CN104462770A - Signal reconstruction method taking sensor performance degradation into consideration - Google Patents
Signal reconstruction method taking sensor performance degradation into consideration Download PDFInfo
- Publication number
- CN104462770A CN104462770A CN201410623331.0A CN201410623331A CN104462770A CN 104462770 A CN104462770 A CN 104462770A CN 201410623331 A CN201410623331 A CN 201410623331A CN 104462770 A CN104462770 A CN 104462770A
- Authority
- CN
- China
- Prior art keywords
- signal
- wavelet
- sensor performance
- layer
- perfect
- 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.)
- Granted
Links
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention relates to a signal reconstruction method taking sensor performance degradation into consideration. The method includes the steps that firstly, an lp norm entropy standard serves as a cost function to select a wavelet function and a decomposition level; then various layers of wavelet detail coefficients of a perfect signal and a performance degradation signal are calculated, related coefficients, corresponding to various layers of wavelet detail signals, between the perfect signal and the performance degradation signal are calculated, and the number of layers with the related coefficients greater than a threshold value is positioned; the related coefficients in the wavelet detail coefficients of the layer in the perfect signal are replaced with the wavelet detail coefficients of the layer in the performance degradation signal, data reconstruction is carried out in a wavelet reconstruction mode, and therefore signal reconstruction taking sensor performance degradation into consideration is finished. The method can effectively utilize useful information in the performance degradation signal, and therefore damage identification and prediction can still be carried out according to existing algorithms when performance of a sensor degrades.
Description
Technical field
The present invention relates to a kind of signal reconfiguring method that sensor performance is degenerated of considering, in particular to a kind of in long term monitoring process due to the performance degradation of sensor cause gather signal to some extent deviation cause judge by accident problem, carried method can be utilized to carry out data reconstruction work.
Background technology
Monitoring structural health conditions mainly installs appropriate sensor at the key/damageable zone of structural system, by carrying out textural anomaly warning, damage reason location and degree of injury prediction to the analysis of Sensor monitoring data.Like this quality of the analysis of SHM measured data, process and damage diagnosis method is just become to weigh monitoring system whether effective crucial.
But in practical structures, structure actual measurement response is often subject to environment (wind, temperature, humidity) with change, work load, sensor performance degenerate/several factors such as fault, damage and the impact that is coupled thereof, and current research and achievement focus mostly in response under consideration environment and load and damage check.Current achievement provides technical support and guarantee for the daily management of great civil engineering structure, maintenance and protection against and mitigation of earthquake disasters, but they obtain on the basis that hypothesis sensor performance well, does not consider its performance degradation and fault.And practical structures detection/monitoring system because funds, number of sensors are huge, external environment impact and the reason such as to work online for a long time makes sensor performance degenerate and even section failure and inefficacy.And the structural response change caused by sensor performance degeneration and fault is often the same order of magnitude with damaging that the structural response that causes changes, therefore be often classified as that damage causes mistakenly, this causes many times wrong early warning, damage check is changed more difficult and unreliable.
Although recently there is some scholars to notice above problem, carry out the work that some are reconstructed sensor performance degraded data and replace, such as utilize PCA, MCPCA etc., but major part does not consider that the signal that it gathers when sensor performance is degenerated contains certain structure self information.
Summary of the invention
Not only the object of the present invention is to provide a kind of signal reconfiguring method that sensor performance is degenerated of considering, thus reach the structure effective information but also the performance degradation information abandoning the sensor self in the signal of collection that retain and gather when sensor generation performance degradation.
For achieving the above object, technical scheme of the present invention is: a kind ofly consider the signal reconfiguring method that sensor performance is degenerated: first, utilizes
lpnorm entropy standard selects the wavelet function Sum decomposition number of plies as cost function; Secondly, calculate each layer wavelet details coefficient of perfect signal and sensor performance degraded signal, and calculating the related coefficient of the correspondence each layer wavelet details signal between perfect signal and sensor performance degraded signal, phase relationship number is greater than the number of plies of threshold value simultaneously; Finally, the wavelet details coefficient of this layer in the related coefficient sensor performance degraded signal in the wavelet details coefficient of this in perfect signal layer is replaced, and utilize wavelet reconstruction means to carry out data reconstruction, thus complete the signal reconstruction considering that sensor performance is degenerated.
