CN104599676B - Eliminate the method that pectination spectral noise influences on vibration level - Google Patents

Eliminate the method that pectination spectral noise influences on vibration level Download PDF

Info

Publication number
CN104599676B
CN104599676B CN201510057963.XA CN201510057963A CN104599676B CN 104599676 B CN104599676 B CN 104599676B CN 201510057963 A CN201510057963 A CN 201510057963A CN 104599676 B CN104599676 B CN 104599676B
Authority
CN
China
Prior art keywords
mrow
msub
frequency
omega
mfrac
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
Application number
CN201510057963.XA
Other languages
Chinese (zh)
Other versions
CN104599676A (en
Inventor
孙涛
高慧中
刘培培
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
705TH RESEARCH INSTITUTE OF CHINA SHIPBUILDING INDUSTRY Corp
Original Assignee
705TH RESEARCH INSTITUTE OF CHINA SHIPBUILDING INDUSTRY Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 705TH RESEARCH INSTITUTE OF CHINA SHIPBUILDING INDUSTRY Corp filed Critical 705TH RESEARCH INSTITUTE OF CHINA SHIPBUILDING INDUSTRY Corp
Priority to CN201510057963.XA priority Critical patent/CN104599676B/en
Publication of CN104599676A publication Critical patent/CN104599676A/en
Application granted granted Critical
Publication of CN104599676B publication Critical patent/CN104599676B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)

Abstract

The invention discloses a kind of method for eliminating pectination spectral noise and being influenceed on vibration level, for solving the technical problem of the method poor practicability of existing elimination pectination spectral noise.Technical scheme is to determine associated technical parameters first, then using Fourier kernel functions as bandpass filter corresponding to basic function construction.The pectinate band-stop filter device of requirement is met by adding frequency-shifting operator again.Then the power spectrum of primary signal and the pectinate band-stop filter device constructed are combined, the periodic interference components in primary signal are filtered out under frequency domain.According to filtered spectral magnitude, power spectral density is calculated, finally calculates third-octave spectrum by frequency range using power spectral density value.This method can filter out the interference signal being spacedly distributed, and vibration level is directly calculated under frequency domain.And reported by mistake by vibration level caused by modifying factor interference noise, the repetition experiment brought by vibration wrong report is avoided, is advantageous to shorten the test period, it is practical.

