CN100561162C - A kind of virtual oscillating table detection signal processing method and equipment thereof - Google Patents

A kind of virtual oscillating table detection signal processing method and equipment thereof Download PDF

Info

Publication number
CN100561162C
CN100561162C CNB2007100292809A CN200710029280A CN100561162C CN 100561162 C CN100561162 C CN 100561162C CN B2007100292809 A CNB2007100292809 A CN B2007100292809A CN 200710029280 A CN200710029280 A CN 200710029280A CN 100561162 C CN100561162 C CN 100561162C
Authority
CN
China
Prior art keywords
signal
frequency
equipment
sigma
computing machine
Prior art date
Application number
CNB2007100292809A
Other languages
Chinese (zh)
Other versions
CN101113936A (en
Inventor
周伦彬
黄吉淘
Original Assignee
广州市计量检测技术研究院
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 广州市计量检测技术研究院 filed Critical 广州市计量检测技术研究院
Priority to CNB2007100292809A priority Critical patent/CN100561162C/en
Publication of CN101113936A publication Critical patent/CN101113936A/en
Application granted granted Critical
Publication of CN100561162C publication Critical patent/CN100561162C/en

Links

Abstract

The present invention relates to shaking table checkout equipment field, purpose is to overcome defective of the prior art, designing a kind of signal processing method makes virtual instrument can be applied to metering field, and design a kind of equipment that is applicable to metering on this basis, say so more specifically a kind of virtual oscillating table detection signal processing method and equipment thereof accordingly.Its method may further comprise the steps: 1. adopt the amount of acceleration of shaking table by sensor, and convert electric signal to; 2. with signal condition equipment signal is nursed one's health; 3. the signal after the conditioning is delivered in the computing machine by the data collecting card collection; 4. by computing machine signal is carried out pre-service; 5. pretreated signal is carried out discrete Fourier transform (DFT) and digital signal frequency domain filtering; 6. at last by an integration output speed component, twice integration exported displacement component.Its structure comprises acceleration transducer, signal condition equipment, data collecting card and computing machine.

Description

A kind of virtual oscillating table detection signal processing method and equipment thereof

Technical field

The present invention relates to shaking table checkout equipment field, say so more specifically a kind of virtual oscillating table detection signal processing method and equipment thereof.

Technical background

In parameter measurement in the past, because the parameter distributions many, measure field of measuring are wide, therefore need multiple separate type surveying instrument such as charge amplifier, voltage table, frequency meter, distortion meter etc., be with great inconvenience to measurement.Adopted digitized frequency meter, dynamic signal analyzer afterwards, but these instruments mainly depend on import, the processing technology complexity manufacture level is required height, and it is higher to buy the used expense of these instruments, and the passage of instrument is few, and use is restricted.That application is maximum at present mainly is B﹠amp; The PULSE vibration test system of K, though it is many that it has function, the analysis precision advantages of higher, but the price that it is expensive, add English operation interface and limited popularization and the use at home of this quasi-instrument, and still adopting artificial calculation mode for some test event, automaticity is not high.

So-called virtual instrument (Virtual Instrument, be called for short VI), be exactly the user on universal computer platform, the test function that designs an apparatus of definition makes the user the same at his self-designed testing tool of operation according to demand.The appearance of virtual instrument, broken traditional instrument producer's definition has been arranged, the mode of operation that the user can't change, make that the user can be according to the demand of oneself, design oneself instrument system, in test macro and instruments design, use software replace hardware as far as possible, make full use of the function that computer technology realized and expanded legacy test system and instrument." software is exactly instrument " is that virtual instrument concept is the simplest, the most essential summary.

Metering is defined as " realizing that unit unification, value are movable accurately and reliably ", and it has accuracy, unitarity, traceability, legal system characteristics.The metering activity can be regarded the test process of verifying character as, and purpose is that the test data of guaranteeing instrument has necessary accuracy and reliability, to satisfy the demand of test.

For traditional instrument, its metrology and measurement is relatively easy, and metrological personnel can be carried out concrete gage work according to the measurement technology rules of country, department or local promulgation.And in national measure base standard device and metering testing system based on virtual instrument technique, although use a computer, application software and instrument moduleization can increase substantially integrated, the robotization and the intelligent degree of device and system, meanwhile brought following problem.

The accuracy or the uncertainty of measurement of each instrument of conventional metered dose standard set-up can be determined by periodic calibration, and by analyzing all kinds of error sources the size of measured uncertainty contribution be assessed.For using the very high measuring system of the integrated degree of virtual instrument technique, the accuracy of data acquisition is the basis of virtual test system, and the accuracy of software module is the soul of virtual test system.Bigger to the accuracy influence of system for hardware module in A/D conversion link that is sent to computing machine by sensor and signal condition link; And the uncertain component that the data algorithm that uses in the virtual instrument software programming brings and any one little careless mistake of appearance all may exert an influence to calibration result, therefore be difficult to assessment and measure the size of each error source, so can't carry out the assessment of uncertainty more according to the conventional method to tested value influence.

Summary of the invention

The objective of the invention is to overcome defective of the prior art, design a kind of signal processing method and make virtual instrument can be applied to metering field, and design a kind of equipment that is applicable to metering accordingly on this basis.

The present invention is achieved through the following technical solutions its purpose.

The present invention has at first designed a kind of virtual oscillating table detection signal processing method, it is characterized in that may further comprise the steps:

1. adopt the amount of acceleration of shaking table by sensor, and convert electric signal to;

2. with signal condition equipment signal is nursed one's health;

3. the signal after the conditioning is delivered in the computing machine by the data collecting card collection;

4. by computing machine signal is carried out pre-service;

5. pretreated signal is carried out discrete Fourier transform (DFT) and digital signal frequency domain filtering;

6. at last by an integration output speed component, twice integration exported displacement component.

Wherein step signal condition process 2. by charge amplifier and anti-aliasing filtering module to signal amplify, isolation, filtering, excitation and linearization.

