 1, wherein: in the step S5, a high-frequency signal and a low-frequency signal are taken as examples:
when the prediction result shows that the signal belongs to the low-frequency signal, the clock frequency is reduced to adapt to the sampling requirement of the low-frequency signal; if the prediction result shows that the signal belongs to the high-frequency signal, the clock frequency is increased to meet the sampling requirement of the high-frequency signal.
3. A method of dynamic signal packet sampling and processing according to claim 2, wherein: in the step S2, the signals in different frequency ranges are sampled in groups by the first clock, wherein in order to realize the sampling of the dynamic signals, a proper highest clock frequency is selected to cover the frequency ranges of all the signals;
the signals are divided into low frequency groups and high frequency groups according to the frequency range of the signals, and the sampling period of each group is determined according to the highest frequency of the signals of the group.
4. A method of dynamic signal packet sampling and processing as claimed in claim 1, wherein: in the step S3, the sharpness of the signal peak is measured by using kurtosis features, that is, the peak degree of the signal distribution.
5. A method of dynamic signal packet sampling and processing as claimed in claim 1, wherein: in the step S5, a frequency range is determined by using the feature prediction result, and a threshold is set according to the actual application requirement and the system design;
when the feature prediction result is smaller than or equal to a threshold value or belongs to a low frequency range, the sampling signal is judged to be a low frequency signal, and when the high frequency signal is sampled, the clock frequency is reduced to adapt to the sampling requirement of the low frequency signal;
when the feature prediction result is greater than the threshold value or belongs to the high frequency range, the sampling signal is judged to be a high frequency signal, and when the high frequency signal is sampled, the clock frequency is increased to meet the sampling requirement of the high frequency signal.
6. A method of dynamic signal packet sampling and processing as claimed in claim 1, wherein: in said step S6, the signals are grouped according to the accurate sampling of the second clock at more accurate time intervals, the number of sampling points in each group being dependent on the sampling frequency and the sampling duration, adapting the adaptive group length to the variation of the signals.
7. A method of dynamic signal packet sampling and processing as claimed in claim 1, wherein: after the signal is sampled in the step S6, a step of monitoring and early warning the waveform is added, and the monitoring and early warning of the waveform is implemented through a signal waveform monitoring system;
the waveform monitoring system comprises a standard waveform database, a waveform comparison module, an image processing module and an alarm module, wherein the waveform comparison module and the image processing module are both in signal connection with the standard waveform database, and the alarm module is in signal connection with the waveform comparison module.
8. A method of dynamic signal packet sampling and processing as claimed in claim 7, wherein: the specific steps of the waveform monitoring and early warning are as follows:
s61, firstly converting the sampled signals into wave bands, obtaining corresponding waveforms, corresponding the wave bands to a standard waveform database, and obtaining the shapes of the corresponding wave bands on the standard waveform database;
s62, comparing the shapes of the signal wave band and the corresponding wave band of the standard waveform through the image processing module, judging the amplitude of the target signal wave through the deviation between the signal wave band and the standard waveform, and when the larger deviation exists between the signal wave band and the standard wave band, indicating that the problem of overlarge or undersize amplitude exists in the target signal wave;
and S63, when the waveform amplitude has a problem, the alarm module alarms staff at the control center, and the staff can implement different plans according to the deviation condition of the waveform.
9. A method of dynamic signal packet sampling and processing as claimed in claim 8, wherein: the standard waveform database is the waveform of a target signal under the condition of no abnormality, and the target signal also comprises a vibration signal, a sound signal, a rotating speed signal, a temperature signal and other signals.
10. A method of dynamic signal packet sampling and processing as claimed in claim 8, wherein: the alarm module comprises an external alarm arranged in the tower and an on-site alarm arranged on the target signal sending equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311544677.7A CN117272030B (en) | 2023-11-20 | 2023-11-20 | Method for sampling and processing dynamic signal packet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311544677.7A CN117272030B (en) | 2023-11-20 | 2023-11-20 | Method for sampling and processing dynamic signal packet |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117272030A CN117272030A (en) | 2023-12-22 |
| CN117272030B true CN117272030B (en) | 2024-01-26 |
Family
ID=89201176
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311544677.7A Active CN117272030B (en) | 2023-11-20 | 2023-11-20 | Method for sampling and processing dynamic signal packet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117272030B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1396711A (en) * | 2002-07-31 | 2003-02-12 | 上海芯华微电子有限公司 | Audio sampling frequency conversion and clock resynchnonizing device with minority coefficient |
| CN113377155A (en) * | 2021-06-08 | 2021-09-10 | 深圳市汇顶科技股份有限公司 | Clock calibration method and device and electronic equipment |
| WO2022062161A1 (en) * | 2020-09-28 | 2022-03-31 | 广东石油化工学院 | Large machine set friction fault analysis method and system based on waveform and dimensionless learning |
| CN116773961A (en) * | 2023-06-16 | 2023-09-19 | 广西电网有限责任公司电力科学研究院 | Transmission line corrosion detection method based on vibration signal high-frequency characteristic analysis |
-
2023
- 2023-11-20 CN CN202311544677.7A patent/CN117272030B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1396711A (en) * | 2002-07-31 | 2003-02-12 | 上海芯华微电子有限公司 | Audio sampling frequency conversion and clock resynchnonizing device with minority coefficient |
| WO2022062161A1 (en) * | 2020-09-28 | 2022-03-31 | 广东石油化工学院 | Large machine set friction fault analysis method and system based on waveform and dimensionless learning |
| CN113377155A (en) * | 2021-06-08 | 2021-09-10 | 深圳市汇顶科技股份有限公司 | Clock calibration method and device and electronic equipment |
| CN116773961A (en) * | 2023-06-16 | 2023-09-19 | 广西电网有限责任公司电力科学研究院 | Transmission line corrosion detection method based on vibration signal high-frequency characteristic analysis |
Non-Patent Citations (1)
| Title |
|---|
| 基于软件同步的采样频率自适应算法及仿真;史彩君;王飞;;内蒙古电力技术(01);第17-20页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117272030A (en) | 2023-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11408797B2 (en) | GIL fault on-line monitoring system based on vibration signals and support vector machine | |
| CN116896481B (en) | Internet of things-based network security data risk assessment system | |
| CN112486097A (en) | Cutting platform wear state monitoring method and system based on model algorithm and readable storage medium | |
| EP3927000B1 (en) | Network element health status detection method and device | |
| CN116660672B (en) | Power grid equipment fault diagnosis method and system based on big data | |
| CN111504450A (en) | Equipment fault alarm method and device | |
| CN117454114B (en) | Subway tunnel tunneling blasting vibration safety monitoring device based on multi-point location distribution | |
| CN115986918B (en) | Intelligent monitoring system for power transmission line | |
| CN116893643A (en) | Intelligent robot driving track safety control system based on data analysis | |
| CN117272030B (en) | Method for sampling and processing dynamic signal packet | |
| CN116086537A (en) | Equipment state monitoring method, device, equipment and storage medium | |
| CN113534727B (en) | Early warning control system of refrigeration equipment for fishing boat based on artificial intelligence platform | |
| CN117291316B (en) | Cabin positioning system for ship unloader based on laser scanner | |
| US20200356668A1 (en) | Event analysis in an electric power system | |
| CN106081958A (en) | A kind of derrick crane on-line monitoring system | |
| CN109635430A (en) | Grid power transmission route transient signal monitoring method and system | |
| CN112114196A (en) | Current harmonic statistic and alarm method and system | |
| CN107390592A (en) | A kind of business electrical monitoring information and mobile phone interaction communication system | |
| Zhong et al. | Fault diagnosis of motor bearing using self-organizing maps | |
| CN111092875B (en) | Transmission and transformation operation and inspection platform Internet of things edge information transmission compression method and system | |
| CN112234601A (en) | Method and system for identifying low-frequency oscillation characteristic parameters of power system on line | |
| CN118233945A (en) | Intelligent wireless data acquisition system | |
| CN117596535B (en) | Intelligent sound box operation detection system based on Internet of things | |
| CN114754860B (en) | Wireless vibration monitoring method, electronic equipment and storage medium | |
| CN117196314B (en) | Building construction safety monitoring system and method based on Internet of things |
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 |