CN105956218A - Steel bridge finite element model correction method based on non-uniform temperature response monitoring value - Google Patents

Abstract

The invention discloses a steel bridge finite element model correction method based on a non-uniform temperature response monitoring value. The method mainly includes the steps of: 1) analyzing the year-round monitoring data of a steel bridge, and finding out the response rule of structure under the action of a non-uniform temperature field, 2) establishing an initial finite element model according to design data, 3) preliminarily determining the horizontal stiffness of a steel bridge bearing by iteration method, 4) performing sensitivity analysis on the steel bridge on the basis of steel bridge temperature response data, and determining a design variable having a high correlation coefficient with actually measured data, and 5) conducting optimization analysis on the steel bridge finite element model by reducing the difference between a finite element calculation result and the actually measured data. Compared with the universally adopted finite element model correction method based on experimental modal data and other dynamic response results, the method provided by the invention has the advantages of simplicity and accuracy, low cost and good safety.

Description

Steel bridge correction method for finite element model based on non-uniform temperature response monitor value

Technical field

The invention belongs to civil engineering numerical simulation analysis field, relate to a kind of based on non-uniform temperature response monitor value Steel bridge correction method for finite element model.

Background technology

Along with the high speed development of national economy, the large bridge construction speed of China is swift and violent.Steel bridge because span ability is strong, execute Degree of the being widely used in bridges such as work speed is fast, play the role of key in local environment, economic and social life.Steel Spanning degree is relatively big, and component is of a great variety, and topology layout is complicated.Unfavorable because of environmental load effect, fatigue effect and material aging etc. The impact of factor, bridge will inevitably occur various defect during long-term use, cause structure partial key structure Part damage accumulation.As the phenomenons such as the rigidity of fruit structure and depression of bearing force can not be found in time, and keeped in repair timely, no Only can affect the normal use of structure, it could even be possible to catastrophic failures such as destroying suddenly or collapse can occur.

It is reasonable prediction bridge safty, durability by finite element numerical simulation technology degree of foundation bridge finite element model With the most universal methods such as anti-seismic performances.But the various defects occurred in life-time service due to bridge so that according to design The FEM (finite element) model that data are set up and structure also exist boundary condition, material and cross section parameter, quality and load assignment equal error, Therefore FEM (finite element) model is difficult to reflect the behavior under structure load effect under arms and duty exactly, it is necessary to based on strong FEM (finite element) model is modified by the response monitor value of health prison examining system, in order to by revised FEM (finite element) model to bridge Carry out further security evaluation.

The dynamic response results such as current commonly used Modal Test data carry out the correction of FEM (finite element) model to steel bridge, in recent years Although it is good to carry out widely used mould measurement safety based on environmental excitation, and do not affect normally using of bridge, but also Existence is a lot of not enough: 1) environmental excitation exists Unknown worm, such as, comprise various noise jamming；2) Modal Parameter Identification is based on mould State theory hypothesis；3) local acknowledgement of structure cannot be determined；4) accurate Modal Parameter Identification requires higher, and data processing amount is relatively Greatly, it is unfavorable for that engineering staff grasps；5) number of sensors requires more.

Steel bridge temperature field on the impact of structure mainly by following three kinds of forms: heat radiation, conduction of heat and thermal convection current.Due to Mutually blocking between the change of solar radiation angle and component, the steel bridge major part whole year moment is mainly by the shadow of Complex Temperature Field Ringing, owing to the structural strain under temperature action and dynamic respond can be obvious, therefore the present invention proposes based on non-uniform temperature The steel bridge correction method for finite element model of response monitor value, the method is simply accurate, and expense is relatively low, and safety is good.

Summary of the invention

Technical problem: the present invention provide a kind of simple accurately, cost is relatively low, good the responding based on non-uniform temperature of safety The steel bridge correction method for finite element model of monitor value.

Technical scheme: the steel bridge correction method for finite element model based on non-uniform temperature response monitor value of the present invention, bag Include following steps:

1) according to the boundary condition error existed when setting up the initial FEM (finite element) model of steel bridge, member section character error, material Error and the structure condition of contact distortion of material character determine that the cloth location of steel bridge temperature sensor and strain, displacement transducer is put；

2) collect and process the temperature sensor data of steel bridge, strain transducer data and displacement transducer data, data Process method particularly includes: utilize WAVELET PACKET DECOMPOSITION technology to reject the dynamic strain composition in strain monitoring data, then to same section The static strain that two strain transducers on face record is averaged, as the axial strain of this component；To many on same bridge pier The displacement data of individual displacement transducer is averaged the displacement as this bridge pier；

3) based on ANSYS FEM software, according to geometrical scale, member section and position in design data Put, material character sets up the initial FEM (finite element) model of steel bridge, and idiographic flow is: first, sets up limited with steel bridge construction node coordinate All nodes of meta-model；Then, all lists of FEM (finite element) model are set up according to design section, material parameter and position Unit；Finally, according to constraints, node is applied coupling and constraint, obtain the initial FEM (finite element) model of steel bridge；

4) the Sensor monitoring value on the date of wind speed w≤[w] level, i.e. steel bridge average temperature data { T} are first filtered out^T, And the displacement structure data { δ } in correspondence moment^TWith strain-responsive data { ε }^T, in the date filtered out, finally choose the temperature difference Big one day, and this day all moment temperature, strain and displacement transducer monitor value are converted into relative value by absolute value, and According to temperature sensor monitors value, do not laid the member temperature value at temperature sensor by linear difference, as finally Temperature field data, wherein, [w] is one day wind speed scale limit value, and the total moment number in this day is T；

5) based on actual measurement support displacement, use iterative method tentatively to revise described step 4) in the initial finite element of steel bridge that obtains The bearing horizontal rigidity of model, idiographic flow is: the bearing horizontal rigidity initial value first setting all bridge piers is K₀, then Reduced the difference of displacement structure and measured value by FEM calculation, obtaining backed horizontal rigidity is K_p；

6) FEM (finite element) model initial to steel bridge carries out sensitive analysis, and idiographic flow is:

(6-a) error variance parameterized probability advantage model is set up: linear expansion coefficient { EX} will be included^T, bullet Property modulus { E}^TWith mass density { ρ }^T, component axial rigidity { K_z}^THorizontal rigidity { K with movable bearing support_s}^TIt is defined as probability to have The input variable of finite element analysis, and the range of variation of specification error variable and probability distribution, will lay displacement transducer The component axial strain of support displacement and laying strain transducer is defined as the output parameter of Finite Element Analysis of Probability；

(6-b) input variable that in Monte Carlo simulation technique, Latin hypercube is specified is utilized at described (6-a) Carry out n times stochastic sampling in range of variation, often complete a stochastic sampling just FEM (finite element) model initial to described steel bridge and carry out one Secondary FEM calculation, obtains stochastic inputs variable { IV}^T=(IV₁, IV₂, IV_x...IV_a) and random output variable { OV}^T= (OV₁, OV₂, OV_y...OV_bLinearly dependent coefficient matrix between), wherein N is frequency in sampling, and a is stochastic inputs total number of variable, B is random output variable sum, { IV}^TFor stochastic inputs variable complete or collected works, IV_xRepresent x-th stochastic inputs variable, { OV}^TFor with Machine output variable complete or collected works, OV_yRepresenting the random output variable of y-th, x is stochastic inputs variable sequence number, and y is random output variable sequence Number；

(6-c) by stochastic inputs variable IV_xWith corresponding random output variable OV_yThe accumulated value of correlation coefficient quadratic sumDescending it is ranked up, wherein r_{X, y}For stochastic inputs variable IV_xWith random output variable OV_yPhase relation Number, takes correlation coefficient accumulated value and optimizes, as next step, the stochastic inputs variable analyzed more than the random input parameters of [Sum]；

7) according to step 4) T load case arranging of the temperature field data FEM (finite element) model initial to steel bridge that finally give Apply temperature load respectively, then according to following steps carry out FEM calculation: optimize the function in module initially with ANSYS Object function is optimized by approximatioss, then uses single order optimizing method to enter object function on the basis of functional approaching optimization Row further optimizes, and obtains revised FEM (finite element) model based on non-uniform temperature response monitor value, the mesh of two suboptimization Scalar functions is:

$Obj=\sqrt{\underset{k}{\overset{T}{\Σ}}\underset{i}{\overset{m}{\Σ}}{\left({\mathrm{\ϵ}}_{cal}\right(i,k)-{\mathrm{\ϵ}}_{mea}(i,k\left)\right)}^2+c\underset{k}{\overset{T}{\Σ}}\underset{j}{\overset{n}{\Σ}}{\left({\mathrm{\δ}}_{cal}\right(j,k)-{\mathrm{\δ}}_{mea}(j,k\left)\right)}^2}$

Wherein m represents the total number of components being disposed with strain transducer, and n represents the total bearing number being disposed with displacement transducer, ε_cal(i, k) and δ_cal(j, the displacement of the strain calculation value and jth bearing that k) represent kth moment i-th component respectively calculates Value, ε_mea(i, k) and δ_mea(j k) represents the strain monitoring value of kth moment i-th component and the displacement prison of jth bearing respectively Measured value, c is the weight of displacement；

The constraints of two suboptimization all includes: the displacement { d} of bearing residing for displacement transducer^T _min≤{d}^T≤{d}^T _max Axial strain { S with strain transducer place rod member_A}^T _min≤{S_A}^T≤{S_A}^T _max。

Further, in the inventive method, described step 4) in by maximum one day all moment different sensors prison of the temperature difference Measured value was converted into relative value's by absolute value method particularly includes: by this day temperature sensor monitors value standard variance minimal instant Data are as initial value: temperature is { T}^T, displacement is { δ₀}^T, strain as { ε₀}^T, the data in other moment and described initial value phase Subtract the result taken as relative value, i.e. temperature gap Δ T_{H, k}, and strain difference ε in corresponding moment_mea(i, k), shift differences δ_mea(j, k), wherein k=1,2,3...T, h represent the h lay temperature sensor component, i represent i-th lay strain pass The component of sensor, j represents jth and lays the bearing of displacement transducer；

Described according to temperature sensor monitors value, do not laid the member temperature at temperature sensor by linear difference Value method particularly includes: by two section temperature values laying temperature sensor, obtain non-cloth by place highly linear difference If the section temperature distribution of temperature sensor, obtain the temperature along depth of section direction by cross section upper and lower side temperature value linear difference Degree.

Further, in the inventive method, described step 5) in the idiographic flow of FEM calculation be:

A () is according to step 4) the temperature field data FEM (finite element) model initial to steel bridge that finally give arrange T load case Apply temperature load respectively, obtain the displacement δ of kth operating mode jth bearing_cal(j, k), k=1,2,3 ... T；

B () is according to the bearing horizontal rigidity under following formula calculating pth step iteration:

Wherein n is the total bearing number being disposed with displacement transducer, K_pFor the horizontal rigidity of bearing during pth time iteration, time initial, take K₀, δ_mea(j k) represents actual monitoring displacement；

C () makes residual errorJudge whether s≤[s] sets up, the most then iteration terminates, by this Time K_pAs bearing horizontal rigidity；If it is not, then return step (b).

Beneficial effect: the present invention compared with prior art, has the advantage that

(1) the steel bridge correction method for finite element model that the present invention proposes makes full use of observed temperature and structural response, input Clearly, method stability and robustness are good in output.The temperature effects that structure is subject to is to change in the moment, the knot that temperature effects produces Structure response also changes in the moment, therefore can be repaiied FEM (finite element) model by the measured data using multiple moment simultaneously Just, with the error avoiding some time data distortion to produce, can also use simultaneously annual multiple moment different temperature fields and Revised FEM (finite element) model is verified by its response results.

(2) steel bridge linear expansion coefficient is big, and under temperature action, change is substantially, and monitoring accuracy is high.Interference, environment due to noise Excitation insufficient and affected by ambient temperature, humidity, wind load, traffic loading environmental factors etc., based on environment swash There is complicated and faint shortcoming in the mould measurement structural vibration response signal encouraged, and the artificial excitations such as hammering method cause bridge Extra damage.The present invention uses the variations in temperature temperature field as steel bridge of the bigger natural law of day and night temperature, and the strain of structure rings Not only should change substantially with dynamic respond, mode surveyed by its cloth also will not bring damage to structure.

(3) temperature effects parameter is relatively low to hardware requirements such as sensors, and data process simplicity, and monitoring scheme is easy to implement. And mould measurement model modification method of based on environmental excitation exists the impalpable problem of modal parameter, temperature effects uses Response parameter (support displacement and Structural Static strain) is the most easily monitored, and higher with temperature dependency.Although strain monitoring Value includes static strain and dynamic strain, but the dynamic strain that vehicle produces can utilize WAVELET PACKET DECOMPOSITION technology to reject easily.

(4) often it is difficult to revise the damage of Local Members based on dynamic response data such as mode, and steel bridge is in annual temperature The strain of field action lower member and the change in displacement of bearing are obvious, and this modification method can prop up for structure rapid wear position and bridge Seat lays monitoring point, it is achieved to steel bridge vulnerable components axial rigidity and the fine correction of bearing horizontal rigidity.

(5) model revised based on temperature control analysis and optimization more can reflect structure temporal behavior, can be follow-up knot Structure health monitoring provides foundation.When some component persistently produces damage accumulation because of the impact of environmental load, its cross-sectional properties and Connection stiffness etc. all may occur deviation again, thus causes Monitoring Data and the FEM (finite element) model meter under the effect of identical temperature field Calculate result and difference occurs, therefore use the safe condition of revised FEM (finite element) model sustainable assessment bridge structure.

Accompanying drawing explanation

Fig. 1 is the flow chart of the inventive method.

Detailed description of the invention

Below in conjunction with Figure of description, the present invention is described in detail.

To steel bridge laying temperature sensor and strain, displacement transducer:

Owing to steel bridge is along less to variations in temperature difference along bridge, temperature sensor answers emphasis high along structure at steel bridge cross section Lay on degree direction different component, as arch rib at the typical sections such as arch bridge vault, arch springing, suspension rod, floorings and All lay temperature sensor on each component of transverse and longitudinal girder system, for tower body at the typical section such as Cable-stayed Bridge Pylon, span centre, hang All lay temperature sensor on each component of rope, floorings and transverse and longitudinal girder system thereof, be used for monitoring steel bridge temperature field along bridge allusion quotation The change in type depth of section direction.Additionally, when each member section lays temperature sensor, should be along depth of section direction upper and lower side Lay respectively, for monitoring the temperature field change along member section direction.

To steel bridge easily occurring component that in corrosion and the component of damage, design data, stress is bigger along depth of section direction Upper and lower side lays strain transducer, and support position each to steel bridge all lays displacement transducer.The error of member section character is main Axial rigidity { K for component_z}^T.Some component is affected by the ambient there is corrosion problem, also has some component growing Occur that local damage is accumulated under phase load effect, cause its section rigidity and connection status etc. all may produce relatively with original state Big difference, and the temperature-responsive of component is affected relatively big by axial strain, therefore in described member section short transverse at upper and lower side Lay strain transducer respectively, the strain monitored is averaged to eliminate the impact of bending strain, to obtain practical structures Actual measurement axial strain under temperature field acts on.The error of boundary condition is mainly the horizontal rigidity { K of movable bearing support_s}^T.Steel bridge All there is certain horizontal rigidity in the rubber support used or spherical bearing etc., but the horizontal rigidity value of this large-scale bearing is usual It is unknown quantity, the result of calculation of FEM (finite element) model is brought the biggest error, needs uniform displacement sensor at each bearing, with Obtain practical structures actual displacement under temperature field acts on.It is to guarantee the effectiveness of data simultaneously, upper and lower at same bridge pier It is both needed to cloth displacement sensor, as mutually correction on trip bearing.Additionally, due to the stress shape of the bridge chord member being connected with bearing State is very big with the dependency of bearing, is also easily subject to the corrosion of the surrounding enviroment such as sea water simultaneously, typically answers cloth on this type of rod member If strain transducer.

2) steel bridge construction temperature, support displacement and the annual Monitoring Data of key member strain, sensor are collected and process Sample frequency typically the highest so that total monitor sample capacity is excessive.Owing to the temperature value in adjacent 20min changes less, can Utilize the monitoring meansigma methods in every 20min to represent the monitor value of this period, therefore can calculate total moment number T=of every day 72.The data processing method of different sensors is: strain data comprises static strain composition and dynamic strain composition simultaneously, should reject Wherein dynamic strain composition.Dynamic strain composition (i.e. spined portion) is mainly caused by train load, and each train can produce through later One bur, and the frequency of frequency static strain the to be far above composition of dynamic strain composition, difference of them is relatively big, therefore first with WAVELET PACKET DECOMPOSITION technology rejects the dynamic strain composition in described strain monitoring data, then to two strain sensings on same cross section The static strain that device records is averaged, as the axial strain of this component, the displacement to displacement transducers multiple on same bridge pier Data are averaged the displacement as this bridge pier.

3) based on ANSYS FEM software, according to geometrical scale, member section and position in design data Put, material character sets up the initial FEM (finite element) model of steel bridge, and idiographic flow is: first, sets up limited with steel bridge construction node coordinate All nodes of meta-model, then, set up all lists of FEM (finite element) model according to design section, material parameter and position Unit, finally, applies coupling and constraint according to constraints to node, obtains the initial FEM (finite element) model of steel bridge.For steel bridge hangs Bar, longeron, crossbeam, truss etc. are proposed with 6DOF beam element, are proposed with 4 node shell units for floorings etc., for Bridge pier is proposed with 8 node hexahedral elements.

4) first filtering out the Sensor monitoring value on the date of wind speed w≤[w] level, [w] is one day wind speed scale limit value, i.e. Steel bridge average temperature data { T}^T, and the displacement structure data { δ } in corresponding moment^TWith strain-responsive data { ε }^T, finally in screening Choose in the natural law gone out that the temperature difference is maximum one day, using the data of this day temperature sensor monitors value standard variance minimal instant as Initial value: temperature is { T}^T, displacement is { δ₀}^T, strain as { ε₀}^T, the data in other moment and described initial value subtract each other the knot taken Really as relative value, i.e. temperature gap Δ T_{H, k}, and strain difference ε in corresponding moment_mea(i, k), shift differences δ_mea(j, k), According to temperature sensor monitors value, do not laid the section temperature distribution of temperature sensor by place highly linear difference, logical Crossing cross section upper and lower side temperature value linear difference and obtain the temperature along depth of section direction, wherein, the total moment number in this day is T, k=1,2,3...T, h represent the h lay temperature sensor component, i represent i-th lay strain transducer component, J represents jth and lays the bearing of displacement transducer.

5) iterative method is used tentatively to revise bearing horizontal rigidity based on actual measurement support displacement.Bridge use spherical bearing or All there is certain horizontal rigidity in rubber support, but its numerical value is difficult to determine, and the biggest on result of finite element impact.For accelerating The follow-up optimization efficiency optimizing analysis, based on actual measurement support displacement, uses iterative method to carry out the bearing horizontal rigidity of steel bridge tentatively Revise, be divided into following step: (5-a) is according to step 4) the steel bridge temperature data that obtains to big across steel bridge initial finite element mould Type arranges T load case and applies temperature load, obtains the displacement δ of kth operating mode jth bearing_cal(j, k), k=1,2,3 ... T；(5-b) according to the bearing horizontal rigidity under following formula calculating pth step iteration: Wherein n represents the total bearing number being disposed with displacement transducer, K_pFor the horizontal rigidity of bearing during pth time iteration, time initial, take K₀, δ_mea(j k) represents actual monitoring displacement；(5-c) residual error is madeJudge whether s≤[s] sets up, institute Stating [s] suggestion value is 0.15 × T, it is believed that be closer to measured displacements, if so, less than the displacement calculated after iteration during this value Then iteration terminates, by K now_pAs bearing horizontal rigidity；If it is not, then return step (5-b).Use iterative method to bearing water After flat rigidity is tentatively revised, then in step 7) optimize in analysis and finely revise.

6) based on non-uniform temperature field structural response measured data, steel bridge is carried out sensitive analysis.Due to boundary condition Error, the error of component physical parameter and structure condition of contact error etc. all may affect the calculating knot of initial FEM (finite element) model Really, it is necessary to determine the above-mentioned error influence degree to result of calculation (support displacement and structural strain) by sensitive analysis. It is divided into following step: (6-a) sets up error variance parameterized probability advantage model: will include linear expansion coefficient {EX}^T, elastic modelling quantity { E}^TWith mass density { ρ }^T, component axial rigidity { K_z}^THorizontal rigidity { K with movable bearing support_s}^TDefinition For the input variable of Finite Element Analysis of Probability, and the range of variation of specification error variable and probability distribution, displacement will be laid The component axial strain of the support displacement of sensor and laying strain transducer is defined as the output parameter of Finite Element Analysis of Probability； (6-b) the input variable range of variation that in Monte Carlo simulation technique, Latin hypercube is specified is utilized at described (6-a) Inside carry out n times stochastic sampling, often complete a stochastic sampling just FEM (finite element) model initial to described steel bridge and carry out a finite element Calculating, described N suggestion takes more than 5000 times, obtains stochastic inputs variable { IV}^T=(IV₁, IV₂, IV_x...IV_a) and random output Variable { OV}^T=(OV₁, OV₂, OV_y...OV_bLinearly dependent coefficient matrix between), wherein N is frequency in sampling, and a is the most defeated Entering total number of variable, b is random output variable sum, { IV}^TFor stochastic inputs variable complete or collected works, IV_xRepresent x-th stochastic inputs to become Amount, { OV}^TFor random output variable complete or collected works, OV_yRepresenting the random output variable of y-th, x is stochastic inputs variable sequence number, y be with Machine output variable sequence number；(6-c) by stochastic inputs variable IV_xWith corresponding random output variable OV_yCorrelation coefficient quadratic sum Accumulated valueDescending it is ranked up, wherein r_{X, y}For stochastic inputs variable IV_xWith random output variable OV_y's Correlation coefficient, takes correlation coefficient accumulated value and optimizes, as next step, the stochastic inputs analyzed more than the random input parameters of [Sum] Variable, described [Sum] by project planner according to the stochastic inputs variable number a chosen in engineering practice and random output variable Number b is set, when | r_{X, y}| closer to 1, illustrate that the dependency of two variablees is the highest.

7) steel bridge FEM (finite element) model is optimized by the difference by reducing result of calculation and measured data.According to step 4) The temperature field data FEM (finite element) model initial to steel bridge finally given arranges T load case and applies temperature load respectively, described FEM calculation follows the steps below: object function is carried out by the functional approaching optimized in module initially with ANSYS Optimize, then use single order optimizing method that object function is further optimized on the basis of functional approaching optimization, obtain Revised FEM (finite element) model based on non-uniform temperature response monitor value, the object function of two suboptimization is:

$Obj=\sqrt{\underset{k}{\overset{T}{\Σ}}\underset{i}{\overset{m}{\Σ}}{\left({\mathrm{\ϵ}}_{cal}\right(i,k)-{\mathrm{\ϵ}}_{mea}(i,k\left)\right)}^2+c\underset{k}{\overset{T}{\Σ}}\underset{j}{\overset{n}{\Σ}}{\left({\mathrm{\δ}}_{cal}\right(j,k)-{\mathrm{\δ}}_{mea}(j,k\left)\right)}^2}$

Wherein m represents the total number of components being disposed with strain transducer, and n represents the total bearing number being disposed with displacement transducer, ε_cal(i, k) and δ_cal(j, the displacement of the strain calculation value and jth bearing that k) represent kth moment i-th component respectively calculates Value, ε_mea(i, k) and δ_mea(j k) represents the strain monitoring value of kth moment i-th component and the displacement prison of jth bearing respectively Measured value, c is as the weight of displacement.

The constraints of two suboptimization all includes: the displacement { d} of bearing residing for displacement transducer^T _min≤{d}^T≤{d}^T _max Axial strain { S with strain transducer place rod member_A}^T _min≤{S_A}^T≤{S_A}^T _max, according to monitoring result by state variable Within limit value is defined on normal range.

8) revised FEM (finite element) model is verified by the measured data using other temperature field moment.Due to the whole year Temperature field is being continually changing, and temperature field the most in the same time therefore can be used to be analyzed revised FEM (finite element) model, will meter Calculate result to contrast with corresponding response monitor value, to verify the correctness of steel bridge FEM (finite element) model.Idiographic flow is: (8-a) exists Temperature field natural law is chosen one day for verifying, it is minimum that initial time chooses temperature sensor monitors value standard variance in a day In the moment, the data in other moment and initial value subtract each other and take relative value, obtain temperature gap Δ T '_{H, k}, and the strain in corresponding moment Difference ε '_mea(j, k), shift differences δ '_mea(j, k), k=1,2,3...T ', i represent i-th lay strain transducer component, J represents jth and lays the bearing of displacement transducer, total moment number that T ' is this day；(8-b) step 7 is used) after the correction set up FEM (finite element) model, arranges the individual load case of T ', each component intensification Δ to laying temperature sensor to steel bridge initial model T′_{H, k}, the section temperature distribution not laying temperature sensor can be pressed by two typical section temperature values laying temperature sensor Place highly linear difference obtains, and the temperature along depth of section direction can carry out linear difference, kth by upper and lower side temperature value Condition calculating value with the residual error of measured value is:

${\mathrm{\ΔR}}_k=\sqrt{\underset{i}{\overset{m}{\Σ}}{\left({\mathrm{\ϵ}}_{cal}^{\′}\right(i,k)-{\mathrm{\ϵ}}_{mea}^{\′}(i,k\left)\right)}^2+c\underset{j}{\overset{n}{\Σ}}{\left({\mathrm{\δ}}_{cal}^{\′}\right(j,k)-{\mathrm{\δ}}_{mea}^{\′}(j,k\left)\right)}^2}$

Wherein m represents the total number of components being disposed with strain transducer, and n represents the total bearing number being disposed with displacement transducer, ε′_cal(i, k) with δ '_cal(j, the displacement of the strain calculation value and jth bearing that k) represent kth moment i-th component respectively calculates Value.

Above-described embodiment is only the preferred embodiment of the present invention, it should be pointed out that: for the ordinary skill of the art For personnel, under the premise without departing from the principles of the invention, it is also possible to making some improvement and equivalent, these are to the present invention Claim improve with equivalent after technical scheme, each fall within protection scope of the present invention.

Claims (3)

1. a steel bridge correction method for finite element model based on non-uniform temperature response monitor value, it is characterised in that the method Comprise the following steps:

1) according to the boundary condition error existed when setting up the initial FEM (finite element) model of steel bridge, member section character error, material The error of matter and structure condition of contact distortion determine that the cloth location of steel bridge temperature sensor and strain, displacement transducer is put；

2) collecting and process the temperature sensor data of steel bridge, strain transducer data and displacement transducer data, data process Method particularly includes: utilize WAVELET PACKET DECOMPOSITION technology to reject the dynamic strain composition in strain monitoring data, then on same cross section The static strain that records of two strain transducers average, as the axial strain of this component；To positions multiple on same bridge pier The displacement data of displacement sensor is averaged the displacement as this bridge pier；

3) based on ANSYS FEM software, according to the geometrical scale in design data, member section and position, material Material character sets up the initial FEM (finite element) model of steel bridge, and idiographic flow is: first, sets up FEM (finite element) model with steel bridge construction node coordinate All nodes；Then, all unit of FEM (finite element) model are set up according to design section, material parameter and position；? After, according to constraints, node is applied coupling and constraint, obtain the initial FEM (finite element) model of steel bridge；

4) the Sensor monitoring value on the date of wind speed w≤[w] level, i.e. steel bridge average temperature data { T} are first filtered out^T, and corresponding The displacement structure data { δ } in moment^TWith strain-responsive data { ε }^T, in the date filtered out, finally choose the one of temperature difference maximum My god, and this day all moment temperature, strain and displacement transducer monitor value are converted into relative value by absolute value, and according to temperature Degree Sensor monitoring value, is not laid the member temperature value at temperature sensor by linear difference, as final temperature Field data, wherein, [w] is one day wind speed scale limit value, and the total moment number in this day is T；

5) based on actual measurement support displacement, use iterative method tentatively to revise described step 4) in the initial FEM (finite element) model of steel bridge that obtains Bearing horizontal rigidity, idiographic flow is: the bearing horizontal rigidity initial value first setting all bridge piers is K₀, then pass through FEM calculation reduces the difference of displacement structure and measured value, and obtaining backed horizontal rigidity is K_p；

6) FEM (finite element) model initial to steel bridge carries out sensitive analysis, and idiographic flow is:

(6-a) error variance parameterized probability advantage model is set up: linear expansion coefficient { EX} will be included^T, elastic modelling quantity {E}^TWith mass density { ρ }^T, component axial rigidity { K_z}^THorizontal rigidity { K with movable bearing support_s}^TIt is defined as probability advantage to divide The input variable of analysis, and the range of variation of specification error variable and probability distribution, will lay a seat of displacement transducer The component axial strain moving and laying strain transducer is defined as the output parameter of Finite Element Analysis of Probability；

(6-b) the input variable variation that in Monte Carlo simulation technique, Latin hypercube is specified is utilized at described (6-a) In the range of carry out n times stochastic sampling, often completing a stochastic sampling just FEM (finite element) model initial to described steel bridge once has Limit unit calculates, and obtains stochastic inputs variable { IV}^T=(IV₁, IV₂, IV_x...IV_a) and random output variable { OV}^T=(OV₁, OV₂, OV_y...OV_bLinearly dependent coefficient matrix between), wherein N is frequency in sampling, and a is stochastic inputs total number of variable, b be with Machine output variable sum, { IV}^TFor stochastic inputs variable complete or collected works, IV_xRepresent x-th stochastic inputs variable, { OV}^TFor the most defeated Go out variable complete or collected works, OV_yRepresenting the random output variable of y-th, x is stochastic inputs variable sequence number, and y is random output variable sequence number；

(6-c) by stochastic inputs variable IV_xWith corresponding random output variable OV_yThe accumulated value of correlation coefficient quadratic sumDescending it is ranked up, wherein r_{X, y}For stochastic inputs variable IV_xWith random output variable OV_yPhase relation Number, takes correlation coefficient accumulated value and optimizes, as next step, the stochastic inputs variable analyzed more than the random input parameters of [Sum]；

7) according to step 4) T load case arranging of the temperature field data FEM (finite element) model initial to steel bridge that finally give distinguish Apply temperature load, then according to following steps carry out FEM calculation: optimize the function approximation in module initially with ANSYS Object function is optimized by method, then uses single order optimizing method to enter object function on the basis of functional approaching optimization The optimization of one step, obtains revised FEM (finite element) model based on non-uniform temperature response monitor value, the target letter of two suboptimization Number is:

$Obj=\sqrt{\underset{k}{\overset{T}{\Σ}}\underset{i}{\overset{m}{\Σ}}{\left({\mathrm{\ϵ}}_{cal}\right(i,k)-{\mathrm{\ϵ}}_{mea}(i,k\left)\right)}^2+c\underset{k}{\overset{T}{\Σ}}\underset{j}{\overset{n}{\Σ}}{\left({\mathrm{\δ}}_{cal}\right(j,k)-{\mathrm{\δ}}_{mea}(j,k\left)\right)}^2}$

Wherein m represents the total number of components being disposed with strain transducer, and n represents the total bearing number being disposed with displacement transducer, ε_cal (i, k) and δ_cal(j, k) represents the strain calculation value of kth moment i-th component and the displacement value of calculation of jth bearing respectively, ε_mea(i, k) and δ_mea(j k) represents strain monitoring value and the displacement monitoring of jth bearing of kth moment i-th component respectively Value, c is the weight of displacement；

The constraints of two suboptimization all includes: the displacement { d} of bearing residing for displacement transducer^T _min≤{d}^T≤{d}^T _maxWith should Become the axial strain { S of sensor place rod member_A}^T _min≤{S_A}^T≤{S_A}^T _max。

Steel bridge correction method for finite element model based on non-uniform temperature response monitor value the most according to claim 1, its Be characterised by, described step 4) in maximum for the temperature difference one day all moment different sensors monitor value are converted into by absolute value relatively Value method particularly includes: using the data of this day temperature sensor monitors value standard variance minimal instant as initial value: temperature is {T}^T, displacement is { δ₀}^T, strain as { ε₀}^T, the data in other moment and described initial value subtract each other the result taken as relative value, I.e. temperature gap Δ T_{H, k}, and strain difference ε in corresponding moment_mea(i, k), shift differences δ_mea(j, k), wherein k=1,2, 3...T, h represents the h component laying temperature sensor, and i represents i-th and lays the component of strain transducer, and j represents jth The bearing of individual laying displacement transducer；

Described according to temperature sensor monitors value, do not laid member temperature value at temperature sensor by linear difference Method particularly includes: by two section temperature values laying temperature sensor, do not laid temperature by place highly linear difference The section temperature distribution of degree sensor, obtains the temperature along depth of section direction by cross section upper and lower side temperature value linear difference.

Steel bridge correction method for finite element model based on non-uniform temperature response monitor value the most according to claim 2, its Be characterised by, described step 5) in the idiographic flow of FEM calculation be:

A () is according to step 4) the temperature field data FEM (finite element) model initial to steel bridge that finally give arrange T load case respectively Apply temperature load, obtain the displacement δ of kth operating mode jth bearing_cal(j, k), k=1,2,3 ... T；

B () is according to the bearing horizontal rigidity under following formula calculating pth step iteration:

Wherein n is the total bearing number being disposed with displacement transducer, K_pFor pth The horizontal rigidity of bearing during secondary iteration, takes K time initial₀, δ_mea(j k) represents actual monitoring displacement；

C () makes residual errorJudge whether s≤[s] sets up, the most then iteration terminates, by K now_p As bearing horizontal rigidity；If it is not, then return step (b).