The signal great majority that obtain from sensor will just can enter data acquisition equipment through conditioning.The method of signal condition comprises amplification, isolation, filtering, excitation, linearization etc.Because different sensors has different characteristics, therefore also will be except these functions according to concrete sensor characteristic and the conditioning functions that requires to select distinctive signal.Introduce respectively below.

1) amplifies

All to amplify improving resolution feeble signal, and the voltage of signals scope after the conditioning is mated with the voltage range of analog to digital converter (ADC).The signal condition module makes signal not be exaggerated before the neighbourhood noise influence that is subjected to transmission signals, thereby obtains good signal-to-noise as much as possible near signal source or sensor.Though also can amplify signal in data acquisition equipment, it also can amplify noise, is unfavorable for the test of back like this, careful usefulness.

2) isolate

When measured signal contained high voltage peak, it may damage computing machine or injury operator, just computing machine and sensor need be isolated for the purpose of safety.The Another reason of isolating is to guarantee that data acquisition equipment is not subjected to the influence of ground potential difference.When data acquisition equipment and signal are not with reference to same physical features point, just circulation over the ground may take place, influence the accuracy of measurement.If the ground potential difference of signal ground and data acquisition equipment is very big, even may damage test macro.Use the isolation module of signal conditioner just can reduce circulation over the ground, guarantee the accuracy of testing.

3) filtering

The purpose of filtering is to remove unwanted composition from measured signal.Most of signal condition modules have low-pass filter, are used for filtering noise.Usually also need frequency overlapped-resistable filter, the signal of all frequencies in the filtered signal more than the interested highest frequency.Some high performance data collecting card self has frequency overlapped-resistable filter.

4) excitation

Signal condition also can need extraneous power supply or current excitation signal such as strain transducer, ICP sensor etc. for some sensor provides required pumping signal.A lot of signal condition modules all provide current source and voltage source so that provide excitation to sensor.

5) linearization

Many sensors are non-linear to measured response, thereby need carry out linearization to its output signal, the error of bringing with compensation sensor.But present trend is that data acquisition system (DAS) can utilize software to solve this problem.

In sum, even sensor is directly exported digital signal, necessity of nursing one's health is arranged also sometimes.Its effect is that the digital signal of sensor output is carried out necessary shaping or level adjustment.Most of digital signal conditioning module also provide some other circuit module, make the user can pass through external units such as the direct control electromagnetic valve of data acquisition card digital I/O, electric light, motor.

In concrete the application, to select the signal condition device type according to the output type of sensor and to the requirement of signal, what sensor came out in the system that the present invention makes up is charge signal, the handy charge amplifier of conditioning device, for fear of the appearance of signal aliasing, conditioning device should contain the anti-aliasing filtering module simultaneously.

Vibration signal pretreatment is the most basic data processing mode that the data of will gather in the vibration-testing are reduced into the actual vibration situation as far as possible really.In vibration-testing, the data that obtain through data acquisition have plenty of digital voltage value, are the integer number amount of unit with resolution and major part provides, therefore at first to carry out engineering calibrating, make it to become digital signal data with respective physical amount unit to these digital quantities.In addition, because the existence of various interference makes the data of test macro collection depart from actual value, therefore eliminate the important content that this deviation also is a vibration signal processing.The present invention at first demarcates conversion with signal, eliminates the polynomial trend item by least square method again, adopts 53 moving average methods or the frequency domain method of average to carry out signal smoothing at last and handles.

The demarcation conversion of general vibration signal is divided into following two steps: calculating voltage amount at first.For the integer digital data, the resolution that multiply by collector is quantization unit, data conversion is become voltage signal, for example for the 16 bit data collectors of input voltage range ± 10v, its full scale voltage is 20v, can obtain resolution with 20 divided by 2 16 powers is 0.000305175V, and it is the vibration signal data of unit that each the shaping data that multiply by the vibration signal that collects with this resolution value respectively can obtain with voltage.Next will carry out the demarcation conversion of physical unit.Usually vibration transducer has charge type and voltage-type, but no matter adopt the sensor of the sort of type, the conditioning of signal all is necessary, different only is the content difference of conditioning, for the charge signal employing is that charge amplifier is nursed one's health, can then nurse one's health for voltage signal with voltage amplifier, because this quasi-instrument has comprised the normalization (sensitivity of sensor is write in the hardware) of hardware in design at present, so its calibration coefficient is exactly the inverse of corresponding yield value, only need to record to multiply by the vibration value that corresponding calibration coefficient can be tried to achieve reality then through the magnitude of voltage after the conditioning.

Vibration signal produces zero point drift, the instability of the outer low frequency signal of sensor frequency scope and the environmental interference around the sensor because amplifier varies with temperature, and tends to depart from baseline, even the size that departs from baseline also can change in time.Depart from the trend term that the time dependent whole process of baseline is called as signal.Trend term directly has influence on the correctness of signal should be with its place to go, and the method for elimination trend term commonly used is the polynomial expression least square method, below simply introduces the principle of this method.

The sampled data of vibration signals measured is { x k(k=1,2,3... n), because sampled data is a constant duration, for simplifying therebetween, makes sampling interval Δ t=1, establishes a polynomial function:

x Λ k = a 0 + a 1 k + a 2 k 2 + a 3 k 3 + . . . + a m k · m (k=1,2,3...,n)

Determine function Each undetermined coefficient a i(i=0,1,2...m), make function With discrete data x kThe error sum of squares minimum promptly: E = Σ k = 1 n ( x Λ k - x k ) 2 = Σ k = 1 n ( Σ i = 0 m a i k i - x k ) 2

The condition that satisfies extreme value is ∂ E ∂ a i = 2 ( Σ k = 1 n k j ( Σ i = 0 m a i k i - x k ) = 0 (j=0,1,...,m)

Get E successively to a iAsk local derviation, can produce a m+1 unit system of linear equations:

Σ k = 1 n ( Σ i = 0 m a i k i - x k ) = 0 (i=0,1,2,...m)

The group of solving an equation.Obtain m+1 undetermined coefficient a i(i=0,1,2...m). top various in, multinomial the make order of m for setting, the scope i ∈ of its value [0, m].

