CN104132791A - Operation mode analysis experiment method and device based on pulse excitation - Google Patents

Abstract

An operation mode analysis experiment method based on pulse excitation comprises the following steps: (1) the endpoint of a beam is chosen as an excitation point, and a steel hammer is used to implement pulse excitation on the beam; (2) response signals generated by a reference point and a response point after pulse excitation are acquired; (3) band-pass filtering is performed on the acquired signals; (5) a cross-correlation function between the reference point and the response point is obtained, and a matrix equation set composed of data at different sampling moments of time of the cross-correlation function is constructed; (5) a coefficient matrix is solved by using the matrix equation set; (6) a system pole is identified, a least square error stabilization graph is established, and a modal shape is solved; and (7) if the modal assurance criterion value is poor, the value at a different sampling moment of time is selected, and the method returns to step (4) until the modal assurance criterion value is within a preset reasonable interval, and the modal parameters of the beam are obtained. The invention further provides an operation mode analysis experiment device based on pulse excitation. The time and the intensity of test can be reduced, and the students' learning effect can be improved.

Description

A kind of operational modal analysis experimental technique and device based on pulse excitation

Technical field

The present invention relates to operational modal analysis technical field, especially a kind of operational modal analysis experimental technique and device.

Background technology

In the teaching process of the courses such as Engineering Testing Technique, Reformation of Mechanical Vibration, add some model analysis experiments, can motivate students' interest in learning, the learning initiative of mobilizing students and initiative, the innovation ability of cultivating student, increases substantially teaching efficiency.Can make by experiment student obtain abundant perceptual knowledge, deepen the understanding of student to physical concept hard to understand, theorem and law, reduce the difficulty that self-learning process middle school student encounter problems; By experiment can image, specifically, describe the problem intuitively, can make abstract knowledge vividly fresh and alive rapidly.

Existing experimental modal analysis system is generally comprised of three parts: 1. excitation system: make system vibration.2. measuring system: with displacement, speed or the acceleration vibration signal on each main position of sensor measurement experimental subjects.3. analytic system: the pumping signal collecting and response signal are recorded in computing machine through digital-to-analog conversion, with the modal parameter of software systems identification vibrational system.The basic step of experiment is as follows: 1) determine empirical model, experiment support structure is got up; 2) mode experiment, generally by exciting hammer hammering method utilization excitation experiment structure, records the response signal of pumping signal and each measuring point; 3) record data are carried out to digital processing, obtain the transport function of each measuring point, and form transfer function matrix; 4) utilize model analysis software to carry out parameter identification; 5) carry out animation demonstration.Mode test is usually for the correctness of extraction, condition monitoring and analysis, nonlinear system research, fault analysis and the checking finite element model of the mode model under actual operating conditions.

From the vibration response signal of the in-service state of structure, extract the operational modal analysis method of modal parameter, the Structure dynamic characteristics of identification more approaches the real kinetic behavior of structure under actual motion condition than test modal analysis, become the active research direction of model analysis field development in recent years.Parameter identification method in operational modal analysis can be divided into time domain, frequency domain and time-frequency domain discrimination method etc., mainly contain time series method, Modal Parameter by Random Decrement (random decrement, RD), environmental excitation method, Random Subspace Method, Empirical mode decomposition (empirical mode decomposition, EMD), peak picking method, frequency domain decomposition method and Using Continuous Wavelet Transform etc., existing analytical approach is many is the hypothesis of zero-mean white noise based on pumping signal, and every kind of method has certain limitation, as more difficult in the model order of time series method definite; Natural excitation method method requirement data sample is long, average time is many; Determining of Random Subspace Method model order is comparatively loaded down with trivial details, and when measuring point is more, Hankel matrix order is very high, easily causes the problem of matrix morbid state; The disadvantage of frequency domain method is that requirement frequency resolution is high and sample is long, and structure is little damping; The response message that Time-frequency Analysis utilizes is less, is a kind of local recognition methods etc.In order to improve student's results of learning, carry out operational modal analysis experimental teaching very necessary, still, there is no at present such experimental provision, therefore for operational modal analysis experimental technique and the device of teaching demand, urgently study.

Summary of the invention

Results of learning when improving Students ' Learning modal analysis technique, the invention provides a kind of can realize quick calculating, degree of accuracy high, there is operational modal analysis experimental technique and device based on pulse excitation that good error is controlled, can be reduced proof strength and time.

The technical solution adopted for the present invention to solve the technical problems is:

An operational modal analysis experimental technique based on pulse excitation, comprises the following steps:

1) select the end points of beam as point of excitation, utilize steel hammer to implement pulse excitation to beam;

Selected distance point of excitation response point near and that response signal amplitude is larger is as a reference point;

Each geometric model node in described reference point and the reflection beam vibration shape is arranged response measuring point;

2) gather the response signal that described reference point and response point produce after pulse excitation;

3) collection signal is carried out to bandpass filtering, its passband is interested structural modal frequency range;

4) ask for the cross correlation function between reference point and response point, and build the matrix equation group that the different sampling instant data of cross correlation function form;

5) utilize described matrix equation group to solve matrix of coefficients;

6) recognition system limit, sets up minimum mean-square error steady state picture, solves Mode Shape;

7) carry out mode and put the calculating of letter criterion matrix value, if it is not good that mode is put letter criterion value, choose different sampling instant values, turn back to step 4) rebuild matrix equation group, until mode is put letter criterion value in default reasonable interval, obtain the modal parameter of beam.

Further, described method is further comprising the steps of: 8) mode animate: draw the Mode Shape vector of each direction of each point, arrange that with measuring point geometric model is corresponding, just obtain describing the Mode Shape animation of the relative amplitude in each measuring point x, y, z direction.

Further, described step 4) in, according to the cross correlation function between formula (1) computation structure response point j and reference point i

$R_{\mathrm{ij}}\left(\mathrm{\τ}\right)={\∫}_o^T{x}_i\left(t\right)x_j(t+\mathrm{\τ})\mathrm{d\τ}---\left(1\right)$

In formula, R _ij(τ) be the cross correlation function between response point j and reference point i, T is the test duration, x _i(t) be the acceleration responsive signal of reference point, x _j(t) be the acceleration responsive signal of response point, τ is the time interval;

To the cross correlation function R between structural response point j and reference point i _ij(τ) according to time interval Δ t sampling, and be expressed as complex mode form

$R_{\mathrm{ij}}=\left(\mathrm{k\Δt}\right)=\underset{r=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{rij}}\exp \left({\mathrm{\λ}}_r\mathrm{k\Δt}\right)---\left(2\right)$

C in formula _rijfor the constant coefficient relevant to r rank mode; N is rank number of mode to be identified; Δ t is sampling time interval; λ _rfor system limit;

By system limit λ _rbe expressed as ξ in formula _rit is r rank damping ratios; ω _rit is r rank mode undamped natural frequency of a mechanical system.

Further, described step 4) in, by the cross correlation function matrix between all response point of each sampling instant and M reference point, form multiple-input and multiple-output matrix, set up constant coefficient finite difference matrix equation formula (4):

A in formula ₀, A ₁... A _mfor matrix of coefficients; R ₁(t ₀), R ₁(t ₁) ... R ₁(t _2N) be that cross correlation function matrix between all measuring points and the first reference point is at t ₀, t ₁... t _2Nvalue constantly, R ₂(t ₁), R ₂(t ₂) ... R ₂(t _2N+1) be that cross correlation function matrix between all measuring points and the second reference point is at t ₁, t ₂... t _2N+1value constantly, R _m(t _2N-1), R _m(t _2N) ... R _m(t _4N-1) be that the response signal cross correlation function matrix of all measuring points and M reference point is at t _2N-1, t _2N... t _4N-1the value of sampling instant, R _m(t _2N), R _m(t _2N+1) ... R _m(t _4N) be that the response signal cross correlation function matrix of all measuring points and M reference point is at t _2N, t _2N+1... t _4Nthe value of sampling instant.

Described step 5) in, utilize the covariance matrix of this system of equations to form compression equation, obtain the least square solution of this overdetermined equation, obtain coefficient matrices A ₀, A ₁... A _mvalue.

Described step 6) in, order: structure following formula

$\underset{k=1}{\overset{2N}{\mathrm{\Σ}}}{A}_k{R}_{\mathrm{ij}}\left(\mathrm{k\Δt}\right)=\underset{k=1}{\overset{2N}{\mathrm{\Σ}}}\left(\underset{r=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{rij}}{A}_k{Q}_r^k\right)=0---\left(6\right)$

Owing at least needing 2N sampled data could determine all N rank mode, therefore get k=0,1,2 ... 2N.As above formula establishment, coefficient A _kmeet Prony rational fraction orthogonal polynomial (7), and this polynomial expression with for characteristic solution.Get A _m=1, obtain

$A_0+A_1{Q}_r^1+\·\·\·+A_{M-1}{Q}_r^{M-1}+Q_r^M=0---\left(7\right)$

By the coefficient matrices A estimating ₀, A ₁... A _m-1substitution formula (7), tries to achieve the limit of system.

Described step 6) in, by cross correlation function matrix representation, be the partial fraction sum of each rank Mode Shape of system and modal participation factors matrix, obtain

$\left[R\left(\mathrm{k\Δt}\right)\right]=\underset{r=1}{\overset{N}{\mathrm{\Σ}}}[\left(V_r\right)\exp \left({\mathrm{\λ}}_r\mathrm{k\Δt}\right)\left(L_r\right)+\left(V_r^{*}\right)\exp \left({\mathrm{\λ}}_r^{*}\mathrm{k\Δt}\right)\left(L_r^{*}\right)]---\left(8\right)$

In formula, V _rfor Mode Shape matrix, L _rfor modal participation factors matrix, be illustrated in the contribution amount of each rank mode in system responses, for the complex-conjugate matrix of Mode Shape matrix, for mode is participated in the complex-conjugate matrix of factor matrix, conjugate complex number for system limit;

By the system limit substitution formula (8) of identification, try to achieve by each rank Mode Shape vector Ψ _rthe Mode Shape matrix V forming _rand modal participation factors matrix L _r, the overall situation that obtains system mode parameter is estimated.

Described step 7), in, mode is put letter criterion matrix value and is:

${\mathrm{MAC}}_{\mathrm{rs}}=\frac{{\left|{{\mathrm{\Ψ}}_r}^{*T}{\mathrm{\Ψ}}_s\right|}^2}{\left({{\mathrm{\Ψ}}_r}^{*T}{\mathrm{\Ψ}}_r\right)\left({{\mathrm{\Ψ}}_s}^{*T}{\mathrm{\Ψ}}_s\right)}---\left(6\right)$

Wherein, Ψ _rit is r rank Mode Shape vector; Ψ _sit is s rank Mode Shape vector; it is the conjugate transpose of r rank Mode Shape vector; it is the conjugate transpose of s rank Mode Shape vector.

A kind of operational modal analysis experimental provision based on pulse excitation, comprise fixed support, elastic threads, steel hammer, acceleration transducer, beam, suspension ring, concentric cable, data acquisition front and operational modal analysis center, described beam is hung by elastic threads, make beam in free boundary condition, described elastic threads one end is connected with fixed support, the other end of described elastic threads is connected with suspension ring, described suspension ring and beam are threaded connection, described acceleration transducer is tested the vibration acceleration-time data of each measuring point, each acceleration transducer is electrically connected to data acquisition front respectively by concentric cable, data acquisition front is electrically connected to operational modal analysis center, acceleration transducer collects after the response signal under pulse excitation, imported into data acquisition front, pass to again operational modal analysis center, the operational modal analysis software module that the vibration response signal data that gather import operational modal analysis center by data acquisition front is carried out analyzing and processing, identification modal parameter.

Beneficial effect of the present invention is mainly manifested in: 1, can realize quick calculating, degree of accuracy high, there is good error and control, can reduce proof strength and time, significantly improve test efficiency; 2, broken through existing experimental modal analysis technical requirement extrinsic motivated response input and the defect to the various pressure hypothesis of excitation input, can realize quickly and easily girder construction is carried out to dynamic analysis; 3, do not need to measure external drive, only measure response data, reduced device requirement, experimentation cost can reduce greatly, for operational modal analysis experimental technique has increased a kind of new method.

Accompanying drawing explanation

Fig. 1 is schematic flow sheet of the present invention.

Fig. 2 is that operational modal analysis experimental provision forms schematic diagram.

Fig. 3 is that girder construction measuring point is arranged schematic diagram.

Fig. 4 is the time domain response oscillogram of reference point.

Fig. 5 is the time domain response oscillogram of measuring point.

Fig. 6 is the cross correlation function figure of reference point and measuring point time domain response.

Fig. 7, for the Mode Shape figure of the beam of identification, wherein, (a) is a first order mode, is (b) second_mode, is (c) three first order modes.

Embodiment

Below in conjunction with accompanying drawing, the invention will be further described.

With reference to Fig. 1～Fig. 7, a kind of axis system operational modal analysis method based on multiple spot pulse excitation, comprises the following steps:

1) select the end points of beam as point of excitation, utilize steel hammer to implement pulse excitation to beam;

Selected distance point of excitation response point near and that response signal amplitude is larger is as a reference point;

Each geometric model node in described reference point and the reflection beam vibration shape is arranged response measuring point;

2) gather the response signal that described reference point and response point produce after pulse excitation;

3) collection signal is carried out to bandpass filtering, its passband is interested structural modal frequency range;

4) ask for the cross correlation function between reference point and response point, and build the matrix equation group that the different sampling instant data of cross correlation function form;

5) utilize described matrix equation group to solve matrix of coefficients;

6) recognition system limit, sets up minimum mean-square error steady state picture, solves Mode Shape;

7) carry out mode and put the calculating of letter criterion matrix value, if it is not good that mode is put letter criterion value, choose different sampling instant values, turn back to step 4) rebuild matrix equation group, until mode is put letter criterion value in default reasonable interval, obtain the modal parameter of beam.

Further, described method is further comprising the steps of: 8) mode animate: draw the Mode Shape vector of each direction of each point, arrange that with measuring point geometric model is corresponding, just obtain describing the Mode Shape animation of the relative amplitude in each measuring point x, y, z direction.

Referring to Fig. 2, operational modal analysis experimental technique and a device based on pulse excitation, comprise fixed support 1, elastic threads 2, steel hammer 3, acceleration transducer 4, beam 5, suspension ring 6, concentric cable 7, count and adopt front end 8, operational modal analysis center 9 (can adopt notebook computer).Described beam 5 is hung by elastic threads 2, excludes the impact of extraneous vibration, and makes beam in free boundary condition.Described elastic threads one end is connected with fixed support 1, one end is connected with suspension ring 6, described suspension ring 6 are threaded connection with beam, vibration acceleration-the time data of described acceleration transducer 4 each measuring points of test, each acceleration transducer 4 is electrically connected to data acquisition front 8 respectively by concentric cable 7, and data acquisition front 8 is electrically connected to operational modal analysis center 9.Acceleration transducer 4 collects after the response signal under pulse excitation, imported into data acquisition front 8, pass to again operational modal analysis center 9, the operational modal analysis software module that the vibration response signal data that gather import operational modal analysis center by data acquisition front is carried out analyzing and processing, identification modal parameter.The concrete operation step of operational modal analysis software module is as follows:

1) select point of excitation

In order to identify the modal parameter of unbonded beam, should input as far as possible a wideband random excitation signal.Auto-power spectrum and the white noise signal of pulse excitation are close, and its spectral density, in lower frequency section close to straight, is comparatively ideal pumping signal.Therefore utilize 3 pairs of beams of steel hammer to apply pulse excitation, to excite each rank mode.

In technical solutions according to the invention, " pulse excitation " refers to and on beam, chooses point of excitation, use steel hammer 3 incentive structures, improves the signal to noise ratio (S/N ratio) of collection signal.Referring to Fig. 3, in beam equal intervals, arrange 9 measuring points.Selecting No. 1 point of end points of beam is point of excitation.

2) select reference point and response point, measure structural vibration response

In the present embodiment, choose No. 1 shop as a reference point on beam to be measured, all the other 8 measuring points are as response point, simultaneously fixing acceleration transducer 4 respectively in reference point and response point.By the vibration acceleration of the acceleration transducer lower reference point of 4 acquisition pulse excitation and response point.The time domain plethysmographic signal of reference point is referring to Fig. 4, and the time domain waveform of No. 3 measuring points is referring to Fig. 5.

3) ask for cross correlation function, and be expressed as complex mode form

Cross correlation function represents be between two time serieses and at the same time sequence in any two degrees of correlation between value in the same time not, be that cross correlation function is to describe random signal x (t), y (t) is in any two degrees of correlation between value in the same time not.According to the cross correlation function between formula (1) computation structure response point j and reference point i

$R_{\mathrm{ij}}\left(\mathrm{\τ}\right)={\∫}_o^T{x}_i\left(t\right)x_j(t+\mathrm{\τ})\mathrm{d\τ}---\left(1\right)$

In formula, R _ij(τ) be the cross correlation function between response point j and reference point i, T is the test duration, x _i(t) be the acceleration responsive signal of reference point, x _j(t) be the acceleration responsive signal of response point, τ is the time interval.

To the cross correlation function R between structural response point j and reference point i _ij(τ) according to time interval Δ t sampling, and be expressed as complex mode form

$R_{\mathrm{ij}}=\left(\mathrm{k\Δt}\right)=\underset{r=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{rij}}\exp \left({\mathrm{\λ}}_r\mathrm{k\Δt}\right)---\left(2\right)$

C in formula _rijfor the constant coefficient relevant to r rank mode; N is rank number of mode to be identified; Δ t is sampling time interval; λ _rfor system limit.

By system limit λ _rbe expressed as ξ in formula _rit is r rank damping ratios; ω _rit is r rank mode undamped natural frequency of a mechanical system.Cross correlation function between No. 3 measuring points and reference point is referring to Fig. 6.

4) build cross correlation function matrix equation

Utilize response signal under pulse excitation to carry out between two computing cross-correlation, by the cross correlation function matrix between all response point of each sampling instant and reference point, form matrix, set up constant coefficient finite difference matrix equation formula (4),

A in formula ₀, A ₁... A _mfor matrix of coefficients; R ₁(t ₀) be that the cross correlation function matrix of all measuring points and the first reference point is at t ₀value constantly, R _m(t _4N) for take response signal cross correlation function matrix that M point is reference point in the value of 4N sampling instant, all the other by that analogy.Utilize the covariance matrix of this system of equations to form compression equation, can obtain the least square solution of this overdetermined equation, obtain coefficient matrices A ₀, A ₁... A _mvalue.

In the present embodiment, select altogether 1 point of excitation, knock 3 times, measure altogether three groups of cross correlation functions, cross correlation function by all measuring points that record and reference point calculates lump cross correlation function, and selection analysis bandwidth is 0-500Hz, and the calculating order of getting finite difference equation is 48.Because selected calculating order is much larger than wish identification physical mode number, for signal noise provides outlet, therefore reduced noise to the impact of true mode, raising Precision of Estimating Modal Parameter.

5) ask for system limit.

For recognition system limit, order: structure following formula

$\underset{k=1}{\overset{2N}{\mathrm{\Σ}}}{A}_k{R}_{\mathrm{ij}}\left(\mathrm{k\Δt}\right)=\underset{k=1}{\overset{2N}{\mathrm{\Σ}}}\left(\underset{r=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{rij}}{A}_k{Q}_r^k\right)=0---\left(6\right)$

Owing at least needing 2N sampled data could determine all N rank mode, therefore get k=0,1,2 ... 2N.As above formula establishment, coefficient A _kmeet Prony rational fraction orthogonal polynomial (7), and this polynomial expression with for characteristic solution.Get A _m=1, obtain

$A_0+A_1{Q}_r^1+\·\·\·+A_{M-1}{Q}_r^{M-1}+Q_r^M=0---\left(7\right)$

By the coefficient matrices A estimating ₀, A ₁... A _m-1substitution formula (7), tries to achieve the limit of system;

6) set up minimum mean-square error steady state picture, solve Mode Shape.By cross correlation function matrix representation, be the partial fraction sum of each rank Mode Shape of system and modal participation factors matrix, obtain

$\left[R\left(\mathrm{k\Δt}\right)\right]=\underset{r=1}{\overset{N}{\mathrm{\Σ}}}[\left(V_r\right)\exp \left({\mathrm{\λ}}_r\mathrm{k\Δt}\right)\left(L_r\right)+\left(V_r^{*}\right)\exp \left({\mathrm{\λ}}_r^{*}\mathrm{k\Δt}\right)\left(L_r^{*}\right)]---\left(8\right)$

In formula, V _rfor Mode Shape matrix, L _rfor modal participation factors matrix, be illustrated in the contribution amount of each rank mode in system responses, for the complex-conjugate matrix of Mode Shape matrix, for mode is participated in the complex-conjugate matrix of factor matrix, conjugate complex number for system limit;

By the system limit substitution formula (8) of identification, try to achieve by each rank Mode Shape vector Ψ _rthe Mode Shape matrix V forming _rand modal participation factors matrix L _r, the overall situation that obtains system mode parameter is estimated.

In the present embodiment, adopt and manyly with reference to least square complex exponential method (pLSCE method), to investigate the different when errors of calculation of Mode Shape of natural frequency corresponding to each rank mode under orders, damping of calculating.In order to realize minimum mean-square error energy Fast Convergent when calculating order increase, the frequency error while setting identification is 2%, and damping ratio error is 5%, and vibration shape error is 2%.If increased, calculate after order, the limit obtaining and residual are substantially constant,, at this frequency place label symbol " S ", if only have frequency constant, note upper " f ", if only have damping ratio constant, mark " d ", only has the constant note of residual upper " V ", obtains minimum mean-square error steady state picture, be chosen on all calculating orders mark " S " and put maximum N to be listed as corresponding frequency be system mode frequency, and calculate thus when Mode Shape of system damping.

7) mode checking and analysis: the verifying correctness that mainly completes operational modal analysis result.Utilize mode to put letter criterion the accuracy of judgement mode estimation.Ψ wherein _rit is r rank Mode Shape vector; Ψ _sit is s rank Mode Shape vector; it is the conjugate transpose of r rank Mode Shape vector; it is the conjugate transpose of s rank Mode Shape vector.By mode, put letter criterion MAC matrix and can judge that modal parameter picks up the correctness of result, thus the accuracy of judgement mode estimation.If there is linear relationship between two Mode Shape, its MAC value is close to 1, if they have nothing to do each other, MAC value is close to zero.Through mode, put the correctness of letter criterion matrix judgement recognition result, if the MAC value between each rank mode is all less than 0.3, each rank mode of identification is true mode, and recognition result is accurate, finishes whole calculating process.If exist the MAC value between certain two rank mode to be greater than 0.3, from step (4), select different sampling instant data to recalculate until meet the requirements.Determined like this each rank modal parameter value, the operational modal analysis core calculations process based on pulse excitation finishes.

8) mode animate: draw the Mode Shape vector of each direction of each point, arrange that with measuring point geometric model is corresponding, just obtain describing the Mode Shape animation of the relative amplitude in each measuring point x, y, z direction, thereby complete whole service model analysis overall process.First three rank Mode Shape figure of the beam of identification is referring to Fig. 7.

Described step 2) in, the vibration acceleration of reference point and response point is measured by acceleration transducer 4, is completed the record of vibration acceleration by data acquisition front 8.

Described step 7), in, utilize mode to put the verifying correctness that letter criterion is carried out recognition result.

The above is only better embodiment of the present invention, and the equivalence of doing according to structure, feature and principle described in patent claim of the present invention therefore all changes or modifies, and is included in patent claim of the present invention.