In embodiments of the present invention, the method specifically comprises the steps:
Step 1: gather acceleration responsive signal, wherein, this acceleration responsive signal containing perfect signal and sensor performance degraded signal, and identifies the sensor performance degraded signal generation moment;
Step 2: the Wavelet Energy Spectrum calculating perfect signal by formula (1)-(3)
lpnorm entropy;
Suppose that the structure acceleration response signal collected is
x(
t), then it can be expressed as:
（1）
Wherein:
x i,j (
t) be WAVELET PACKET DECOMPOSITION reconstruction coefficients,
ifor the WAVELET PACKET DECOMPOSITION number of plies, and
i=3,4 ... 10,
jfor decomposing the position node at stratum place, and
j=0,1,2 ..., 2
i-1
;
tit is time series;
So wavelet packet exists
ilayer
jthe Wavelet Packet Energy Spectrum of node
e i,j then can be represented by formula (2):
（2）
Wavelet Packet Energy Spectrum
lpnorm entropy
s k (
e i ) calculating formula see formula (3):
（3）
Wherein:
kfor db race wavelet function db
nin small echo
n, and
n=3,4 ... 20;
ifor the WAVELET PACKET DECOMPOSITION number of plies;
e for Wavelet Packet Energy Spectrum;
sfor the norm entropy of Wavelet Packet Energy Spectrum; And 1≤
p≤ 2;
Step 3: select the wavelet function Sum decomposition number of plies according to formula (4):
（4）
Step 4: calculate perfect signal and each layer detail signal of small echo of sensor performance degraded signal and the related coefficient of approximate signal, that is:
（5）
In formula:
x l,
y l be respectively perfect signal and sensor performance degraded signal in the structural response abnormal front and back cycle;
lfor in the cycle
lindividual sampled point;
nfor the sampling number in the cycle;
rthe each layer signal related coefficient of small echo of each layer signal of the small echo for perfect signal and sensor performance degraded signal;
Step 5: the signal that a sampling period of abnormal front and back according to the wavelet function Sum decomposition number of plies selected by step 3, sensor signal is occurred respectively carries out wavelet decomposition;
Step 6: the detail coefficients related coefficient calculated in step 4 being greater than the sensor performance degraded signal of this layer of threshold value replaces the detail coefficients of this layer in perfect signal;
Step 7: the detail coefficients finally formed according to step 6 carries out wavelet reconstruction by wavelet reconstruction technology.
In embodiments of the present invention, in described step 6, described threshold value gets 0.9.
Compared to prior art, the present invention has following beneficial effect:
1, the technology of the present invention directly utilizes acceleration signal, succinctly facilitates;
2, the technology of the present invention proposes a kind of signal reconfiguring method that sensor performance is degenerated of considering;
3, contemplated by the invention the structure effective information gathered when performance degradation occurs sensor self;
4, the technology of the present invention is utilized can to continue the damnification recognition method using forefathers to carry.
Accompanying drawing explanation
Fig. 1 is the process flow diagram that the present invention considers the signal reconfiguring method that sensor performance is degenerated.
Fig. 2 is the truss finite element model figure in the embodiment of the present invention.
Fig. 3 is failure-free data and the fault data of the signal reconstruction result figure of the inventive method.
Fig. 4 is the failure-free data of the signal reconstruction result figure of the inventive method and recovers data.
Embodiment
Below in conjunction with accompanying drawing, technical scheme of the present invention is specifically described.
As shown in Figure 1, the present invention is a kind of considers the signal reconfiguring method that sensor performance is degenerated: the method considers the spontaneous raw performance degradation of sensor and starts, and the structural response information that the performance degradation of its signal gathered both containing sensor causes is also containing structure self response message.For this reason, first, utilize
lpnorm entropy standard selects the suitable wavelet function Sum decomposition number of plies as cost function; Secondly, the while of calculating each layer wavelet details coefficient of perfect signal and sensor performance degraded signal and calculate the related coefficient of the correspondence each layer wavelet details signal between perfect signal and sensor performance degraded signal, phase relationship number is greater than the number of plies of threshold value; Finally, the wavelet details coefficient of this layer in the related coefficient sensor performance degraded signal in the wavelet details coefficient of this in perfect signal layer is replaced, and utilize wavelet reconstruction means to carry out data reconstruction, thus complete the signal reconstruction considering that sensor performance is degenerated.The method can effectively utilize the useful information in performance degradation signal, thus reaches and still can carry out non-destructive tests, prediction according to existing algorithm when sensor generation performance degradation.
Consider in described method that the signal reconfiguring method that sensor performance is degenerated comprises following steps:
Step 1: gather acceleration responsive signal, wherein, this acceleration responsive signal containing perfect signal and (abnormal signal) sensor performance degraded signal, and identifies the sensor performance degraded signal generation moment;
Step 2: the Wavelet Energy Spectrum calculating perfect signal by formula (1)-(3)
lpnorm entropy;
Suppose that the structure acceleration response signal collected is
x(
t), then it can be expressed as:
（1）
Wherein:
x i,j (
t) be WAVELET PACKET DECOMPOSITION reconstruction coefficients,
ifor the WAVELET PACKET DECOMPOSITION number of plies (
i=3,4 ... 10),
jfor decomposing the position node at stratum place,
j=0,1,2 ..., 2
i-1
;
tit is time series;
So wavelet packet exists
ilayer
jthe Wavelet Packet Energy Spectrum of node
e i,j then can be represented by formula (2):
（2）
Wavelet Packet Energy Spectrum
lpnorm entropy
s k (
e i ) calculating formula see formula (3):
（3）
Wherein:
kfor db race wavelet function db
nin small echo
n, (
n=3,4 ... 20);
ifor Decomposition order;
e for Wavelet Packet Energy Spectrum;
sfor the norm entropy of Wavelet Packet Energy Spectrum; 1≤
p≤ 2;
Step 3: select the wavelet function Sum decomposition number of plies according to formula (4):
（4）
Step 4: calculate perfect signal and each layer detail signal of small echo of abnormal signal and the related coefficient of approximate signal, that is:
（5）
In formula:
x l,
y l be respectively perfect signal and abnormal signal in the structural response abnormal front and back cycle;
lfor in the cycle
lindividual sampled point;
nfor the employing in the cycle is counted.
rthe each layer signal related coefficient of small echo of each layer signal of the small echo for perfect signal and abnormal signal;
Step 5: the signal that a sampling period of abnormal front and back according to the wavelet function Sum decomposition number of plies selected by step 3, sensor signal is occurred respectively carries out wavelet decomposition;
Step 6: the detail coefficients related coefficient calculated in step 4 being greater than the performance degradation signal of this layer of threshold value (0.9) replaces the detail coefficients of this layer in perfect signal;
Step 7: the detail coefficients finally formed according to step 6 carries out wavelet reconstruction by wavelet reconstruction technology.
Concrete, with the truss model of a numerical simulation, this algorithm is described.The size of model as shown in Figure 2.This model adopts Newmark integration to carry out model configuration dynamic response, the parameter wherein related to
γ=1/2,
β=1/4.Acquire the vertical dynamic response acceleration of node 1-5.Sample frequency is 5000Hz, gathers the acceleration responsive under transient excite, treats that acceleration responsive is substantially steady and just start to encourage close to null value next time after excitation input each time.The acceleration sampled result of performance degradation employing drift mode and sensor 3 is carried out following change and is carried out analog sensor performance degradation.
Wherein
a=0.3,
b=0.0001, for this sensor performance degeneration operating mode, gather acceleration responsive 20 times altogether, each collection 500 sample points.
(1) first consider that in hammering method, percussion power is not quite similar, therefore the acceleration collected carried out standardization by following formula:
Wherein
x 0for the acceleration responsive after process,
xfor the acceleration responsive gathered,
for acceleration responsive average.
(2) perfect signal that utilizes utilizing the present invention to propose selects the wavelet function Sum decomposition number of plies.Result is db10, decomposes 5 layers.
(3) calculate perfect signal and each layer detail signal of small echo of abnormal signal and the related coefficient of approximate signal, the results are shown in Table 1.
(4) can find out from the related coefficient of table 1 and can the coefficient detail coefficients of the 1st layer to the 3rd layer be replaced.
(5) wavelet reconstruction: carry out small echo according to the wavelet coefficient after replacement and reconstruct layer by layer, Fig. 3 is reconstruction result figure.
Be more than preferred embodiment of the present invention, all changes done according to technical solution of the present invention, when the function produced does not exceed the scope of technical solution of the present invention, all belong to protection scope of the present invention.
Claims (3)
1. consider to it is characterized in that the signal reconfiguring method that sensor performance is degenerated: first, utilize
lpnorm entropy standard selects the wavelet function Sum decomposition number of plies as cost function; Secondly, calculate each layer wavelet details coefficient of perfect signal and sensor performance degraded signal, and calculating the related coefficient of the correspondence each layer wavelet details signal between perfect signal and sensor performance degraded signal, phase relationship number is greater than the number of plies of threshold value simultaneously; Finally, the wavelet details coefficient of this layer in the related coefficient sensor performance degraded signal in the wavelet details coefficient of this in perfect signal layer is replaced, and utilize wavelet reconstruction means to carry out data reconstruction, thus complete the signal reconstruction considering that sensor performance is degenerated.
2. the signal reconfiguring method of consideration sensor performance degeneration according to claim 1, is characterized in that: specifically comprise the steps:
Step 1: gather acceleration responsive signal, wherein, this acceleration responsive signal containing perfect signal and sensor performance degraded signal, and identifies the sensor performance degraded signal generation moment;
Step 2: the wavelet energy calculating perfect signal by formula (1)-(3)
lpspectral norm entropy;
Suppose that the structure acceleration response signal collected is
x(
t), then it can be expressed as:
（1）
Wherein:
x i,j (
t) be WAVELET PACKET DECOMPOSITION reconstruction coefficients,
ifor the WAVELET PACKET DECOMPOSITION number of plies, and
i=3,4 ... 10,
jfor decomposing the position node at stratum place, and
j=0,1,2 ..., 2
i-1
;
tit is time series;
So wavelet packet exists
ilayer
jthe Wavelet Packet Energy Spectrum of node
e i,j then can be represented by formula (2):
（2）
Wavelet Packet Energy Spectrum
lpnorm entropy
s k (
e i ) calculating formula see formula (3):
（3）
Wherein:
kfor db race wavelet function db
nin small echo
n, and
n=3,4 ... 20;
ifor the WAVELET PACKET DECOMPOSITION number of plies;
e for Wavelet Packet Energy Spectrum;
sfor the norm entropy of Wavelet Packet Energy Spectrum; And 1≤
p≤ 2;
Step 3: select the wavelet function Sum decomposition number of plies according to formula (4):
（4）
Step 4: calculate perfect signal and each layer detail signal of small echo of sensor performance degraded signal and the related coefficient of approximate signal, that is:
（5）
In formula:
x l,
y l be respectively perfect signal and sensor performance degraded signal in the structural response abnormal front and back cycle;
lfor in the cycle
lindividual sampled point;
nfor the sampling number in the cycle;
rthe each layer signal related coefficient of small echo of each layer signal of the small echo for perfect signal and sensor performance degraded signal;
Step 5: the signal that a sampling period of abnormal front and back according to the wavelet function Sum decomposition number of plies selected by step 3, sensor signal is occurred respectively carries out wavelet decomposition;
Step 6: the detail coefficients related coefficient calculated in step 4 being greater than the sensor performance degraded signal of this layer of threshold value replaces the detail coefficients of this layer in perfect signal;
Step 7: the detail coefficients finally formed according to step 6 carries out wavelet reconstruction by wavelet reconstruction technology.
3. the signal reconfiguring method of consideration sensor performance degeneration according to claim 2, it is characterized in that: in described step 6, described threshold value gets 0.9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410623331.0A CN104462770B (en) | 2014-11-10 | 2014-11-10 | Signal reconstruction method taking sensor performance degradation into consideration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410623331.0A CN104462770B (en) | 2014-11-10 | 2014-11-10 | Signal reconstruction method taking sensor performance degradation into consideration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104462770A true CN104462770A (en) | 2015-03-25 |
| CN104462770B CN104462770B (en) | 2017-04-12 |
Family
ID=52908799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410623331.0A Active CN104462770B (en) | 2014-11-10 | 2014-11-10 | Signal reconstruction method taking sensor performance degradation into consideration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104462770B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104880217A (en) * | 2015-06-17 | 2015-09-02 | 卢伟 | Fault sensor information reconstruction method based on measured value association degree |
| CN111880117A (en) * | 2020-07-28 | 2020-11-03 | 北京交通大学 | Fault diagnosis method and device for energy-fed power supply device and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060029282A1 (en) * | 2004-08-09 | 2006-02-09 | Electric Picture Company, Ltd. | Image Blocking Artifact Reduction Via Transform Pair |
| US7672717B1 (en) * | 2003-10-22 | 2010-03-02 | Bionova Technologies Inc. | Method and system for the denoising of large-amplitude artifacts in electrograms using time-frequency transforms |
| CN102072829A (en) * | 2010-11-04 | 2011-05-25 | 同济大学 | Iron and steel continuous casting equipment oriented method and device for forecasting faults |
-
2014
- 2014-11-10 CN CN201410623331.0A patent/CN104462770B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7672717B1 (en) * | 2003-10-22 | 2010-03-02 | Bionova Technologies Inc. | Method and system for the denoising of large-amplitude artifacts in electrograms using time-frequency transforms |
| US20060029282A1 (en) * | 2004-08-09 | 2006-02-09 | Electric Picture Company, Ltd. | Image Blocking Artifact Reduction Via Transform Pair |
| CN102072829A (en) * | 2010-11-04 | 2011-05-25 | 同济大学 | Iron and steel continuous casting equipment oriented method and device for forecasting faults |
Non-Patent Citations (2)
| Title |
|---|
| 左宪章 等: "《基于小波包最优基子带能量的裂纹特征提取》", 《机械强度》 * |
| 李昕昱，张颖: "《小波变换不同替换方法对图像融合的影响》", 《地理空间信息》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104880217A (en) * | 2015-06-17 | 2015-09-02 | 卢伟 | Fault sensor information reconstruction method based on measured value association degree |
| CN111880117A (en) * | 2020-07-28 | 2020-11-03 | 北京交通大学 | Fault diagnosis method and device for energy-fed power supply device and storage medium |
| CN111880117B (en) * | 2020-07-28 | 2022-04-08 | 北京交通大学 | Fault diagnosis method and device for energy-fed power supply device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104462770B (en) | 2017-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Klis et al. | Vibration monitoring via spectro-temporal compressive sensing for wireless sensor networks | |
| KR101942806B1 (en) | Moving type defect diagnosis system for photovoltaic power generation equipment | |
| CN102269655B (en) | Method for diagnosing bearing fault | |
| Zonzini et al. | Model-assisted compressed sensing for vibration-based structural health monitoring | |
| CN104880217B (en) | A kind of fault sensor signal reconstruct method based on the measured value degree of association | |
| CN104239736A (en) | Structure damage diagnosis method based on power spectrum and intelligent algorithms | |
| CN104265577B (en) | Wind generating set abnormal detection method based on compressed sensing | |
| CN204113701U (en) | A kind of mine fan on-line monitoring and fault diagnosis system | |
| CN105379186A (en) | Determining response similarity neighborhoods | |
| CN106021719A (en) | Unscented Kalman filter algorithm-based method for predicting residual life of bearing | |
| CN104316167B (en) | Structural response method for detecting abnormality under a kind of low signal-to-noise ratio | |
| Hartmann et al. | Coupling sensor-based structural health monitoring with finite element model updating for probabilistic lifetime estimation of wind energy converter structures | |
| Xie et al. | Data fault detection for wireless sensor networks using multi-scale PCA method | |
| CN103235568B (en) | A kind of nuclear power plant equipment condition monitoring wireless detection device | |
| CN105491614A (en) | Wireless sensor network abnormal event detection method and system based on secondary mixed compression | |
| CN107656156A (en) | A kind of equipment fault diagnosis and running status appraisal procedure and system based on cloud platform | |
| Moeinifard et al. | Lost vibration test data recovery using convolutional neural network: a case study | |
| CN104462770A (en) | Signal reconstruction method taking sensor performance degradation into consideration | |
| CN102323049B (en) | Structural abnormality detection method based on consistent data replacement under incomplete data | |
| CN116662920B (en) | Abnormal data identification method, system, equipment and medium for drilling and blasting method construction equipment | |
| Carratù et al. | Earthquake early detection as an IEEE1451 transducer network trigger for urban infrastructure monitoring and protection | |
| KR20180125276A (en) | Fault Signal Recovery System and Method | |
| CN106197759A (en) | A kind of temperature sensor control system | |
| CN115436800A (en) | Motor remote state monitoring method based on multiple modules | |
| CN113077628B (en) | Algorithm of composite geomagnetic vehicle detector |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |