CN104732098A - Early warning method for deterioration of bearing capacity of railway steel truss arched bridge girder - Google Patents

Abstract

The invention discloses an early warning method for deterioration of bearing capacity of a railway steel truss arched bridge girder. The method comprises the following steps that 1, temperature data and strain data are collected; 2, components, influenced by train loads, in the strain data are removed; 3, main components of the temperature data and temperature difference data are extracted; 4, a mathematic model between the main components of the strain data and the main components of the temperature data and the temperature difference data is built; 5, the mathematic model is utilized for calculating a residual error between measured strain data and analog strain data; 6, the bearing capacity of the girder is evaluated through the strain residual error. According to the early warning method, the early warning in the deterioration process of the bearing capacity of the railway steel truss arched bridge girder is achieved, and good project practical value is achieved.

Description

A kind of method for early warning of Railway Steel Truss arch bridge girder load-resisting capacity degradation

Technical field

The present invention relates to science of bridge building, particularly a kind of load carrying capacity of bridge degeneration method for early warning.

Background technology

In long-span bridges, strain and cause primarily of temperature load, there are some researches show that diurnal variation temperature field and seasonal variations temperature field can cause significant strain level, significantly more than the strain level caused by vehicular load.Therefore there is certain correlativity between strain and temperature load, this correlativity can be used for the load-carrying properties characterizing long-span bridges girder; If this correlativity generation ANOMALOUS VARIATIONS, then characterize the degeneration of long-span bridges girder load-carrying properties.The multiple linear regression model that existing researchist utilizes the correlativity between strain and temperature to assess to set up bridge main beam load-carrying properties.

But there is following defect in existing achievement in research at present: (1) at present existing achievement in research only pays close attention to the correlation properties between strain and temperature, and does not consider to strain the correlation properties between the temperature difference.But in fact the span centre position of steel arch purlin bridge construction form exists very large rise, this rise certainly exists the comparatively large vertical temperature difference, the bar member section of steel arch purlin bridge construction simultaneously also exists the very large temperature difference, causes the strain caused by the temperature difference can not be ignored; (2) research at present is only confined to strain the correlation analysis with single temperature influence factor, and the strain of in fact a certain measuring point is produced by the joint effect of different measuring points temperature and the temperature difference; (3) early warning during Railway Steel Truss arch bridge girder load-resisting capacity degradation is not yet realized at present.

Therefore, be necessary to study a kind of Railway Steel Truss arch bridge girder load-resisting capacity degradation method for early warning, the method can take into full account the correlation properties between strain and temperature, the vertical temperature difference, section temperature difference, and the acting in conjunction of other measuring point temperature and the temperature difference is on the impact analyzing measuring point strain.

Summary of the invention

Goal of the invention: in order to overcome the deficiencies in the prior art, the invention provides a kind of Railway Steel Truss arch bridge girder load-resisting capacity degradation method for early warning, considering not comprehensively and the technical matters lacking early warning for solving the existing research to the correlativity between bridge strain and temperature field.

Technical scheme: for achieving the above object, the technical solution used in the present invention is:

A method for early warning for Railway Steel Truss arch bridge girder load-resisting capacity degradation, comprises the following steps that order performs:

Step 1, collecting temperature data and strain data:

Choose the top boom of steel truss arched bridge girder span centre position, diagonal web member, lower boom and bridge floor chord member four components, the equal set temperature sensor of upstream side in the centre position of these four components and downstream and strain transducer carry out temperature acquisition and strain acquirement, in order to ensure that the temperature data gathered can reflect Seasonal variation preferably, acquisition time length L should be more than or equal to eight months, sample frequency selects a fixed frequency point in 1Hz to 10Hz, both ensure that data cover was wide and data total amount is unlikely to too huge;

Wherein the temperature acquisition result in top boom upstream side and downstream adopts T ₁and T ₂represent, the temperature acquisition result in diagonal web member upstream side and downstream adopts T ₃and T ₄represent, the temperature acquisition result in lower boom upstream side and downstream adopts T ₅and T ₆represent, the temperature acquisition result in bridge floor chord member upstream side and downstream adopts T ₇and T ₈represent; The temperature difference between different acquisition position adopts T _ijrepresent, definition of T _ij=T _i-T _j, wherein i=1,2 ..., 8, j=1,2 ..., 8, and i ≠ j;

Wherein the strain acquirement result in top boom upstream side and downstream adopts Y ₁and Y ₂represent, the strain acquirement result in diagonal web member upstream side and downstream adopts Y ₃and Y ₄represent, the strain acquirement result in lower boom upstream side and downstream adopts Y ₅and Y ₆represent, the strain acquirement result in bridge floor chord member upstream side and downstream adopts Y ₇and Y ₈represent;

The composition affected by train load in step 2, rejecting strain data:

Because strain data is caused jointly by train load and temperature load, the strain affected by train load becomes to belong to radio-frequency component, the strain affected by temperature load becomes to belong to low-frequency component, therefore adopts Wavelet Packet Technique to reject the strain composition affected by train load.Wavelet Packet Technique is a kind of analytical approach that becomes more meticulous frequency band and signal can being selected adaptively to match according to characteristics of signals and analysis requirement, the low-frequency component of signal and radio-frequency component can be decomposed, be used widely at present.So utilize Wavelet Packet Technique to the often group strain data Y gathered _kcarry out eight layers of wavelet packet Scale Decomposition, obtain 2 altogether ⁸individual WAVELET PACKET DECOMPOSITION coefficient, because first WAVELET PACKET DECOMPOSITION coefficient is in low-frequency range, therefore chooses first WAVELET PACKET DECOMPOSITION coefficient and reconstructs, and obtains and rejects the strain data S that train load affects data _k, and S _kwith Y _kone_to_one corresponding, here k=1,2 ..., 8;

The major component of step 3, Extracting temperature data and temperature difference data:

Definition temperature data vector T=[T ₁, T ₂, T ₃, T ₄, T ₅, T ₆, T ₇, T ₈] ' and temperature difference data vector D=[T ₁₂, T ₃₄, T ₅₆, T ₇₈, T ₁₃, T ₁₅, T ₁₇, T ₃₅, T ₃₇, T ₅₇, T ₂₄, T ₂₆, T ₂₈, T ₄₆, T ₄₈, T ₆₈] ';

Because the data volume of temperature data vector T and temperature difference data vector D is more, therefore in order to simplify subsequent analysis work, principal component analysis (PCA) is adopted to simplify.Principal component analysis (PCA) is the thought utilizing dimensionality reduction, be a few generalized variable and major component by multiple variables transformations, wherein each major component is the linear combination of original variable, uncorrelated mutually between each major component, thus these major components can reflect most information of original variable, and contained information non-overlapping copies.Principal component analysis (PCA) is used widely in each ambit.

Utilize principal component analysis (PCA) to carry out principal component decomposition to temperature data vector T, obtain eight major components of temperature data vector T, choose front M major component P of temperature data vector T _m, m=1 here, 2 ..., M, wherein M is the minimum number making the variance contribution ratio sum of the major component of temperature data vector T reach more than 95%; When specifically determining M, eight major components of vector T have respective method contribution rate, first be arranged in order according to the large young pathbreaker of variance contribution ratio eight major components, add up successively the variance contribution ratio of each major component again from first major component after arrangement, and when variance contribution ratio begins to exceed 95%, corresponding major component number is M;

Utilize principal component analysis (PCA) to carry out principal component decomposition to temperature difference data vector D, obtain 16 major components of temperature difference data vector D, choose the top n major component R of temperature difference data vector D _n, n=1 here, 2 ..., N, wherein N is the minimum number making the variance contribution ratio sum of the major component of temperature difference data vector D reach more than 95%; The concrete defining method of N can refer to M;

Step 4, set up mathematical model between strain data and temperature data major component, temperature difference data major component:

Definition temperature data principal component vector P=[P ₁, P ₂..., P _m] ', temperature difference data principal component vector R=[R ₁, R ₂..., R _n] ' and reject the strain data vector S=[S of train load impact ₁, S ₂..., S ₈] ';

Using temperature data principal component vector P and temperature difference data principal component vector R as independent variable, the strain data vector S rejecting train load impact, as dependent variable, sets up the multiple linear regression equations between (1) S and P shown in formula, R:

S＝λP+γR+c (1)

(1) in formula:

λ, γ and c are respectively the performance parameter vector of temperature data major component, the performance parameter vector of temperature difference data major component and constant term, and have

$\mathrm{\λ}=\left[\begin{array}{c}{\mathrm{\λ}}_{11},{\mathrm{\λ}}_{12},...,{\mathrm{\λ}}_{1M}\\ {\mathrm{\λ}}_{21},{\mathrm{\λ}}_{22},...,{\mathrm{\λ}}_{2M}\\ .........\\ {\mathrm{\λ}}_{81},{\mathrm{\λ}}_{82},...,{\mathrm{\λ}}_{8M}\end{array}\right],$ $\mathrm{\γ}=\left[\begin{array}{c}{\mathrm{\γ}}_{11},{\mathrm{\γ}}_{12},...,{\mathrm{\γ}}_{1N}\\ {\mathrm{\γ}}_{21},{\mathrm{\γ}}_{22},...,{\mathrm{\γ}}_{2N}\\ .........\\ {\mathrm{\γ}}_{81},{\mathrm{\γ}}_{82},...,{\mathrm{\γ}}_{8N}\end{array}\right],$ $c=\left[\begin{array}{c}{c}_1\\ c_2\\ ...\\ c_8\end{array}\right]$

Utilize the concrete value of least square method determination parameter lambda, γ and c;

Step 5, calculated with mathematical model is utilized to survey strain residual error between strain data and simulated strain data:

In order to the temperature data of front L month and strain data are used for founding mathematical models as training data, temperature data after L month and strain data are used for the use of girder load-bearing capacity Performance Evaluation as assessment data, therefore after step 1 continuous collecting was by L month, the sample frequency identical with step 1 is kept to continue collecting temperature data and strain data, acquisition time length H is more than or equal to 10 days, acquisition time length H cross that I haven't seen you for ages and cause girder degradation information catch because being flooded by noise less than, it is more suitable empirical values that the lower limit of H gets 10, then the temperature data principal component vector P in method acquisition acquisition time this time period of length H of step 2 and step 3 is repeated ^a, temperature difference data principal component vector R ^as vectorial with the strain data of rejecting after train load impact ^a,

By P ^aand R ^a(2) formula of bringing into obtains simulated strain data vector S ^b:

S ^b＝λP ^a+γR ^a+c (2)

Further utilization (3) formula obtains strain residual vector D:

D＝S ^b-S ^a(3)

(3) in formula:

D=[d ₁, d ₂..., d ₈] ', d _krepresent kth group strain residual error data, k=1,2 ..., 8;

Step 6, utilization strain residual error assessment girder load-bearing capacity:

Due to eight groups of strain residual error d ₁, d ₂, d ₃, d ₄, d ₅, d ₆, d ₇and d ₈smooth performance be related to the load-bearing capacity of Railway Steel Truss arch bridge girder, therefore to eight groups of strain residual error d ₁, d ₂, d ₃, d ₄, d ₅, d ₆, d ₇and d ₈variation tendency carry out ADF unit root test.ADF unit root test is a kind of statistical analysis technique judging time series stationarity, be also called augmentation Dickey-Fu Le to check, specifically obtain assay by the function command adftest in Calling MATLAB mathematical software, if assay does not refuse the null hypothesis of an existence unit root, then show that time series has non-stationary; If the null hypothesis of an assay refusal existence unit root, then show that time series has stationarity.

Assay can be divided into following two kinds of situations:

If often the ADF unit root test result of group strain residual error all refuses the null hypothesis of an existence unit root, then Railway Steel Truss arch bridge girder load-bearing capacity is in shape;

If the ADF unit root test result of at least one group of strain residual error does not refuse the null hypothesis of an existence unit root, then show that Railway Steel Truss arch bridge girder load-bearing capacity is degenerated, early warning should be made in time.

Beneficial effect:

The method for early warning of a kind of Railway Steel Truss arch bridge girder load-resisting capacity degradation of the present invention is at consideration temperature data T ₁, T ₂, T ₃, T ₄, T ₅, T ₆, T ₇, T ₈on the basis of impact, have chosen again 12 groups of vertical temperature difference data T ₁₃, T ₁₅, T ₁₇, T ₃₅, T ₃₇, T ₅₇, T ₂₄, T ₂₆, T ₂₈, T ₄₆, T ₄₈, T ₆₈with 4 groups of section temperature difference data T ₁₂, T ₃₄, T ₅₆, T ₇₈, therefore taken into full account the impact that the vertical temperature difference and section temperature difference produce the strain of steel arch trusses component, and the present invention is considered more comprehensively accurately;

And set up new multiple linear regression equations further, reflect the correlation properties between the temperature data principal component vector of strain data vector and the multiple measuring point analyzing measuring point and temperature difference data principal component vector, the multiple linear regression model therefore set up is more accurate;

Therefore, a kind of Railway Steel Truss arch bridge girder load-resisting capacity degradation method for early warning that the present invention proposes, it is more accurately thorough to consider, and has good engineering practical value, can obtain extensive promotion and application.

Accompanying drawing explanation

Fig. 1 is the side elevational view of Foundations of Dashengguan Changjiang River Bridge in the embodiment of the present invention;

Fig. 2 is the arrangenent diagram of temperature sensor and strain transducer 1-1 section in FIG in the embodiment of the present invention;

Fig. 3 is the arrangenent diagram of temperature sensor and strain transducer 2-2 section in fig. 2 in the embodiment of the present invention;

Fig. 4 is the arrangenent diagram of temperature sensor and strain transducer 3-3 section in figure 3 in the embodiment of the present invention;

Fig. 5 is the arrangenent diagram of temperature sensor and strain transducer 4-4 section in figure 3 in the embodiment of the present invention;

Fig. 6 is the arrangenent diagram of temperature sensor and strain transducer 5-5 section in figure 3 in the embodiment of the present invention;

Fig. 7 is the arrangenent diagram of temperature sensor and strain transducer 6-6 section in figure 3 in the embodiment of the present invention;

Fig. 8 is embodiment of the present invention T ₁monthly time-history curves;

Fig. 9 is embodiment of the present invention T ₁odd-numbered day in time-history curves;

Figure 10 is embodiment of the present invention T ₁₂monthly time-history curves;

Figure 11 is embodiment of the present invention T ₁₂odd-numbered day in time-history curves;

Figure 12 is embodiment of the present invention Y ₁monthly time-history curves;

Figure 13 is embodiment of the present invention Y ₁odd-numbered day in time-history curves;

Figure 14 is embodiment of the present invention Y ₂monthly time-history curves;

Figure 15 is embodiment of the present invention Y ₂odd-numbered day in time-history curves;

Figure 16 is embodiment of the present invention S ₁monthly time-history curves;

Figure 17 is embodiment of the present invention S ₁odd-numbered day in time-history curves;

Figure 18 is embodiment of the present invention S ₂monthly time-history curves;

Figure 19 is embodiment of the present invention S ₂odd-numbered day in time-history curves;

Figure 20 is the explained variance histogram of embodiment of the present invention vector T eight major components;

Figure 21 is the explained variance histogram of embodiment of the present invention vector D eight major components;

Figure 22 is embodiment of the present invention strain residual error d ₁the variation tendency of data;

Figure 23 is embodiment of the present invention strain residual error d ₂the variation tendency of data;

Figure 24 is embodiment of the present invention strain residual error d ₃the variation tendency of data;

Figure 25 is embodiment of the present invention strain residual error d ₄the variation tendency of data;

Figure 26 is embodiment of the present invention strain residual error d ₅the variation tendency of data;

Figure 27 is embodiment of the present invention strain residual error d ₆the variation tendency of data;

Figure 28 is embodiment of the present invention strain residual error d ₇the variation tendency of data;

Figure 29 is embodiment of the present invention strain residual error d ₈the variation tendency of data.

Embodiment

Close steel truss arched bridge to win completely below, and by reference to the accompanying drawings the present invention is further described.

Win completely close steel truss arched bridge be Beijing-Shanghai High-Speed Railway and Shanghai Chinese Rong railway share across Jiang Tongdao, bridge carries the two-wire subway in Nanjing simultaneously.Main Bridge of Dashengguan Changjiang River Bridge is continuous truss arch bridge, and its main span reaches 336 meters, and as shown in Figure 1, continuous truss arch bridge is encircleed by steel truss and steel bridge deck is formed.In addition, steel truss arch comprises box-type section chord member (top boom, lower boom, bridge floor chord member), I-shaped cross-section diagonal web member, montant and horizontal and vertical support etc., as shown in Figures 2 and 3.Steel bridge deck is made up of top board and lateral stiffening beam, as shown in Figure 2.

A degeneration method for early warning for Railway Steel Truss arch bridge girder load-bearing capacity, this method for early warning comprises the steps:

Step 1, collecting temperature data and strain data:

Eight temperature sensors are arranged on respectively top boom, diagonal web member, the upstream side in lower boom and these four component centre positions of bridge floor chord member and downstream and carry out temperature acquisition, its position as shown in Figure 4 to 7, sample frequency is set as 1Hz, and sampling time length L is 8 months.The temperature acquisition result in top boom upstream side and downstream adopts T ₁and T ₂represent, the temperature acquisition result in diagonal web member upstream side and downstream adopts T ₃and T ₄represent, the temperature acquisition result in lower boom upstream side and downstream adopts T ₅and T ₆represent, the temperature acquisition result in bridge floor chord member upstream side and downstream adopts T ₇and T ₈represent.The temperature difference between four component different acquisition positions adopts T _ijrepresent, T _ij=T _i-T _j, i=1,2 ..., 8, j=1,2 ..., 8, and i ≠ j.T ₁and T ₁₂time-history curves respectively as shown in Figure 3 and Figure 4;

Strain acquirement is carried out in the upstream side and the downstream that eight fiber Bragg grating strain sensors are arranged on respectively four component centre positions, its position as shown in Figure 4 to 7, sample frequency is set as 1Hz, sampling time length is eight months, and wherein the strain acquirement result in top boom upstream side and downstream adopts Y ₁and Y ₂represent, the strain acquirement result in diagonal web member upstream side and downstream adopts Y ₃and Y ₄represent, the strain acquirement result in lower boom upstream side and downstream adopts Y ₅and Y ₆represent, the strain acquirement result in bridge floor chord member upstream side and downstream adopts Y ₇and Y ₈represent.Y ₁time-history curves respectively as shown in Figure 12 and Figure 13, Y ₂time-history curves respectively as shown in Figure 14 and Figure 15;

Although temperature and the strain of same point should be surveyed in theory, but because temperature sensor and strain transducer all occupy certain space, therefore make temperature sensor and strain transducer close to measuring point as much as possible when practical operation, so adopt the position in Fig. 4 ~ Fig. 7 to place;

The composition affected by train load in step 2, rejecting strain data:

Wavelet Packet Technique is adopted to reject the strain composition affected by train load, by the often group strain data Y gathered _k(k=1,2 ... 8) carry out eight layers of wavelet packet Scale Decomposition, obtain 2 altogether ⁸individual WAVELET PACKET DECOMPOSITION coefficient.Choose first WAVELET PACKET DECOMPOSITION coefficient to be reconstructed, obtain the strain sequence S rejecting train load impact _k, S _kwith Y _kone_to_one corresponding.S ₁time-history curves respectively as shown in Figure 16 and Figure 17, S ₂time-history curves respectively as shown in Figure 18 and Figure 19, can find out and adopt WAVELET PACKET DECOMPOSITION technology can effectively reject the static strain caused by train load;

Major component in step 3, Extracting temperature data and temperature difference data:

Temperature data has eight groups, forms temperature data vector T=[T ₁, T ₂, T ₃, T ₄, T ₅, T ₆, T ₇, T ₈] '; Temperature difference data have 16 groups, form temperature difference data vector D=[T ₁₂, T ₃₄, T ₅₆, T ₇₈, T ₁₃, T ₁₅, T ₁₇, T ₃₅, T ₃₇, T ₅₇, T ₂₄, T ₂₆, T ₂₈, T ₄₆, T ₄₈, T ₆₈] ';

Utilize principal component analysis (PCA) to carry out principal component decomposition to temperature data vector T, obtain eight major components of temperature data vector T, the explained variance histogram of eight major components as shown in figure 20, chooses the 1st major component P of temperature data vector T ₁the variance contribution ratio sum of temperature data vector T major component can be made to reach more than 95%;

Equally, utilize principal component analysis (PCA) to carry out principal component decomposition to temperature difference data vector D, obtain 16 major components of temperature difference data vector D, the explained variance histogram of 16 major components as shown in figure 21, chooses front 3 major component R of temperature difference data vector D ₁, R ₂and R ₃the explained variance contribution rate sum of temperature difference data vector D major component can be made to reach more than 95%;

Step 4, set up mathematical model between strain data and temperature data major component, temperature difference data major component:

Definition temperature data principal component vector P=[P ₁] ', temperature difference data principal component vector R=[R ₁, R ₂, R ₃] ', reject the strain data vector S=[S of train load impact ₁, S ₂..., S ₈] ';

Using temperature data principal component vector P and temperature difference data principal component vector R as independent variable, except the strain data vector S of train load impact is as dependent variable, set up the multiple linear regression equations between S and P, R:

S＝λP+γR+c (1)

And $\mathrm{\λ}=\left[\begin{array}{c}{\mathrm{\λ}}_{11}\\ {\mathrm{\λ}}_{21}\\ ...\\ {\mathrm{\λ}}_{81}\end{array}\right],$ $\mathrm{\γ}=\left[\begin{array}{c}{\mathrm{\γ}}_{11},{\mathrm{\γ}}_{12},{\mathrm{\γ}}_{13}\\ {\mathrm{\γ}}_{21},{\mathrm{\γ}}_{22},{\mathrm{\γ}}_{23}\\ .........\\ {\mathrm{\γ}}_{81},{\mathrm{\γ}}_{82},{\mathrm{\γ}}_{83}\end{array}\right],$ $c=\left[\begin{array}{c}{c}_1\\ c_2\\ ...\\ c_8\end{array}\right]$

In formula, λ, γ and c are respectively the performance parameter vector of temperature data major component, the performance parameter vector of temperature difference data major component and constant term.Vectorial P, R and S are brought into formula (1), and the concrete value of parameter lambda, γ and c is as follows to utilize least square method to determine:

$\mathrm{\λ}=\left[\begin{array}{c}0.454\\ 1.585\\ 0.301\\ 0.350\\ 0.179\\ 0.062\\ -0.367\\ 0.611\end{array}\right],$ $\mathrm{\γ}=\left[\begin{array}{ccc}-4.711,& -2.597,& -4.944\\ 0.013,& 4.104,& -1.247\\ 3.416,& -2.098,& 2.968\\ -0.734,& 0.351,& 0.948\\ -4.219,& 1.827,& 3.165\\ -3.245,& 8.242,& -5.398\\ 2.411,& -0.598,& 5.531\\ 2.188,& -1.130,& 2.018\end{array}\right],$ $c=\left[\begin{array}{c}-46.806\\ -70.238\\ -6.884\\ -23.597\\ -22.049\\ -22.862\\ 17.075\\ -32.799\end{array}\right]$

Step 5, calculated with mathematical model is utilized to survey residual error between strain data and simulated strain data:

Continue the temperature data after collection eight month and strain data, this time acquisition time length H is 19 days.Repetition step 2 and step 3 obtain temperature data principal component vector P in these 19 days afterwards ^a, temperature difference data principal component vector R ^as vectorial with the strain data of rejecting after train load impact ^a;

By P ^aand R ^abring following formula into and obtain simulated strain data vector S ^b:

S ^b＝λP ^a+γR ^a+c (2)

Following formula is utilized to obtain strain residual vector D further:

D＝S ^b-S ^a(3)

In formula, D=[d ₁, d ₂..., d ₈] ', d _krepresent kth group strain residual error data, its variation tendency is as shown in Figure 22 ~ Figure 29.

Step 6, utilization strain residual error assessment girder load-bearing capacity:

Eight groups of strain residual error d ₁, d ₂, d ₃, d ₄, d ₅, d ₆, d ₇and d ₈smooth performance be related to the load-bearing capacity of Railway Steel Truss arch bridge girder, therefore ADF unit root test is carried out to the variation tendency of eight groups of strain residual errors, assay is that the assay often organizing strain residual error all refuses the null hypothesis of an existence unit root, then win that to close steel truss arched bridge girder load-bearing capacity in shape completely.

The above is only the preferred embodiment of the present invention; be noted that for those skilled in the art; under the premise without departing from the principles of the invention, can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (2)

1. a method for early warning for Railway Steel Truss arch bridge girder load-resisting capacity degradation, is characterized in that: comprise the following steps that order performs:

Step 1, collecting temperature data and strain data:

Choose the top boom of steel truss arched bridge girder span centre position, diagonal web member, lower boom and bridge floor chord member four components, the equal set temperature sensor of upstream side in the centre position of these four components and downstream and strain transducer carry out temperature acquisition and strain acquirement, and acquisition time length L is more than or equal to eight months;

Wherein the temperature acquisition result in top boom upstream side and downstream adopts T ₁and T ₂represent, the temperature acquisition result in diagonal web member upstream side and downstream adopts T ₃and T ₄represent, the temperature acquisition result in lower boom upstream side and downstream adopts T ₅and T ₆represent, the temperature acquisition result in bridge floor chord member upstream side and downstream adopts T ₇and T ₈represent; The temperature difference between different acquisition position adopts T _ijrepresent, definition of T _ij=T _i-T _j, wherein i=1,2 ..., 8, j=1,2 ..., 8, and i ≠ j;

Wherein the strain acquirement result in top boom upstream side and downstream adopts Y ₁and Y ₂represent, the strain acquirement result in diagonal web member upstream side and downstream adopts Y ₃and Y ₄represent, the strain acquirement result in lower boom upstream side and downstream adopts Y ₅and Y ₆represent, the strain acquirement result in bridge floor chord member upstream side and downstream adopts Y ₇and Y ₈represent;

The composition affected by train load in step 2, rejecting strain data:

Utilize Wavelet Packet Technique to the often group strain data Y gathered _kdecompose, choose first WAVELET PACKET DECOMPOSITION coefficient and reconstruct, obtain and reject the strain data S that train load affects data _k, and S _kwith Y _kone_to_one corresponding, here k=1,2 ..., 8;

The major component of step 3, Extracting temperature data and temperature difference data:

Definition temperature data vector T=[T ₁, T ₂, T ₃, T ₄, T ₅, T ₆, T ₇, T ₈] ' and temperature difference data vector D=[T ₁₂, T ₃₄, T ₅₆, T ₇₈, T ₁₃, T ₁₅, T ₁₇, T ₃₅, T ₃₇, T ₅₇, T ₂₄, T ₂₆, T ₂₈, T ₄₆, T ₄₈, T ₆₈] ';

Utilize principal component analysis (PCA) to carry out principal component decomposition to temperature data vector T, obtain eight major components of temperature data vector T, choose front M major component P of temperature data vector T _m, m=1 here, 2 ..., M, wherein M is the minimum number making the variance contribution ratio sum of the major component of temperature data vector T reach more than 95%;

Utilize principal component analysis (PCA) to carry out principal component decomposition to temperature difference data vector D, obtain 16 major components of temperature difference data vector D, choose the top n major component R of temperature difference data vector D _n, n=1 here, 2 ..., N, wherein N is the minimum number making the variance contribution ratio sum of the major component of temperature difference data vector D reach more than 95%;

Step 4, set up mathematical model between strain data and temperature data major component, temperature difference data major component:

Definition temperature data principal component vector P=[P ₁, P ₂..., P _m] ', temperature difference data principal component vector R=[R ₁, R ₂..., R _n] ' and reject the strain data vector S=[S of train load impact ₁, S ₂..., S ₈] ';

Using temperature data principal component vector P and temperature difference data principal component vector R as independent variable, the strain data vector S rejecting train load impact, as dependent variable, sets up the multiple linear regression equations between (1) S and P shown in formula, R:

S＝λP+γR+c (1)

(1) in formula:

λ, γ and c are respectively the performance parameter vector of temperature data major component, the performance parameter vector of temperature difference data major component and constant term, and have

$\mathrm{\λ}=\left[\begin{array}{c}{\mathrm{\λ}}_{11},{\mathrm{\λ}}_{12},...,{\mathrm{\λ}}_{1M}\\ {\mathrm{\λ}}_{21},{\mathrm{\λ}}_{22},...,{\mathrm{\λ}}_{2M}\\ .........\\ {\mathrm{\λ}}_{81},{\mathrm{\λ}}_{82},...,{\mathrm{\λ}}_{8M}\end{array}\right],\mathrm{\γ}=\left[\begin{array}{c}{\mathrm{\γ}}_{11},{\mathrm{\γ}}_{12},...,{\mathrm{\γ}}_{1N}\\ {\mathrm{\γ}}_{21},{\mathrm{\γ}}_{22},...,{\mathrm{\γ}}_{2N}\\ .........\\ {\mathrm{\γ}}_{81},{\mathrm{\γ}}_{82},...,{\mathrm{\γ}}_{8N}\end{array}\right],c=\left[\begin{array}{c}{c}_1\\ c_2\\ ...\\ c_8\end{array}\right]$

Utilize the concrete value of least square method determination parameter lambda, γ and c;

Step 5, calculated with mathematical model is utilized to survey strain residual error between strain data and simulated strain data:

After step 1 collection is terminated, keep the sample frequency identical with step 1 to continue collecting temperature data and strain data, acquisition time length H is more than or equal to 10 days, and the method repeating step 2 and step 3 obtains the temperature data principal component vector P in acquisition time this time period of length H ^a, temperature difference data principal component vector R ^as vectorial with the strain data of rejecting after train load impact ^a;

By P ^aand R ^a(2) formula of bringing into obtains simulated strain data vector S ^b:

S ^b＝λP ^a+γR ^a+c (2)

Further utilization (3) formula obtains strain residual vector D:

D＝S ^b-S ^a(3)

(3) in formula:

D=[d ₁, d ₂..., d ₈] ', d _krepresent kth group strain residual error data, k=1,2 ..., 8;

Step 6, utilization strain residual error assessment girder load-bearing capacity:

To eight groups of strain residual error d ₁, d ₂, d ₃, d ₄, d ₅, d ₆, d ₇and d ₈variation tendency carry out ADF unit root test:

If often the ADF unit root test result of group strain residual error all refuses the null hypothesis of an existence unit root, then Railway Steel Truss arch bridge girder load-bearing capacity is in shape;

If the ADF unit root test result of at least one group of strain residual error does not refuse the null hypothesis of an existence unit root, then show that Railway Steel Truss arch bridge girder load-bearing capacity is degenerated, early warning should be made in time.

2. the method for early warning of Railway Steel Truss arch bridge girder load-resisting capacity degradation according to claim 1, is characterized in that: in step 1, sample frequency selects a fixed frequency point in 1Hz to 10Hz.