CN113946989B - Structural damage detection method based on generalized pattern search algorithm - Google Patents
Structural damage detection method based on generalized pattern search algorithm Download PDFInfo
- Publication number
- CN113946989B CN113946989B CN202110529763.5A CN202110529763A CN113946989B CN 113946989 B CN113946989 B CN 113946989B CN 202110529763 A CN202110529763 A CN 202110529763A CN 113946989 B CN113946989 B CN 113946989B
- Authority
- CN
- China
- Prior art keywords
- damage
- moment
- acceleration
- displacement
- structural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000006378 damage Effects 0.000 title claims abstract description 138
- 238000001514 detection method Methods 0.000 title claims abstract description 71
- 238000010845 search algorithm Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 115
- 230000001133 acceleration Effects 0.000 claims abstract description 46
- 238000006073 displacement reaction Methods 0.000 claims abstract description 41
- OXURYBANZVUSFY-UHFFFAOYSA-N 2-[3-(diaminomethylideneamino)propyl]butanedioic acid Chemical compound NC(N)=NCCCC(C(O)=O)CC(O)=O OXURYBANZVUSFY-UHFFFAOYSA-N 0.000 claims description 27
- 102100032610 Guanine nucleotide-binding protein G(s) subunit alpha isoforms XLas Human genes 0.000 claims description 27
- 101001014590 Homo sapiens Guanine nucleotide-binding protein G(s) subunit alpha isoforms XLas Proteins 0.000 claims description 27
- 101001014594 Homo sapiens Guanine nucleotide-binding protein G(s) subunit alpha isoforms short Proteins 0.000 claims description 27
- 101001014610 Homo sapiens Neuroendocrine secretory protein 55 Proteins 0.000 claims description 27
- 101000797903 Homo sapiens Protein ALEX Proteins 0.000 claims description 27
- 239000011159 matrix material Substances 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 230000004044 response Effects 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 13
- 238000001914 filtration Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 abstract description 18
- 238000004458 analytical method Methods 0.000 abstract description 16
- 238000004088 simulation Methods 0.000 abstract description 15
- 238000005457 optimization Methods 0.000 abstract description 7
- 238000004364 calculation method Methods 0.000 abstract description 6
- 238000005259 measurement Methods 0.000 abstract description 2
- 238000009825 accumulation Methods 0.000 abstract 1
- 238000003745 diagnosis Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 8
- 230000005284 excitation Effects 0.000 description 6
- 230000009471 action Effects 0.000 description 5
- 238000005452 bending Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000005070 sampling Methods 0.000 description 4
- 239000011150 reinforced concrete Substances 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000013016 damping Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000036039 immunity Effects 0.000 description 2
- 230000001737 promoting effect Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000004567 concrete Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Geometry (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Architecture (AREA)
- Civil Engineering (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention belongs to the technical field of building structure damage detection, and particularly relates to a structural damage detection method based on a generalized pattern search algorithm. According to the method, based on a statistical moment theory derived by a single-degree-of-freedom system, a generalized mode search algorithm is integrated, a fused displacement fourth-order moment and an acceleration eighth-order moment are used as indexes, numerical simulation is carried out on the condition that a 12-layer structure numerical model has noise or not, the accuracy and the noise resistance of the method are analyzed, the method is compared with a statistical moment two-norm optimization method and a Bayesian-like thought method under the same indexes, superiority of the method in damage detection precision, stability and quick detection calculation time is verified, 3 typical working conditions are selected to respectively carry out damage diagnosis through 12-layer standard frame vibration table test data, and the method is compared with actual measurement report analysis, so that the result shows that the method can reflect unit damage degree changes which are shown by accumulation of test working conditions.
Description
Technical Field
The invention belongs to the technical field of building structure damage detection, and particularly relates to a structural damage detection method based on a generalized pattern search algorithm.
Background
The traditional nondestructive or semi-destructive building structure has the defects of large detection workload and high cost, and people or test instruments at certain parts of the structure cannot reach the structure and cannot detect the structure; and only damage on or near the surface of the structure can be detected, real-time comprehensive information of the structure is difficult to obtain, and the detection of the whole damage based on the sampling result often depends on personnel experience and subjective judgment. Compared with the traditional detection method of the building structure, the method for detecting the structural damage based on the dynamic test is a hot spot of academic research, and is expected to solve the problems.
At present, the identification methods for structural power detection are mainly divided into a model-free damage detection method and a model damage detection method. Liu Wenfeng and the like use a simply supported beam and a clamped beam as research objects, and propose a model-free damage detection method for positioning the damage position of a structure based on the change rate of the natural frequency of a first three-order structure. Mo Xiaopeng and the like, which take a cantilever beam as an analysis object, propose to take a model-free index of the first order vibration mode change rate as a structural damage detection index, and apply a neural network to identify the damage position. Cao Hui, establishing a simply supported beam and continuous beam finite element model, and identifying the structural damage position based on the mode flexibility difference model-free index of the first vibration mode. Yang Yang et al propose a recognition method for quickly judging whether a structure has damage or not and the damage position by directly using an output signal by using the structural acceleration eighth moment as a model-free damage index in a time domain.
The model-free method is simple and quick in identifying structural damage, limited in accuracy and large in difference from practical application; the model methods such as the residual force vector method, the strain mode method, the direct stiffness method, the characteristic value sensitivity method, the wavelet analysis method and the like need to analyze the integral parameters of the structure, and the method can be relatively accurate but takes longer time and is not efficient when applied to detecting the local damage position. Zhang et al propose a three-layer planar structure model damage detection method with a fourth-order acceleration statistical moment as a damage index; furthermore, yang Yang and the like compare the characteristics of the acceleration statistical moment and the displacement statistical moment on the basis, and propose a conventional two-norm and Bayesian-like damage detection method using the fused displacement fourth-order moment and acceleration eighth-order moment as damage indexes, and preliminary verification is carried out through numerical calculation examples and experimental researches, but the accuracy and the high efficiency under the noise level still remain to be solved.
Disclosure of Invention
Aiming at the defects existing in the prior art, the invention provides a structural damage detection method based on a generalized pattern search algorithm, the invention combines the generalized pattern search algorithm on the basis of a statistical moment theory, integrates the characteristics of both model-free and model-free methods, the method comprises the steps of firstly, rapidly detecting a structural damage region by a model-free method of combining displacement fourth-order moment and acceleration eighth-order moment damage indexes, and then rapidly and accurately identifying the positions and the degrees of damaged components of the structural damage region by a generalized pattern search algorithm based on a model method aiming at a local region, so that a novel simple, rapid and efficient practical power detection method is formed. The novel method can effectively identify the damage position and damage degree, and has considerable advantages in identification efficiency, precision and calculation time as other methods. According to the invention, 3 indexes and the research and comparison under the 3-class damage detection method are carried out aiming at different damage working conditions of the 12-layer standard frame model and 3 typical working conditions of the vibration table test, so that the reliability of the invention in engineering detection is verified.
In order to solve the technical problems, the invention adopts the following technical scheme:
The structural damage detection method based on the generalized pattern search algorithm comprises the following steps:
the first step: detecting a damaged area by using a model-free method;
And a second step of: the damaged member and extent were detected for the damaged area using GPSA method.
Further, the first step specifically includes:
1) Arranging the horizontal acceleration sensor at the same vertical position in different horizontal areas of the structure;
2) Calculating a displacement fourth-order moment M 4,d and an acceleration eighth-order moment M 8,a of each measuring point after filtering the first-order frequency according to the displacement and acceleration time-course response of each measuring point of the structure;
3) According to the model-free method, the change amounts of the displacement fourth-order moment and the acceleration eighth-order moment and the reference value are respectively calculated by using the formula (3), and when the statistical moment change amount of a certain floor is larger than the reference value, the floor area is judged to be damaged.
Further, the formula (3) is:
wherein M 4,d,u,M8,a,u is displacement fourth-order moment and acceleration eighth-order statistical moment under a nondestructive working condition, and M 4,d,d,M8,a,d is acceleration eighth-order statistical moment under a damage working condition.
Further, the second step specifically includes:
4) Firstly, establishing a finite element model of an actual structure, and giving an initial nondestructive rigidity value E 0, wherein an initial step length delta 0 is more than 0;
5) Calculating an objective function value, i.e., a two-norm value of the residual, |f (E k)||2,||F(Ek)||2＝||M-M(Ek)||2, where k=1, 2,., n;
6) Determining a direction matrix search step s k by the mode matrix P k and the step size parameter delta 0; calculating a difference value rho k＝||F(Ek)||2-||F(Ek+sk)||2;
7) If the difference ρ k > 0 is judged to be true, E k+1＝Ek+sk is caused, otherwise E k+1＝Ek is caused;
8) Updating the generation matrix C k and the step size parameter delta k, and when the algorithm is terminated, obtaining an objective function value F (E k)||2 is minimum, and the structural rigidity E k is the structural real estimated value at the moment), thereby judging the specific damaged member and the damage degree in the structural floor area.
Compared with the prior art, the invention has the following beneficial effects:
1. the displacement fourth-order moment and the acceleration eighth-order moment are fused, so that the method is suitable for rapid damage detection indexes of a model-free method and GPSA algorithm, and has good accuracy and noise immunity.
2. Compared with the result of numerical simulation and experimental detection of the GPSA method, the two-norm optimization method and the Bayesian-like thought method, the method has more advantages in recognition accuracy, time and stability.
3. The method has the advantages of combining the rapid detection of the damaged area without the model with the rapid and accurate detection of the position and degree of the damaged component by GPSA algorithm, and is favorable for promoting the combination of the method without the model and the method with the model to be applied to practical engineering.
Drawings
FIG. 1 is a diagram of a twelve-layer structural model in an embodiment of a structural damage detection method based on a generalized pattern search algorithm according to the present invention;
FIG. 2 is a diagram of the result of performing numerical simulation analysis on a single damage condition 2 in an embodiment of the structural damage detection method based on a generalized pattern search algorithm under the conditions of no noise, 40DB and 30DB signal to noise ratio;
FIG. 3 is a diagram showing the results of numerical simulation analysis of a plurality of damage conditions 4 in an embodiment of the structural damage detection method based on a generalized pattern search algorithm under the conditions of no noise, 40DB and 30DB signal to noise ratio;
FIG. 4 is a diagram of the detection result of the damage of working condition 2 in the embodiment of the structural damage detection method based on the generalized pattern search algorithm of the present invention;
FIG. 5 is a diagram of the result of detection of damage in condition 4 (beam integral unit) in an embodiment of the structural damage detection method based on the generalized pattern search algorithm of the present invention;
FIG. 6 is a diagram of the result of detecting damage in condition 4 (right column unit) in an embodiment of the method for detecting structural damage based on a generalized pattern search algorithm according to the present invention;
FIG. 7 is a diagram of the result of detecting damage in condition 4 (left column unit) in an embodiment of a method for detecting structural damage based on a generalized pattern search algorithm according to the present invention;
FIG. 8 is a graph of SD versus time (no noise) for each method;
FIG. 9 is a graph (40 db) of SD versus time for each method;
FIG. 10 is a graph (30 db) of the SD versus time for each method;
FIG. 11 is a graph of SD versus time for each method (versus time for each method);
FIG. 12 is a diagram of a test pattern of a vibrating table in an embodiment of a structural damage detection method based on a generalized pattern search algorithm of the present invention;
FIG. 13 is a graph showing a twelfth layer displacement time-course response under the action of white noise in an embodiment of a structural damage detection method based on a generalized pattern search algorithm according to the present invention;
FIG. 14 is a graph showing the detection result of the bending stiffness ratio of each layer of beam under the action of working condition 1 in the embodiment of the structural damage detection method based on the generalized pattern search algorithm;
FIG. 15 is a graph showing the detection result of the bending stiffness ratio of each layer of beam under the action of the working condition 9 in the embodiment of the structural damage detection method based on the generalized pattern search algorithm;
FIG. 16 is a graph showing the results of the detection of the flexural rigidity ratio of each layer of beams under the action of the working condition 16 in the embodiment of the structural damage detection method based on the generalized pattern search algorithm.
Detailed Description
In order that those skilled in the art will better understand the present invention, the following technical scheme of the present invention will be further described with reference to the accompanying drawings and examples.
Wherein the drawings are for illustrative purposes only and are shown in schematic, non-physical, and not intended to be limiting of the present patent; for the purpose of better illustrating embodiments of the invention, certain elements of the drawings may be omitted, enlarged or reduced and do not represent the size of the actual product; it will be appreciated by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
1 Theory of
1.1 Statistical moment theory
For multi-degree of freedom building structures, the equation of motion of the structure can be expressed as
Wherein M, C, K are the mass matrix, damping matrix and stiffness matrix of the structure respectively; Acceleration, velocity and displacement time-course responses of the structure under external excitation; f (t) is an external stimulus, f (t) = [ f 1(t),f2(t),...,fN(t)]T ]. Solving to obtain structural acceleration response And a displacement response x j＝[x1j,x2j,...,xNs,j ], whereinFor the sampling data set of the acceleration response of the jth measuring point, x j is the sampling data set of the displacement response of the jth measuring point, N s is a sampling point, and the statistical moment is expressed as follows:
the expression of the change amount of the displacement fourth-order moment and the acceleration statistical moment is as follows:
wherein M 4,d,u,M8,a,u is displacement fourth-order moment and acceleration eighth-order statistical moment under a nondestructive working condition, and M 4,d,d,M8,a,d is acceleration eighth-order statistical moment under a damage working condition.
In the prior art, it is proposed that an acceleration eighth-order moment is used as a damage index, a damage area is judged by comparing the magnitude of a reference value BS a (i.e. the sum of a mean value mu a and a standard deviation sigma a of 1.645 times) of the statistical moment variation of each layer of the acceleration eighth-order statistical moment variation eta a, and when the measured floor statistical moment variation is larger than the reference value, the floor area is considered to be damaged. However, as only the structural acceleration is considered and the structural displacement time-course response is not considered, the method is insensitive to damage detection of damage working conditions of multiple areas. The invention adds the variable quantity eta d of the displacement fourth-order moment as a model-free damage index, and the numerical simulation below shows that the displacement fourth-order moment can be effectively identified for damage working conditions of a plurality of areas. Therefore, the invention takes the change amounts of the displacement fourth-order moment and the acceleration eighth-order moment of each floor area as damage indexes, and judges that the floor area of the structure is damaged when one index statistical moment change amount is larger than a corresponding reference value.
In summary, the displacement fourth-order moment and the acceleration eighth-order moment are fused to serve as damage indexes, and the relation between the variation quantity of a single index and the reference value of the displacement fourth-order moment and the acceleration eighth-order moment is combined, so that damage areas of single damage working conditions and multiple damage working conditions can be accurately identified, and compared with single index identification, the time is not increased.
1.2 Generalized pattern search algorithm
GPSA is a special subset group in the direct search algorithm, mainly extracting an objective function from a special direction set, and finding out a descending direction by comparing the magnitudes of function values so as to solve the optimization problem. Because the method does not need to calculate any derivative, the convergence of the algorithm can be obtained, and compared with other algorithms, the method is easy to realize and has high calculation efficiency. The search step s k of the GPSA algorithm is jointly determined by the pattern matrix P k (determining the direction of s k) and the step size control parameter Δ k (determining the step size of s k). Defining a mode matrix:
Pk＝BCk＝[BMk,-BMk,BLk]＝[BΓk,BLk] (4)
wherein: the base matrix B is any nonsingular matrix The iteration process is generally kept unchanged; the generator matrix C k∈Zn×p (p >2 n) is arranged as follows:
Ck＝[Mk,-Μk,Lk]＝[Γk,Lk] (5)
Wherein: m k∈M∈Zn×n (M is a finite set of non-singular matrices). L k∈Zn×(p-2n) contains at least one 0 column so that the search can be returned to the initial point as soon as possible when it fails. Since the ranks of B and C k are both n, P k constitutes Is a positive base of (a). Therefore, at any iteration step k, a heuristic point x i k＝xk+Si k＝xk+ΔkBci k centered on x k is defined, where the search step isSearch direction Bc i k(Ck＝[ci k…cp k). The exploratory move for GPSA has the following two assumptions: s k∈ΔkPk≡ΔkB Ck≡Δk[BΓk,BLk ]; if it isThen there is f (x k+sk)＜f(xk). When the step control parameter Δ k is sufficiently small, the algorithm is considered to have converged (at this time, the gradient in all directions is 0, and no feasible descent direction can be found), and the algorithm is terminated.
Based on the theory, the statistical moment theory is combined to minimize the second norm of the residual error of the objective function I F (E) I 2, and the formula of F (E) is as follows:
wherein M (E) is an analog statistical moment, Is the measured statistical moment.
And finally obtaining the rigidity E of the structure under the condition of minimum objective function to be the true estimated value of the structure, thereby detecting the damage degree of the structure.
In combination with GPSA statistical moment numerical simulation research, the method has the advantages that GPSA algorithm is performed on a local floor area to identify the position and degree of a specific floor damaged member in the damage detection process taking fusion displacement fourth-order moment and acceleration eighth-order moment as indexes, the robustness is good, the method is superior to other indexes, the damage degree of a damaged unit can be accurately identified in a targeted manner, and the error is small. The situation that the damage position and the damage degree of the structure are recognized in a global optimization mode based on GPSA, the time is long, and misjudgment is easily caused by damage influence of adjacent areas is avoided. Therefore, the invention provides a novel method for rapidly identifying the damaged area by a model-free method and rapidly and accurately identifying the position and the damage degree of a specific component in the damaged area by a GPSA method.
1.3 New method calculation flow
Based on a statistical moment theory, a generalized mode search algorithm is integrated, the novel method detects the damage position and the damage degree of the structure according to a two-step method, and the main calculation flow is as follows:
the first step: detection of damaged areas using model-free methods
1) The horizontal acceleration sensor is arranged at the same vertical position in different horizontal areas of the structure
2) Calculating a displacement fourth-order moment M 4,d and an acceleration eighth-order moment M 8,a of each measuring point after filtering the first-order frequency according to the displacement and acceleration time-course response of each measuring point of the structure;
3) According to the model-free method, the change amounts of the displacement fourth-order moment and the acceleration eighth-order moment and the reference value are calculated respectively by using the formula (3), and when the statistical moment change amount of a certain floor is larger than the reference value, the floor area is judged to be damaged;
And a second step of: detecting damaged member and extent for damaged area using GPSA method
4) Firstly, establishing a finite element model of an actual structure, giving an initial nondestructive rigidity value E 0, and setting an initial step length delta 0 to be more than 0
5) Calculating an objective function value, i.e., a two-norm value of the residual, |f (E k)||2,||F(Ek)||2＝||M-M(Ek)||2, where k=1, 2,., n;
6) Determining a direction matrix search step s k by the mode matrix P k and the step size parameter delta 0; calculating a difference value rho k＝||F(Ek)||2-||F(Ek+sk)||2;
7) If the difference ρ k > 0 is judged to be true, E k+1＝Ek+sk is caused, otherwise E k+1＝Ek is caused;
8) The generator matrix C k and the step size parameter Δ k are updated. When the algorithm is terminated, the obtained objective function value F (E k)||2 is the smallest, and the structural rigidity E k is the structural real estimated value at this time), and thus the specific damaged component and the damage degree in the structural floor area are judged.
2 Numerical simulation
The numerical simulation analysis model of the invention is a 12-layer planar frame model, the heights of all layers are 3m, the span of frame beams is 6m, the elastic modulus of frame beams and columns is 3X 10 10N·m2, and the linear densities are respectivelyDamping ratio ζ i =0.05 (i=1, 2, 3..n), planar frame is shown in fig. 1.
According to the invention, firstly, the simulation structure carries out numerical simulation under Gaussian white noise excitation and under the condition of no consideration of environmental noise, and the damage simulation working condition is set as follows. The damage degree of the invention is the degree of reduction of the elastic modulus E.
Working condition 1: the structure is not provided with damage;
Working condition 2: the first layer beam unit is damaged by 10%;
Working condition 3: the first layer beam unit is damaged by 20%;
Working condition 4: the first layer beam unit and the second layer beam unit are damaged by 20 percent simultaneously;
working condition 5: the first and third layer beam units are damaged by 20% at the same time.
2.1 Determining model-free method damage index
On the basis of a model-free method, the structural damage position is judged by using displacement fourth-order moment and acceleration eighth-order moment as indexes. And carrying out numerical simulation analysis on the single damage working condition 2 and the multiple damage working conditions 4 under the noise-free, 40DB and 30DB signal-to-noise ratios, wherein the results are respectively shown in fig. 2 and 3, the 1 st to 12 th points in the figures are floors, and the XIII point on the abscissa is a reference value.
For the single damage working condition, as can be seen from fig. 2, the statistical moment variation of the acceleration eighth moment layer 1 is larger than the reference value, namely the damage of the layer 1 area is identified; the change of the layer 1 statistical moment of the displacement fourth-order moment is obviously larger than the reference value, the change of the layer 2 is slightly smaller than the reference value, misjudgment is easy to cause, but the floor area 1 can be judged to be the damage position by combining two index results, and the damage position is consistent with the assumption of the original working condition 2.
For a plurality of damage working conditions, as can be seen from fig. 3, the acceleration eighth moment variation is smaller than the reference value, i.e. the damage region of the structure cannot be identified. And the statistical moment variation of the 1 st and 2 nd floors of the displacement fourth-order moment is larger than a reference value, namely the 1 st and 2 nd floors of the structure are identified as damaged areas. Is in accordance with the assumption of a plurality of damage conditions 4.
In summary, the acceleration eighth moment is used as an index to identify a damaged area of a single damaged working condition, and insensitive phenomenon exists in the identification of the damaged area of a plurality of damaged working conditions; the displacement fourth-order moment is sensitive to multiple damage working conditions, and the single damage working condition is poor in sensitivity compared with the acceleration eighth-order moment. Therefore, the invention adopts the displacement fourth moment and the acceleration eighth moment as the damage index to judge the damage area of the structure.
2.2 New method injury detection
The damage detection of the new method adopts a two-step method: the first step uses a model-free method to rapidly detect the damaged area, and the second step uses GPSA method to identify the position and degree of a specific component aiming at the damaged area.
The first step: and identifying a single damage working condition 2 by using a fourth-order displacement moment and an acceleration eighth-order moment as indexes by using a model-free method, and judging that the 1 st layer of the structure is a damage area through the knowledge of the upper section.
And a second step of: the damage degree is identified by GPSA method for the damaged area, i.e. the layer 1 beam unit, the left column unit, and the right column unit, as shown in fig. 4. The figure shows that the first layer beam unit has 20% damage, the left column and the right column have no damage, the error is within 0.5% under the noiseless working condition, the error is within 1% under the 40DB working condition, the error is within 3% under the 30DB working condition, and the working condition accords with the original assumed working condition.
And similarly, performing damage detection on a plurality of damage working conditions 4. The first step: and identifying a damaged area by using a model-free method, wherein the 1 st layer and the 2 nd layer are the damaged areas. And a second step of: the method GPSA is used to identify the extent and extent of damage to the damaged area, i.e., layers 1 and 2 of the structure, as shown in fig. 5, 6 and 7. As can be seen from fig. 5, 6 and 7, the first and second beam units have 20% of damage, the left and right columns have no damage, the error is within 0.5% under the noiseless working condition, the error is within 1% under the 40DB working condition, the error is within 3% under the 30DB working condition, and the working conditions are consistent with the original assumed working conditions.
In conclusion, the novel method is proved to have better accuracy and noise resistance by detecting the structural damage region by using a model-free method and identifying the position and degree of a specific component of the structural damage region by using a GPSA method, and has important significance in practical engineering.
2.3 Comparing with other methods with the same index
Compared with other statistical moment methods, the novel method provided by the invention does not need to carry out damage detection analysis on each component in the structural global, and combines a model-free damage area detection method to carry out damage detection on damaged components more quickly, so that repeated identification on the damaged components is reduced, and the damage detection time is greatly shortened. The invention compares and analyzes the recognition accuracy and the recognition time by comparing GPSA method with the statistical moment two-norm optimization method and applying Bayesian-like thought method, and considers the comparison under the conditions of no noise, 40DB noise and 30DB noise, and the results are shown in figures 8, 9, 10 and 11. Defining the damage detection error condition of the standard deviation SD reaction structure damaged unit:
Wherein i is the number of the damage units to be detected, x i is the value obtained by numerical simulation, The larger SD is the structural unit damage set value, which indicates that the larger the damage detection result error is.
Also selecting the working conditions 1 to 5 for comparison analysis, and considering the influence of different noise levels, wherein the results are shown in fig. 8, 9, 10 and 11;
As shown in fig. 8, 9 and 10, in the damage detection process considering no noise, 40DB noise and 30DB noise level, the standard deviation SD value of the damage detection method GPSA is generally smaller than that of the two-norm method and the bayesian-like concept method, and the error of the damage detection standard deviation SD value is close to 0 under the noise-free working condition; under the 40DB noise working condition, the SD value error of the GPSA method is also controlled within a range of 1 percent and is far smaller than the SD value of the identification error standard deviation of the other two methods; under the 30DB noise working condition, the error of the SD value of the GPSA method is about 3% at maximum and is generally smaller than the SD value of other two methods, and the advantages are obvious in the double damage working condition identification. In terms of time contrast, as shown in fig. 11, the GPSA method takes time that is not much different from the bayesian-like ideas method, but less than the statistical moment two-norm method. In combination, the method of the invention has better accuracy and can be applied to rapid damage detection under the condition of adopting the same fourth-order displacement statistical moment and acceleration index as the other two methods.
3 Test analysis
3.1 12-Layer laboratory frame model
A 12-layer standard frame vibration table test in the prior art was selected, and the test model was a single span 12-layer Reinforced Concrete (RC) structure, as shown in fig. 12. The structure is a 1/10 reduced scale model, and the model material is particulate concrete and galvanized iron wires. The actual structural fitment and 50% live load are considered, and the mass block is arranged on the plate for counterweight. 19.4kg of weight was placed on each of the standard layers, and 19.7kg of weight was placed on the roofing layer. The test places accelerometers and strain sensors at different locations on the basis of the model structure, 2 layers, 4 layers, 6 layers, 8 layers, 10 layers and 12 layers, respectively. In the test process, four seismic waves of Shanghai artificial wave, EI Centro wave, kobe wave and Shanghai bedrock wave and white noise are adopted as load excitation, and the specific loading mode and the detection device setting prior art are described in detail and are limited to spread.
3.2 Comparison of calculated results
3.2.1 Theoretical model verification comparison
According to related information of the prior art model, an X-direction plane model is established according to a consistent mass matrix and a consistent stiffness matrix, the first 3-order frequency of the structure is basically matched with the actual measurement frequency through analysis, the excitation of white noise and seismic waves to the theoretical model is simulated, the displacement time-course response and the displacement time-course response of the structure are compared, and as shown in fig. 13, under the excitation action of the white noise, the result curves of the two are basically coincident, so that the simplified plane model can be used for optimizing analysis.
3.2.2 Lateral analysis of test Condition
The test is carried out for 62 working conditions in total, as shown in table 1, according to the description of the test phenomenon in the prior art, no crack is found on the surface of the structure under the first 8 working conditions, wherein the working condition 1 is the first white noise input and is regarded as a lossless working condition; after the 9 th working condition of the offshore artificial wave, a fine vertical crack which develops from top to bottom and from bottom to top is formed at the beam end of the fourth layer of frame beam parallel to the X vibration direction, and the width of the crack is smaller than 0.05mm; after the third white noise input, the working condition 16 is that the beam ends of the 4-6 layers of frame beams parallel to the X vibration direction are provided with vertical cracks, the width of the crack is about 0.08mm, no crack is observed in each column, and no crack is found in the frame beam columns parallel to the Y vibration direction. Therefore, the invention selects 3 representative working conditions corresponding to the typical working condition 1 (lossless working condition), the working condition 9 (working condition in which the crack is observed for the first time) and the working condition 16 (working condition in which the crack is observed for 3 places) for analysis, and compares the test results of the novel method, the statistical moment two-norm method and the Bayesian idea-like method in the identification of the working conditions.
TABLE 1 list of 62 conditions
Note that:
EL-El Centro wave (X unidirectional);
EY-El Centro wave (X, Y bi-directional);
EZ-El Centro wave (X, Y, Z three directions);
KB-Kobe wave (X unidirectional);
KY-Kobe wave (X, Y bi-directional);
KZ-Kobe wave (X, Y, Z three directions);
SH-Shanghai artificial wave (X unidirectional);
sj—Shanghai bedrock wave (X unidirectional);
X∶Y∶Z＝1∶0.85∶0.5。
In the working condition 1, the bending rigidity of the beam and column members of each layer of the 12-layer frame is identified according to the theoretical detection by analyzing the displacement time-course response of the corresponding floor sensor, and the final detection result of each layer of the beam is shown in fig. 14. As can be seen from fig. 14, the difference between the optimized result and the theoretical stiffness value of the structure is within 3%, and the detection result is consistent with the experimental phenomenon within 5% of engineering precision.
The working condition 9 is the working condition that cracks are observed for the first time, the bending stiffness of each layer of beam is detected by analyzing the actually measured response data after the input of the artificial wave in the Shanghai, and the ratio of the identified stiffness to the theoretical stiffness is shown in figure 15. As can be seen from FIG. 15, the bending stiffness of the fourth layer beam is reduced maximally, the damage degree is close to 11%, the maximum error value of damage detection at other positions is close to 4%, and the maximum error value is smaller than the error range of 5% of the engineering progress, and the maximum error value is basically consistent with the test phenomenon.
The working condition 16 is a working condition that cracks are observed in the 4 th, 5 th and 6 th layers of beams in the X vibration direction under the third white noise excitation, and the ratio of the identification stiffness to the theoretical stiffness of each layer of beams is obtained through analysis by using measured response data, and is shown in fig. 16. The identification result shows that the damage degree of the fourth layer beam is most obvious, and the damage degree is close to 25%; the beams of the 5 th layer and the 6 th layer are damaged to a certain extent, wherein the damage degree of the beam of the fifth layer is about 21%, the damage degree of the beam of the sixth layer is about 19%, the maximum error of the rigidity ratio of the other beams is 5.5%, and the range of the stiffness ratio approaches to the engineering progress by 5%, and the beam is basically consistent with the test phenomenon.
As can be seen from comparison of the final analysis results of the test methods shown in FIGS. 14 to 16, the method of the present invention is superior to the other two methods in terms of recognition accuracy and stability, and the method is shown to be better applicable to actual engineering detection.
Conclusion 4
Based on a statistical moment theory, the method utilizes fusion displacement fourth-order moment and acceleration eighth-order moment as indexes, combines a model-free method and GPSA optimization method, verifies the feasibility and superiority of the method relative to other methods with the same indexes through numerical simulation, simulates a vibration table test of a 12-layer reinforced concrete structure for damage detection, and obtains the following conclusion:
1) The displacement fourth-order moment and the acceleration eighth-order moment are fused, so that the method is suitable for rapid damage detection indexes of a model-free method and GPSA algorithm, and has good accuracy and noise immunity.
2) Compared with the result of numerical simulation and experimental detection of the GPSA method, the two-norm optimization method and the Bayesian-like thought method, the method has more advantages in recognition accuracy, time and stability.
3) The method has the advantages of combining the rapid detection of the damaged area without the model with the rapid and accurate detection of the position and degree of the damaged component by GPSA algorithm, and is favorable for promoting the combination of the method without the model and the method with the model to be applied to practical engineering.
It should be noted that, the Generalized model search Algorithm (generally PATTERN SEARCH Algorithm, GPSA for short) can rapidly compare the magnitudes of the adjacent point function values in a local range, and compared with other algorithms for deriving objective function values or gradient information, the Algorithm has the advantages of reaching convergence with fewer iterations and evaluation times, and shows better robustness in engineering application.
The foregoing is merely exemplary of the present application, and specific structures and features well known in the art will not be described in detail herein, so that those skilled in the art will be aware of all the prior art to which the present application pertains, and will be able to ascertain the general knowledge of the technical field in the application or prior art, and will not be able to ascertain the general knowledge of the technical field in the prior art, without using the prior art, to practice the present application, with the aid of the present application, to ascertain the general knowledge of the same general knowledge of the technical field in general purpose. It should be noted that modifications and improvements can be made by those skilled in the art without departing from the structure of the present application, and these should also be considered as the scope of the present application, which does not affect the effect of the implementation of the present application and the utility of the patent.
The same or similar reference numbers in the drawings of embodiments of the invention correspond to the same or similar components; in the description of the present invention, it should be understood that, if the terms "upper", "lower", "left", "right", "inner", "outer", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, only for convenience in describing the present invention and simplifying the description, rather than indicating or implying that the apparatus or elements being referred to must have a specific orientation, be constructed and operated in a specific orientation, so that the terms describing the positional relationships in the drawings are merely for exemplary illustration and should not be construed as limiting the present patent, and that the specific meaning of the terms described above may be understood by those of ordinary skill in the art according to specific circumstances.
In the description of the present invention, unless explicitly stated and limited otherwise, the term "coupled" or the like should be interpreted broadly, as it may be fixedly coupled, detachably coupled, or integrally formed, as indicating the relationship of components; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between the two parts or interaction relationship between the two parts. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
Claims (2)
1. The structural damage detection method based on the generalized pattern search algorithm is characterized by comprising the following steps of:
the first step: detecting a damaged area by using a model-free method;
And a second step of: detecting a damaged member and a degree of damage to the damaged area by using GPSA method;
The first step specifically comprises the following steps:
1) Arranging the horizontal acceleration sensor at the same vertical position in different horizontal areas of the structure;
2) Calculating a displacement fourth-order moment M 4,d and an acceleration eighth-order moment M 8,a of each measuring point after filtering the first-order frequency according to the displacement and acceleration time-course response of each measuring point of the structure;
3) According to the model-free method, the change amounts of the displacement fourth-order moment and the acceleration eighth-order moment are calculated respectively by using the formula (3), and when the statistical moment change amount of a certain floor is larger than a reference value, the floor area is judged to be damaged;
The formula (3) is:
wherein M 4,d,u,M8,a,u is displacement fourth-order moment and acceleration eighth-order statistical moment under a nondestructive working condition, and M 4,d,d,M8,a,d is acceleration eighth-order statistical moment under a damage working condition.
2. The structural damage detection method based on a generalized pattern search algorithm according to claim 1, wherein: the second step specifically comprises the following steps:
4) Firstly, establishing a finite element model of an actual structure, and giving an initial nondestructive rigidity value E 0, wherein an initial step length delta 0 is more than 0;
5) Calculating an objective function value, i.e., a two-norm value of the residual, |f (E k)||2,||F(Ek)||2＝||M-M(Ek)||2, where k=1, 2,., n;
6) Determining a direction matrix search step s k by the mode matrix P k and the step size parameter delta 0; calculating a difference value rho k＝||F(Ek)||2-||F(Ek+sk)||2;
7) If the difference ρ k > 0 is judged to be true, E k+1＝Ek+sk is caused, otherwise E k+1＝Ek is caused;
8) Updating the generation matrix C k and the step size parameter delta k, and when the algorithm is terminated, obtaining an objective function value F (E k)||2 is minimum, and the structural rigidity E k is the structural real estimated value at the moment), thereby judging the specific damaged member and the damage degree in the structural floor area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110529763.5A CN113946989B (en) | 2021-05-14 | 2021-05-14 | Structural damage detection method based on generalized pattern search algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110529763.5A CN113946989B (en) | 2021-05-14 | 2021-05-14 | Structural damage detection method based on generalized pattern search algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113946989A CN113946989A (en) | 2022-01-18 |
| CN113946989B true CN113946989B (en) | 2024-09-27 |
Family
ID=79327340
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110529763.5A Active CN113946989B (en) | 2021-05-14 | 2021-05-14 | Structural damage detection method based on generalized pattern search algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113946989B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118362161B (en) * | 2024-04-17 | 2024-10-18 | 重庆大学 | Intelligent global monitoring and alarming method for construction scaffold |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110568069A (en) * | 2019-09-11 | 2019-12-13 | 重庆大学 | beam bridge damage identification method based on average value of acceleration ratio and GPSA (general purpose analysis System) algorithm |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110567661B (en) * | 2019-09-11 | 2021-02-09 | 重庆大学 | Bridge damage identification method based on generalized pattern search algorithm and axle coupling |
-
2021
- 2021-05-14 CN CN202110529763.5A patent/CN113946989B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110568069A (en) * | 2019-09-11 | 2019-12-13 | 重庆大学 | beam bridge damage identification method based on average value of acceleration ratio and GPSA (general purpose analysis System) algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113946989A (en) | 2022-01-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108226399B (en) | A kind of beam-string structure damage combined recognising method based on modal parameter | |
| Li et al. | Analysis and assessment of bridge health monitoring mass data—progress in research/development of “Structural Health Monitoring” | |
| CN103048102B (en) | A kind of state evaluating method of beam bridge | |
| Esfandiari et al. | Finite element model updating using frequency response function of incomplete strain data | |
| CN112985671B (en) | Damage judgment method based on full-cable-system cable force test and error adaptive analysis | |
| CN102829898B (en) | Internal force detecting method for hanger rod with shock absorber | |
| CN108280294B (en) | A kind of cable arch structure damage combined recognising method based on modal parameter | |
| JP6001740B1 (en) | High precision evaluation of structure transfer functions, earthquake response prediction, deterioration diagnosis system and method | |
| CN108875178A (en) | For reducing the probabilistic sensor arrangement method of distinguishing structural mode | |
| CN113946989B (en) | Structural damage detection method based on generalized pattern search algorithm | |
| CN106323575B (en) | The reinforced concrete frame structure performance test methods of bottom steel core concrete column | |
| CN107246944B (en) | A kind of Structural Damage Identification based on statistical moment theory | |
| KR102480382B1 (en) | Wind load estimation system based on artificial intelligence | |
| CN110427716B (en) | Statistical moment-based high-rise structure model-free damage identification method | |
| KR101227776B1 (en) | Designing method for displacement estimation on construction using ann and estimating method for displacement of construction | |
| CN103984856A (en) | System and method for predicting longitudinal crack width of wall body based on neural network | |
| RU2699918C1 (en) | Diagnostic method of technical state of buildings and structures | |
| CN110362902A (en) | A kind of single source Dynamic Load Identification method analyzed based on section by dimension | |
| CN117610374A (en) | Method for rapidly evaluating earthquake resistance of reinforced concrete structure after earthquake damage | |
| CN110501127A (en) | A kind of uniform beam damnification recognition method based on faulted condition inclination angle slope | |
| Zhang et al. | Structural damage identification with output-only measurements using modified Jaya algorithm and Tikhonov regularization method | |
| Skjaerbaek et al. | Identification of damage in reinforced concrete structures from earthquake records—optimal location of sensors | |
| CN112818577B (en) | Method for identifying post-fire damage of laminated beam based on deep learning theory | |
| CN115046727A (en) | Bridge nonlinear vortex-induced force accurate identification method based on Hilbert transform | |
| CN113935206B (en) | Damage identification method based on statistical moment theory and MH-GIBBS mixed sampling Bayesian algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |