CN112461152A - Large-scale industrial structure deformation monitoring and analyzing method - Google Patents

Abstract

A large-scale industrial structure deformation monitoring and analysis method belongs to the field of large-scale industrial structure deformation monitoring, and is characterized in that: a measuring platform is built and a space rectangular coordinate system of the built measuring platform is established; The moving platform is provided with a camera; the camera includes a measuring camera and a transfer camera; solving the motion change parameters of the measuring platform; solving the external parameters of the measuring camera; solving the coordinates of the point to be measured and determining the large-scale industrial structure Deformation size; finally, safety assessment of large industrial structures. The method can achieve lower measurement costs with fewer measurement cameras; meanwhile, the safety assessment method based on evidence-based reasoning can realize the transparency of the assessment process and ensure the interpretability of the assessment results.

Description

A Deformation Monitoring and Analysis Method for Large Industrial Structures

technical field

The invention belongs to the field of large-scale industrial structure deformation monitoring, and in particular relates to a large-scale industrial structure deformation monitoring and analysis method.

Background technique

At present, countries have accelerated the pace of exploration in outer space, and large-scale industrial structures such as launch vehicles, space shuttles, space stations, and artificial satellites are constantly emerging. For example, as an important carrier for deep space exploration, large liquid carrier rockets have attracted more and more attention, and the frequency of launching rockets by countries has also increased year by year. However, in parallel with the development of emerging technologies, accidents in the aerospace sector have also increased. Lack of reliable safety monitoring is one of the important reasons for accidents. As one of the important indicators of safety monitoring, structural deformation frequently occurs in large industrial structures, which mainly comes from the extrusion of the internal materials of the structure and the vibration of the external environment.

Judging from the research status of large-scale industrial structural deformation monitoring technology, the current main methods are: methods based on acoustic emission, methods based on ultrasonic image processing, methods based on fiber grating sensing, and methods based on camera/scene measurement. Among them, the method based on acoustic emission can effectively locate the deformation, but cannot identify its size and degree; the method based on ultrasonic image processing can effectively identify the size and degree of deformation, but there are also shortcomings such as low efficiency in locating the deformation; based on fiber grating The sensing method can effectively obtain the position and degree of deformation, but the embedded sensor will have an impact on the mechanical properties of the structure; the method based on optical camera/scene measurement has the advantages of non-contact, high-precision online measurement and convenient operation. .

After obtaining the deformation variables, further safety assessment of large-scale industrial structures is required. However, due to the complexity of large-scale industrial structures, the diversity of internal instruments and equipment, and the existence of certain electromagnetic interference and coupling between different equipment, the analytical model-based safety assessment method is difficult to achieve accurate mechanism analysis and cannot meet the requirements of accurate construction. mold requirements. In addition, although the traditional data-driven neural network method can avoid the complex problems caused by modeling, its evaluation process is not interpretable and traceable, losing the physical meaning of mechanism analysis and some monitoring parameters, and the results are difficult to Convincing.

SUMMARY OF THE INVENTION

The present invention aims to solve the above-mentioned problems, and provides a method for realizing accurate positioning and identification of the deformation of large-scale industrial structures, which can meet the high reliability and high-precision requirements for structural deformation monitoring, and can meet the requirements of effective and interpretable large-scale industrial structure deformation monitoring. Industrial structural deformation monitoring and analysis methods.

The large-scale industrial structure deformation monitoring and analysis method of the present invention builds a measuring platform and establishes a space rectangular coordinate system of the built measuring platform; a moving platform is arranged in the measuring platform; a camera is arranged on the moving platform; The camera includes a measuring camera and a transfer camera; solve the motion change parameters of the measuring platform; solve the external parameters of the measuring camera; solve the coordinates of the point to be measured and determine the deformation size of the large industrial structure; Sexual assessment.

Preferably, the large-scale industrial structure deformation monitoring and analysis method of the present invention, the described building a measuring platform and establishing the space rectangular coordinate system of the built measuring platform, including:

Select a moving platform, fix the measuring camera and the transfer camera on the selected running platform in turn, and at the same time make the moving platform move around the surface of the large industrial structure to be monitored;

Select the space origin, establish the world coordinate system O-XYZ, and at the same time establish two sets of moving measuring camera coordinate systems O _D,i -X _D,i Y according to the positions of the measuring cameras at different times t _i and t _{i+1 D,i} Z _D,i and O _D,i+1 -X _D,i+1 Y _D,i+1 Z _D,i+1 , and establish the transfer camera coordinate system O _T -X _T Y _T Z _T with the motion platform coordinate system OP -X _P Y _P Z _{P .}

Preferably, the large-scale industrial structure deformation monitoring and analysis method of the present invention, the motion variation parameters of the solution measurement platform, including:

Select the measurement point S (X, Y, Z) and the corresponding image point coordinates s _i (x _i , y _i ), s _i+1 (x _i+1 , y _i+1 ) to establish the measurement point S and the image point Homogeneous coordinate relationship corresponding to s _i and s _i+1 ;

Establish the relationship between P in O-XYZ system and O _P -X _P Y _P Z _P system, and between O _P -X _P Y _P Z _P system and O _T- X _T Y _T Z _T system , and further establish the relationship between S in the O-XYZ system and the O _T -X _T Y _T Z _T system;

Using the transfer camera to shoot the artificial marking points in real time, according to the relationship between the marking points and the corresponding image points, the rotation matrix and translation vector of the O-XYZ system converted to the O _T -X _T Y _T Z _T system are obtained;

Establish the conversion relationship between the O-XYZ system and the O _T -X _T Y _T Z _T system at the time t _i and t i ₊ 1, and then solve the motion change parameters of the platform between the time t _i and t _i+1 .

Preferably, in the method for monitoring and analyzing the deformation of large industrial structures according to the present invention, the solving of the external parameters of the measuring camera includes: respectively establishing the O _P - X _P Y _P Z _P system and O at t _i and t _i+1 times Homogeneous coordinate relationship of _D,i -X _D,i Y _D,i Z _D,i system, O _D,i+1 -X _D,i+1 Y _D,i+1 Z _D,i+1 system;

Establish the relationship between the O-XYZ system and the O _D,i -X _D,i Y _D,i Z _D,i system at different times;

Select the initial moment to measure the camera coordinate system O _{D, 0} -X _{D, 0} Y _{D, 0} Z _{D, 0} , and obtain any moment O _{D, i} -X _{D, i} Y _{D, i} Z _{D, i} system through iteration The relationship between the position and attitude relative to the O-XYZ system, that is, the external parameters of the camera are measured.

Preferably, in the method for monitoring and analyzing the deformation of a large-scale industrial structure according to the present invention, the solving of the coordinates of the point to be measured and the determination of the deformation of the large-scale industrial structure include: establishing a position equation of the measurement point S at time t _i ; establishing an initial time Measure the position equation of the camera, and then obtain the world coordinates of the measurement point S; repeat the above measurement steps to obtain the world coordinates of the monitoring points at different positions at different times; Solve for the corresponding deformation variable.

Preferably, in the method for monitoring and analyzing the deformation of large industrial structures of the present invention, the safety assessment for large industrial structures includes: determining deformation weights based on the coefficient of variation method; calculating reliability based on distances; Security of industrial structures.

The large-scale industrial structure deformation monitoring and analysis method of the invention can realize the non-contact and high-precision measurement of the large-scale industrial structure deformation; the mobile camera measurement process only needs to calibrate the external parameters of the transfer camera once at the initial moment, And can convert all measurement results to the same coordinate system. The method of the invention is suitable for deformation monitoring of large-scale industrial structure surface area, and can achieve lower measurement costs with fewer measurement cameras; at the same time, the safety evaluation method based on evidence reasoning can realize the transparency of the evaluation process, and It ensures that the evaluation results are interpretable.

Description of drawings

1 is a schematic diagram of the measurement of a mobile camera according to the present invention;

2 is a schematic diagram of the mobile camera measurement under the transfer camera according to the present invention;

FIG. 3 is a frame diagram of the safety assessment of the large-scale industrial structure according to the present invention.

