A civil engineering structure damage detection method based on step limit technology

Abstract

The application discloses a kind of civil engineering structure damage detection methods based on step length limiting technique, according to the objective function of damage detection problem of time domain test data construction, it is solved by Gauss-Newton algorithm iteration, singular value decomposition is carried out to sensitivity matrix in iterative process, and it is adaptively constrained iteration increment in combination with preset step length limit value, amplify the minimum singular value of sensitivity matrix, effectively improve the ill-posedness of structural damage detection problem.The application can guarantee that each step iteration damage detection parameter obtained has physical meaning, avoids the solution distortion problem caused by error in test data.Verified by Benchmark structure test, the method can accurately detect the structure damage position, has strong anti-interference ability, and the calculation result has strong stability and rationality, and is suitable for damage detection and health assessment of various civil engineering buildings, bridges and other structures.

Description

Technical Field

[0001] This invention relates to a method for detecting damage to civil engineering structures based on limited step length technology, and belongs to the category of damage detection methods for civil engineering structures. Background Technology

[0002] Structural damage detection is a core technology for the health monitoring of civil engineering structures. By monitoring the dynamic response data of the structure, changes in physical parameters such as structural stiffness and mass can be inverted to assess structural damage. This is of great significance for ensuring the safe service of civil engineering structures, predicting structural defects, and formulating operation and maintenance plans.

[0003] Current mainstream technologies for structural damage detection in civil engineering rely on physical parameters such as structural mass, stiffness, damping, and boundary constraints to construct analytical models, transforming structural damage detection into an objective function optimization problem based on time-domain monitoring data. The Gauss-Newton algorithm, due to its advantages of fast convergence and high accuracy, is widely used in solving such optimization problems. It can iteratively search for the optimal solution that minimizes the objective function in the parameter vector space, thereby enabling the detection of structural damage parameters.

[0004] Limiting step size techniques are currently mainly used in the field of signal processing. Their core function is to optimize the iteration step size and improve the speed and stability of solving the objective function. However, the unavoidable test errors in the measured data have a significant impact on the ill-posedness of limiting step size techniques. Furthermore, the specific requirements for the step size limit are different for the accurate solution of damage detection problems formed by different civil engineering structures. As a result, the application of limiting step size techniques in the field of damage detection of civil engineering structures is very limited, and a mature iterative constraint and damage detection scheme has not yet been formed.

[0005] Damage detection in civil engineering structures is characterized by ill-posedness, meaning that even minor noise and measurement errors in the detection data can lead to significant deviations in the final solution, severely impacting the accuracy of damage detection. To mitigate this ill-posedness, existing techniques generally employ regularization to process iterative increments, which can suppress solution bias to some extent.

[0006] However, in actual engineering testing scenarios, the testing data often contains large errors. Even after regularization, there are still some problems with excessively large iteration increments in some iteration steps. This causes the damage detection parameters obtained through iterative solutions to deviate significantly from the actual physical parameters of the structure, resulting in the loss of physical meaning of the detection parameters. Ultimately, this leads to inaccurate damage status and location, failing to meet the needs of damage detection in engineering structures. Summary of the Invention

[0007] Purpose of the Invention: To overcome the shortcomings of existing technologies, this invention provides a method for detecting damage in civil engineering structures based on a step-size limit technique. It establishes an innovative step-size limit iterative constraint mechanism by pre-setting a step-size limit value greater than 0 and less than 1, amplifying the minimum singular value of the sensitivity matrix, effectively suppressing the adverse effects of test errors on the step-size limit technique, significantly improving the ill-posedness of structural damage detection problems, precisely limiting the iteration increment of each iteration step, ensuring that the iteration parameters of each step have physical meaning, and improving the rationality of the damage detection results. Furthermore, the step-size limit set by this invention, around 0.3, can accurately solve damage detection problems for different civil engineering structures, significantly improving the adaptability of this invention.

[0008] Technical Solution: To solve the above-mentioned technical problems, the present invention provides a method for detecting damage to civil engineering structures based on limited step length technology, comprising the following steps:

[0009] S1. Based on the time-domain test data of civil engineering structures, construct the objective function for the structural damage detection problem;

[0010] S2. The Gauss-Newton algorithm is used to iteratively solve the above objective function minimization problem. The sensitivity matrix required for each iteration is calculated, and the sensitivity matrix is ​​decomposed into singular values.

[0011] S3. Calculate the initial iteration increment based on the sensitivity matrix, and combine the preset step size limit and the singular value decomposition result to adaptively correct and obtain the final iteration increment, thereby realizing the limit constraint of the iteration step size.

[0012] S4. Repeated iterative calculation and step size constraint process, after iterative convergence, to obtain stable damage detection parameters, thus completing the damage detection of civil engineering structures.

[0013] Preferably, step S1 specifically involves the following steps:

[0014] S1.1 Collect time-domain test data of the structure and select the structural acceleration response data as the benchmark dataset;

[0015] S1.2 Establish a numerical analysis model of the civil engineering structure to be tested, define a damage detection parameter vector containing structural stiffness parameters, and calculate the structural simulation response corresponding one-to-one with the measured time domain data through the numerical model.

[0016] S1.3 Construct response residuals based on the difference between measured and simulated responses, integrate the response residuals of all measurement points, and form a residual vector that can characterize the overall response;

[0017] S1.4 Based on the least squares method, with the minimum residual vector as the optimization objective, a continuously differentiable objective function is constructed to complete the construction of the structural damage detection problem.

[0018] The specific steps for establishing a numerical analysis model of the civil engineering structure to be tested are as follows:

[0019] S1.2.1 Structural Discretization Process

[0020] Building structures: Discrete by floor, each floor is equivalent to a concentrated mass block, and the total mass of each floor is calculated from the floor dead load, live load and component self-weight;

[0021] Inter-floor connection: Horizontal spring units are used to simulate lateral stiffness between adjacent floors, and the spring stiffness is the damage detection parameter;

[0022] Bridges / complex structures: Finite element mesh discretization is performed using beam elements and shell elements, and element stiffness is used as a damage detection parameter.

[0023] S1.2.2 Define the fundamental matrix of dynamics

[0024] The quality matrix M is obtained by assembling the mass of each layer / unit and remains constant during model iteration.

[0025] Damping matrix C: Constructed using Rayleigh damping, calculated based on the damping ratio of the first two modes of the structure;

[0026] Stiffness matrix K ( ): Assembled from the initial stiffness of the elements, introducing a damage parameter vector. =[ 1, 2,..., n ],

[0027] i This represents the stiffness reduction factor for the i-th layer element.

[0028] Damage mechanism: i A decrease in numerical value indicates that the component at that location has cracked or the connection has failed, thus achieving a quantitative characterization of the location and extent of damage.

[0029] S1.2.3 Apply boundary constraints and external excitations

[0030] Boundary constraints: The bottom is fixed, the sliding support, the elastic support, etc. are strictly matched to the actual constraints of the project;

[0031] Excitation loading: Input the time-domain time history of the field measured environmental excitation, earthquake excitation, and vehicle excitation to ensure that the excitation is synchronized with the measured acceleration data.

[0032] S1.2.4 Define the damage detection parameter vector

[0033] The horizontal stiffness reduction coefficients of all layers or elements are combined into a parameter vector to be inverted, which is used as the solution variable for subsequent Gauss-Newton iterations. The vector dimension is equal to the number of damage regions to be identified.

[0034] S1.2.5 Time-domain dynamic solution, output matching simulation response

[0035] Based on the multi-degree-of-freedom vibration control equation: The Newmark-β method, a step-by-step integration method, is used to perform time series calculations, outputting the simulated acceleration time history for each measurement point and each sampling time. This ensures that the sampling time step, the number of measurement points, and the location of the measurement points correspond completely one-to-one with the actual measured time-domain acceleration data, thus obtaining the simulated response sequence required for S1.3.

[0036] Preferably, the minimization problem constructed in step S1 is:

[0037]

[0038] in, This is the result of damage detection calculation. The definition is shown in the following formula.

[0039]

[0040] In the formula, It is a damage detection parameter vector, and its dimension N is the number of damage detection parameters; It is a 2-norm operator; s[i] is the time-domain test data at time i, i=1,…,τ; z[i; Therefore The simulated response data corresponding to s[i] is calculated by the analytical model for the parameters, where τ is the total number of time-domain test data, and the analytical model is the existing interlayer lumped mass shear dynamics model; r is the result of s[i] - z[i; The calculated residual vector; superscript T This indicates the transpose operation.

[0041] Preferably, the sensitivity matrix S required for the k-th iteration in step S2 is... (k) Constructed using the Gauss-Newton algorithm, for S (k) The singular value decomposition formula is as follows:

[0042]

[0043] In the formula, matrix S (k) U (k) V (k) The number of rows × columns are τ×N, τ×N, and N×N, respectively. , They are U (k) V(k) The column vector, where N is the number of damage detection parameters; the diagonal matrix ∑ (k) It contains N diagonal elements, arranged in descending order from top left to bottom right, namely: , ... .

[0044] Preferably, the specific steps of step S3 are as follows:

[0045] S3.1. Based on the damage detection parameter vector from the (k-1)th iteration, calculate the initial iteration increment for the kth iteration. Let the damage detection parameter vector calculated in the (k-1)th iteration be... , , They are , The j-th element, where j = 1, ..., N;

[0046] S3.2. Let α be a pre-set step size limit greater than 0 and less than 1, so that the damage detection parameters obtained in each iteration step have physical meaning. Calculate the iteration step size correction coefficient M. (k) = Compared to M (k) With respect to the magnitude of α;

[0047] S3.3, when M (k) When M > α, the initial iteration increment is corrected to obtain the final iteration increment; when M (k) At time α, the initial iteration increment is retained as the final iteration increment, i.e.

[0048]

[0049] Preferably, step S4 involves repeating steps S2-S3 above to obtain a convergent sequence: Then, by the iterative formula (k) = (k-1) + (k) The following convergent sequence is formed: Where k=1,...,N i N i It is the number of iterations required for convergence. (0) These are the initial values ​​for the iterative calculation of the structural damage detection parameters; the damage detection calculation result is... .

[0050] Beneficial Effects: The civil engineering structural damage detection method based on limited step-size technology of this invention significantly reduces the ill-posedness of structural damage detection problems by amplifying the minimum singular value of the sensitivity matrix through step-size limiting. Simultaneously, it effectively limits the iterative increment in the structural damage detection calculation process, ensuring that the damage detection parameters obtained in each iteration step have physical meaning and guaranteeing the rationality of the calculation results. Using the method of this patented invention, the damage state of the benchmark structure designed by the International Structural Control Association and the American Society of Civil Engineers under the first-stage working condition can be accurately detected. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the process of the present invention.

[0052] Figure 2 This is a configuration diagram of the Benchmark structure.

[0053] Figure 3 The diagram shows the iterative calculation process of the damage detection results obtained in Table 2, where (a) is the curve of the residual modulus changing with the number of iterations, and (b) is the step size correction coefficient M. (k) The curves are shown in (c) and (d). (c) shows the curves before and after amplification of the minimum singular value of the sensitivity matrix. Detailed Implementation

[0054] The invention will now be further described with reference to the accompanying drawings.

[0055] like Figure 1 As shown, the present invention provides a method for detecting damage to civil engineering structures based on limited step length technology, comprising the following steps:

[0056] Step S1: Construct the objective function for the structural damage detection problem step by step.

[0057] The specific step-by-step process is as follows:

[0058] S1.1 Time-Domain Monitoring Data Preprocessing and Dataset Construction

[0059] Sensors are deployed at key floors and components of civil engineering structures to collect time-domain acceleration response data of the structure under normal service conditions or excitation conditions. The raw data is preprocessed to remove anomalous and abrupt data, repair missing time-series data, and align the time series of all monitoring data to ensure continuity and uniformity in the measured time dimension, ultimately forming a complete dataset of the structure's measured time-domain response.

[0060] S1.2 Structural Numerical Model Construction and Simulation Response Solution

[0061] Based on the design parameters, boundary constraints, material properties, and inter-story mass and stiffness distribution of the civil engineering structure under test, a refined numerical analysis model of the structure is established. A physical parameter vector is defined to characterize the structural damage state, with the horizontal stiffness of each story being the preferred parameter. The numerical changes in the parameter vector characterize the damage at different degrees and locations of the structure. The numerical model is then used for time-series calculations to obtain the simulated structural response corresponding one-to-one with the sampling time of the measured data, ensuring a complete match between the simulated and measured data in both time and measurement point dimensions.

[0062] S1.3 Construction of Temporal Residual Vectors

[0063] Based on the difference between the measured and numerical responses of the structure, a response residual is defined. This residual directly reflects the difference between the model calculation results and the actual structural response under the current structural parameters. By sequentially traversing all measurement points, the response residuals of all measurement points are obtained, and these residuals are assembled to obtain a residual vector that comprehensively characterizes the overall structural response.

[0064] S1.4 Establishment of the Objective Function Minimization Problem

[0065] Based on the least squares optimization criterion, the vector 2-norm is used to numerically represent the global residual vector. To ensure the objective function is smooth, continuous, differentiable, and iterable, and to adapt to the solution logic of the Gauss-Newton iterative algorithm, the square of the residual 2-norm is selected as the minimization objective. With the physical meaning of structural parameters as constraints and the minimization of the global residual as the minimization objective, a minimization problem for iteratively calculating structural damage parameters is constructed.

[0066] Through the above four steps, the formula for minimizing the damage detection objective function of this invention is finally obtained:

[0067] , In the formula, It is a damage detection parameter vector, and its dimension N is the number of damage detection parameters; It is a 2-norm operator; s[i] is the time-domain test data at time i, i=1,…,τ; z[i; Therefore The simulated response data corresponding to s[i] is calculated by the analytical model for the parameters, where τ is the total number of time-domain test data; r is the result of s[i] - z[i; The calculated residual vector; superscript T Indicates the transpose operation;

[0068] Step S2: Use the Gauss-Newton algorithm to iteratively solve the above objective function minimization problem, and complete the sensitivity matrix calculation and singular value decomposition step by step.

[0069] Based on the least-squares minimization problem for structural damage detection constructed in step S1, the Gauss-Newton iterative method is used to iteratively optimize the damage detection parameter vector. This invention constructs the sensitivity matrix and completes singular value decomposition in each iteration through step-by-step iterative calculation, providing iterative feature parameters for subsequent step-size increment correction. The specific subdivision steps are as follows:

[0070] S2.1 Initialize the basic conditions for iteration

[0071] Read the physical parameters of the structure in its undamaged state, and set the initial iteration values ​​of the damage detection parameter vector as the initial values ​​for the first iteration. Define the total number of iterations, the iteration convergence threshold, and the dimension of the damage parameter vector to establish the basic boundary conditions for the iterative calculation, ensuring that the dimension of subsequent iterations is consistent and the loop logic is closed.

[0072] S2.2 Calculate the residual gradient information for the current iteration step.

[0073] For the k-th iteration, the damage detection parameter vector updated in the previous iteration is used as the calculation benchmark and substituted into the structural numerical model to solve for the structural simulation response at the corresponding time. By comparing with the measured time-domain response data, the gradient of the residual vector with respect to the damage detection parameters is calculated to obtain the gradient information of the residual as a function of each structural stiffness parameter.

[0074] S2.3 Assemble the sensitivity matrix of the Gauss-Newton algorithm

[0075] Based on gradient information, and following the correspondence between the temporal sampling dimension and the parameter dimension, gradient data is integrated row by row and column by column to systematically assemble the sensitivity matrix corresponding to the current iteration step. This matrix can fully reflect the coupling relationship between the structural response residual and the damage detection parameters, and is the core foundational matrix for the Gauss-Newton algorithm to solve for the iterative increment.

[0076] S2.4 Perform singular value decomposition on the sensitivity matrix for each iteration step.

[0077] The sensitivity matrix is ​​decomposed into singular value decomposition, which breaks it down into a combination of two sets of orthogonal matrices and a diagonal singular value matrix. This decouples and separates the matrix feature information, and accurately extracts the main feature components and weak feature components of the matrix.

[0078] S2.5 Sort and store iterative feature parameters

[0079] All singular values ​​obtained from the decomposition are sorted in descending order, and the orthogonal matrix and the sorted diagonal matrix of singular values ​​are stored simultaneously. This fully preserves all matrix feature information of the current iteration step, providing core data support for subsequent iteration increment calculation, step size coefficient solution and adaptive amplitude limiting correction.

[0080] After completing the above step-by-step operations, the matrix decomposition work for this iteration step is finished, and the corresponding singular value decomposition mathematical expression is: In the formula, For the k-th iteration step The first-order sensitivity matrix; for orthogonal matrix of order 1; for orthogonal matrix of order 1; It is a singular value diagonal matrix, with the diagonal elements being N singular values ​​arranged in descending order.

[0081] Step S3: Calculation of iterative incremental correction with limited step size

[0082] According to the The damage detection parameter vector obtained in the next iteration Calculate the first The initial iteration increment for the next iteration. Define the iteration step size correction coefficient. A preset step size limit greater than 0 and less than 1. Greater than 1 and less than positive integers .

[0083] contrast With step size limit The numerical value is used to adaptively adjust the initial iteration increment: when When the initial iteration increment is reduced, the final iteration increment is obtained; when In this case, the initial iteration increment does not need to be corrected and is directly used as the final iteration increment. This mechanism achieves precise limiting of the iteration step size, avoiding parameter distortion caused by excessively large increments.

[0084]

[0085] Step S4: Iteratively converge to solve for the damage detection parameters

[0086] Through iterative formula (k) = (k-1) + (k) Update the damage detection parameter vector, where (k) For the first The final iteration increment. Repeat steps S2 and S3 for matrix decomposition and step size correction, continuously iterating and updating parameters until the convergence condition is met. The iteration terminates. The parameter vector after the final iteration converges is the structural damage detection result. The damage degree of each structural layer is calculated based on the parameter changes, completing the damage detection calculation. That is, repeating steps S2~S3 above yields the converged sequence: Then, by the iterative formula (k) = (k-1) + (k) The following convergent sequence is formed: Where k=1,...,N i N i It is the number of iterations required for convergence. (0) These are the initial values ​​for the iterative calculation of the structural damage detection parameters; the damage detection calculation result is... Specifically, the International Structural Control Association and the American Society of Civil Engineers have proposed a... Figure 2 The benchmark structure shown serves as an experimental research platform for comparing various structural damage detection methods, allowing researchers to apply different methods to this structure and evaluate their merits.

[0087] Due to space limitations, the example given here is the first stage problem of the Benchmark structure. The original monitoring data used are all simulated monitoring data calculated by MATLAB software, which includes white noise error with a noise-to-signal ratio of 20%. The damage is simulated by the diagonal supports in the structure. That is, the state in which all 8*4=32 diagonal supports of each layer of the structure are present corresponds to the undamaged state; removing different numbers of diagonal supports in different layers corresponds to different damage states.

[0088] The parameters of the Benchmark structure are detailed in Table 1; the horizontal stiffness k1~k4 of the first to fourth layers are used as the detection parameters to form a detection parameter vector. =[k1,k2,k3,k4] T That is, the number of parameters N=4; using the top-level acceleration test data, the objective function of the minimization problem is formed.

[0089] Damage mode two involves removing all diagonal supports from the first and third floors. The structural parameters are listed in Table 1 under the damaged state. Initial values ​​for structural damage detection parameters are also provided. (0) Take the stiffness parameters in the non-destructive state from Table 1, and set the step size limit. Take 0.3, Figure 1 In step 6, N t Take 3; the damage detection results obtained using this patented technology are shown in Table 2, and the iterative calculation process is shown in [the table below]. Figure 3 .

[0090] in, Figure 3 (a) is the modulus ||r of the residual column vector. (k) The change with the number of iterations k, according to Figure 1 The given criterion for ending the iterative calculation is that when k=12, ||r(k) -r (k-1) ||≤0.001 The iteration ends; Figure 3 (b) is Figure 1 The M obtained in step 5 (k) The changes are shown by the horizontal red dashed line, which represents the step size limit. It can be seen that M is at k=1-4 and 6. (k) > When k=5, 7-12, M (k) < ; Figure 3 (c) is Figure 1 The result of step ③ With corresponding M (k) / The changes are due to Figure 1 The increment obtained in step ⑥ for the k-th iteration According to the definition, when the iteration increment is large, the iteration steps k=1-4, 6 Increased M (k) / When the iteration increment is small, such as k=5 or 7-12, the iteration increment is multiplied by 1. It did not increase in size; Figure 3 (d) shows the changes in the four detection parameters.

[0091] Table 1 Parameters of the Benchmark Structure

[0092] ;

[0093] Table 2 shows the damage detection results for k1 to k4 obtained using this patented technology.

[0094] ; The damage levels corresponding to the four stiffness parameter detection results in Table 2 are: -90.09%, 2.82%, -83.73%, and -4.49%. Here, the damage level refers to the difference between the damage detection result and the stiffness parameter value under the undamaged state, divided by the stiffness parameter value under the undamaged state. The exact solutions are -71%, 0, -71%, and 0. Therefore, the structural damage state can be accurately detected, and the damage location can be determined to be on the 1st and 3rd floors.

[0095] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for detecting damage to civil engineering structures based on limited step length technology, characterized in that, Includes the following steps: S1. Based on the time-domain test data of civil engineering structures, construct the objective function for the structural damage detection problem; S2. The Gauss-Newton algorithm is used to iteratively solve the above objective function minimization problem. The sensitivity matrix required for each iteration is calculated, and the sensitivity matrix is ​​decomposed into singular values. S3. Calculate the initial iteration increment based on the sensitivity matrix, and combine the preset step size limit and the singular value decomposition result to adaptively correct and obtain the final iteration increment, thereby realizing the limit constraint of the iteration step size. S4. Repeated iterative calculation and step size constraint process, after iterative convergence, to obtain stable damage detection parameters, thus completing the damage detection of civil engineering structures.

2. The method for detecting damage to civil engineering structures based on limited step length technology according to claim 1, characterized in that: The specific steps of step S1 are as follows: S1.1 Collect time-domain test data of the structure and select the structural acceleration response data as the benchmark dataset; S1.2 Establish a numerical analysis model of the civil engineering structure to be tested, define a damage detection parameter vector containing structural stiffness parameters, and calculate the structural simulation response corresponding one-to-one with the measured time domain data through the numerical model. S1.3 Construct response residuals based on the difference between the measured response and the simulated response at the same moment, integrate the response residuals of all measurement points, and form a residual vector that can characterize the overall response; S1.4 Based on the least squares method, with the minimum residual vector as the optimization objective, a continuously differentiable objective function is constructed to complete the construction of the structural damage detection problem.

3. The method for detecting damage to civil engineering structures based on limited step length technology according to claim 2, characterized in that: The minimization problem constructed in step S1 is as follows: ,in, This is the result of damage detection calculation. The definition is shown in the following formula: In the formula, It is a damage detection parameter vector, and its dimension N is the number of damage detection parameters; It is a 2-norm operator; s[i] is the time-domain test data at time i, i=1,…,τ; z[i; Therefore The simulated response data corresponding to s[i] is calculated by the analytical model for the parameters, where τ is the total number of time-domain test data; r is the result of s[i] - z[i; The calculated residual vector; superscript T This indicates the transpose operation.

4. The method for detecting damage to civil engineering structures based on limited step length technology according to claim 1, characterized in that: The sensitivity matrix S required for the k-th iteration in step S2 is... (k) Constructed using the Gauss-Newton algorithm, for S (k) The singular value decomposition formula is as follows: In the formula, matrix S (k) U (k) V (k) The number of rows × columns are τ×N, τ×N, and N×N, respectively. , They are U (k) V (k) The column vector, where N is the number of damage detection parameters; the diagonal matrix ∑ (k) It contains N diagonal elements, arranged in descending order from top left to bottom right, namely: , ... .

5. The method for detecting damage to civil engineering structures based on limited step length technology according to claim 1, characterized in that: The specific steps of step S3 are as follows: S3.

1. Based on the damage detection parameter vector from the (k-1)th iteration, calculate the initial iteration increment for the kth iteration. Let the damage detection parameter vector calculated in the (k-1)th iteration be... , , They are , The j-th element, where j = 1, ..., N; S3.

2. Let α be a pre-set step size limit greater than 0 and less than 1, where each iteration step's damage detection parameters have physical meaning. positive integers Calculate the iteration step size correction coefficient M (k) = Compared to M (k) With respect to the magnitude of α; S3.3, when M (k) When M > α, the initial iteration increment is corrected to obtain the final iteration increment; when M (k) At time α, the initial iteration increment is retained as the final iteration increment, i.e. 。 6. The method for detecting damage to civil engineering structures based on limited step length technology according to claim 1, characterized in that: The specific steps of step S4 are as follows: Repeat steps S2 to S3 above to obtain the convergent sequence: Then, by the iterative formula (k) = (k-1) + (k) The following convergent sequence is formed: Where k=1,...,N i N i It is the number of iterations required for convergence. (0) These are the initial values ​​for the iterative calculation of the structural damage detection parameters; the damage detection calculation result is... .

Images