The above is only the preferred embodiment of the present invention, and protection scope of the present invention is not limited to above-described embodiment, and all technical schemes belonging under thinking of the present invention all belong to protection scope of the present invention.It should be pointed out that for those skilled in the art, some improvements and modifications without departing from the principles of the present invention, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (8)

1. the operational modal analysis experimental technique based on pulse excitation, is characterized in that: said method comprising the steps of:

1) select the end points of beam as point of excitation, utilize steel hammer to implement pulse excitation to beam;

Selected distance point of excitation response point near and that response signal amplitude is larger is as a reference point;

Each geometric model node in described reference point and the reflection beam vibration shape is arranged response measuring point;

2) gather the response signal that described reference point and response point produce after pulse excitation;

3) collection signal is carried out to bandpass filtering, its passband is interested structural modal frequency range;

4) ask for the cross correlation function between reference point and response point, and build the matrix equation group that the different sampling instant data of cross correlation function form;

5) utilize described matrix equation group to solve matrix of coefficients;

6) recognition system limit, sets up minimum mean-square error steady state picture, solves Mode Shape;

7) carry out mode and put the calculating of letter criterion matrix value, if it is not good that mode is put letter criterion value, choose different sampling instant values, turn back to step 4) rebuild matrix equation group, until mode is put letter criterion value in default reasonable interval, obtain the modal parameter of beam.

2. a kind of operational modal analysis experimental technique based on pulse excitation as claimed in claim 1, it is characterized in that: described method is further comprising the steps of: 8) mode animate: the Mode Shape vector that draws each direction of each point, arrange that with measuring point geometric model is corresponding, just obtain describing the Mode Shape animation of beam.

3. a kind of operational modal analysis experimental technique based on pulse excitation as claimed in claim 1 or 2, is characterized in that: described step 4), according to the cross correlation function between formula (1) computation structure response point j and reference point i

$R_{\mathrm{ij}}\left(\mathrm{\τ}\right)={\∫}_o^T{x}_i\left(t\right)x_j(t+\mathrm{\τ})\mathrm{d\τ}---\left(1\right)$

In formula, R _ij(τ) be the cross correlation function between response point j and reference point i, T is the test duration, x _i(t) be the acceleration responsive signal of reference point, x _j(t) be the acceleration responsive signal of response point, τ is the time interval;

To the cross correlation function R between structural response point j and reference point i _ij(τ) according to time interval Δ t sampling, and be expressed as complex mode form

$R_{\mathrm{ij}}=\left(\mathrm{k\Δt}\right)=\underset{r=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{rij}}\exp \left({\mathrm{\λ}}_r\mathrm{k\Δt}\right)---\left(2\right)$

C in formula _rijfor the constant coefficient relevant to r rank mode; N is rank number of mode to be identified; Δ t is sampling time interval; λ _rfor system limit;

By system limit λ _rbe expressed as ξ in formula _rit is r rank damping ratios; ω _rit is r rank mode undamped natural frequency of a mechanical system.

4. a kind of operational modal analysis experimental technique based on pulse excitation as claimed in claim 3, it is characterized in that: described step 4), by the cross correlation function matrix between all response point of each sampling instant and M reference point, form multiple-input and multiple-output matrix, set up constant coefficient finite difference matrix equation formula (4):

A in formula ₀, A ₁... A _mfor matrix of coefficients; R ₁(t ₀), R ₁(t ₁) ... R ₁(t _2N) be that cross correlation function matrix between all measuring points and the first reference point is at t ₀, t ₁... t _2Nvalue constantly, R ₂(t ₁), R ₂(t ₂) ... R ₂(t _2N+1) be that cross correlation function matrix between all measuring points and the second reference point is at t ₁, t ₂... t _2N+1value constantly, R _m(t _2N-1), R _m(t _2N) ... R _m(t _4N-1) be that the response signal cross correlation function matrix of all measuring points and M reference point is at t _2N-1, t _2N... t _4N-1the value of sampling instant, R _m(t _2N), R _m(t _2N+1) ... R _m(t _4N) be that the response signal cross correlation function matrix of all measuring points and M reference point is at t _2N, t _2N+1... t _4Nthe value of sampling instant;

Described step 5) in, utilize the covariance matrix of this system of equations to form compression equation, obtain the least square solution of this overdetermined equation, obtain coefficient matrices A ₀, A ₁... A _mvalue.

5. a kind of operational modal analysis experimental technique based on pulse excitation as claimed in claim 4, is characterized in that: described step 6), and order: structure following formula

$\underset{k=1}{\overset{2N}{\mathrm{\Σ}}}{A}_k{R}_{\mathrm{ij}}\left(\mathrm{k\Δt}\right)=\underset{k=1}{\overset{2N}{\mathrm{\Σ}}}\left(\underset{r=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{rij}}{A}_k{Q}_r^k\right)=0---\left(6\right)$

Owing at least needing 2N sampled data could determine all N rank mode, therefore get k=0,1,2 ... 2N, as above formula establishment, coefficient A _kmeet Prony rational fraction orthogonal polynomial (7), and this polynomial expression with for characteristic solution, get A _m=1, obtain

$A_0+A_1{Q}_r^1+\·\·\·+A_{M-1}{Q}_r^{M-1}+Q_r^M=0---\left(7\right)$

By the coefficient matrices A estimating ₀, A ₁... A _m-1substitution formula (7), tries to achieve the limit of system.

6. a kind of operational modal analysis experimental technique based on pulse excitation as claimed in claim 5, it is characterized in that: described step 6), by cross correlation function matrix representation, be the partial fraction sum of each rank Mode Shape of system and modal participation factors matrix, obtain

$\left[R\left(\mathrm{k\Δt}\right)\right]=\underset{r=1}{\overset{N}{\mathrm{\Σ}}}[\left(V_r\right)\exp \left({\mathrm{\λ}}_r\mathrm{k\Δt}\right)\left(L_r\right)+\left(V_r^{*}\right)\exp \left({\mathrm{\λ}}_r^{*}\mathrm{k\Δt}\right)\left(L_r^{*}\right)]---\left(8\right)$

In formula, V _rfor Mode Shape matrix, L _rfor modal participation factors matrix, be illustrated in the contribution amount of each rank mode in system responses, for the complex-conjugate matrix of Mode Shape matrix, for mode is participated in the complex-conjugate matrix of factor matrix, conjugate complex number for system limit;

By the system limit substitution formula (8) of identification, try to achieve by each rank Mode Shape vector Ψ _rthe Mode Shape matrix V forming _rand modal participation factors matrix L _r, the overall situation that obtains system mode parameter is estimated.

7. a kind of operational modal analysis experimental technique based on pulse excitation as claimed in claim 6, is characterized in that: described step 7), mode is put letter criterion matrix value and is:

${\mathrm{MAC}}_{\mathrm{rs}}=\frac{{\left|{{\mathrm{\Ψ}}_r}^{*T}{\mathrm{\Ψ}}_s\right|}^2}{\left({{\mathrm{\Ψ}}_r}^{*T}{\mathrm{\Ψ}}_r\right)\left({{\mathrm{\Ψ}}_s}^{*T}{\mathrm{\Ψ}}_s\right)}---\left(6\right)$

Wherein, Ψ _rit is r rank Mode Shape vector; Ψ _sit is s rank Mode Shape vector; it is the conjugate transpose of r rank Mode Shape vector; it is the conjugate transpose of s rank Mode Shape vector.

8. the operational modal analysis experimental provision based on pulse excitation, it is characterized in that: described device comprises fixed support, elastic threads, steel hammer, acceleration transducer, beam, suspension ring, concentric cable, data acquisition front and operational modal analysis center, described beam is hung by elastic threads, make beam in free boundary condition, described elastic threads one end is connected with fixed support, the other end of described elastic threads is connected with suspension ring, described suspension ring and beam are threaded connection, described acceleration transducer is tested the vibration acceleration-time data of each measuring point, each acceleration transducer is electrically connected to data acquisition front respectively by concentric cable, data acquisition front is electrically connected to operational modal analysis center, acceleration transducer collects after the response signal under pulse excitation, imported into data acquisition front, pass to again operational modal analysis center, the operational modal analysis software module that the vibration response signal data that gather import operational modal analysis center by data acquisition front is carried out analyzing and processing, identification modal parameter.