Description

Eliminate the method that pectination spectral noise influences on vibration level
Technical field
The present invention relates to a kind of method for eliminating pectination spectral noise, more particularly to a kind of pectination spectral noise that eliminates is to vibrational energy The method that level influences.
Background technology
Third-octave spectrum is a kind of frequency-domain analysis method constructed according to human ear to the response characteristic of sound, has spectral line The characteristics of few bandwidth, it is widely used in the analyzing and processing of acoustics and vibration signal, the level value (vibration level/sound level) of third-octave It is the leading indicator for judging vibration noise size.As concern of the people to noise and vibration increasingly increases, octave method Constantly improved.The Chinese patent authorized in 2014《Directly obtain the multi-measuring point variable working condition octave spectrum battle array near same operating mode Method (patent No. is ZL 201110012751.1)》Comprehensive review has been carried out to the present Research of domestic and international octave method, Indicate that early-stage Study is mainly improved from computational accuracy and wave filter digitlization to standard octave algorithm, can only directly locate Manage multi-measuring point either simplex condition vibration noise data, it is impossible to multi-measuring point variable working condition vibration noise data are carried out at octave exactly Reason, and a kind of method list for the multi-measuring point variable working condition octave spectrum battle array that can be directly obtained and be arranged near same operating mode is proposed accordingly Measuring point variable working condition octave composes battle array and composes battle array with the multi-measuring point variable working condition octave near operating mode.Block measuring point spectrum battle array is determined first With block time matrix, next establishes operating mode time matrix and performance of operating condition mean parameter matrix, then determines each stable work again Start-stop data block block number corresponding to condition, finally determine single measuring point variable working condition octave spectrum battle array and become with the multi-measuring point near operating mode Operating mode octave composes battle array.Data processing time is greatly reduced compared to original method, the method, improves treatment effeciency. Nevertheless, such third-octave processing method is all still using test signal as normal signal, not comprising significantly interfering with to be preceding Carry, thus a certain degree of technological deficiency be present in processing band noise signal.
Live measured signal usually contains the periodic frequency as caused by the local defect of the parts such as gear, bearing Composition, or even the PERIODIC INTERFERENCE spectral line that not clear interference source introduces, if do not carried out to these periodic interfering frequency compositions Processing just directly calculates vibration level, necessarily causes result of calculation bigger than normal, influences the accuracy judged, and gently then repeats to test, weight Then qualitative is defective work, greatly increases development cost.In order to eliminate PERIODIC INTERFERENCE spectral line, people construct a variety of pectination filters Ripple device, comb filter extensive application at present is in mechanical fault diagnosis.It is published within 2010 periodical coal mine machinery volume 31 the 2nd The paper of phase《Improve application of the pectination envelope-demodulation method in rolling bearing fault diagnosis》To conventional comb filter structure The method of making is reviewed, it is indicated that Morlet small echos are " clock " shape curve that top magnitude is not 1 under frequency domain, thus are being filtered When can produce signal amplitude change.Such non-flat-top feature says the information for causing can not accurately to extract signal spectrum, filtering Effect is poor, so as to the popularization of method for limiting.And by the wave filter of Fourier Kernels, there is " box well under frequency domain Shape " feature, on the basis of this characteristic, it is possible to by the passband bandpass filter and comb band-pass filter constructed, Directly effectively design the pectinate band-stop filter device that can filter out PERIODIC INTERFERENCE spectral line.
The content of the invention
In order to overcome the shortcomings of the method poor practicability of existing elimination pectination spectral noise.The present invention provides a kind of elimination pectination The method that spectral noise influences on vibration level.Distribution feelings of this method first by observing pectinate noise in original signal spectrum Condition, associated technical parameters are determined, then using Fourier kernel functions as bandpass filter corresponding to basic function construction.It is basic herein On, the pectinate band-stop filter device of requirement is met by adding frequency-shifting operator.Then frequency domain phase is equivalent to according to convolution The principle multiplied, the power spectrum of primary signal and the pectinate band-stop filter device constructed are combined, original letter is filtered out under frequency domain Periodic interference components in number.According to filtered spectral magnitude, power spectral density is calculated, is carried with reference to international standard value Data are divided frequency range by the center frequency value of confession, and finally calculating third-octave by frequency range using power spectral density value composes, and with This be according to judge disappear it is dry after signal vibration level it is whether up to standard, it is practical.
The technical solution adopted for the present invention to solve the technical problems is：One kind eliminates pectination spectral noise to vibration level shadow Loud method, it is characterized in comprising the following steps：
Step 1, if primary signal is x (t), 0 ＜ t ＜ T.Fourier's change is carried out to x (t) within the period [0, T] first Change, obtain corresponding frequency domain distribution curve X (ω),
Step 2, observe in frequency spectrum and whether there is pectinate noise spectral profile.If in the presence of determining week according to spectrum distribution The centre frequency and interval width of phase property noise spectral line, determine the centre frequency f of wave filtercWith the bandwidth B of wave filter；
Step 3, the flat-top bandpass filter s (t) that amplitude is 1 is constructed by basic function of Fourier kernel functions,
Wherein, ω0>=0 represents the cut-off frequency of institute's construction bandpass filter.
Frequency domain distribution characteristic is corresponding to s (t)
Filtered by setting frequency-shifting operator to construct the pectination based on Fourier basic functions being made up of multiple bandpass filters Ripple device；
It is f to construction fundamental frequencyb, with a width of 2 ω0Comb filter hb(t), then by setting frequency displacement because obtaining：
hh(t) can only be f by frequencybAnd its comb band-pass filter of frequency multiplication.Frequency faIn the range of the passband that constructs Bandpass filter ha(t),
Obtained h will be constructedaAnd h (t)b(t), Fourier is carried out to convert to obtain Ha(ω) and Hb(ω).According to Fourier Knowable to the characteristic of conversion, comb filter Hc(ω)=Ha(ω)-Hb(ω) will be only to fbAnd its frequency multiplication produces band-stop response, and There is bandpass characteristics to remaining frequency content.
Step 4, the principle of frequency domain multiplication is equivalent to according to convolution, the signal spectrum W (ω) after comb filtering is by W (ω)=X (ω) × Hc(ω) is calculated.
Step 5, power spectral density G (ω) is calculated according to frequency spectrum W (ω)
Thus obtain, with fcCentered on frequency h-th of frequency range [fl,fu] in, the average value of power spectrum data is
In formula, n is spectral line number in frequency range.Centre frequency fcInternational Organization for standardization ISO recommendation is derived from, in frequency range Lower limit flAnd fuFollowing relation should be met
If with a0For zero shellfish a reference value, then ahEnergy level be expressed as
Lh=20log (ah/a0) (9)。
The beneficial effects of the invention are as follows：First by observing the distribution situation of pectinate noise in original signal spectrum, it is determined that Associated technical parameters, then using Fourier kernel functions as bandpass filter corresponding to basic function construction.On this basis, pass through Addition frequency-shifting operator is met the pectinate band-stop filter device of requirement.The original of frequency domain multiplication is then equivalent to according to convolution Reason, the power spectrum of primary signal and the pectinate band-stop filter device constructed are combined, filtered out under frequency domain in primary signal Periodic interference components.According to filtered spectral magnitude, power spectral density is calculated, in being provided with reference to international standard value Data are divided frequency range by frequency of heart value, finally using power spectral density value by frequency range calculate third-octave spectrum, and as according to According to judge disappear it is dry after signal vibration level it is whether up to standard.This method can filter out the interference signal being spacedly distributed, and then in frequency Vibration level is directly calculated under domain, obtains the indeed vibrations energy level of subject body.Pass through vibrational energy caused by modifying factor interference noise Level wrong report, the repetition experiment brought by vibration wrong report is avoided, is advantageous to shorten the test period, it is practical.
The present invention is elaborated with reference to the accompanying drawings and detailed description.
Brief description of the drawings
Fig. 1 is the flow chart that the present invention eliminates the method that pectination spectral noise influences on vibration level.
Fig. 2 (a) is the analog signal frequency spectrum profile of the inventive method before processing.
Fig. 2 (b) is pectinate band-stop filter device frequency spectrum profile.
Fig. 2 (c) is the analog signal spectrogram after filtering process.
Fig. 2 (d) is the octave comparison diagram before and after analog signal comb filtering.
Fig. 3 (a) is the original vibration signal time domain waveform and spectrogram before filtering process.
Fig. 3 (b) is the power density spectrum comparison diagram before and after original vibration signal comb filtering.
Fig. 3 (c) is primary signal octave and the octave comparison diagram of the signal after comb filtering.
Embodiment
Reference picture 1-3 (c).The method that the present invention eliminates pectination spectral noise and influenceed on vibration level comprises the following steps that：
Embodiment 1：
This example is to emulate obtained data containing noise at equal intervals according to vibration data and performance data, by institute State and be filtered processing containing the pectination spectral noise at intervals of 10Hz, and calculate the vibration level after eliminating interference.The simulation is shaken Dynamic data sampling frequency is 100Hz, sampling number 2048.Third-octave spectrum centre frequency f usedcDistribution is as follows,
fc=[1,1.25,1.6,2,2.5,3.15,4,5,6.3,8,10,12.5,16,20,25,31.5,40,50]
Comprise the following steps that：
Step 1, to emulating signal x (t), according to
Carry out FFT and whether spectrum analysis, the regularity of distribution for observing abnormal spectral line meet the characteristic being spacedly distributed, according to The composition of analog signal can be found that its spectrogram is furnished with equally spaced spectral line in 10,20,30,40Hz punishment, and amplitude is long-range In the value of remaining frequency range.It is considered that pectinate noise in signal be present, it is necessary to which noise filter is carried out to it by comb filter Remove.
Step 2, spectral line frequency and interval width determine bandpass filter according to corresponding to comb spectrum in analog signal frequency spectrum Centre frequency fc=10,20,30,40Hz, bandwidth B=2Hz, based on Fourier basic functions, construction amplitude be 1 it is flat Heading tape bandpass filter, may be defined as
Wherein cut-off frequency ω0>=0, the ω in this analog signal0=2 π.Because s (t) is the symmetrical of independent variable zero crossing Function, if directly taking t=0 to calculate, correct numerical solution can not be obtained, and then limit side can be used in clear and definite proposition in the present invention Method solves, i.e.,
Step 3, on the basis of the bandpass filter that step 2 designs, setting 10,20,30,40Hz frequency-shifting operator exp (j2 π k10t) (j=1,2,3,4), construct the pectination based on Fourier basic functions corresponding with core frequency at equal intervals Bandpass filter, wherein containing four passbands, i.e.,
Now designed wave filter hb(t) can only by width near 10,20,30,40Hz be 2Hz comb spectrum.For Filter out these frequency band signals and retain remaining signal, can be constructed in 0~50Hz analysis frequency range a width 0~ 46Hz passband bandpass filter
Next, to haAnd h (t)b(t) according to
FFT is carried out, finally gives respective spectrum distribution Ha(ω) and Hb(ω)。
According to Fourier convert characteristic, by under frequency domain according to
Hc(ω)=Ha(ω)-Hb(ω) (8)
It can be obtained by the pectinate band-stop filter device under frequency domain.
Step 4, the spectrum H of pectinate band-stop filter device is being obtainedcAfter (ω), frequency domain is equivalent to according to convolution and is multiplied Principle, filtering can complete directly under frequency domain, i.e.,：
W (ω)=X (ω) × Hc(ω) (9)
So as to filtered out the signal spectrum W (ω) after noise, reference picture 2 (b) can be seen that, periodicity spectral line obtains bright It is aobvious to suppress.
Step 5, according to power spectral density calculation formula,Obtain filtering in duration 20.47s The length of signal is 1024 power spectral density value G after ripple2Sequence,
G2=[6.2e-6,1.7e-5 ..., 5.5e-11,1.3e-10 ..., 2.0e-10] (10)
It is similar therewith, using same procedure can obtain primary signal length be 1024 power spectral density value G1Sequence Row,
G1=[6.2e-6,1.7e-5 ..., 1.4e-3,1.9e-3 ..., 2.0e-10] (11)
Then in 1~40Hz analysis frequency, average decibel value of two groups of power spectral density values in each centre frequency section RespectivelyWithWherein n is the spectral line number in each centre frequency section, is specifically distributed as
N=[5,7,8,11,13,17,21,27,34,41,51,71,82,102,133,174] (12)
If with a0=3*10-4For zero shellfish a reference value, then primary signal L1With filtered signal L2Vibration level be
L1=20log (ah1/a0) (13)
L2=20log (ah2/a0) (14)
By the way that the vibration level sequential value of two groups of signals is contrasted one by one, you can verify that filtered signal vibration has been dropped It is low.It can be clearly seen that in the 3rd centre frequency section reciprocal, vibration level has obtained significant suppression from Fig. 2 (d).
Embodiment 2：
The power test of power set is to completely examine the maximally effective research technique of power set performance by land.Once certain Performance assessment criteria is not up to standard, gently then reforms experiment, increases cost, delay progress, it is heavy then it is qualitative be defective work, caused to factory Heavy economic losses.Vibration is limited, vibration is surveyed as one of most important performance assessment criteria of power test by on-the-spot test condition Various interference are inevitably included in amount data.
In certain field test, the continuous power test evaluation point vibration signal several times wherein in a period of time all includes At intervals of 100Hz PERIODIC INTERFERENCE.Propulsion machine motion-activated frequency is equal to the fraction times or integer that engine turns frequency Times, when engine is when turning frequency change, the mechanical movement driving frequency of power set necessarily changes therewith, therefore, judges to assess The spike cluster of point rumble spectrum is not as caused by the mechanical movement excitation of tool component, and comes from external disturbance Source.In order to avoid repeating to test, testing expenses are saved, it is necessary to which the influence to interference is assessed.
Step 1, original signal strength 128010, sample frequency 25600Hz.FFT is carried out to it first, obtains frequency domain spectra Line X (ω), i.e.,：
Wherein durationAbout 5s.By observation it can be found that the pectination in problem disturbs spectral line to concentrate out In the range of present 4500Hz~6100Hz, about 100Hz spectral line at equal intervals be present, thus will mainly be adopted for this segment data Take filtering process and analyzed.
Step 2, with reference to spectral profile, each centre frequency f of comb filter is determinedc=4547Hz, 4647Hz ..., 6047Hz, bandwidth B=20Hz.The band logical for meeting to require is constructed using Fourier basic functions on this basis Wave filterWherein filter cutoff frequency ω0=5*2 π, then s (0)=10 is obtained according to limit algorithm.
Step 3, on the basis of the bandpass filter constructed, by frequency-shifting operator exp (the j2 π for adding 100Hz K100t), it can obtain the bandpass filter corresponding with centre frequency
N is the pectination spectral line number in 4500Hz~6100Hz frequency ranges in formula.
In order to further obtain trying to achieve pectinate band-stop filter device, one is constructed in 4500~6100Hz analysis frequency range Passband bandpass filter of the individual width in 4500~6100Hz
Then still converted using Fourier, obtain the frequency domain distribution of two wave filters
Pass through Hc(ω)=Ha(ω)-Hb(ω) relation, it is possible to be met the bandstop filter H of requirementcThe frequency of (ω) Domain representation form.
Step 4, H is being obtainedcAfter (ω), the principle of frequency domain multiplication is equivalent to according to convolution, filtering can lead to The multiplication realization crossed under frequency domain,
W (ω)=X (ω) × Hc(ω) (6)
After the frequency domain distribution W (ω) after having obtained filtering out interference, so that it may further calculate the work(of corresponding frequency spectrum Rate spectral density valueAnd by its with non-filtered original signal power spectral density value in Fig. 3 (b) Contrasted.
Step 5, finally according to power spectral density value and third-octave frequency domain algorithm
L=20log (ah/a0) (8)
The vibration level of filtered signal is calculated, and it is being schemed with the vibration level of non-filtered primary signal Contrasted in 3 (c).
By carrying out finding after effectively analyzing to institute's gathered data, eliminating the vibration level calculated after PERIODIC INTERFERENCE expires The corresponding standard of foot, avoids and reforms experiment, saved testing expenses, shorten the test period, while has proved what this was invented Validity.