FIG. 4 is an experimental diagram of the mobile camera measurement according to the present invention.

FIG. 5 is a position diagram of the mobile camera measurement identification point according to the present invention.

FIG. 6 is a schematic diagram of the measurement error according to the present invention.

FIG. 7 is a graph showing the deformation measurement result of the large-scale industrial structure according to the present invention.

FIG. 8 is a graph of the safety confidence distribution of the large industrial structure according to the present invention.

Fig. 9 is the utility diagram of the safety level of the large-scale industrial structure according to the present invention;

Among them, 1-measurement camera, 2-transmission camera, 3-track, 4-marker.

Detailed ways

The following describes the deformation monitoring and analysis method of the large-scale industrial structure of the present invention in detail with reference to the accompanying drawings and embodiments.

In the embodiment of the present disclosure, a large-scale industrial structure is taken as the object. FIG. 1 is a schematic diagram of the measurement of the moving camera. The left and right are the positions of the moving camera at time t _i and t _i+1 , respectively. For O _D,i -X _D,i Y _D,i Z _D,i and O _D,i+1 -X _D,i+1 Y _D,i+1 Z _D,i+1 , transfer camera 2 coordinates The system is O _T -X _T Y _T Z _T , the motion platform coordinate system is O _P -X _P Y _P Z _P , and the world coordinate system is O-XYZ.

和

The rotation matrix and translation vector of the camera coordinate system relative to the world coordinate system are M _i , V _i and M _i+1 , V _i+1 respectively; the rotation matrix and translation vector between the camera coordinate system at two adjacent moments are respectively Denoted as M _i,i+1 ,V _i,i+1 . The target point collected by the camera, that is, the deformation monitoring point of the large industrial structure, is S(X,Y,Z), and the image coordinate points at the corresponding moment are s _i (x _i , y _i ), s _i+1 (x _i+1 , y _i+1 ), the homogeneous coordinate relationship corresponding to the three points is expressed as S,

and

According to the geometric relationship of camera imaging, the relationship between the measurement point S and the image points _si and _si+1 can be expressed as:

In formula (1), γ _i and γ _i+1 are proportional coefficients, and Q is the parameter matrix in the camera. The different positions of the cameras at adjacent moments can be equivalent to a binocular camera measurement system.

Figure 2 is a schematic diagram of the mobile camera measurement under the transfer camera 2, which includes the measurement camera 1, the transfer camera 2, the track 3 and the marker 4 plane. The entire measurement process consists of the following five steps:

Step 1: Fix the measuring camera 1 and the transfer camera 2 on the platform respectively, and the platform moves around the surface of the structure.

Step 2: Use the transfer camera 2 to shoot real-time artificial cooperation identification points with accurately known coordinates pasted on the ground, and obtain the real-time external parameters of the transfer camera 2 according to the relationship between the identification points and the corresponding image points.

Step 3: Using the relative fixed positional relationship between the transfer camera 2, the motion platform, and the measurement camera 1, calculate the state transition relationship between adjacent moments of the motion platform according to the external parameters of the transfer camera 2.

Assuming that S _T and S _P represent the homogeneous coordinates of the measurement point S (X, Y, Z) in the O _T -X _T Y _T Z _T and O _P -X _P Y _P Z _P coordinate systems respectively, then S is in O The relationship between -XYZ and O _P -X _P Y _P Z _P coordinate system and the relationship between O _P -X _P Y _P Z _P and O _T -X _T Y _T Z _T coordinate system can be expressed as:

Among them, M _WP , V _WP respectively represent the rotation matrix and translation vector of the O-XYZ coordinate system converted to the OP-X _P Y _P Z _P coordinate system _; M _PT , V _PT represent the OP-X _P Y _P Z _respectively The rotation matrix and translation vector for converting the _P coordinate system to the O _T -X _T Y _T Z _T coordinate system can be obtained from step 2. The relationship between the O-XYZ coordinate system and the O _T -X _T Y _T Z _T coordinate system can be expressed as:

S _T =M _WT S+V _WT (3)

in:

In formula (4), M _WT , V _WT respectively represent the rotation matrix and translation vector transformed from the O-XYZ coordinate system to the O _T -X _T Y _T Z _T coordinate system, which can be obtained by means of manual identification points.

Assuming that the homogeneous coordinates of the measurement point S in the motion platform coordinate system at times t _i and t _j are S _Pi and S _Pj , respectively, the two satisfy the relation:

S _Pj =M _ij S _Pi +V _ij (5)

In formula (5), M _ij and V _ij respectively represent the rotation from the O _P - X _P Y _P Z _P coordinate system at time t _i to the O _P - X _P Y _P Z _P coordinate system at time t _j Matrix and translation vector. Assuming that the homogeneous coordinates of S in the O _T -X _T Y _T ZT coordinate system at times t _i and t _j are S _Ti and S _Tj respectively, there are:

In _formula ( ₆ ), _{M WTi} , _{V WTi} and _{M WTj} , _{V WTj} respectively represent the rotation matrix and Translation vector.

According to formula (2) and formula (6), we can get:

Then, the relative external parameters between the O _P -X _P Y _P Z _P coordinate system at time t _i and t _j are as follows:

Step 4: Solve the external parameters of the measuring camera.

In the case of establishing the measuring camera 1 coordinate system O _C -X _C Y _C Z _C , the homogeneous coordinate relationship between the motion platform coordinate system and the measuring camera 1 coordinate system are:

S _C =M _PC S _P +V _PC (9)

In formula (9), M _PC , V _PC respectively represent the rotation matrix and translation vector of the O _P -X _P Y _P Z _P coordinate system converted to the O _T -X _T Y _T Z _T coordinate system, which can be obtained by calibration . Similarly, there is a relationship between the coordinate system of the motion platform and the coordinate system of the measuring camera 1 at time t _i and t _j :

According to equations (5) and (10), the world coordinate system and the coordinate system of the surveying camera 1 at different times satisfy the following relationship:

Through the iterative formula (11), the external parameters of the measurement camera 1 can be obtained.

Step 5: According to the state transition relationship between the adjacent moments of the motion platform and the external parameters of the measuring camera 1, the position coordinates of the measuring point are obtained in real time.

According to formula (1), point S at time t _i satisfies:

Among them, the initial camera position satisfies:

By repeating the above steps, the world coordinates of the measurement point _Si at different monitoring moments can be obtained. Compare the position of the measurement point and the identification point, obtain the deformation variables Δx _i , Δy _i , Δzi _i of each measurement point along the x, y, z directions, and obtain the comprehensive deformation corresponding to each measurement result through simple weighting , as shown in formula (14):

Figure 3 shows the framework for the safety assessment of large-scale industrial structures. The entire assessment process mainly includes the following three steps:

Step 1: Use the above camera measurement method to perform repeated L rounds of measurement, each round of measurement period is T time points, obtain the deformation variable data δ ₁ ,...,δ _L according to formula (14), and use the coefficient of variation method to obtain each group of measurements. The weights corresponding to the data. Suppose the deformation variable data measured in the i-th round can be expressed as:

δ _i (t)=[δ _i (tT),δ _i (t-T+1),…,δ _i (t-1)]

和均方差分别可表示为：the corresponding mean

and the mean square error can be expressed as:

The weight corresponding to the result of the i-th round of measurement is:

是一个归一化因子，可以表征第i轮测量结果的相对 变化，进而反映该轮测试的形变量数据的波动性。因此，

越大，w_i(T)就 越大。in,

is a normalization factor, which can characterize the relative change of the i-th round of measurement results, thereby reflecting the volatility of the deformation data of this round of testing. therefore,

The larger the value, the larger the _wi (T).

的距离可表示为：Step 2: According to the deformation variable test data, a distance-based method is used to obtain the reliability corresponding to each group of measurement data. δ _i (t) and

The distance can be expressed as:

The average distance of the deformation variable data measured in the i-th round in T time points is:

Then the reliability corresponding to the result of the i-th round of measurement is:

Step 3: According to the test data of deformation variables, the input information conversion method based on utility is used to convert the deformation variables into the form of confidence distribution, that is, the initial evidence. Assuming that the deformation variables can be divided into N safety evaluation levels, namely {H ₁ ,...,H _n ,...,H _N }, the corresponding reference values are {h ₁ ,...,h _n ,...,h _N }. If rank H _n+1 is better than rank H _n if and only if h _n+1 > h _n , where n,n+1∈[1,N]. Let h _N and h ₁ be the maximum and minimum reference values, respectively, the deformation variable δ _i can be transformed into a confidence distribution as shown below:

S(δ _i )={(H _n ,β _n,i ),n=1,2,...,N,i=1,2,...,L} (21)

in

The above information is fused by the method of evidence reasoning, and its analytical expression is as follows:

w _i = _wi /(1+ _wi -r _i ) (25)

where β _n represents the confidence assigned to the evaluation level H _n , and _wi and _ri are given by equations (17) and (20), respectively. At this time, the safety assessment result G(T) of the structure can be expressed as:

G(T)={(H _n , β _n (t)), n=1,2,...,N} (26)

According to the safety evaluation result given by formula (26), the safety status of large industrial structures can be effectively judged, and then reasonable maintenance measures can be taken.

The present invention designs a mobile camera measurement experiment as shown in FIG. 4 to illustrate the effectiveness of the present invention. In the measurement experiment, the transfer camera 2 and the measurement camera 1 are respectively fixed on the platform track 3, and 15 points are pasted on the surface of a large industrial structure model as the points to be measured. The side of the parallel track 3 is attached with some coordinates with known coordinates. Manual identification points. In the experiment, the resolutions of the transfer camera 2 and the measurement camera 1 are both 1536 pixels×1024 pixels, and the frame rates are both 3 frame/s, and the motion speed of the platform track 3 is 20 mm/s. According to the measurement process, the positions of 15 points to be measured are obtained, and the results are shown in Figure 5. The results can intuitively reflect the positions of the preset identification points on the surface of the structure. Obviously, in this experiment, with the assistance of the transfer camera 2, the mobile camera can complete the dynamic measurement and obtain the position of the marking point on the structure surface in real time, which makes up for the defect that the mobile camera cannot obtain external parameters. After distance calculation, the errors between the 15 points to be measured and their actual positions are obtained, as shown in Figure 6. The root mean square error is calculated for the data in Fig. 6, and the result is 0.0432mm, which shows that the present invention can meet the dynamic measurement requirements of structural deformation.

According to the same method, the 20 identification points at the bottom, middle and top of the structure were measured by camera respectively, and these three parts were used as three indicators, and the deformation measurement results were obtained as shown in Figure 7. The structural deformation size is divided into three grades: "small", "medium" and "large", and the corresponding reference values are shown in Table 1:

Table 1 Deformation reference value of measuring point (unit: mm) semantic value Small middle Big bottom 0 3 6 Central 0 2 4 top 0 1 2

The security of the structure is divided into three levels of "high", "medium" and "low", and the corresponding quantitative utility values are shown in Table 2:

Table 2 Structural safety reference values semantic value high middle Low Reference 1 0.5 0

According to the aforementioned coefficient of variation method and the method based on distance, the weights and reliability of the three indicators are calculated respectively, and the results are w ₁ =0.7764, w ₂ =0.7824, w ₃ =0.8993, r ₁ =0.4088, r ₂ =0.4604, r ₃ = 0.5588.

Evidence reasoning method is used to evaluate the safety of the structure, and the safety confidence distribution form obtained from the evaluation is shown in Figure 8. The confidence distribution is then converted into safety utility, and the results are shown in Figure 9.

According to Fig. 9, the safety of the structure gradually decreases with the increase of the number of measurements, which is consistent with the trend of increasing wear degree and deformation of the surface during the working process of large industrial structures. Therefore, the present invention is effective.

Claims (6)

1. a large-scale industrial structure deformation monitoring and analysis method, it is characterized in that: build a measuring platform and establish the space rectangular coordinate system of the built measuring platform; be provided with a motion platform in the described measuring platform; be provided with on the described motion platform camera; the camera includes a measuring camera and a transfer camera; solving the motion change parameters of the measuring platform; solving the external parameters of the measuring camera; solving the coordinates of the point to be measured and determining the deformation size of the large industrial structure; Safety assessment of industrial structures. 2. The large-scale industrial structure deformation monitoring and analysis method according to claim 1, characterized in that, said building a measuring platform and establishing a space rectangular coordinate system of the built measuring platform, comprising: Select a moving platform, fix the measuring camera and the transfer camera on the selected running platform in turn, and at the same time make the moving platform move around the surface of the large industrial structure being monitored; Select the space origin, establish the world coordinate system O-XYZ, and at the same time establish two sets of moving measuring camera coordinate systems O _D,i -X _D,i Y according to the positions of the measuring cameras at different times t _i and t _{i+1 D,i} Z _D,i and O _D,i+1 -X _D,i+1 Y _D,i+1 Z _D,i+1 , and establish the transfer camera coordinate system O _T -X _T Y _T Z _T with the motion platform coordinate system OP -X _P Y _P Z _{P .} 3. The large-scale industrial structure deformation monitoring and analysis method according to claim 2, wherein the motion variation parameter of the solution measurement platform comprises: Select the measurement point S (X, Y, Z) and the corresponding image point coordinates s _i (x _i , y _i ), s _i+1 (x _i+1 , y _i+1 ) to establish the measurement point S and the image point Homogeneous coordinate relationship corresponding to s _i and s _i+1 ; Establish the relationship between P in O-XYZ system and O _P -X _P Y _P Z _P system, and between O _P -X _P Y _P Z _P system and O _T- X _T Y _T Z _T system , and further establish the relationship between S in the O-XYZ system and the O _T -X _T Y _T Z _T system; Using the transfer camera to shoot the artificial marking points in real time, according to the relationship between the marking points and the corresponding image points, the rotation matrix and translation vector of the O-XYZ system converted to the O _T -X _T Y _T Z _T system are obtained; Establish the conversion relationship between the O-XYZ system and the O _T -X _T Y _T Z _T system at the time t _i and t i ₊ 1, and then solve the motion change parameters of the platform between the time t _i and t _i+1 . 4. the deformation monitoring and analysis method of large-scale industrial structure according to claim 3, is characterized in that, described solving the external parameter of measuring camera, comprises: establish t _i and t _i+1 moment O _P -X _P Y _P respectively Alignment of Z _P series with O _D,i -X _D,i Y _D,i Z _D,i series and O _D,i+1 -X _{D,i+ 1} Y _D,i+1 Z _D,i+1 series Secondary coordinate relationship; Establish the relationship between the O-XYZ system and the O _D,i -X _D,i Y _D,i Z _D,i system at different times; Select the initial moment to measure the camera coordinate system O _D,0 -X _D,0 Y _D,0 Z _D,0 , and obtain the system O _D,i -X _D,i Y _D,i Z _D,i at any moment through iteration The relationship between the position and attitude relative to the O-XYZ system, that is, the external parameters of the camera are measured. 5. The deformation monitoring and analysis method of a large-scale industrial structure according to claim 4, wherein the solution for solving the coordinates of the point to be measured and determining the deformation size of the large-scale industrial structure comprises: Establish the position equation of the measurement point S at time t _i ; Establish the position equation of the measurement camera at the initial moment, and then obtain the world coordinate of the measurement point S; Repeat the above measurement steps to obtain the world coordinates of the monitoring points at different positions at different times; According to the measurement position of the monitoring point and the actual position of the identification point, the corresponding deformation variable is solved. 6. The large-scale industrial structure deformation monitoring and analysis method according to claim 5, characterized in that, the safety assessment of the large-scale industrial structure comprises: determining the deformation weight based on the coefficient of variation method; calculating the reliability of the deformation based on the distance; Evidence-based reasoning to assess the safety of large industrial structures.

Images