CN106289691A - A kind of bridge block impact vibration detection method based on microwave radar device and detection device - Google Patents

Abstract

The invention discloses a kind of bridge block impact vibration detection method based on microwave radar device and detection device, wherein detection method step is: bridge is divided into a minor structure, and arranges impulsive force input point and displacement output point in each minor structure；Act on impulsive force successively in impulsive force input point in bridge piecemeal minor structure, synchronize the displacement time course data using microwave radar device to gather minor structure displacement output node；Seek minor structure frequency response function；According to the frequency response function of each minor structure, identify the modal parameter of each minor structure；Build the MAC matrix of structure, identify k rank Mode Shape direction coefficient before each minor structure；Use self-adapted genetic algorithm, identify the Mode Shape direction coefficient on the whole rank of each minor structure；The flexibility matrix of full structure calculates and amount of deflection is predicted.The present invention can realize vibration-testing known to low cost, high efficiency bridge structure input power conscientiously, to be applied to detection and the safety generaI investigation of bridge structure.

Description

A kind of bridge block impact vibration detection method based on microwave radar device and detection Device

Technical field

The present invention relates to microwave interference measurement technology and bridge health monitoring field, particularly to a kind of bridge structure The detection method of piecemeal vibration-testing and detection device.

Background technology

China is currently under novel urbanization and industrialization fast development period, and infrastructure investment is to account for national product The ratio of total value about 15%-20% grows steadily, and large quantities of great infrastructure have completed or built.On the other hand, China Unsafe bridge is more than 90,000 at present.In 2007-2012, the whole nation has 37 bridge blocks and collapses, and wherein 13 in bridge construction beam generation thing Therefore, cause 182 people to die altogether, 177 people are injured.There are 7.4 " taking life bridge by force " every year on average, i.e. averagely just had one less than two months The accident of rising occurs.The most the Eleventh Five-Year Plan period, need the bridge quantity of Measuring error every year, account for highway network bridge sum After 15%, especially 2006,2,007 two years statistics rural highways, quantity and the proportion thereof of unsafe bridge are all substantially increased.State Family's bridge inspection and maintenance diagnosis market is huge, but traditional method is to be manually main time-consuming present situation, in the urgent need to simple and efficient Bridge fast evaluation method.

Structural health monitoring technology is the most gradually being applied to numerous civil engineering structure after the development of nearly 30 years In security diagnostics and regular maintenance.Environmental vibration testing is the Main Means of existing structure health monitoring, it utilize wind load and The natural conditions excitation bridges such as wagon flow, have easy to operate advantage relative to artificial excitation test, but due to civil structure Complexity and the existence of the challenge such as incompleteness of observation data, existing environmental vibration testing method mainly exports frequency The structure basic parameter such as rate and the vibration shape, also cannot directly support bridge maintenance and management decision-making.Impact shock is tested at home and abroad Also there is a certain degree of application, but existing impact shock device is heavy or function is limited, cause on-the-spot test cost height, bridge Closedown time that is open to traffic is long.

Compared to the somewhat expensive of the most time-consuming of periodic sensing approach He structural health monitoring technology, recent domestic Scholar has been presented for multiple bridge method for rapidly testing based on mobile vehicle, by installing accelerometer on vehicle and gathering Acceleration information when it travels on bridge floor is to identify the basic parameters such as bridge frequency, and can carry out bridge tapping scanning Vehicle and by observation vehicle reaction identify bridge damnification.Between said method is convenient and swift, but they rely on bridge Connect measurement to be analyzed, be only capable of identifying the basic parameter such as bridge frequency and the vibration shape and carrying out preliminary non-destructive tests.

Based on considerations above, bridge based on impact shock is without reference point piecemeal method for testing vibration, in conjunction with bridge intelligence The utilization of diagnosis car, makes up the deficiency of conventional impact method for testing vibration, fully while realizing the convenient and swift test of bridge Play the basic advantage of impact shock, really realize comprehensively identifying and effective assessment of performance of structural parameters.But tradition punching Hit sensor in vibration-testing and be often arranged in whole bridge floor, cause that demand number of sensors is many, data transmission wire length, construction Time length, testing cost are expensive；Additionally, bridge intelligent diagnostics car self-sensing system has limitation, the office of bridge can only be covered Region, portion.In tradition piecemeal vibration-testing, its algorithm only accounts for minimum potential energy principal, can solve the lower mode of structure. But due to the potential energy convergence to rank number of mode, the high order mode identification causing structure is inaccurate.Simultaneously because algorithm is for enumerating Method, can cause when situation number is more, and computational efficiency is low, even cannot calculate.

Summary of the invention

The deficiency existed with technology for above-mentioned existing method, present invention one to be provided is applicable to the impact of bridge piecemeal and shakes Dynamic detection method and microwave radar detection device, thus realize the low cost of bridge, the monitoring of high efficiency structural behaviour and assessment.

For solving above-mentioned technical problem, the technical solution used in the present invention is:

A kind of bridge block impact vibration detection method based on microwave radar device, it is characterised in that step is:

Step one, bridge is divided into the minor structure of m (m >=2), and one or more rushing is set in each minor structure Hit power input point and whole displacement output point；

Act on impulsive force in step 2, successively the impulsive force input point in bridge piecemeal minor structure, synchronize employing micro- Ripple radar equipment gathers the displacement time course data that minor structure displacement output node is caused by the impulsive force of impulsive force input node；

Step 3, seek minor structure frequency response function H₁(ω), H₂(ω) ..., H_m(ω): gather each minor structure respectively is defeated The displacement output point displacement time course data of ingress impulsive force and microwave radar collection carries out signal processing, then uses arbitrary frequency response Function Estimation algorithm estimates the displacement frequency response function of each minor structure；

Step 4, frequency response function according to each minor structure, identify the modal parameter of each minor structure: respectively to each minor structure The frequency response function H estimated₁(ω), H₂(ω) ..., H_m(ω) modal parameter CMIF method of identification is used to do Modal Parameter Identification, it is thus achieved that The modal parameter of each minor structure: system poleMode zoom factorThe displacement modes vibration shapeAnd mode participation coefficientWherein, subscript s be minor structure label (s=1,2 ..., m)；Subscript r be identify minor structure rank number of mode (r=1, 2,…,n)；

Step 5, the MAC matrix of structure structure, before identifying each minor structureRank Mode Shape direction coefficient；

Step 6, employing self-adapted genetic algorithm, identify the Mode Shape direction coefficient on each minor structure whole n rank；

Step 7, the flexibility matrix of full structure calculate and amount of deflection prediction: in the Mode Shape direction asking for each minor structure be After number, obtain integrally-built Mode Shape according to minor structure modal combination formula；Calculate according to flexibility computing formula further Draw the integrally-built flexibility matrix for predicting deflection of bridge span.

Described step 5 method particularly includes:

Arranging variables D iffMAC, it is 2 that structure operating mode number is helped in corresponding minor structure fusion^(m-1)k, filter out DiffMAC The operating mode that little value is corresponding:

$DiffMAC=\left|\right|MAC|-{\mathrm{\Π}}_{i=1}^k{\mathrm{MAC}}_{ii}|---\left(1\right)$

In formula, MAC is the metric characteristic in vibration shape matrix array space,

MAC_ijThe i-th rank displacement modes vibration shape and jth component level for each operating mode Move the included angle cosine between Mode Shape, { φ_iIt is the i-th rank displacement modes vibration shape of each operating mode, { φ_jIt is the jth rank of each operating mode The displacement modes vibration shape)；| MAC | is the value of MAC matrix determinant；Value for MAC matrix diagonals line element；

Calculating the potential energy value that the operating mode of MAC matrix screening is corresponding further, corresponding operating mode number is 2^m-1, potential energy minimum person It is the correct vibration shape that k rank before full structure are corresponding.

Described step 6 method particularly includes:

A, stochastic generation represent the chromosome of substructure mode vibration shape direction coefficient:

X={x₁₁ x₁₂…x_1nx₂₁ x₂₂…x_2n…x_m1 x_m2…x_(m-1)n}_1×(m-1)n (2)

In formula, x_srIn the label that subscript s is minor structure, except target minor structure (s=1,2 ..., m-1)；Subscript r generation The rank number of mode of each minor structure of table (r=1,2 ..., n)；

The size according to minor structure division number m and rank number of mode n of minor structure of choosing of population number S is determined, takes 50 ～200；

B, according to displacement modes vibration shape direction coefficient result in k rank before the MAC matrix identification structure of structure, revise stochastic generation Chromosome before the value of (m-1) k element, amended chromosome is shown as:

C, fitness value calculation also differentiate, choose maximum adaptation angle value and store globally optimal solution；

F (x)=max (-(1-p) Π_p) (4)

In formula, p is penalty,For each work The maximum of off diagonal element in the MAC matrix that condition calculates；Pc is penalty factor (pc=10)；

D, carry out roulette selection operation further, selective factor B P is set_S∈ (0.7-0.9), enters the gene selecting pond Carry out next step intersection and variation:

Crossover probability P_cWith mutation probability P_mSize be:

$P_c=\left\{\begin{array}{cc}{P}_{c1}-\frac{P_{c1}-P_{c2}}{F_{\max }-\stackrel{\&OverBar;}{F}}(F^{\&prime;}-\stackrel{\&OverBar;}{F}),& F^{\&prime;}\&GreaterEqual;\stackrel{\&OverBar;}{F}\\ P_{c1},& F^{\&prime;}<\stackrel{\&OverBar;}{F}\end{array}\right.---\left(5\right)$

$P_m=\left\{\begin{array}{cc}{P}_{m1}-\frac{P_{m1}-P_{m2}}{F_{\max }-\stackrel{\&OverBar;}{F}}(F-\stackrel{\&OverBar;}{F}),& F\&GreaterEqual;\stackrel{\&OverBar;}{F}\\ P_{c1},& F<\stackrel{\&OverBar;}{F}\end{array}\right.---\left(6\right)$

In formula, P_c1Higher limit for crossover probability；P_c2Lower limit for crossover probability；P_m1Higher limit for mutation probability； P_m2Lower limit for mutation probability；F_maxFor the maximum of ideal adaptation degree in population；F is every average fitness value for population； F ' is fitness value bigger in two individualities to be intersected；F is the individual fitness value that to make a variation；

Calculate the fitness value of new gene after making a variation after e, variation according to formula 4 and differentiate, if it is more than mesh Front globally optimal solution then substitutes, if less than globally optimal solution, putting into the iteration entering next step in population pond.

The higher limit of the described intersection factor is 0.9, and the lower limit of the intersection factor is 0.7；Described mutagenic factor higher limit is 0.2, the lower limit of mutagenic factor is 0.001.

The displacement time course data of described minor structure displacement output node is measured by microwave radar systems.

A kind of bridge block impact vibration microwave radar detection device, including impact shock device, microwave radar systems with And data processing unit, described data processing unit uses above-mentioned bridge block impact vibration detection method, shakes described impact Dynamic device acts on the displacement time course data of the impulsive force on bridge and described microwave radar systems collection position and processes, and draws For predicting the integrally-built flexibility matrix of deflection of bridge span.

Described microwave radar systems includes: reception antenna, transmitting antenna, transmitter, receiver, signal processor, electronics Magnetic compass, GPS module, laser designator and range finder using laser, described signal processor is mainly for the treatment of described transmitter Zero intermediate frequency signals after the microwave signal mixing that the microwave signal sent and receiver receive, draws required displacement time-histories letter Number.

Microwave radar systems of the present invention energy high accuracy, multi-point and lon g-distance monitor micro-vibration of bridge structure in real time, and carry out Deflection of bridge structure is predicted, effectively incorporates bridge piecemeal vibration-testing performance evaluation system；Bridge structure piecemeal of the present invention impacts Method for detecting vibration, uses and can effectively solve tradition piecemeal vibration-testing higher order accuracy difference without reference point optimized algorithm and calculate effect The problem that rate is low, can realize low cost, in high precision conscientiously, and the quickly test of high efficiency bridge structure health performance is commented Estimate.

Beneficial effect

1, a kind of microwave radar being applicable to the test of bridge block impact vibration that the present invention provides, at bridge routine health Monitoring and evaluation stage achieve monitors fast and efficiently, compensate for deficiency and the traditional axle beam of tradition bridge health monitoring The defect of structure piecemeal vibration-testing.

2, the method that bridge micrometric displacement surveyed by the microwave radar in the present invention, relatively traditional monitoring bridge sensor is compared and is significantly subtracted Lack in required number of sensors, and test and need not close traffic, can conscientiously realize the quick survey of the micro-vibration of bridge structure Examination.

3, the substructure mode vibration shape direction system of the test of the consideration structure MAC matrix bridge block impact vibration in the present invention Count the more traditional bridge structure block impact vibration of method of identification to test without reference point diagnostic method, improve structure high order mode identification Precision.

4, the substructure mode vibration shape direction-adaptive heredity of the test of the bridge structure block impact vibration in the present invention is calculated Method method of identification, more traditional bridge structure block impact vibration test, can accurately, efficiently without reference point method of discrimination (enumerative technique) Ask for the whole Mode Shape of each minor structure, thus quickly identify modal parameter and the Modal Flexibility of bridge structure, and carry out amount of deflection Prediction, substantially increases computational efficiency, can conscientiously realize Bridge performance efficient, quick assessment.

Accompanying drawing explanation

Fig. 1 is that real bridge schematic diagram surveyed by microwave radar；

Fig. 2 is that embodiment of the present invention Bridge structon structure divides schematic diagram；

Fig. 3 is that in the embodiment of the present invention, the impact of impulsive force input node 9 is tried hard to；

Fig. 4 is the displacement time-histories figure of displacement output node 9 in the embodiment of the present invention；

Fig. 5 is embodiment of the present invention Bridge structure recognition MAC matrix diagram；

Fig. 6 is self-adapted genetic algorithm evolution graph in the embodiment of the present invention, and wherein (a) is for identifying 12 rank, and (b) is identification 15 Rank；

Fig. 7 is 16 first order mode comparison diagram before the structure of fusant structure in the embodiment of the present invention；

Fig. 8 is the flexibility matrix figure of embodiment of the present invention Bridge structure；

Fig. 9 is embodiment of the present invention Bridge structure amount of deflection prognostic chart under action of static load.

Detailed description of the invention:

In order to be better understood from the present invention, below in conjunction with the accompanying drawings, technical scheme is described in detail.

First bridge is carried out impacting without reference point piecemeal by the present invention, utilizes microwave radar equipment to gather bridge test point Displacement time course data, each minor structure displacement time course data gathered according to microwave radar equipment and the impact of impulsive force input node Power, the frequency response function carrying out minor structure is estimated and the Modal Parameter Identification of each minor structure.In bridge structural health monitoring, real Existing purpose is monitoring and the output of full structure, for merging each minor structure, the substructure mode vibration shape need to asked for it further Zoom in and out and adjust and the differentiation of Mode Shape direction coefficient.According to the traditional axle girder construction piecemeal described in foregoing invention background Impact shock test substructure mode vibration shape direction coefficient method of discrimination, it only accounts for minimum potential energy principal, can solve structure Lower mode.But due to the potential energy convergence to rank number of mode, the high order mode identification causing structure is inaccurate.Simultaneously because Algorithm is enumerative technique, can cause when situation number is more, and computational efficiency is low, even cannot calculate.Therefore, the invention provides Consider that the MAC matrix of bridge structure carries out structure lower mode recognition of vibration, furthermore utilize self-adapted genetic algorithm identification structure Whole Mode Shape, thus obtain the flexibility matrix of structure and carry out amount of deflection prediction, it is achieved the fast monitored of complete bridge with Assessment.

Three parts are comprised altogether based on the said process present invention.

One, the substructure mode vibration shape direction MAC matrix method of identification of bridge structure block impact vibration test

Though bridge structure piecemeal vibration-testing solves, conventional impact vibrating device is heavy or function is limited, causes on-the-spot survey The problem that examination cost is high, the time length that is open to traffic closed by bridge.But traditional axle girder construction block impact vibration test substructure mode is shaken Type direction coefficient method of discrimination, due to the potential energy convergence to rank number of mode, it only accounts for minimum potential energy principal and may result in knot The high order mode identification of structure is inaccurate, thus causes the precision with traditional monitoring method that wastes time and energy to reduce problem.For realizing bridge The low cost of girder construction, high-precision configuration performance monitoring and assessment, solve traditional axle girder construction block impact vibration testing algorithm Deficiency.Therefore the present invention proposes consideration structure MAC matrix identification substructure mode vibration shape discriminating direction algorithm.Its particular content For:

As the instrument of evaluation structure modal vector orthogonality, MAC matrix is the metric characteristic in vibration shape matrix array space. Its expression is:

${\mathrm{MAC}}_{ij}=\frac{{\left({\left\{{\mathrm{\&phi;}}_i\right\}}^T\right\{{\mathrm{\&phi;}}_j\left\}\right)}^2}{\left({\left\{{\mathrm{\&phi;}}_i\right\}}^T\right\{{\mathrm{\&phi;}}_i\left\}\right)\left({\left\{{\mathrm{\&phi;}}_i\right\}}^T\right\{{\mathrm{\&phi;}}_j\left\}\right)}---\left(1\right)$

In formula, MAC_ijMore than the angle between the i-th rank displacement modes vibration shape and the jth rank displacement modes vibration shape of each operating mode String；{φ_iIt it is the i-th rank modal vector of each operating mode；{φ_jIt it is the jth rank modal vector of each operating mode；

In MAC matrix, the contribution margin of each first order mode is identical, i.e. high order mode, low order mode sensitivity is the same, thus Can effectively make up in minimum potential energy principal, the defect that high order mode sensitivity is more weak.Full structure is divided into m minor structure, respectively The rank number of mode of minor structure is n rank, then have 2^(m-1)nIndividual modal combination scheme, understands according to above formula (1), MAC matrix non-right Angle element M AC_ijBeing closer to zero, the value of MAC matrix determinant is closer to the product of diagonal element.Therefore, variable is set DiffMAC is

$DiffMAC=\left|\right|MAC|-{\mathrm{\&Pi;}}_{i=1}^k{\mathrm{MAC}}_{ii}|---\left(2\right)$

In formula, | MAC | is the value of MAC matrix determinant；Value for MAC matrix diagonals line element.

2^(m-1)nIn individual operating mode number, corresponding 2^(m-1)nIndividual DiffMAC value, then minima min (DiffMAC) of DiffMAC The corresponding vibration shape is the target vibration shape, or the vibration shape antipodal with the target vibration shape.For differentiating from min (DiffMAC) further Combination in determine the correct target vibration shape, then differentiate further combined with minimum potential energy principal.

The flexibility matrix F obtained is combined with any one vibration shape direction regulation coefficient, can pre-geodesic structure meaning in office node Displacement structure vector u under load vector f effect, i.e.

$F=\underset{r=1}{\overset{n}{\&Sigma;}}(\frac{{\mathrm{\&phi;}}_r{\left({\mathrm{\&phi;}}_r\right)}^T{Q}_r}{-{\mathrm{\&lambda;}}_r}+\frac{{\mathrm{\&phi;}}_r^{*}{\left({\mathrm{\&phi;}}_r^{*}\right)}^T{Q}_r^{*}}{-{\mathrm{\&lambda;}}_r^{*}})---\left(3\right)$

U=Ff (4)

Thus, the combination of different directions coefficient also will obtain different motion vectors under same load vector f effect u.According to minimum potential energy principal, occur real displacement u structurally will make the potential energy minimalization of system.So correct shakes The deformation of type combination correspondence is that structure real displacement under this load case can make the potential energy of structure take minima.Namely The potential energy of modal combination after MAC matrix differentiates, the vibration shape group that potential energy minimum is corresponding under certain load case can be calculated Conjunction is exactly correct modal combination.To discrete elastomer, potential energy is calculated as follows,

${\mathrm{\&Pi;}}_p=\frac{1}{2}{u}^{T}Ku-u^{T}f---\left(5\right)$

In formula, K is integrally-built stiffness matrix, is unknown quantity for practical structures, so also cannot be straight by above formula Connect the potential energy of computation structure.Formula 4 is substituted into formula 5, and considers that displacement flexibility matrix is symmetrical matrix, and be stiffness matrix Inverse matrix, can obtain following formula in conjunction with formula 3

${\mathrm{\&Pi;}}_p=-\frac{1}{2}{f}^T\&lsqb;\underset{r=1}{\overset{m}{\&Sigma;}}(\frac{{\mathrm{\&phi;}}_r{\left({\mathrm{\&phi;}}_r\right)}^T{Q}_r}{-{\mathrm{\&lambda;}}_r}+\frac{{\mathrm{\&phi;}}_r^{*}{\left({\mathrm{\&phi;}}_r^{*}\right)}^T{Q}_r^{*}}{-{\mathrm{\&lambda;}}_r^{*}})\&rsqb;f---\left(6\right)$

Owing to civil engineering structure is mostly little damping structure, the vibration shape of identification is the real vibration shape, soAbove formula Can be as follows with abbreviation,

${\mathrm{\&Pi;}}_p=-\frac{1}{2}{f}^T\&lsqb;\underset{r=1}{\overset{m}{\&Sigma;}}{\mathrm{\&phi;}}_r{\left({\mathrm{\&phi;}}_r\right)}^T(\frac{Q_r}{-{\mathrm{\&lambda;}}_r}+\frac{Q_r^{*}}{-{\mathrm{\&lambda;}}_r^{*}})\&rsqb;f=-\frac{1}{2}\underset{r=1}{\overset{m}{\&Sigma;}}\mathrm{\&lsqb;}{f}^T{\mathrm{\&phi;}}_r{\left({\mathrm{\&phi;}}_r\right)}^{T}f(\frac{Q_r}{-{\mathrm{\&lambda;}}_r}+\frac{Q_r^{*}}{-{\mathrm{\&lambda;}}_r^{*}})\&rsqb;---\left(7\right)$

In above formula, f^Tφ_r(φ_r)^TF is a number and equal, and institute's above formula can be with abbreviation,

${\mathrm{\&Pi;}}_p=-\frac{1}{2}\underset{r=1}{\overset{m}{\&Sigma;}}\&lsqb;{\left(f^T{\mathrm{\&phi;}}_r\right)}^2(\frac{Q_r}{-{\mathrm{\&lambda;}}_r}+\frac{Q_r^{*}}{-{\mathrm{\&lambda;}}_r^{*}})\&rsqb;---\left(8\right)$

Load column vector f in above formula is that the load of the upper each node acting on total is (if some node is not Imposed load, the value that this load vector is corresponding at this node can be taken as 0).By load vector f by the segment partition scheme shape of minor structure Formula can be written as following formula,

$f=\left[\begin{array}{c}{f}^1\\ f^2\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ f^m\end{array}\right]---\left(9\right)$

In formula, f¹For acting on the load column vector in minor structure 1；f²For acting on the load column vector in minor structure 2； f^mFor acting on the load column vector in minor structure m.

Understanding, according to least potential energy theory, the displacement that the correct vibration shape calculates is that structure real displacement makes the potential energy of structure Little, therefore potential energy is ranked up from small to large, all first order mode combination sides of all minor structures corresponding to potential energy minimum person Case is correct modal combination scheme, thus may determine that the value of direction coefficient.

Two, the substructure mode vibration shape direction-adaptive genetic algorithm in identification method of bridge structure block impact vibration test

Substructure mode vibration shape direction according to consideration structure MAC matrix bridge block impact vibration test described above Coefficient method of identification understands, and the method belongs to enumerative technique, particularly divides more in minor structure, i.e. m is relatively big, and structural modal rank Number is higher, when i.e. n is bigger, easily occurs that situation number is too many, thus causes computational efficiency low.Though computational accuracy improves, but also The problem the most effectively solving the quick health evaluating of bridge structure.Therefore the present invention proposes the test of bridge structure block impact vibration The method in self-adapted genetic algorithm identification substructure mode vibration shape direction.Itself particularly as follows:

In substructure mode Parameter fusion, for minor structure divide more, solve rank number of mode higher time, it is considered to structure The low problem of substructure mode vibration shape direction coefficient method of identification computational efficiency of MAC matrix bridge block impact vibration test, Use for reference genetic algorithm (Genetic Algorithm) i.e. one class and use for reference evolution laws (survival of the fittest, the survival of the fittest something lost of biosphere Pass mechanism) develop and next randomization searching method.Its main thought is the operation in mimic biology system to gene: replicate, Intersect and variation etc. generates the optimal solution of a problem.Thus the problems of value of each minor structure vibration shape direction coefficient can be turned The problem turning to solve the structure potential energy correct vibration shape correspondence minimum potential energy of search at different vibration shape direction coefficient load cases combination.Cause This, based on self-adapted genetic algorithm, substructure mode Parameter fusion Optimized model is:

With structure potential energy corresponding to different vibration shape direction coefficient values as object function, with structure potential energy minimum for optimizing mesh Mark, i.e.

${\mathrm{min\&Pi;}}_p=\min (-\frac{1}{2}\underset{r=1}{\overset{m}{\&Sigma;}}\{{\&lsqb;{\left(f^1\right)}^T{\mathrm{\&phi;}}_r^1+{\left(f^2\right)}^T{\mathrm{\&phi;}}_r^2\sqrt{\frac{Q_r^2}{Q_r^1}}{\mathrm{\&eta;}}_r^2+...+{\left(f^m\right)}^T{\mathrm{\&phi;}}_r^m\sqrt{\frac{Q_r^m}{Q_r^1}}{\mathrm{\&eta;}}_r^m\&rsqb;}^2\left(\frac{Q_r^1}{-{\mathrm{\&lambda;}}_r}+\frac{Q_r^{1*}}{-{\mathrm{\&lambda;}}_r^{*}}\right)\left\}\right)---\left(10\right)$

In formula, η_rFor minor structure r first order mode direction regulation coefficient.

At the substructure mode vibration shape direction coefficient method of identification considering the test of structure MAC matrix bridge block impact vibration In, take MAC matrix into consideration and minimum potential energy principal has differentiated.Therefore in self-adapted genetic algorithm, use penalty function Form considers the MAC matrix of structure, and as evaluating individual excellent degree, the formula that fitness function uses is:

F (x)=max (-(1-p) Π_p) (11)

In formula, p is penalty,max(MAC_i≠j) it is each operating mode meter The maximum of off diagonal element in the MAC matrix calculated；Pc is penalty factor (pc=10)；

Simultaneously, it is contemplated that the computational efficiency of algorithm and the enforcement purpose of the present invention, for different vibration shape direction coefficients it is Optimized variable, randomly selects the binary coding scheme of gene, and the present invention proposes according to structure MAC matrix k rank before structure Vibration shape direction coefficient differentiates, and then can be imported by the correct vibration shape direction coefficient that differentiate in the encoding gene of genetic algorithm, Thus it is effectively reduced Population Size and population quantity, reach to improve the purpose of computational efficiency.

Particularly as follows:

X is made to represent vibration shape direction coefficient η_r, exist

$x={\mathrm{\&eta;}}_r=\left\{\begin{array}{c}-1\\ 1\end{array}\right.---\left(12\right)$

Then can directly utilize variable x for this problem to encode, i.e.

X={x₁₁ x₁₂…x_1nx₂₁ x₂₂…x_2n…x_m1 x_m2…x_(m-1)n}_1×(m-1)n (13)

In formula, x_srIn the label that subscript s is minor structure, except target minor structure, s=1,2 ..., m-1；Subscript r represents The rank number of mode of each minor structure, r=1,2 ..., n；

Amended gene representation is:

Simultaneously, it is contemplated that convergence and search precision, method proposes adaptive strategy actual according to population Situation adjusts crossover probability P at random_cWith mutation probability P_mSize.I.e.

$P_c=\left\{\begin{array}{cc}{P}_{c1}-\frac{P_{c1}-P_{c2}}{F_{\max }-\stackrel{\&OverBar;}{F}}(F^{\&prime;}-\stackrel{\&OverBar;}{F}),& F^{\&prime;}\&GreaterEqual;\stackrel{\&OverBar;}{F}\\ P_{c1},& F^{\&prime;}<\stackrel{\&OverBar;}{F}\end{array}\right.---\left(5\right)$

$P_m=\left\{\begin{array}{cc}{P}_{m1}-\frac{P_{m1}-P_{m2}}{F_{\max }-\stackrel{\&OverBar;}{F}}(F-\stackrel{\&OverBar;}{F}),& F\&GreaterEqual;\stackrel{\&OverBar;}{F}\\ P_{c1},& F<\stackrel{\&OverBar;}{F}\end{array}\right.---\left(16\right)$

In formula, P_c1Higher limit for crossover probability；P_c2Lower limit for crossover probability；P_m1Higher limit for mutation probability； P_m2Lower limit for mutation probability；F_maxFor the maximum of ideal adaptation degree in population；F is every average fitness value for population； F ' is fitness value bigger in two individualities to be intersected；F is the individual fitness value that to make a variation；From formula (15) and formula (16) It can be seen that for fitness higher than the individuality of population average fitness value, give relatively low probability of crossover and mutation probability, make Obtain individual obtaining and protect the entrance next generation；For fitness value less than the individuality of population average fitness value, give higher miscellaneous Hand over probability and mutation probability, make this individuality eliminate.

After obtaining the Mode Shape direction coefficient of each sub-block, just can obtain integrally-built Mode Shape, it is achieved thereby that The fusion of each sub-block Mode Shape, can calculate integrally-built flexibility matrix.Deflection of bridge span can be predicted by this matrix, Thus bridge health monitoring and rapid evaluation can be carried out.

Three, bridge displacement collection and microwave radar data processing system

The purpose of bridge health monitoring is to realize monitoring and the assessment of full structure, bridge based on above two method identification Girder construction piecemeal substructure mode vibration shape direction, and be wirelessly transferred the input impulsive force of collection, then one data of demand process Full structure is identified and health evaluating by the output displacement after system may utilize above-mentioned known input power and the differentiation of each minor structure. Be the rapid evaluation realizing bridge structure, then simultaneously demand high accuracy, non-contacting monitoring device.Therefore the present invention considers that introducing is suitable Microwave radar systems for the test of bridge block impact vibration.Its particular content is:

Its described equipment mainly includes radar signal processor and monitoring unit two large divisions.Mainly comprise parts for receiving Antenna, launch antenna, transmitter, receiver, signal processor, high-precision electronic magnetic compass, GPS module, laser designator, swash Ligh-ranging machine, aobvious control unit, self-test unit and power supply unit.Being 1.5ms by the microwave radar transmitting cycle, initial frequency is 15.85GHz, termination frequency are the FM signal of 16.15GHz, are reflected back, thunder after the microwave signal launched runs into target The reception antenna reached receives echo information.

Assume to produce echo-signal at distance radar R, then, after this echo arrives radar, echo delay is τ₀, and(c is the light velocity), then echo-signal is represented by:

$s\left(t\right)=A_r{e}^{j\{2{\mathrm{\&pi;f}}_0(t-{\mathrm{\&tau;}}_0)+\frac{k{(t-{\mathrm{\&tau;}}_0)}^2}{2}\}}---\left(17\right)$

In formula, A_rFor receiving the amplitude of signal；τ₀For echo delay time.Exported by frequency mixer in receiver further The zero intermediate frequency signals of Quadrature double path:

Zero intermediate frequency Beat Signal is obtained through mixing:

$s_{\mathrm{\&Delta;}}\left(t\right)={\mathrm{AA}}_r{e}^{j(2{\mathrm{\&pi;f}}_0{\mathrm{\&tau;}}_0+kt\mathrm{\&tau;}-\frac{{\mathrm{kt}}^2}{2})}---\left(18\right)$

In receiver, the Quadrature double path zero intermediate frequency signals of frequency mixer output, after anti-confusion filtering, amplification, send two-way AD Changer, it is thus achieved that discrete echo data send signal processor to carry out FFT process.Wherein, a frequency sweep cycle of signal is launched In a corresponding sampling period, the data in a sampling period are a snap, a fast beat of data is N point FFT, obtains difference The echo phase information of distance objective, is the FFT of M snap continuously, data is lined up M row N row, more every column data is FFT Conversion, the phase of echo fluctuation information of available each range gate target.

After determining target, based on phase interference method principle, obtain target x by receiver_iEcho, by resolving, Show that measured target is at t eventually₁And t₂The phase contrast in moment isThen fine motion displacement is based on obtaining impact point x_iEcho Signal phase difference and obtain, it may be assumed that

In order to obtain the vertical displacement of the body of a bridge, need the radial displacement value obtained by phase calculation is carried out true value projection meter Calculate.By comparing the phase place between twice measurement signal, target can be accurately obtained relative to radar radially micro-displacement, Yong Huguan The heart be target vertical to displacement y (x_i, t):

$y(x_i,t)=\frac{y_R(x_i,t)}{{\mathrm{sin\&alpha;}}_i}=y_R(x_i,t)\frac{R_i}{h_i}---\left(20\right)$

In formula, R_iFor radar to impact point x_iRadial distance；h_iFor radar to impact point x_iVertical dimension；α_iFor radar Transmitted wave and the angle of horizontal plane.

In conjunction with above-mentioned bridge structure block impact vibration test substructure mode vibration shape direction MAC matrix method of identification and Self-adapted genetic algorithm method of identification, and the displacement of above-mentioned microwave radar equipment Real-time Collection and be wirelessly transferred input impact Power, the present invention propose the microwave radar systems being applicable to bridge piecemeal vibration-testing, can effective integration said method provide The modal parameter of each minor structure, can quickly carry out the Modal Parameter Identification of full structure, on the monitoring of bridge routine health and assessment rank Section achieves monitoring and evaluation fast and efficiently.

Embodiment one

Bridge structural model as shown in Figure 1, this bridge carries out one-lane traffic, whole concrete bridge in each direction Face width 6.5m, overall length 15.54m, wherein both sides are respectively arranged with the footpath that 1.07m is wide.On armored concrete deck, three freely-supporteds are cold rolling The spacing of steel I-beam is 2.18m.Microwave radar test is as it is shown in figure 1, gather full knot by microwave radar device transform test point The structural response that structure is evoked by impact load.Being embodied as step is:

(1) minor structure splitting scheme determines

Site environment according to tested bridge structure and bridge structure form, be divided into bridge structure 21 monitorings and click on Whole monitoring points are divided into 2 minor structures simultaneously and are monitored by row monitoring, and splitting scheme is as shown in Figure 2.Minor structure 1 is main Comprise displacement output node 1/2/3/4/8/9/10/15/16/17, wherein choose node 2/4/9/16 for impulsive force input node； Minor structure 2 mainly comprises displacement output node 5/6/7/11/12/13/14/18/19/20/21, wherein chooses node 11/13/ 18/20 is impact force action node.

(2) radar placement location is selected

According to tested bridge structure site environment and minor structure splitting scheme, in conjunction with radar emission and the wave beam of reception antenna Angle, determines radar equipment placement location.Owing to bridge structure is divided into 2 minor structures, therefore choose the both sides bridge of bridge structure Under pier, place is as radar set-point.Radar placement location schematic diagram refers to Fig. 1.

(3) test of bridge piecemeal minor structure 1/2 impact shock and data acquisition

Microwave radar equipment is positioned over the set-point chosen, opens power supply unit and aobvious control unit is ready for data acquisition Collection.Use the percussion mechanism impulsive force input point in bridge piecemeal minor structure successively to act on impulsive force, synchronize to use microwave Radar equipment gathers the displacement time course data that the displacement output node of minor structure is caused by the impulsive force of impulsive force input node.Its In, the impulsive force of collection uses the form being wirelessly transferred to be transferred to microwave radar data processing unit, and microwave radar gathers displacement The displacement time course data of output node is transmitted directly to microwave radar data processing unit.As a example by minor structure 1, impulsive force inputs The impulsive force of node 9 refers to Fig. 3, and the displacement time course data of displacement output node 9 refers to Fig. 4.

(4) data of minor structure 1/2 process

At microwave radar data processing unit, it is known that the input power of minor structure 1 and minor structure 2 and each displacement output node Displacement time course data.After the impulsive force time course data of each sub-block and displacement output node time course data are carried out windowing, filtering, Estimate the frequency response function H of each sub-block respectively₁(ω), H₂(ω).According to minor structure 1 and the frequency response function of minor structure 2, use CMIF Modal Parameter Identification algorithm identifies the modal parameter of each minor structure respectively, and the modal parameter of the minor structure 1/2 of identification refers to table 1:

Table 1

(5) fusion of minor structure 1/2 modal parameter

Under same order frequency, owing to minor structure 1 is different with the Mode Shape scaling of minor structure 2, in fusant structure 1 During with the Mode Shape of minor structure 2, need to use unified scaling standard that the Mode Shape of each minor structure is zoomed in and out adjustment.With Minor structure 1 is target minor structure, according to the modal parameter of each minor structure asked for, zooms in and out the Mode Shape of minor structure 2 Adjust and the differentiation of Mode Shape direction coefficient.(6) son considering the test of structure MAC matrix bridge block impact vibration is used to tie Structure Mode Shape direction coefficient method of identification, it determines the front k=7 rank Mode Shape direction coefficient of minor structure 2, concrete operations are:

Whole circumstances number 2 is calculated according to formula 1^7×(2-1)The MAC value of=128；And then filter out according to formula 2 (operating mode number is 2 to operating mode corresponding to DiffMAC minima^(2-1)=2), furthermore, use formula 7 to calculate the work from the screening of MAC matrix The potential energy value that condition is corresponding, potential energy minimum person is the accurate vibration shape.The front k=7 first order mode direction coefficient of minor structure 1/2 refers to table 2:

Table 2

Exponent number	1	2	3	4	5	6	7
								Minor structure 1	1	1	1	1	1	1	1
Minor structure 2	‐1	‐1	‐1	‐1	1	‐1	1

(7) the substructure mode vibration shape direction-adaptive genetic algorithm in identification of bridge structure block impact vibration test is used Method, identifies the Mode Shape direction coefficient that minor structure 2 is whole, and concrete operations are:

Represent the chromosome of substructure mode vibration shape direction coefficient according to formula 13 stochastic generation, situation sum is chosen for S =50.Further according to formula 14, front 7 rank of the chromosome of amendment stochastic generation are MAC matrix and minimum potential energy principal differentiation Right value；Amended gene representation is:

X '={-1-1-1-1 1-1 1 x₂₈ x₂₉…x_2n}

Next step calculates the fitness value of all situations number according to formula 11, and differentiates, chooses maximum adaptation angle value Storage is to globally optimal solution.

Carrying out roulette selection operation further, arranging select probability is P_s=0.8, enter under selecting the gene in pond to carry out The intersection of one step and variation.Wherein, this algorithm uses single-point Crossover Strategy and adaptive crossover mutation, chooses friendship according to formula 15 Fork probabilistic upper bound value is 0.9, and crossover probability lower limit is 0.7.Formed new after calculating intersection according to formula 11 after intersection The fitness value of gene also differentiates, if it is more than current globally optimal solution, substitutes, if less than globally optimal solution, putting into In population pond, and then carrying out next step variation, according to formula 16, this algorithm uses self-adaptive mutation, arranges variation general Rate higher limit is 0.2, and mutation probability lower limit is 0.001.After calculating variation according to formula 11 after variation, new gene fits Answer angle value and differentiate, if it is more than current globally optimal solution, substituting, if less than globally optimal solution, putting in population pond Entering next step iteration, it is 200 that this algorithm arranges iterations.

Identify that the evolution graph of the self-adapted genetic algorithm on each substructure mode exponent number 12 rank and 15 rank refers to accompanying drawing 6, it is known that Employing self-adapted genetic algorithm is identified, and 12 rank and 15 rank have been restrained in 8 generations and 14 generations respectively, and computational efficiency significantly carries High.

(8) overall structure flexibility matrix calculates and amount of deflection is predicted

In microwave radar data processing unit, after asking for the Mode Shape direction coefficient of minor structure 1 and minor structure 2, root Just can obtain integrally-built Mode Shape according to minor structure modal combination formula, before the full structure wherein identified, 16 first order modes refer to Fig. 7, the structure MAC matrix that the correct vibration shape of identification calculates refers to Fig. 5；Just can calculate according to flexibility computing formula further Integrally-built flexibility matrix, the full Structure Flexibility Matrix solved refers to Fig. 8；Deflection of bridge span can be predicted by this matrix, The deflection of bridge structure calculated based on microwave radar systems under the effect of evenly load 4445KN refers to Fig. 9, by this amount of deflection predictive value Compare with calculated value and can realize bridge security investigation.

Claims (9)

1. a bridge block impact vibration detection method based on microwave radar device, it is characterised in that step is:

Step one, bridge is divided into m minor structure, and one or more impulsive force input points is set in each minor structure With whole displacement output points, wherein, m >=2；

Act on impulsive force in step 2, successively the impulsive force input point in bridge piecemeal minor structure, synchronize to use microwave thunder Reach device and gather the displacement time course data that minor structure displacement output node is caused by the impulsive force of impulsive force input node；

Step 3, seek minor structure frequency response function H₁(ω), H₂(ω) ..., H_m(ω): the input joint respectively each minor structure gathered The displacement output node displacement time-histories of some impulsive force and microwave radar collection carries out signal processing, then uses arbitrary frequency response function to estimate Calculating method estimates the displacement frequency response function of each minor structure；

Step 4, frequency response function according to each minor structure, identify the modal parameter of each minor structure: estimate each minor structure respectively Frequency response function H₁(ω), H₂(ω) ..., H_m(ω) CMIF Modal Parameter Identification method is used to do Modal Parameter Identification, it is thus achieved that each sub-knot The modal parameter of structure: system poleMode zoom factorThe displacement modes vibration shapeAnd mode participation coefficientIts In, subscript s is minor structure label, s=1,2 ..., m；Subscript r is the rank number of mode identifying minor structure, r=1,2 ..., n；

Step 5, the MAC matrix of structure structure, identify k rank Mode Shape direction coefficient before each minor structure,

Step 6, employing self-adapted genetic algorithm, identify the Mode Shape direction coefficient on each minor structure whole n rank；

Step 7, the flexibility matrix of full structure calculate and amount of deflection prediction: after asking for the Mode Shape direction coefficient of each minor structure, Integrally-built Mode Shape is obtained according to minor structure modal combination formula；Use is calculated further according to flexibility computing formula Integrally-built flexibility matrix in prediction deflection of bridge span.

Bridge block impact vibration detection method the most according to claim 1, it is characterised in that described step 5 concrete Method is:

Arranging variables D iffMAC, it is 2 that structure operating mode number is helped in corresponding minor structure fusion^(m-1)k, filter out DiffMAC minima Corresponding operating mode:

$DiffMAC=\left|\right|MAC|-{\mathrm{\&Pi;}}_{i=1}^k{\mathrm{MAC}}_{ii}|---\left(1\right)$

In formula,MAC_ijThe i-th rank displacement modes vibration shape and jth component level for each operating mode Move the included angle cosine between Mode Shape, { φ_iIt is the i-th rank displacement modes vibration shape of each operating mode, { φ_jIt is the jth rank of each operating mode The displacement modes vibration shape)；| MAC | is the value of MAC matrix determinant；Value for MAC matrix diagonals line element；

Calculating the potential energy value that the operating mode of MAC matrix screening is corresponding further, corresponding operating mode number is 2^m-1, potential energy minimum person be The correct vibration shape that before full structure, k rank are corresponding.

Bridge block impact vibration detection method the most according to claim 2, it is characterised in that described step 6 concrete Method is:

A, stochastic generation represent the chromosome of substructure mode vibration shape direction coefficient:

X={x₁₁ x₁₂ … x_1nx₂₁ x₂₂ … x_2n… x_m1 x_m2 … x_(m-1)n}_1×(m-1)n (2)

In formula, x_srIn the label that subscript s is minor structure, except target minor structure, s=1,2 ..., m-1；Subscript r represents each son The rank number of mode of structure, r=1,2 ..., n；

The size choosing rank number of mode n dividing number m and minor structure according to minor structure of population number S is determined, take 50～ 200；

B, according to displacement modes vibration shape direction coefficient result in k rank before the MAC matrix identification structure of structure, amendment stochastic generation dyeing The value of (m-1) k element before body, amended chromosome is expressed as:

C, fitness value calculation also differentiate, choose maximum adaptation angle value and store globally optimal solution；

F (x)=max (-(1-p) Π_p) (4)

In formula, p is penalty,For each operating mode meter The maximum of off diagonal element in the MAC matrix calculated；Pc is penalty factor (pc=10)；

D, carry out roulette selection operation further, selective factor B P is set_S∈ (0.7-0.9), enters and selects the gene in pond to carry out Next step intersection and variation:

Crossover probability P_cWith mutation probability P_mSize be:

$P_c=\left\{\begin{array}{cc}{P}_{c1}-\frac{P_{c1}-P_{c2}}{F_{\max }-\stackrel{\&OverBar;}{F}}\left(F^{\&prime;}-\stackrel{\&OverBar;}{F}\right),& F^{\&prime;}\&GreaterEqual;\stackrel{\&OverBar;}{F}\\ P_{c1},& F^{\&prime;}<\stackrel{\&OverBar;}{F}\end{array}\right.---\left(5\right)$

$P_m=\left\{\begin{array}{cc}{P}_{m1}-\frac{P_{m1}-P_{m2}}{F_{\max }-\stackrel{\&OverBar;}{F}}\left(F-\stackrel{\&OverBar;}{F}\right),& F\&GreaterEqual;\stackrel{\&OverBar;}{F}\\ P_{c1},& F<\stackrel{\&OverBar;}{F}\end{array}\right.---\left(6\right)$

In formula, P_c1Higher limit for crossover probability；P_c2Lower limit for crossover probability；P_m1Higher limit for mutation probability；P_m2For The lower limit of mutation probability；F_maxFor the maximum of ideal adaptation degree in population；For often for the average fitness value of population；F′ For fitness value bigger in two individualities to be intersected；F is the individual fitness value that to make a variation；

E, intersect, make a variation after calculate according to formula 4 intersect after and variation after the fitness value of new gene differentiating, if It then substitutes more than current globally optimal solution, if less than globally optimal solution, putting into the iteration entering next step in population pond.

Bridge block impact vibration detection method the most according to claim 3, it is characterised in that described crossover probability upper Limit value is 0.9, and the lower limit of crossover probability is 0.7；Described mutation probability higher limit is 0.2, and the lower limit of mutation probability is 0.001。

Bridge block impact vibration detection method the most according to claim 1, it is characterised in that in described step 3, point The displacement output node displacement time-histories of the other input node impulsive force gathering each minor structure and microwave radar collection carries out signal Process includes windowing and Filtering Processing.

Bridge block impact vibration detection method the most according to claim 1, it is characterised in that in described step 3, appoints One frequency response function algorithm for estimating is H1, H2 or Hv method.

7. according to the arbitrary described bridge block impact vibration detection method of claim 16, it is characterised in that minor structure displacement The displacement time course data of output node is to be measured by microwave radar systems.

8. the microwave radar detection device of a bridge block impact vibration based on microwave radar device, it is characterised in that: bag Including impact shock device, microwave radar systems and data processing unit, described data processing unit uses claim 17 institute State bridge block impact vibration detection method, described impact shock device is acted on the impulsive force on bridge and described radar system The displacement time course data that system gathers processes, and draws the full Structure Flexibility Matrix for predicting deflection of bridge span.

Bridge block impact vibration microwave radar detection device the most according to claim 8, it is characterised in that: described microwave Radar system includes: reception antenna, launch antenna, transmitter, receiver, signal processor, Electronic Megnetic Compass, GPS module, swash Optical indicator and range finder using laser, described signal processor processes microwave signal and the receiver reception that described transmitter sends Microwave signal mixing after zero intermediate frequency signals, draw required displacement TIME HISTORY SIGNAL.