The trend term of trying to achieve when m=0 is a constant, has

Σ k = 1 n a 0 k 0 - Σ k = 1 n x k k 0 = 0

Solve an equation,

a 0 = 1 n Σ k = 1 m x k

As can be seen, the trend term when m=0 is the arithmetic mean of signal sample data, and elimination constant trend term gets computing formula and is y k = x k - x Λ k = x k - a 0 (k=1,2...n)

When m=1, be the linear trend item, have

Σ k = 1 n a 0 k 0 + Σ k = 1 n a 1 k 1 - Σ k = 1 n x k k 0 = 0 Σ k = 1 n a 0 k + Σ k = 1 n a 1 k 2 - Σ k = 1 n x k k = 0

The group of solving an equation,

a 0 = 2 ( 2 n + 1 ) Σ k = 1 n x k - 6 Σ k = 1 n x k k n ( n - 1 ) a 1 = 12 ( Σ k = 1 k = n x k k - 6 ( n - 1 ) Σ k = 1 n x k ) n ( n - 1 ) ( n + 1 )

The computing formula of eliminating the linear trend item is:

y k = x k - x Λ k = x k - ( a 0 - a 1 k ) (k=1,2,3...,n)

M 〉=2 o'clock are the curvilinear trend item, in the vibration signal processing of reality, get m=1~3 usually and come sampled data is carried out the processing that the polynomial trend item is eliminated.

The vibration signal that obtains by the data acquisition unit sampling often is superimposed with noise signal.Noise signal also has irregular random disturbance signal except periodic jamming signals such as power frequency that 50Hz is arranged and octave thereof.Because the frequency band broad of undesired signal immediately, the proportion that accounts for of radio-frequency component is bigger sometimes, makes the discrete data that collects plot and presents many burrs on the oscillating curve, and is very rough.In order to weaken the influence of undesired signal, improve the smoothness of oscillating curve, usually need sampled data is carried out smoothing processing.Certainly, the irregular trend item that the smoothing processing of data also can erasure signal.Thinking is to adopt moving average method that this signal is carried out repeatedly data smoothing to handle, and obtains a smooth trend term.Deduct the irregular trend term that this trend term gets final product erasure signal with raw data.Below highlight 53 moving average methods and the frequency domain method of average used herein.

5 triple smoothings are to utilize principle of least square method that discrete data is carried out the method that three least square polynomial expressions are changed frequently, and it is as follows that 5 triple smoothings get computing method:

Smoothing processing as time domain and frequency-region signal.This Processing Algorithm mainly is the high frequency random noise that reduces in the vibration signal for the effect of time domain data, then is to make smooth that spectral line becomes for the effect of frequency domain data.Certain this smoothing processing tends to make the peak value in the spectral line to reduce, and the bodily form broadens, and may cause the error of parameter recognition to increase, so level and smooth number of times is difficult for too much.

In the collection of digital signal and handling, all have in various degree by noise, as electrical noise, mechanical noise etc., pollution problems.This noise may come from test structure itself, also may be from the power supply of testing tool and environment on every side, adopt averaging can reduce The noise, should select mean type and average time according to the purpose of research and the characteristics of measured signal in actual applications.

1, Pu linear averaging

This is a most basic a kind of mean type.When adopting this mean type, FFT and other computings done one by one in the record of each given length, the spectrum value to each Frequency point waits the power linear averaging respectively promptly then

Y i = i - 1 i Y i - 1 + 1 i X i (i=1,2,…m)

X wherein iBe the record of i given length, Y iBe i since 1 progressive mean value, m is an average time, and X can represent from spectrum, cross-spectrum, effective value spectrum etc.

For the Measurement and analysis of stationary stochastic process, increase average time and can reduce relative standard deviation.For steady-state signal, get the equal weight weighting to participating in each average sample, average influence with the extraneous random noise in place to go.

2, Pu exponential average

Exponential average is different with linear averaging, and it gives bigger weighting to new subclass, and old subclass is given more little weighting, and the algorithm of exponential average is as follows:

Y i = n - 1 n Y i - 1 + 1 n X i (i=1,2,…)

X wherein iBe the record of i given length, Y iBe i since 1 progressive mean value, n represents weight, and (when i=1, n=1), n is set up on their own according to the characteristics of signal by the user usually, and we establish n=10 in the present invention, and X shows from spectrum, cross-spectrum, effective value spectrum etc.

Exponential average is generally used for the analysis of non-stationary process.Because adopt this average mode, can investigate the feature of " up-to-date " measuring-signal, again can by with " old " the deviation that on average the reducing of measured value measured or improve signal to noise ratio (S/N ratio).

3, peak value keeps

It in fact is not real average that peak value keeps, keep function will keep the maximal value of every spectral line reading in the analytic process, but the not necessarily synchronization appearance of these peak values.The algorithm that peak value keeps is as follows:

The FFT spectrum MAX X · X *

Power spectrum MAX (XX *)

X is the complex values after signal changes through FFT in the formula, X *Be its conjugate complex numerical value.

In the use of reality, linear averaging is average according to certain number of times equal weight, and average time arrives post analysis to be stopped immediately.And exponential average and peak value keep all not stipulating average time.

Signal after the pre-service is carried out filtering, and frequency domain filtering is handled and is also referred to as spectrum analysis, is that the time-frequency conversion that is based upon on the Fourier transform basis is handled.Resulting result is to be the function of variable with the frequency.It is Fourier transform (FFT) that frequency domain is handled main method, can derive many application by it, receives analysis, transport function etc. as amplitude spectrum, phase spectrum, auto-power spectrum, degree of distortion analysis, letter.Different with time-domain analysis, the frequency domain representation signal is more terse, and the analysis problem is also more deep.The frequency domain method of digital filtering is to utilize the FFT fast algorithm that the sampled data of input signal is carried out discrete Fourier transformation with analysis spectrum, requirement according to filtering, with the frequency part that needs filtering directly be arranged to zero or add the gradual transition frequency band after be arranged to zero again, for example between passband and stopband, add the transitional zone of one section cosine class window function, and then data are carried out inverse discrete Fourier transform and are recovered time-domain signal after utilizing the IFFT fast algorithm to Filtering Processing.Frequency domain method has frequency selectivity and dirigibility preferably, because the frequency characteristic of Fourier spectrum and wave filter is the relation of simply multiplying each other, its arithmetic speed is more a lot of soon than calculating time domain convolution of equal value, and produces time shift unlike time-domain filtering method.

The described step 5. input/output relation of middle digital signal frequency domain filtering is:

y ( r ) = Σ k = 0 N - 1 H ( k ) X ( k ) e j 2 πk / N

Wherein: X is the discrete Fourier transformation of input signal x, and H is the frequency response function of wave filter.

Frequency response function H is in the wave filter conduct:

During low-pass filter be H ( k ) = 1 . . . . . . . . ( kΔf ≤ f u ) 0 . . . . . . . . ( else )

During Hi-pass filter be H ( k ) = 1 . . . . . . . . ( kΔf ≥ f d ) 0 . . . . . . . . ( else )

During bandpass filter be H ( k ) = 1 . . . . . . . . ( f d ≤ kΔf ≤ f u ) 0 . . . . . . . . . ( else )

During rejection filter be H ( k ) = 1 . . . . . . . . ( kΔf ≥ f u , kΔf ≤ f d ) 0 . . . . . . . . ( else )

F wherein uFor the upper limit by frequency, f dFor lower limit by frequency, Δ f is a frequency resolution.

The characteristics of the frequency domain method of digital filtering are that method is simple, and computing velocity is fast, and filtered band control accuracy height can be used for designing any response filter that comprises many comb filter.

In the frequency domain method of digital filtering because blocking to signal, the signal limiting of an endless for the limit for length is arranged, even the signal averaging beyond the finite interval is zero, be equivalent to multiply by signal with a rectangular window, show on the frequency domain to be exactly spectrum leakage, Here it is spectrum leakage to other Frequency points.

The method that solves spectrum leakage is that elder generation estimates the amplitude and the frequency of the spectrum peak of signal before carrying out the digital signal frequency domain filtering:

Peak value place Frequency Estimation input/output relation is as follows:

freqPeak = Σ i = i - span 2 i + span 2 autoSpectrum j j × df Σ i = i - span 2 i + span 2 autoSpectrum j

Peak value place Amplitude Estimation input/output relation is as follows:

powerPeak = Σ i = i - span 2 i + span 2 a utoSpectrum i enbw

Wherein i is the index of search rate (searchFreq)

Df is a frequency interval;

Enbw is the equivalent noise bandwidth of selected window function

The function parameter meaning

Input parameter:

AutoSpecturm: one-sided power spectrum;

N: the input power spectrum array comprises the number of first number;

SearchFreq: expectation estimated frequency point, unit is Hz usually, if searchFreq smaller or equal to 0, PowerFrequencyEstimate can find the spectrum peak maximum point automatically, and estimates the frequency and the amplitude of this point according to above-mentioned formula.

WindowConstants: window function constant (structure).Wherein enbw is the equivalent noise bandwidth of selected window function.

Df: frequency interval;

Span: estimated frequency point place spectral line number.

Output:

FrePeak: the estimated frequency at the Frequency point place of trying to achieve.

PowerPeak: the estimated amplitude at calculated rate point place.

To do Fourier transform through the signal after the Filtering Processing, then the result after the conversion be carried out integral operation in frequency domain, after inverse Fourier transform obtains integrated signal.

According to the formula of inverse Fourier transform, acceleration signal can be expressed as in the Fourier transform amount of optional frequency

a(t)=Ae jwt

In the formula: a (t) is the Fourier components of acceleration signal at frequency w.

The initial velocity component is 0 o'clock, can obtain speed component to the time integral of signal for faster component, promptly

v ( t ) = ∫ 0 t a ( τ ) dτ = ∫ 0 t Ae jwτ dτ = A jw e jwt = Ve jwt

In the formula: v (t) is the Fourier components of rate signal at frequency w, and V is the coefficient of the v (t) of correspondence.

So the relational expression of an integration in frequency domain is V = A jw

Initial velocity and just the displacement component be at 0 o'clock, can obtain displacement component to twice integration of Fourier components of acceleration signal:

x ( t ) = ∫ 0 t [ ∫ 0 t a ( λ ) dλ ] dτ = ∫ 0 t Ve jwτ dτ = V jw e jwt = - A w 2 e jwt = Xe jwt

In the formula: x (t) is the Fourier components of rate signal at frequency w, and X is the coefficient of the v (t) of correspondence.

So the relational expression of twice integration in frequency domain is: X = - A w 2

In the digital signal of reality was used, the input/output relation of an integration was:

y ( r ) = Σ k = 0 N - 1 1 j 2 πkΔf H ( k ) X ( k ) e j 2 πkr N

The input/output relation of quadratic integral is:

y ( r ) = Σ k = 0 N - 1 - 1 ( 2 πkΔf ) 2 H ( k ) X ( k ) e j 2 πkr N

Wherein f lAnd f hBe respectively lower limit by frequency and the upper limit by frequency, X (k) is the Fourier transform of x (r), Δ f is the frequency resolution of signal.

Compare with the time-domain integration method, the Frequency Domain Integration method is promptly thorough simply again for the processing of the trend term of integral result, only needs to be lower than the part zero setting under the useful frequency, uses certain frequency spectrum alignment technique simultaneously, can realize the accurate Calculation of displacement.

The present invention has also designed a kind of virtual oscillating table checkout equipment, its structure comprises acceleration transducer, signal condition equipment, data collecting card and computing machine, the said equipment connects successively, described signal condition equipment and data collecting card are installed in independently and form the external hanging type data collector in the housing, are connected with computing machine by USB interface.Acceleration transducer comprises a three-dimensional acceleration transducer and at least four unidirectional acceleration transducers, described three-dimensional acceleration transducer and one of them unidirectional acceleration transducer are installed on the center of shaking table again, and at least three unidirectional acceleration transducers are rigidly attached on the shaking table in addition.

The deficiency that is faced in view of current shaking table testing tool, at the demand of market for low price, high-performance shaking table parameter detection device, the present invention has adopted advanced signal processing technology and USB interface-based Portable Data Acquisition Device, greatly reduce testing cost, the robotization that has incorporated data is simultaneously handled, under the prerequisite that guarantees measuring accuracy and reliability, improved testing efficiency.

Description of drawings

Fig. 1 is a system architecture synoptic diagram of the present invention;

Fig. 2 position system module figure of the present invention;

Fig. 3 is the synoptic diagram of acceleration transducer installation site on shaking table.

Embodiment

The present invention is described further below in conjunction with accompanying drawing.

Shaking table is a kind of device that various vibration environments are investigated the product functional reliability of simulating, and can be divided into electrodynamic type vibration table and mechanical type vibration table according to structural principle.Usually meet the requirement of national relevant art standard in order to ensure the operation conditions of shaking table, to carry out the relevant parameters test at various shaking table, the test index of its medium frequency, acceleration, speed, displacement, aspect ratio, degree of distortion, seven parameters of homogeneity is to weigh the key of shaking table quality, so the shaking table testing synthesis parameter is just at the test of above-mentioned seven parameters.

The shaking table testing synthesis parameter at first will be installed on the tested shaking table 4 by degree of will speed up sensor, specifically as shown in Figure 3: for frequency (f), acceleration (a), speed (V), displacement (D), the test of degree of distortion (S), only need a unidirectional acceleration transducer 1 to be installed, select different Frequency points to test respectively then in center position; Test request for aspect ratio is installed a three-dimensional acceleration transducer 2 in center position, choose a plurality of frequency values according to octave, and under given amplitude, vibrate, get the acceleration amplitude of three directions from the vialog successively, and calculate aspect ratio T according to following formula:

T = a x 2 + a y 2 a z × 100 %

In the formula: a x, a yThe component of two mutually perpendicular acceleration amplitudes of-vertical and the main direction of shaking, m/s 2

a z-be the acceleration amplitude of the main direction of shaking, m/s 2

Then need again to measure the acceleration amplitude of each position then successively, and to measure its uniformity coefficient N with the connection of four unidirectional acceleration transducer 3 rigidity as shown in the figure for uniformity test according to following formula.

N = | Δa | a × 100 %

In the formula: during the a-homogeneous is measured, the acceleration amplitude of central point, m/s 2

| during Δ a|-homogeneous is measured, the maximum deviation (absolute value) of each point Jia Dudu and central point acceleration, m/s 2

According to relevant technologies standard-requireds such as JJG 189-97, JJG 190-97, JJG298-2005, its maximum should be surveyed characteristic parameter and contain: vibration displacement D, vibration velocity V, vibration acceleration a, vibration frequency f, vibrational waveform degree of distortion S, vibration acceleration homogeneity N, shaking table transverse acceleration T, it is as follows to require system to satisfy the measured parameter technical indicator:

1. test frequency scope 5HZ-4KHz, require frequency measurement accuracy be better than ± 0.1%

2. testing acceleration scope 0.1-2000m/s 2, the acceleration amplitude precision is better than ± 3%, and displacement is better than ± and 5%

Require system to have vibrating data collection, analysis, storage, demonstration, printing reports function simultaneously, and be easy to carry.

Virtual instrument has constituted different architectures according to different classification forms, has passed through the development of decades, and virtual instrument has formed following several important architectures: GPIB, PXI, VXI, PC-DAQ.Do with comparison with regard to the relative merits of various architectures below, and, select a kind of suitable architecture in conjunction with requirement of the present invention.

GPIB (HP-IB or IEEE488)-universal serial port bus is the interface of computing machine and traditional instrument.The various instruments that include gpib interface can be coupled together by gpib bus, thereby can realize automatic test based on traditional instrument.Its advantage is utilized traditional instrument exactly, can reduce the expense of building of system, but, because the data capability of handling up limited (1Mb/s) of bus itself, and its brick pattern mode of building makes that test macro itself is huger, be difficult to satisfy the needs of on-the-spot test, therefore use to be subjected to certain restriction.

VXI(VME?Extension?for?Instrumentation)。Its version is with various standard modules such as signals collecting, the signal condition standard PC case of packing into, and cabinet is by inserting card or the embedded controller and the computing machine communication of computing machine.VXI observes the software specifications and the VXI alliance relevant hardware standard of VPP alliance defined, and system has very high reliability, compatibility and integrated level, but because the selling at exorbitant prices of VXI is mainly used in most advanced and sophisticated field tests such as Aeronautics and Astronautics.

PXI(PCI?Extension?for?Instrumentation)。PXI is the instrument expansion of pci bus.Its version and VXI are basic identical, and difference is that bus difference (transfer rate reaches 132Mb/s) and price are more acceptant, and present NI company tries hard to recommend the product of PXI bus, and it has broad application prospects, but price is still than higher.

The PC-DAQ architecture is to be simple and easy to a kind of of usefulness most in the virtual instrument architecture, and its implementation has following two kinds.A kind of is directly to insert general data collecting card (being called the interpolation type data collecting card again) in general PC slot, finishes the function of test macro by software programming control data capture card.Its advantage is that the system constructing cost is minimum, is convenient to use in the laboratory, and shortcoming is that system noise is bigger, is not easy to carry, and Electro Magnetic Compatibility and system reliability are relatively poor.Another is the external hanging type data collector that utilizes the USB of computing machine, it is big that it has a data throughout, and low price such as is convenient for carrying at advantage, adopt the data collector of the type both to have the measurement quality of top-grade instrument, can satisfy the diversity of measurement demand again.To most of demands, this scheme is not only practical, and has very high cost performance, is a kind of virtual instrument allocation plan that is particularly suitable for China's national situation.

Described as shown in Figure 1 signal condition equipment and data collecting card are installed in independently and form external hanging type data collector 8 in the housing, are connected with computing machine 7 by USB interface, and input is used to gather the signal of acceleration transducer on the shaking table 4.

Vibration signal detects and is realized by sensor, usually the most frequently used acceleration transducer has following several: piezoelectric acceleration transducer, ICP (Integrated Circuits Piezoelectric) formula acceleration transducer, capacitance acceleration transducer and piezoresistive accelerometer, wherein capacitance acceleration transducer is used for the vibratory impulse test more, piezoresistive transducer is used for measuring shock pulse signal when long more, the ICP sensor also is widely used in the vibration-testing field, but just the range of using at present adopts existing piezoelectric acceleration transducer or main flow.Piezoelectric acceleration sensor is a kind of easy for installation, use vibration detecting sensor widely, it has that volume is little, in light weight, mechanism is simple, reliable operation, signal to noise ratio (S/N ratio) height, advantage such as easy for installation are taken all factors into consideration various factors, and this paper selects for use piezoelectric acceleration sensor as vibration detecting sensor.

The charge signal that comes from sensor can't be measured with data acquisition equipment, topmost problem is the requirement that the electric charge parameter does not reach Computer Processing, and very easily be subjected to The noise, and also may there be very high kurtosis in some signal, therefore before converting them to digital quantity, to amplify, pre-service such as filtering, and require the signal condition module to have antijamming capability.Abandoning tradition charge amplifier of the present invention in addition is the shortcoming of control manually, requires charge amplifier to include programmable interface (as RS232).The multi-functional advance signal conditioning device 5 of YE5864 of for this reason selecting for use Yangzhou Radio No.2 Factory to produce, can electric charge, multiple input mode inputs (convenient later upgrading) such as voltage, ICP, be provided with double integration circuit simultaneously, the anti alias filter that gear is adjustable, this signal condition equipment 5 can be realized program control to signal condition equipment 5 by computing machine 7 serial ports, as shown in the figure.

Combined system is as shown in Figure 2 then according to above-mentioned steps apolegamy equipment, acceleration transducer 1,2,3 is gathered the vibration signal of shaking table 4, after the conditioning of signal condition equipment 5, enter data collecting card 6, data collecting card 6 is finished and after corresponding A/D transforms data is passed to computing machine by usb bus, computing machine 7 is responsible for data are carried out discrete Fourier transform (DFT) and digital signal frequency domain filtering, carry out pre-service by the signal behind the frequency domain filtering, by an integration output speed component, twice integration exported displacement component then.Simultaneous computer 7 can be realized the control to conditioning device 5 by the parameter of RS232 mouth signalization conditioning device 5.

The test value that below is the dynamic signal analyzer ATS of present device and Audio Precision company production is done to compare.

Test result deck watch

The amplitude-frequency test error that can be calculated system by last table is as follows:

The test error deck watch

From table we as can be seen, the precision of system testing frequency can satisfy fully<0.1% target, its frequency test precision is better than ATS system testing precision, amplitude error also is better than 1%.

The storage effect test

Owing to have integral relation between acceleration, speed and the displacement, so the computational accuracy of speed and displacement signal depends on the precision (amplitude-frequency precision) of acceleration signal.In above-mentioned test, provided the amplitude-frequency precision of acceleration signal, be reference value so can set the amplitude-frequency of signal for faster, and provide standard value, adopted time-domain integration method and Frequency Domain Integration method to provide measured value then respectively by Theoretical Calculation.Usually displacement measurement is used for the occasion of low-frequency vibration more, has therefore only got the signal of frequency less than 200Hz in test, and its test data is as follows:

Integration test result deck watch

Can find that by last table the data and the theoretical value that adopt the software integral method to obtain are more approaching, and the precision of Frequency Domain Integration method is a little more than the time-domain integration method, in the Signal Processing process, also avoided the problem of baseline correction, therefore in the present invention, final integral algorithm has still been selected the Frequency Domain Integration method for use.

The degree of distortion test

Degree of distortion is used for investigating the size that system under test (SUT) is introduced nonlinear distortion, needs the system that obtains to introduce the size of humorous wave amplitude in calculating.Harmonic distortion is the amplitude of harmonic component and the ratio of fundamental voltage amplitude.The computing formula of total harmonic distortion is as follows

THD = A 2 2 + A 3 2 + . . . + A n 2 A 1

A wherein 1Be the amplitude of first-harmonic, A 2... A nRepresent the amplitude of nth harmonic.Below we by testing the validity that test is described.

Signal frequency Standard distortion degree (%) System testing value of the present invention (%) ATS system testing value ??20 ??0.36 ??0.38 ??0.376

??80 ??0.36 ??0.38 ??0.376 ??160 ??0.36 ??0.37 ??0.377 ??400 ??0.36 ??0.38 ??0.378

Claims (6)

1. virtual oscillating table detection signal processing method is characterized in that may further comprise the steps:
1. adopt the amount of acceleration of shaking table by sensor, and convert electric signal to;
2. with signal condition equipment signal is nursed one's health;
3. the signal after the conditioning is delivered in the computing machine by the data collecting card collection;
4. by computing machine signal is carried out pre-service;
5. pretreated signal is carried out discrete Fourier transform (DFT) and digital signal frequency domain filtering, elder generation estimates the amplitude and the frequency of the spectrum peak of signal before carrying out the digital signal frequency domain filtering:
Peak value place Frequency Estimation input/output relation is as follows:
freqPeak = Σ i = i - span 2 i + span 2 autoSpectrum i j × df Σ i = i - span 2 i + span 2 autoSpectru m i
Peak value place Amplitude Estimation input/output relation is as follows:
powerPeak = Σ i = i - span 2 i + span 2 autoSpectru m i enbw
Wherein i is the index of search rate;
Df is a frequency interval;
Enbw is the equivalent noise bandwidth of selected window function;
AutoSpecturm: one-sided power spectrum;
N: the input power spectrum array comprises the number of first number;
SearchFreq: expectation estimated frequency point;
WindowConstants: window function constant;
Df: frequency interval;
Span: estimated frequency point place spectral line number;
FrePeak: the estimated frequency at the Frequency point place of trying to achieve;
PowerPeak: the estimated amplitude at calculated rate point place;
The input/output relation of digital signal frequency domain filtering is:
y ( r ) = Σ k = 0 N - 1 H ( k ) X ( k ) e j 2 πk / N
Wherein: X is the discrete Fourier transformation of input signal x, and H is the described frequency response function H of the frequency response function of wave filter in the wave filter conduct:
During low-pass filter be H ( k ) = 1 . . . . . . . . ( kΔf ≤ f u ) 0 . . . . . . . . . ( else )
During Hi-pass filter be H ( k ) = 1 . . . . . . . . ( kΔf ≥ f d ) 0 . . . . . . . . . ( else )
During bandpass filter be H ( k ) = 1 . . . . . . . . ( f d ≤ kΔf ≤ f u ) 0 . . . . . . . . . ( else )
During rejection filter be H ( k ) = 1 . . . . . . . . ( kΔf ≥ f u , kΔf ≤ f d ) 0 . . . . . . . . . ( else )
F wherein uFor the upper limit by frequency, f dFor lower limit by frequency, Δ f is a frequency resolution;
6. at last by an integration output speed component, twice integration exported displacement component.
2. virtual oscillating table detection signal processing method according to claim 1, it is characterized in that step integral process 6. is specific as follows: at first will need the signal of integration to do Fourier transform, then the result after the conversion is carried out integral operation in frequency domain, after inverse Fourier transform obtains integrated signal.
3. virtual oscillating table detection signal processing method according to claim 2 is characterized in that the input/output relation of a described integration is:
y ( r ) = Σ k = 0 N - 1 1 j 2 πkΔf H ( k ) X ( k ) e J 2 πkr N
The input/output relation of quadratic integral is:
y ( r ) = Σ k = 0 N - 1 - 1 ( 2 πkΔf ) 2 H ( k ) X ( k ) e J 2 πkr N
Wherein f lAnd f hBe respectively lower limit by frequency and the upper limit by frequency, X (k) is the Fourier transform of x (r), Δ f is the frequency resolution of signal.
4. virtual oscillating table detection signal processing method according to claim 1, it is characterized in that step signal condition process 2. by charge amplifier and anti-aliasing filtering module to signal amplify, isolation, filtering, excitation and linearization.
5. virtual oscillating table detection signal processing method according to claim 1, it is characterized in that in the step preprocessing process 4., at first signal is demarcated conversion, eliminate the polynomial trend item by least square method again, adopt 53 moving average methods or the frequency domain method of average to carry out signal smoothing at last and handle.
6. virtual oscillating table checkout equipment, it is characterized in that comprising acceleration transducer, signal condition equipment, data collecting card and computing machine, the said equipment connects successively, described signal condition equipment and data collecting card are installed in independently and form the external hanging type data collector in the housing, are connected with computing machine by USB interface; Described acceleration transducer comprises a three-dimensional acceleration transducer and at least four unidirectional acceleration transducers, described three-dimensional acceleration transducer and one of them unidirectional acceleration transducer are installed on the center of shaking table, and at least three unidirectional acceleration transducers are rigidly attached on the shaking table in addition.
CNB2007100292809A 2007-07-20 2007-07-20 A kind of virtual oscillating table detection signal processing method and equipment thereof CN100561162C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2007100292809A CN100561162C (en) 2007-07-20 2007-07-20 A kind of virtual oscillating table detection signal processing method and equipment thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2007100292809A CN100561162C (en) 2007-07-20 2007-07-20 A kind of virtual oscillating table detection signal processing method and equipment thereof

Publications (2)

Publication Number Publication Date
CN101113936A CN101113936A (en) 2008-01-30
CN100561162C true CN100561162C (en) 2009-11-18

Family

ID=39022393

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2007100292809A CN100561162C (en) 2007-07-20 2007-07-20 A kind of virtual oscillating table detection signal processing method and equipment thereof

Country Status (1)

Country Link
CN (1) CN100561162C (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103310172A (en) * 2013-05-13 2013-09-18 天津通广集团谷津高科技有限公司 Echo frequency discrimination method for passive radio frequency identification temperature measurement system
CN106415549A (en) * 2014-05-06 2017-02-15 贝克曼考尔特公司 Method and system for forming site network
US10317275B2 (en) 2016-01-20 2019-06-11 Simmonds Precision Products, Inc. Vibration monitoring systems

Families Citing this family (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100562718C (en) * 2008-03-13 2009-11-25 西南石油大学 Virtual type vibrating screen dynamic parameter detector
CN101819084A (en) * 2010-04-29 2010-09-01 哈尔滨工程大学 Self-adaptive harmonic suppression method of electro-hydraulic-type vibration simulation platform
CN102279084B (en) * 2011-05-03 2013-08-07 西安工程大学 Transmission line oscillation positioning method based on micro inertial measurement combination
CN103196407B (en) * 2012-01-09 2015-06-17 中联重科股份有限公司 Method, device and system for vibration displacement measurement of cantilever crane of pump truck and engineering machinery device
CN102608527A (en) * 2012-03-14 2012-07-25 南京因泰莱配电自动化设备有限公司 Data filtering processing method for dynamic characteristic test for circuit breaker
CN102680082A (en) * 2012-05-24 2012-09-19 中冶连铸技术工程股份有限公司 Amplitude testing method and device of crystallizer
CN103048106B (en) * 2012-12-26 2015-07-01 成都发动机(集团)有限公司 Aeroengine vibration transducer double-wire checking and detecting method and implementation system thereof
CN103419671B (en) * 2013-07-23 2015-09-16 南通大学 Plug-in hybrid electric vehicle lithium battery real-time monitoring system
CA2920699A1 (en) * 2013-08-14 2015-02-19 Eyesense Ag Method and apparatus for determining a smoothed data point within a stream of data points
CN103398846B (en) * 2013-08-16 2016-06-22 大连美恒时代科技有限公司 A kind of decelerator health analysis method and analysis platform system thereof
CN103528671A (en) * 2013-09-30 2014-01-22 黑龙江八一农垦大学 Effective method for removing background vibration in vibration observation signal
CN103528711A (en) * 2013-10-22 2014-01-22 天津七一二通信广播有限公司 Label bandwidth expansion measurement and analysis method based on Internet of Things electric power temperature measurement equipment
CN103674221B (en) * 2013-11-29 2015-11-04 北京兴华机械厂 Hydrostatic liquid bearing PIGA motor vibrations input, analysis and screening technique
CN104198281B (en) * 2014-08-13 2016-10-19 中国石油天然气集团公司 A kind of wide plate test machine data collecting system
CN106524967B (en) * 2016-11-07 2019-08-02 重庆理工大学 A kind of automobile core wheel actual travel displacement measurement and extracting method
CN108734060A (en) * 2017-04-18 2018-11-02 香港理工大学深圳研究院 A kind of recognition methods of high-speed EMUs wheel polygonization and device
CN107727227B (en) * 2017-09-30 2019-05-21 南京大学 Ultra-high-tension power transmission line icing based on Φ-OTDR waves monitoring method
CN108181486B (en) * 2018-01-25 2019-12-03 中国科学院电子学研究所 The processing method and processing device of acceleration signal
CN108398222A (en) * 2018-02-28 2018-08-14 武汉优泰电子技术有限公司 A method of passing through acceleration acquisition speed and displacement
CN108593095A (en) * 2018-04-26 2018-09-28 盐城博鸣信息科技有限公司 A kind of Vibration Fault Signal acquiring and processing method of converter power transformer tap switch
CN109029888A (en) * 2018-08-14 2018-12-18 广西电网有限责任公司电力科学研究院 A kind of three combined environment test case vibration parameters setting methods
CN109540545B (en) * 2018-11-30 2020-04-14 厦门大学 Abnormal sound diagnosis signal acquisition device and processing method for power output assembly of tractor
CN110657935A (en) * 2019-09-24 2020-01-07 浙江大学 Seismic wave acceleration integral processing method and system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
一种新型电动机械振动台测控系统. 刘小勇,施仁.仪器仪表学报,第24卷第3期. 2003 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103310172A (en) * 2013-05-13 2013-09-18 天津通广集团谷津高科技有限公司 Echo frequency discrimination method for passive radio frequency identification temperature measurement system
CN106415549A (en) * 2014-05-06 2017-02-15 贝克曼考尔特公司 Method and system for forming site network
CN106415549B (en) * 2014-05-06 2020-02-11 贝克曼考尔特公司 Method and system for forming site network
US10317275B2 (en) 2016-01-20 2019-06-11 Simmonds Precision Products, Inc. Vibration monitoring systems

Also Published As

Publication number Publication date
CN101113936A (en) 2008-01-30

Similar Documents

Publication Publication Date Title
US20200217708A1 (en) Ultrasonic flow meter with subsampling of ultrasonic transducer signals
US3229198A (en) Eddy current nondestructive testing device for measuring multiple parameter variables of a metal sample
CN100543481C (en) Detection is delivered to the method for the complex impedance of the RF power of load and this load
CN1673768B (en) Calibration method and apparatus
CN101825660B (en) High-efficiency measurement method for sinusoidal signal frequency in undersampling and implementation device
US10775212B2 (en) Ultrasonic flow meter
Crama et al. Initial estimates of Wiener and Hammerstein systems using multisine excitation
Stewart et al. Ferroelectric hysteresis measurement and analysis
CN101807795B (en) Method for forming electric energy metering simulation system and device thereof
CN102183366A (en) Device and method for vibration measurement and failure analysis of rolling bearing
CN100495037C (en) Static induction space filter measuring method for gas-solid two phase tube flow particle speed
Boyle Uncertainty in hybrid gravitational waveforms: Optimizing initial orbital frequencies for binary black-hole simulations
CN101403774B (en) Harmonic wave analysis method based on non-synchronous sampling
CN101488031B (en) High-precision magnetic bearing axial control method based on interference observer
US7383140B2 (en) Capacitance, inductance and impedance measurements using multi-tone stimulation and DSP algorithms
CN101819252B (en) Analog-circuit fault diagnosis method
JP2009543050A (en) Modal parameter estimation method and apparatus
CN101135723A (en) Universal electric power mutual-inductor verify device
AU2007260584B2 (en) A system for complex impedance measurement
Tomic et al. A new power system digital harmonic analyzer
CN104991142B (en) A kind of signal analyzer, device and processing method
US6801873B1 (en) Analysis of rotating machines
CN104991210B (en) The evaluation method and caliberating device of a kind of local discharge detection device
Lataire et al. Non-parametric estimate of the system function of a time-varying system
Hessling Dynamic metrology—an approach to dynamic evaluation of linear time-invariant measurement systems

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20091118

Termination date: 20120720