Claims (1)

  1. A kind of 1. method for eliminating pectination spectral noise and being influenceed on vibration level, it is characterised in that comprise the following steps：
    Step 1, if primary signal is x (t), 0 ＜ t ＜ T；Fourier transformation is carried out to x (t) within the period [0, T] first, obtained To corresponding frequency domain distribution curve X (ω),
    <mrow> <mi>X</mi> <mrow> <mo>(</mo> <mi>&amp;omega;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msubsup> <mo>&amp;Integral;</mo> <mn>0</mn> <mi>T</mi> </msubsup> <mi>x</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <msup> <mi>e</mi> <mrow> <mo>-</mo> <mi>j</mi> <mi>&amp;omega;</mi> <mi>t</mi> </mrow> </msup> <mi>d</mi> <mi>t</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>
    Step 2, observe in frequency spectrum and whether there is pectinate noise spectral profile；If in the presence of, according to spectrum distribution determine periodically The centre frequency and interval width of noise spectral line, determine the centre frequency f of wave filtercWith the bandwidth B of wave filter；
    Step 3, the flat-top bandpass filter s (t) that amplitude is 1 is constructed by basic function of Fourier kernel functions,
    <mrow> <mi>s</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <mi>sin&amp;omega;</mi> <mn>0</mn> </msub> <mi>t</mi> </mrow> <mrow> <mi>&amp;pi;</mi> <mi>t</mi> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow>
    Wherein, ω0>=0 represents the cut-off frequency of institute's construction bandpass filter；
    Frequency domain distribution characteristic is corresponding to s (t)
    <mrow> <mi>S</mi> <mrow> <mo>(</mo> <mi>&amp;omega;</mi> <mo>)</mo> </mrow> <mo>=</mo> <mo>{</mo> <mrow> <mtable> <mtr> <mtd> <mrow> <mn>1</mn> <mo>,</mo> <mo>|</mo> <mi>&amp;omega;</mi> <mo>|</mo> <mo>&amp;le;</mo> <msub> <mi>&amp;omega;</mi> <mn>0</mn> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mn>0</mn> <mo>,</mo> <mi>&amp;omega;</mi> <mo>&gt;</mo> <msub> <mi>&amp;omega;</mi> <mn>0</mn> </msub> <mo>,</mo> <mi>&amp;omega;</mi> <mo>&lt;</mo> <mo>-</mo> <msub> <mi>&amp;omega;</mi> <mn>0</mn> </msub> </mrow> </mtd> </mtr> </mtable> <mo>-</mo> <mo>-</mo> <mo>-</mo> </mrow> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow>
    By setting frequency-shifting operator to construct the comb filtering based on Fourier basic functions being made up of multiple bandpass filters Device；
    It is f to construction fundamental frequencyb, with a width of 2 ω0Comb filter hb(t), then by setting frequency-shifting operator to obtain：
    <mrow> <msub> <mi>h</mi> <mi>b</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <mi>sin&amp;omega;</mi> <mn>0</mn> </msub> <mi>t</mi> </mrow> <mrow> <mi>&amp;pi;</mi> <mi>t</mi> </mrow> </mfrac> <mo>&amp;times;</mo> <msubsup> <mi>&amp;Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </msubsup> <mi>exp</mi> <mrow> <mo>(</mo> <mi>j</mi> <mn>2</mn> <msub> <mi>&amp;pi;kf</mi> <mi>b</mi> </msub> <mi>t</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow>
    hb(t) can only be f by frequencybAnd its comb band-pass filter of frequency multiplication；Frequency faIn the range of construct passband band logical filter Ripple device ha(t),
    <mrow> <msub> <mi>h</mi> <mi>a</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <mi>s</mi> <mi>i</mi> <mi>n</mi> <mn>2</mn> <mi>&amp;pi;</mi> <mrow> <mo>(</mo> <msub> <mi>f</mi> <mi>a</mi> </msub> <mo>/</mo> <mn>2</mn> <mo>)</mo> </mrow> <mi>t</mi> </mrow> <mrow> <mi>&amp;pi;</mi> <mi>t</mi> </mrow> </mfrac> <mo>&amp;times;</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mi>j</mi> <mn>2</mn> <msub> <mi>&amp;pi;kf</mi> <mi>a</mi> </msub> <mi>t</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow>
    Obtained h will be constructedaAnd h (t)b(t), Fourier is carried out to convert to obtain H2(ω) and Hb(ω)；Converted according to Fourier Characteristic understand, comb filter Hc(ω)=H2(ω)-Hb(ω) will be only to fbAnd its frequency multiplication produces band-stop response, and to it Remaining frequency content has bandpass characteristics；
    Step 4, the principle of frequency domain multiplication is equivalent to according to convolution, the signal spectrum W (ω) after comb filtering by W (ω)= X(ω)×Hc(ω) is calculated；
    Step 5, power spectral density G (ω) is calculated according to frequency spectrum W (ω)
    <mrow> <mi>G</mi> <mrow> <mo>(</mo> <mi>&amp;omega;</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>2</mn> <mi>T</mi> </mfrac> <mo>|</mo> <mi>W</mi> <mrow> <mo>(</mo> <mi>&amp;omega;</mi> <mo>)</mo> </mrow> <msup> <mo>|</mo> <mn>2</mn> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow>
    Thus obtain, with fcCentered on frequency h-th of frequency range [fl,fu] in, the average value of power spectrum data is
    <mrow> <msub> <mi>a</mi> <mi>h</mi> </msub> <mo>=</mo> <mfrac> <mrow> <mi>G</mi> <mrow> <mo>(</mo> <mi>&amp;omega;</mi> <mo>)</mo> </mrow> </mrow> <mi>n</mi> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>
    In formula, n is spectral line number in frequency range；Centre frequency fcIt is derived from International Organization for standardization ISO recommendation, frequency range bound flAnd fuFollowing relation should be met
    <mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>f</mi> <mi>c</mi> </msub> <mo>=</mo> <msqrt> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>&amp;CenterDot;</mo> <msub> <mi>f</mi> <mi>l</mi> </msub> </mrow> </msqrt> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>/</mo> <msub> <mi>f</mi> <mi>l</mi> </msub> <mo>=</mo> <msup> <mn>2</mn> <mrow> <mn>1</mn> <mo>/</mo> <mn>3</mn> </mrow> </msup> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>8</mn> <mo>)</mo> </mrow> </mrow>
    If with a0For zero shellfish a reference value, then ahEnergy level be expressed as
    Lh=20log (ah/a0) (9)。
CN201510057963.XA 2014-10-08 2015-02-04 Eliminate the method that pectination spectral noise influences on vibration level Active CN104599676B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510057963.XA CN104599676B (en) 2014-10-08 2015-02-04 Eliminate the method that pectination spectral noise influences on vibration level

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
CN201410525500 2014-10-08
CN2014105255007 2014-10-08
CN201510057963.XA CN104599676B (en) 2014-10-08 2015-02-04 Eliminate the method that pectination spectral noise influences on vibration level

Publications (2)

Publication Number Publication Date
CN104599676A CN104599676A (en) 2015-05-06
CN104599676B true CN104599676B (en) 2017-12-12

Family

ID=53125409

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510057963.XA Active CN104599676B (en) 2014-10-08 2015-02-04 Eliminate the method that pectination spectral noise influences on vibration level

Country Status (1)

Country Link
CN (1) CN104599676B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109716664B (en) * 2016-10-10 2020-06-16 华为技术有限公司 Comb filtering noise elimination method and device and frequency domain self-adaptive equalization device

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5903615A (en) * 1998-03-30 1999-05-11 3Com Corporation Low complexity frequency estimator and interference cancellation method and device
CN1222994A (en) * 1996-03-13 1999-07-14 艾利森公司 Noise suppressor circuit and associated method for suppressing periodic interference component portions of communication signal
CN1552146A (en) * 2001-09-28 2004-12-01 Apparatus and method for inhibiting periodic interference signals
CN101256094A (en) * 2007-11-30 2008-09-03 西安交通大学 Signal noise-extinguishing method for extracting special periodic signal component
CN102759448A (en) * 2012-06-26 2012-10-31 西安瑞特快速制造工程研究有限公司 Gearbox fault detection method based on flexible time-domain averaging
CN104133982A (en) * 2014-06-27 2014-11-05 中天启明石油技术有限公司 Elimination method of slurry pulse signal pump stroke noise

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1222994A (en) * 1996-03-13 1999-07-14 艾利森公司 Noise suppressor circuit and associated method for suppressing periodic interference component portions of communication signal
US5903615A (en) * 1998-03-30 1999-05-11 3Com Corporation Low complexity frequency estimator and interference cancellation method and device
CN1552146A (en) * 2001-09-28 2004-12-01 Apparatus and method for inhibiting periodic interference signals
CN101256094A (en) * 2007-11-30 2008-09-03 西安交通大学 Signal noise-extinguishing method for extracting special periodic signal component
CN102759448A (en) * 2012-06-26 2012-10-31 西安瑞特快速制造工程研究有限公司 Gearbox fault detection method based on flexible time-domain averaging
CN104133982A (en) * 2014-06-27 2014-11-05 中天启明石油技术有限公司 Elimination method of slurry pulse signal pump stroke noise

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
一种多测点变工况倍频程快速处理方法;孙涛等;《鱼雷技术》;20111231;第19卷(第6期);455-458 *
梳状滤波器在滚动轴承早期故障诊断中的应用;赵俊龙等;《振动与冲击》;20081231;第27卷(第12期);171-174 *

Also Published As

Publication number Publication date
CN104599676A (en) 2015-05-06

Similar Documents

Publication Publication Date Title
CN107505135B (en) Rolling bearing composite fault extraction method and system
CN107356432B (en) Fault Diagnosis of Roller Bearings based on frequency domain window experience small echo resonance and demodulation
CN104297112B (en) The measuring method of a kind of wet-steam region droplet particles and device
EP2085902B1 (en) Order tracking method and system
CN104375111B (en) The method that quick high accuracy refinement correction is carried out to intensive spectrum
Zhang et al. Energy operator demodulating of optimal resonance components for the compound faults diagnosis of gearboxes
CN108764073B (en) Acceleration noise filtering and integrating method combined with spectrum energy form fitting
CN102980945A (en) Frequency mixing nonlinear ultrasonic detection method for structure microcracks based on bispectrum analysis
CN104931197A (en) EEMD-based automatic balancing machine vibration signal processing method
CN107957566A (en) Magnetic resonance depth measurement method for extracting signal based on frequency selection singular spectrum analysis
CN104820786A (en) Method for analyzing instantly weighted synchronous extrusion wavelet bispectrum
CN105865793B (en) A kind of method for improving more rotor vibration monitoring of aero-engine precision
CN109374119A (en) Transformer vibration signal Characteristic Extraction method
CN109765052B (en) Planetary gearbox early fault diagnosis method based on GOA-ASR
CN105092249A (en) Rolling bearing fault diagnosis method based on Gabor filter
CN104729677B (en) A kind of time-domain digital weighted method of nonstationary noise signal
CN104599676B (en) Eliminate the method that pectination spectral noise influences on vibration level
CN111736222A (en) Single-shot data signal-to-noise ratio determining method and device
CN111487318A (en) Time-varying structure instantaneous frequency extraction method
Matania et al. Algorithms for spectrum background estimation of non-stationary signals
Ong et al. An inconsistent phase selection assessment for harmonic peaks elimination in operational modal testing
CN103926097A (en) Method for collecting and extracting fault feature information of low-speed and heavy-load device
CN211293127U (en) Partial discharge measurement system under harmonic voltage effect
CN114065562A (en) EMD-based structural part fatigue simulation analysis method, device, terminal and medium
US20140028283A1 (en) Method of extracting zero crossing data from full spectrum signals

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant