Bridge anti-seismic model online correction method and system based on parameter identification

Abstract

The invention relates to the technical field of bridge engineering data processing, and particularly provides an online correction method and system for a bridge anti-seismic model based on parameter identification, and the method comprises the steps: continuously collecting a multi-source mixed response signal of a bridge in real vibration; automatically generating self-adaptive feature slices according to a vibration energy transfer path and a structure local nonlinear behavior, and outputting a recombined feature set; constructing a non-uniform evolution trajectory for each independent fluctuation unit in the recombination feature set; generating a directional disturbance vector field in a bridge anti-seismic model parameter space; and the directional disturbance vector field takes the sensitive domain intensity as a disturbance weight, and asymmetric fine tuning is carried out on original bridge anti-seismic model parameters. The system comprises a lossless stripping and recombination module, a time evolution trajectory modeling module and a parameter space directional disturbance module. According to the method, the time domain limitation of traditional offline correction is broken through, and the anti-seismic model ecological system with the autonomous evolution capability is formed through dynamic coupling of response feature-parameter disturbance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bridge engineering data processing, in particular to a bridge anti-seismic model online correction method and system based on parameter identification. BACKGROUND

[0002] With the development of modern bridge engineering towards large-span and high-pier, the seismic performance of bridge structure becomes more and more important in design. Traditional seismic design relies on static model or finite element simulation, but in actual earthquake, the dynamic response of bridge is often affected by time-varying factors such as material aging, environmental erosion and local damage, resulting in significant deviation between design model and actual response. Especially under strong earthquake, this deviation may cause disastrous consequences. The current engineering community urgently needs a technology that can capture the changes of bridge state in real time and dynamically correct the seismic model.

[0003] Prior art one, Chinese patent, application number 202310283973.X discloses a bridge seismic vulnerability evaluation method based on influence line correction model, including the following steps: finite element model correction; seismic vulnerability evaluation. The present application takes the measured bridge structure strain influence line as the target parameter, uses artificial neural network to correct the bridge finite element model; carries out seismic vulnerability analysis on the optimized model of the corrected bridge finite element model, imports 20 seismic records and 1 artificial seismic wave, extracts the displacement response of the bridge structure under different peak ground accelerations, draws the structure exceeding probability and calculates the vulnerability matrix, and then constructs the structure seismic damage index expression based on the beta probability distribution. Although the seismic damage index expectation value is used to draw the seismic damage comparison curve, the accurate evaluation of the seismic resistance of the existing bridge structure is realized; but only through strain influence line for offline model correction, the transient nonlinear response under seismic action cannot be captured; the defect of seismic vulnerability evaluation depends on the preset seismic wave.

[0004] The prior art two, Chinese patent, application number 202311218504.6 discloses a bridge damage identification method, comprising the steps of: establishing an initial finite element model of the target bridge to perform vehicle-bridge coupling analysis to obtain a simulated deflection response, combining the measured deflection response to optimize the initial finite element model to obtain a modified finite element model; collecting traffic flow vehicle load data of the target bridge within a certain time, statistically analyzing to determine the physical parameter information of each vehicle type and establishing a standard vehicle physical model, using a sampling method to perform random sampling to generate a random traffic sample sequence; designing different damage conditions of the target bridge, performing vehicle-bridge coupling analysis under random traffic according to different damage conditions and the modified finite element model to obtain bridge simulation deflection response data; constructing a training data set with different damage conditions and corresponding bridge simulation deflection response data to establish a damage identification model to identify the damage of the target bridge. Although real-time online bridge damage positioning and quantification can be achieved; but the deflection response based on vehicle load cannot effectively reflect the structural seismic performance degradation; the damage condition library needs to be preset.

[0005] The prior art three, Chinese patent, application number 202111015298.X discloses a bridge structure digital twin based on BIM-FEM and a method, comprising a BIM modeling module, a full life cycle operation and maintenance management module, a digital twin basic module, a digital twin data conversion and optimization module, and a digital twin application module. Information switching and interaction can be realized between the modules. The method comprises the following steps: establishing a BIM three-dimensional information model of the bridge structure; constructing a digital twin basic layer, converting and optimizing digital twin data; and building an intelligent operation and maintenance management platform based on the digital twin. Although based on bridge construction, operation and maintenance stage structure state monitoring information collection, data processing and analysis, through synchronous updating and parameter correction optimization of the bridge information model (BIM) and the finite element analysis model (FEM), the structure state online comprehensive evaluation and response prediction are completed, thereby realizing real-time monitoring and control of the full life cycle bridge construction safety and healthy operation; but the parameter update depends on periodic data synchronization, and the full life cycle data integration is focused on but lacks real-time seismic analysis capability.

[0006] The prior art one, the prior art two and the prior art three have the problems of mismatch between static correction and dynamic seismic excitation, preset damage mode cannot adapt to unknown seismic damage, and discrete update causes insufficient real-time performance. Therefore, the present application provides a bridge seismic model online correction method and system based on parameter identification. SUMMARY

[0007] To achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows:

[0008] In one aspect of the present application, an online correction method for bridge seismic model based on parameter identification is provided, comprising: decomposing the bridge seismic response signal into independent fluctuation units and generating adaptive feature slices containing amplitude, time-varying phase and spatial weight; analyzing the feature evolution trajectory based on the phase coupling strength matrix and the energy attenuation gradient field to automatically identify abnormal coupling nodes and gradient mutation areas as parameter sensitive domains;

[0009] Based on the spatial positioning and intensity distribution of the parameter sensitive domain, a directional disturbance vector field is generated in the parameter space of the bridge seismic model; the directional disturbance vector field takes the sensitive domain intensity as the disturbance weight and the phase coupling direction in the evolution trajectory as the disturbance axis direction to implement asymmetric fine-tuning of the original bridge seismic model parameters.

[0010] In an optional implementation, the process of implementing asymmetric fine-tuning of the original bridge seismic model parameters comprises the following steps:

[0011] The spatial positioning and intensity distribution of the parameter sensitive domain are input into a space-parameter mapper to establish a parameter influence weight matrix according to the spatial membership relationship between the parameter sensitive domain boundary coordinates and the bridge seismic model parameters; the sensitive domain intensity value is distributed to the corresponding parameter group according to the weight matrix, and a parameter sensitivity distribution map with intensity annotation is output;

[0012] The phase coupling direction in the non-uniform evolution trajectory is input into a direction-parameter converter to extract the principal axis direction of the phase coupling direction in the structure space, and according to the structure dynamics constitutive relation, the spatial axis direction is mapped to the parameter correlation direction vector in the parameter space, and a parameter coupling axis direction set defining the parameter disturbance direction is output;

[0013] The parameter sensitivity distribution map and the parameter coupling axis direction set are input into a field synthesis operator to generate a disturbance amplitude scalar field based on the intensity value of the parameter sensitivity distribution map and construct a disturbance direction tensor field based on the direction vector of the parameter coupling axis direction set, and the Kronecker product operation is performed on the disturbance amplitude scalar field and the disturbance direction tensor field, and a directional disturbance vector field with spatial heterogeneity is output, which meets the intensity non-uniformity and direction asymmetry in vector distribution.

[0014] In an optional implementation, the process of outputting the directional disturbance vector field with spatial heterogeneity comprises the following steps:

[0015] The intensity value of the parameter sensitivity distribution map is input into an amplitude field generator to establish an intensity interpolation grid based on the parameter space coordinates, and the discrete intensity value is converted into a continuous scalar distribution through a radial basis function, and a disturbance amplitude scalar field covering the entire parameter space is output, and the field strength gradient is consistent with the spatial energy anomaly degree distribution of the sensitive domain;

[0016] The parameter coupling axial set direction vector is input into a direction field generator, the effect anchor point of each axial vector is marked in the parameter space, the direction vector diffusion flow is generated based on the anchor point position, the non-conflict tensor field is formed through the divergence-curl constraint equation, and the perturbation direction tensor field with smooth continuity is output, and the streamline topology inherits the spatial characteristics of the phase transmission path breaking mode;

[0017] The perturbation amplitude variable field and the perturbation direction tensor field are input into a field coupler, spatial alignment of the two fields is performed under the isomorphic coordinate system in the parameter space, orthogonal projection multiplication is performed on the perturbation amplitude scalar and the perturbation direction tensor of each space point, the action dimension of the amplitude scalar is expanded along the main axis of the direction tensor, and the full-parameter space directional perturbation vector field that integrates the amplitude and the direction attribute is output.

[0018] In an alternative embodiment, the process of inputting the perturbation amplitude variable field and the perturbation direction tensor field into the field coupler comprises the following steps:

[0019] The perturbation amplitude scalar field and the perturbation direction tensor field are used to establish a unified dimension grid based on the parameter space coordinates, the topological position matching is performed between the field strength sampling points of the perturbation amplitude scalar field and the streamline control points of the perturbation direction tensor field, and an isomorphic reference field pair with spatial correspondence is output, and the grid nodes simultaneously carry the amplitude value and the direction tensor;

[0020] The isomorphic reference field pair is used to extract the main action axis of the direction tensor of each space point, the orthogonal projection coordinate system is established along the main action axis, the amplitude scalar is decomposed into axial action components, and the projection amplitude component set with direction constraint is output.

[0021] The projection amplitude component set is used to construct a vector generation framework based on the main axis of the direction tensor, the axial scalar-vector conversion is performed on the amplitude components along the framework, the conversion results of all space points are aggregated, and the full-space directional perturbation vector field that meets the asymmetric constraint is output.

[0022] In an alternative embodiment, the process of aggregating the conversion results of all space points comprises the following steps:

[0023] The main axis of the direction tensor is used to construct an orthogonal zero-return auxiliary surface in a vertical plane with the main axis direction as a reference axis, a three-dimensional non-commutative framework containing the main direction constraint is output, and the reference axis direction inherits the spatial directivity of the phase transmission path breaking feature;

[0024] The amplitude component value of each space point of the projection amplitude component set is assigned to the reference axis of the three-dimensional non-commutative framework, and a zero-value constraint condition is applied to the orthogonal zero-return auxiliary surface, and a directional action quantity that only acts along the reference axis direction is output.

[0025] The directed action quantity is sequentially indexed with the parameter space coordinates, the discrete directed action quantity is connected according to the space adjacency topological rule, and the connection mutation is eliminated through the smoothness condition of the manifold; and a continuous directed disturbance vector field covering the whole parameter space is output.

[0026] In an alternative embodiment, the process of eliminating the connection mutation through the smoothness condition of the manifold comprises the following steps:

[0027] The directed action quantity is sequentially indexed with the parameter space coordinates, the discrete directed action quantity is connected according to the space adjacency topological rule, and the connection mutation is eliminated through the smoothness condition of the manifold; and a continuous directed disturbance vector field covering the whole parameter space is output.

[0028] The topologically ordered action quantity chain is extracted, the action quantity values of the adjacent two points in the chain and the reference axis direction are extracted, the linear transition vector segment connecting the two points is generated by taking the energy transmission impedance ratio between the two points as the weight, and a segmented linear vector field covering all adjacent point pairs is output, wherein the connection direction strictly follows the spatial trend of the energy path.

[0029] The segmented linear vector field is detected, the direction mutation angle at the turning point of the vector segment is detected, the curvature continuous boundary condition is reconstructed by taking the cotangent value of the mutation angle as the smoothness adjustment coefficient, and a globally continuous directed disturbance vector field eliminating local mutation is output.

[0030] In an alternative embodiment, the transient response verification is performed after each disturbance using the current seismic input adaptive feature, the parameter sensitivity feedback loop is formed by comparing the convergence speed changes of the feature slices in the energy attenuation gradient field before and after the disturbance, and the model parameter set is output after online correction, wherein the correction direction is autonomously guided by the feature evolution trajectory, and the correction amplitude is dynamically constrained by the sensitivity domain strength.

[0031] In an alternative embodiment, the multi-source mixed response signals of the bridge in real seismic vibration are continuously collected, the environmental noise and known steady-state interference components are stripped, the remaining mixed response signal stream is decomposed into several independent wave units, the adaptive feature slices are automatically generated according to the seismic energy transmission path and the local nonlinear behavior of the structure, each adaptive feature slice contains three attributes of amplitude, time-varying phase and spatial correlation weight, and the reorganized feature set is output; the reorganized feature set implicitly contains the spatiotemporal distribution information of the potential damage or parameter drift of the structure, and the reorganization logic is dynamically determined by the instantaneous propagation characteristics of seismic energy in the structure.

[0032] In an alternative embodiment, a non-uniform evolution trajectory is constructed for each independent wave unit in the reorganized feature set on a continuous time axis; the non-uniform evolution trajectory quantifies the variation of the characteristic unit of each independent wave unit with the duration of the vibration by establishing a phase coupling strength matrix and an energy attenuation gradient field between adjacent slices; an abnormal coupling node in the phase coupling strength matrix and a mutation region in the gradient field are automatically marked as a parameter sensitive domain.

[0033] In another aspect of the present application, a parameter identification based online correction system for a bridge seismic model is provided, which includes:

[0034] A lossless stripping and reorganization module is used to continuously collect multi-source mixed response signals of a bridge in real vibration, strip environmental noise and known steady-state interference components, decompose the remaining mixed response signal stream into several independent wave units, and output a reorganized feature set according to self-adaptive feature slices automatically generated according to the vibration energy transmission path and the local nonlinear behavior of the structure.

[0035] A time evolution trajectory modeling module is used to construct a non-uniform evolution trajectory for each independent wave unit in the reorganized feature set on a continuous time axis; the non-uniform evolution trajectory quantifies the variation of the characteristic unit of each independent wave unit with the duration of the vibration by establishing a phase coupling strength matrix and an energy attenuation gradient field between adjacent slices; an abnormal coupling node in the phase coupling strength matrix and a mutation region in the gradient field are automatically marked as a parameter sensitive domain.

[0036] A parameter space directional disturbance module is used to generate a directional disturbance vector field in the parameter space of the bridge seismic model based on the spatial positioning and intensity distribution of the parameter sensitive domain; the directional disturbance vector field takes the sensitive domain intensity as the disturbance weight and the phase coupling direction in the evolution trajectory as the disturbance axis, and performs asymmetric fine tuning on the original bridge seismic model parameters; after each disturbance, the current vibration input is used to verify the transient response of the adaptive feature, and the parameter sensitivity feedback loop is formed by comparing the convergence speed change of the feature slices in the energy attenuation gradient field before and after the disturbance; the output model parameter set is corrected, the correction direction is guided by the feature evolution trajectory, and the correction amplitude is dynamically constrained by the sensitive domain intensity.

[0037] The present invention achieves autonomous adaptive correction of the bridge seismic model through the synergistic effect of three stages: dynamic decomposition of multi-source signals, self-organizing mapping of nonlinear features, and directional perturbation in parameter space. A nonlinear mapping channel from physical response to model parameters is established. The local nonlinear behavior implicit in the bridge vibration response is converted into the spatiotemporal distribution characteristics of the parameter sensitive domain through adaptive feature slicing generation and non-uniform evolution trajectory construction of mixed signals. The joint analysis of the phase coupling intensity matrix and the energy attenuation gradient field achieves quantitative spatial positioning of structural parameter drift and potential damage, overcoming the traditional method's dependence on preset damage patterns. Based on the directional perturbation vector field generated by the sensitive domain intensity distribution, an adaptive matching relationship between the parameter correction amount and the degree of physical response anomaly is established. Through closed-loop feedback of asymmetric fine-tuning and transient response verification, the model parameter update is always carried out along the optimal direction guided by the feature evolution trajectory, ensuring that the correction process responds to local nonlinear mutations while maintaining overall stability. The dynamic generation mechanism of the reorganized feature set and the real-time calibration function of the parameter sensitive domain ensure that the model correction is synchronized with the vibration input; the spatial correlation weights of the feature slices automatically adapt to changes in the energy propagation path, so that the correction process can capture the transient nonlinear effects under short-term strong earthquakes and the progressive drift of parameters in long-term service. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0039] Figure 1 This is a flow chart of the online correction method for the bridge seismic model based on parameter identification provided in Example 1 of the present invention;

[0040] Figure 2 This is a process diagram of outputting a recombinant feature set provided in Example 2 of the present invention;

[0041] Figure 3 This is a process diagram of automatically marking parameter-sensitive domains provided in Example 3 of the present invention;

[0042] Figure 4 This is a process diagram for implementing asymmetric fine-tuning of the original bridge seismic model parameters provided in Example 4 of the present invention;

[0043] Figure 5 This is a block diagram of an online correction system for a bridge seismic model based on parameter identification provided in Example 9 of the present invention;

[0044] Figure 6 A block diagram of the electronic device provided by the present invention;

[0045] Figure 7 A block diagram of a computer-readable storage medium provided by the present invention. DETAILED DESCRIPTION

[0046] The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all of them.

[0047] Hereinafter, the terms "first", "second", and the like are only used for the convenience of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second", and the like can be explicitly or implicitly included one or more features. In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0048] In the present application, unless otherwise explicitly specified and limited, the term "connection" should be understood broadly, for example, "connection" can be a fixed mechanical connection, or a detachable mechanical connection, or integrated; or "connection" can be direct connection, or indirect connection through intermediate medium. In addition, unless otherwise explicitly specified and limited, the term "coupling" should be understood broadly, for example, "coupling" can be direct electrical connection, for example, physical contact and electrical conduction between two components, or can be understood as electrical connection between different components in a circuit structure through printed circuit board (PCB) copper foil or wire, etc. to transmit electrical signals; or "coupling" can be indirect electrical connection between two components through intermediate medium; or "coupling" can be electrical connection between two components through space / non-contact, for example, capacitive coupling between two components to transmit electrical signals.

[0049] In the embodiments of the present application, the orientation terms such as "up", "down", "left", "right", etc. can include but not limited to the orientation defined by the relative position of the components in the drawings. It should be understood that these directional terms can be relative concepts, which are used for relative description and clarification, and can be changed accordingly according to the change of the position of the components in the drawings.

[0050] Embodiment 1:

[0051] As shown in Figure 1 The present application provides a bridge seismic resistance model online correction method based on parameter identification, comprising the following steps:

[0052] Step S100: continuously collect the multi-source mixed response signals of the bridge in real vibration, strip the environmental noise and known steady-state interference components; decompose the remaining mixed response signal stream into several independent wave units, automatically generate adaptive feature slices according to the vibration energy transmission path and the local nonlinear behavior of the structure, each adaptive feature slice contains three attributes of amplitude, time-varying phase and spatial correlation weight, and output the reorganized feature set; the reorganized feature set has the spatial and temporal distribution information of the potential damage or parameter drift of the structure, and the reorganization logic is dynamically determined by the instantaneous propagation characteristics of the vibration energy in the structure;

[0053] Step S200: construct a non-uniform evolution track of each independent wave unit in the reorganized feature set on a continuous time axis; the non-uniform evolution track quantifies the change rule of the characteristic unit of each independent wave unit with the duration of vibration by establishing the phase coupling strength matrix and the energy attenuation gradient field between adjacent slices; the abnormal coupling nodes in the phase coupling strength matrix and the mutation regions in the gradient field will be automatically marked as parameter sensitive domains;

[0054] Step S300: based on the spatial positioning and intensity distribution of the parameter sensitive domain, generate a directional disturbance vector field in the bridge seismic model parameter space; the directional disturbance vector field takes the sensitive domain intensity as the disturbance weight and the phase coupling direction in the evolution track as the disturbance axis, and implements asymmetric fine tuning on the original bridge seismic model parameters; after each disturbance, the current seismic input adaptive feature is used for transient response verification, and by comparing the convergence speed change of the feature slices in the energy attenuation gradient field before and after the disturbance, a parameter sensitivity feedback closed loop is formed; output the model parameter set after online correction, and the correction direction is guided by the feature evolution track and the correction amplitude is dynamically constrained by the sensitive domain intensity.

[0055] Among the above-mentioned embodiments, this embodiment achieves autonomous adaptive correction of the bridge seismic model through the synergistic effect of three stages: dynamic decomposition of multi-source signals, self-organizing mapping of nonlinear features, and directional perturbation in parameter space. A nonlinear mapping channel from physical response to model parameters is established. The local nonlinear behavior implicit in the bridge vibration response is converted into spatiotemporal distribution characteristics of the parameter sensitive domain through adaptive feature slicing generation and non-uniform evolution trajectory construction of mixed signals. Joint analysis of the phase coupling intensity matrix and the energy attenuation gradient field enables quantitative spatial localization of structural parameter drift and potential damage, overcoming the traditional method's dependence on preset damage patterns. Based on the directional perturbation vector field generated from the sensitive domain intensity distribution, an adaptive matching relationship between the parameter correction amount and the degree of physical response anomaly is established. Through closed-loop feedback of asymmetric fine-tuning and transient response verification, model parameter updates are always carried out in the optimal direction guided by the feature evolution trajectory, ensuring that the correction process responds to local nonlinear mutations while maintaining overall stability. The dynamic generation mechanism of the reorganized feature set and the real-time calibration function of the parameter sensitive domain ensure that the model correction is synchronized with the vibration input; the spatial correlation weights of the feature slices automatically adapt to changes in the energy propagation path, so that the correction process can capture the transient nonlinear effects under short-term strong earthquakes and the progressive drift of parameters in long-term service.

[0056] In summary, this implementation overcomes the time-domain limitations of traditional offline corrections. By dynamically coupling response characteristics with parameter perturbations, it forms a seismic model ecosystem with autonomous evolutionary capabilities. Without requiring pre-defined damage patterns or manual intervention, this approach enables continuous tracking and autonomous optimization of the seismic model's time-varying parameters, significantly improving the real-time accuracy of bridge response predictions under strong earthquakes.

[0057] Example 2:

[0058] like Figure 2 As shown, based on Example 1, the process of outputting the recombined feature set in step S100 provided in this embodiment of the present invention includes the following steps:

[0059] Step S101: The remaining mixed response signal is input into the energy conduction topology network. Based on the bridge geometry and material constitutive relations, the propagation delay differences of the vibration wave in the structure are dynamically mapped. The output is primary wave clusters sorted by the arrival time of the vibration energy. Each wave cluster is marked with its physical propagation path number and initial energy density.

[0060] Step S102: Input the primary wave cluster into the local behavior modulator, and identify asymmetric waveform distortion caused by the structural hinge region or material yield point by detecting the fluctuation amplitude mutation point and curvature extreme value. For each wave segment within the distortion interval, extract its phase offset relative to the ideal elastic waveform, and generate a distortion-phase binding pair containing the coordinates of the distortion position.

[0061] Step S103: input the distortion-phase binding pairs into the energy density space distributor, calculate the energy receiving proportion of all sensors in the influence area of each binding pair, fit the energy density space attenuation surface according to the energy receiving proportion distribution, generate the space correlation weight vector with the maximum gradient direction of the energy density space attenuation surface as the axis direction, and output the adaptive feature slice of the fused amplitude, time-varying phase and space correlation weight.

[0062] wherein the expression for calculating the energy receiving proportion of all sensors in the influence area of each binding pair is:

[0063]

[0064] wherein, represents the normalized energy receiving proportion of the kth distortion-phase binding pair at the jth sensor; ξ represents the Frobenius norm adjustment factor of the energy attenuation coefficient matrix; represents the energy density gradient tensor of the ith distortion interval in the parameter space; represents the material constitutive inverse matrix of the jth sensor; N s represents the total number of sensors; represents the ray propagation dissipation rate of the kth binding pair in the mth sensor direction; represents the position coordinate tensor (including direction cosine) of the mth sensor; Γ K represents the wavefront curvature covariance matrix of the kth distortion-phase binding pair; α represents the nonlinear distribution index; represents the frequency response function vector of the jth sensor to the kth distortion-phase binding pair; ζ i represents the phase shift characteristic kernel of the ith distortion interval; represents the energy diffusion variance of the kth distortion-phase binding pair in the ith distortion interval; this formula system establishes a strict mapping from microscopic waveform distortion to macroscopic energy distribution: the molecular term is attenuated by the exponential term quantifies the energy dissipation on the wave conduction path, and the deterministic operator in the denominator embodies the geometric constraint of the sensor spatial configuration on energy capture; Hilbert space norm Hadamard product of the phase shift kernel ζ i characterizes the energy redistribution effect caused by inelastic deformation; gradient tensor contraction operation with the material inverse matrix reflects the energy conduction direction preference in anisotropic medium; the relevance between the formulas is reflected in that Γ K simultaneously affects the denominator term and gradient calculation of the main calculation formula, and the value of is determined by the gradient norm The computational framework satisfies the law of energy conservation in continuum mechanics and implements nonlinear distribution through the α exponent to adapt to the plastic deformation scenario of the material.

[0065] In the above-mentioned embodiments, this embodiment establishes a precise correspondence between the timing of vibration wave transmission and the physical path through dynamic mapping of the energy conduction topology network, quantifies the phase offset of the nonlinear distortion region of the structure by the local behavior modulator, and generates a spatial weight vector by the energy density spatial distributor, ultimately achieving the conversion of multi-source mixed response signals into adaptive feature slices with clear physical meaning. This conversion process fully preserves the spatiotemporal characteristics of vibration energy propagation and the local nonlinear behavior characteristics of the structure. The output recombined feature set can accurately reflect the distribution pattern of potential damage or parameter drift in the structure, providing high-fidelity input for subsequent parameter sensitive domain identification. The entire process ensures the physical interpretability and engineering applicability of feature extraction through the synergistic effect of signal conduction timing analysis, nonlinear distortion detection, and spatial energy distribution calculation.

[0066] Example 3:

[0067] like Figure 3 As shown, based on Example 1, the process of automatically marking as a parameter sensitive domain in step S200 provided in this embodiment of the present invention includes the following steps:

[0068] Step S201: Input the phase coupling strength matrix into the topological dispersion analyzer, calculate the phase transfer consistency coefficient between each node in the phase coupling strength matrix and all adjacent nodes, and generate a coupling dispersion scalar field based on the standard deviation of the phase transfer consistency coefficient; output a dispersion distribution cloud map representing the phase transfer anomaly, and its peak area indicates the phase transfer path breakpoint;

[0069] Step S202: Input the coupled discreteness scalar field into the curvature field generator, calculate the energy decay acceleration within a unit distance along the gradient direction, perform curvature tensor transformation on the acceleration field, and output a curvature intensity field that reveals the sudden change properties of energy flow. The negative curvature extreme value area corresponds to the energy dissipation domain;

[0070] Step S203: Input the dispersion distribution cloud map and the curvature intensity field into the spatiotemporal convolution kernel, perform a dual-field intensity product operation under the condition of spatial coordinate alignment, perform anisotropic Gaussian diffusion processing on the product field, and extract the connected areas in the diffusion field whose intensity exceeds the adaptive threshold; the output is a parameter sensitive domain map with spatial weight attributes, whose boundary is naturally defined by the isosurface of the dual-field intensity product.

[0071] In the aforementioned embodiments, this embodiment integrates an intelligent parameter-sensitive domain calibration system based on multimodal field coupling analysis. By cascading the phase transfer consistency coefficient and the coupled discreteness scalar field, a node-level topological anomaly detection mechanism is established. While preserving the topological characteristics of the original phase-coupled network, multi-scale localization of transmission path breakpoints is achieved, providing a spatial reference framework for subsequent energy analysis. The gradient-directed energy decay acceleration analysis introduced by the curvature field generator transforms discreteness characteristics into energy dynamics indicators. The curvature tensor transformation reveals hidden dissipative structures that are difficult to capture using traditional scalar field analysis, particularly the strong correlation between negative curvature extremes and system vulnerability. The dual-field fusion algorithm of the spatiotemporal convolution kernel innovatively utilizes a product operation to couple topological anomalies and energy mutation characteristics. Anisotropic Gaussian diffusion is used to achieve nonlinear superposition of characteristic fields. Its adaptive threshold mechanism ensures that the sensitive domain boundaries conform to the true energy threshold characteristics of the physical system. The resulting spatial weight map is essentially a probability density representation of the propagation path of system parameter perturbations, and its isosurface boundaries strictly satisfy the energy conservation constraints of the Hamiltonian system.

[0072] In summary, this embodiment realizes cross-scale correlation analysis from microscopic phase transfer anomalies to macroscopic energy dissipation characteristics. The resulting parameter sensitivity domain map has both topological integrity and physical interpretability, providing a causal intervention target positioning capability for stability optimization of complex systems.

[0073] Example 4:

[0074] like Figure 4 As shown, based on Example 1, the process of asymmetric fine-tuning of the original bridge seismic model parameters in step S300 provided in this embodiment of the present invention includes the following steps:

[0075] Step S301: Input the spatial location and intensity distribution of the parameter sensitive domain into the space-parameter mapper. According to the spatial membership relationship between the boundary coordinates of the parameter sensitive domain and the parameters of the bridge seismic model, a parameter influence weight matrix is ​​established. The intensity values ​​of the sensitive domains are assigned to the corresponding parameter groups according to the weight matrix, and the output is a parameter sensitivity distribution map with intensity annotations. The intensity values ​​of the sensitive domains are derived from the spatial energy anomaly of the parameter sensitive domains.

[0076] Step S302: Input the phase coupling direction in the non-uniform evolution trajectory into a direction-parameter converter, extract the main axis of the phase coupling direction in the structural space, and map the spatial axis to a parameter-related direction vector in the parameter space according to the structural dynamics constitutive relation. The output is a parameter coupling axis set that defines the parameter perturbation direction, whose direction is determined by the fracture mode of the phase transfer path.

[0077] Step S303: input the parameter sensitivity distribution map into the parameter coupling axis set field synthesis operator, generate a perturbation amplitude scalar field based on the intensity value of the parameter sensitivity distribution map, construct a perturbation direction tensor field based on the direction vector of the parameter coupling axis set, perform a Kronecker product operation on the perturbation amplitude scalar field and the perturbation direction tensor field, and output a directional perturbation vector field with spatial heterogeneity, which meets the intensity non-uniformity and direction asymmetry simultaneously.

[0078] In the above embodiment, the present embodiment realizes multi-dimensional collaborative optimization of the bridge seismic model parameters through coupling analysis of parameter sensitive domain spatial positioning and non-uniform evolution trajectory. A bidirectional mapping mechanism of parameter space-structure response is established, the spatial-energy distribution of the structure sensitive domain and the phase-coupling dynamic characteristics are converted into mathematical representations in the parameter space through cascaded operations of the space-parameter mapper and the direction-parameter converter, and a composite perturbation field including the intensity gradient and the direction vector is constructed. The mechanism breaks through the limitations of traditional symmetric parameter adjustment and realizes accurate inversion of the structure response characteristics to the parameter space. A directional perturbation field with physical constraints is generated, the field synthesis operator fuses the sensitivity scalar field and the coupling direction tensor field through the Kronecker product, and the perturbation vector field formed meets the following conditions simultaneously: intensity dimension: inherits the energy anomaly distribution characteristics of the parameter sensitive domain, direction dimension: follows the fracture mode constraint of phase coupling transmission, and spatial dimension: maintains the constitutive correlation characteristics between parameters; and the physical interpretability of parameter optimization is realized.

[0079] In summary, the present embodiment maintains the consistency of parameter adjustment and physical mechanism through the constitutive relationship of structure dynamics, and the asymmetry of the perturbation vector directly reflects the difference in energy dissipation paths of the structure system, so that the parameter adjustment has a clear engineering physical meaning.

[0080] Embodiment 5

[0081] On the basis of embodiment 4, the process of outputting the directional perturbation vector field with spatial heterogeneity in step S303 provided by the present embodiment comprises the following steps:

[0082] Step S3031: input the intensity value of the parameter sensitivity distribution map into the amplitude field generator, establish an intensity interpolation grid based on the parameter space coordinates, convert the discrete intensity value into a continuous scalar distribution through a radial basis function, and output a perturbation amplitude scalar field covering the entire parameter space, the field intensity gradient of which is consistent with the spatial energy anomaly distribution of the sensitive domain;

[0083] Step S3032: input the direction vector of the parameter coupling axial set into the direction field generator, mark the action anchor point of each axial vector in the parameter space, generate a direction vector diffusion flow based on the anchor point position, form a non-conflict tensor field through the divergence-curl constraint equation, and output a perturbation direction tensor field with smooth continuity, and the streamline topology of the perturbation direction tensor field inherits the spatial characteristics of the phase transmission path breaking mode;

[0084] Step S3033: input the perturbation amplitude variable field and the perturbation direction tensor field into the field coupler, perform spatial alignment of the two fields in the parameter space isomorphism coordinate system, implement orthogonal projection multiplication of the perturbation amplitude scalar and the perturbation direction tensor for each spatial point, expand the action dimension of the amplitude scalar along the main axis of the direction tensor, and output a full-parameter space directional perturbation vector field that fuses the amplitude and direction attributes, and the vector distribution satisfies: the module length is uniquely determined by the field strength of the perturbation amplitude variable field, and the pointing direction is uniquely constrained by the streamline direction of the perturbation direction tensor field.

[0085] In the above embodiment, the present embodiment constructs a dynamic perturbation generation system with spatial adaptive characteristics through joint mapping of parameter sensitivity distribution and coupling axial set. The system converts discrete parameter sensitivity data into a continuous differentiable scalar field, simultaneously generates a non-conflict vector topological structure based on the parameter coupling relationship, and finally realizes tensor product fusion of scalar field strength and vector direction through orthogonal projection multiplication. The technical essence is to establish a mathematical isomorphism relationship between perturbation energy distribution and phase transmission path in the parameter space, and the output directional perturbation vector field simultaneously satisfies: the perturbation intensity strictly follows the gradient distribution law of the parameter sensitivity domain, and the perturbation direction completely retains the manifold geometric characteristics of the parameter coupling network; the fusion field realizes non-uniform directional propagation of perturbation action in the parameter space, and the vector module length and direction respectively encode the quantization information of the parameter sensitivity and the topological constraint of the coupling relationship, thereby providing a spatial navigation field with physical consistency for subsequent parameter optimization.

[0086] Embodiment 6:

[0087] On the basis of embodiment 5, the process of inputting the perturbation amplitude variable field and the perturbation direction tensor field into the field coupler in step S3033 provided by the present embodiment includes the following steps:

[0088] Step S30331: based on the parameter space coordinates, establish a unified dimension grid for the perturbation amplitude scalar field and the perturbation direction tensor field, perform topological position matching between the field strength sampling points of the perturbation amplitude scalar field and the streamline control points of the perturbation direction tensor field, and output an isomorphism reference field pair with spatial correspondence relationship, and the grid nodes simultaneously carry the amplitude value and the direction tensor;

[0089] Step S30332: for each spatial point direction tensor, extract the principal action axis, establish an orthogonal projection coordinate system along the principal action axis; decompose the amplitude scalar into an axial action component; output a projection amplitude component set with direction constraints, whose component values retain the original amplitude but the action dimension is limited to the principal axis direction;

[0090] Step S30333: construct a vector generation framework with the projection amplitude component set as the basis of the principal axis of the direction tensor; perform axial scalar-vector conversion on the amplitude component along the framework, aggregate the conversion results of all spatial points, and output a full-space directional disturbance vector field that satisfies the asymmetric constraint, whose physical characteristics are: the vector length is equal to the component value of the projection amplitude component set, and the vector direction is parallel to the principal action axis of the disturbance direction tensor field.

[0091] In the above embodiment, the embodiment realizes accurate projection and vectorization reconstruction of disturbance energy in the parameter coupling direction by establishing an isomorphic space mapping of scalar field and tensor field; first, a double-field space alignment basis is constructed through topological matching, then an orthogonal projection framework is established based on the principal axis characteristics of the direction tensor, and finally the vectorization conversion of the amplitude scalar in the constrained direction is completed. The core is to strictly constrain the energy distribution information of the scalar field in the manifold geometry defined by the direction tensor, and the generated directional disturbance vector field is mathematically represented as: the vector of each spatial point is a linear mapping of the original scalar field strength in the local tensor principal direction, which not only retains the disturbance intensity information determined by the parameter sensitivity, but also strictly satisfies the phase transfer direction constraint specified by the parameter coupling network; this field coupling mechanism essentially constructs a differential homeomorphism transformation in the parameter space, so that the unstructured global disturbance energy distribution is converted into a structured vector flow that follows the bottom physical constraint, providing a navigation field for parameter optimization that combines energy fidelity and direction compliance.

[0092] Embodiment 7:

[0093] On the basis of embodiment 6, the process of aggregating the conversion results of all spatial points in step S30333 provided by the embodiment of the application comprises the following steps:

[0094] Step S303331: construct an orthogonal zero-return auxiliary surface in the vertical plane with the principal axis of the direction tensor as the reference axis; output a three-dimensional non-commutative framework containing the principal direction constraint, whose reference axis direction inherits the spatial directivity of the phase transfer path breaking feature;

[0095] Step S303332: assign the amplitude component value of each spatial point of the projection amplitude component set to the reference axis of the three-dimensional non-commutative framework; apply a zero-value constraint condition on the orthogonal zero-return auxiliary surface; output a directional action quantity that only acts along the reference axis direction, which is essentially a scalar with a single-dimensional action attribute;

[0096] Step S303333: The directional action quantity is sequentially indexed by the parameter space coordinates, the discrete directional action quantity is connected according to the spatial adjacency topological rule, and the connection mutation is eliminated through the flow smoothing condition; and a continuous directional disturbance vector field covering the whole parameter space is output, and the characteristics meet: the vector length at any point is equal to the directional action value at the point; and the vector direction at any point is parallel to the reference axis.

[0097] In the above embodiment, the present embodiment inherits the phase transmission path breaking feature through the reference axis in the main axis direction, ensures that the vector direction of all space points strictly maintains the topological constraint characteristics of the original tensor field, and forms the direction transmission consistency with physical meaning. The directional projection of the amplitude component is realized by using a three-dimensional non-interchange frame, the multi-dimensional data is compressed to a single-dimensional action quantity through the zero value constraint of the orthogonal zero auxiliary surface, the original amplitude information is retained and the non-main axis direction interference is eliminated, and a strict correspondence between the scalar field and the vector field is established. Based on the adjacency topological rule of the sequential indexing of the parameter space, the continuity conversion of the discrete action quantity is realized through the flow smoothing condition, and the generated vector field meets the dual constraints of length accuracy (strictly corresponding to the original scalar value) and direction consistency (parallel to the reference axis in the whole space). The finally output continuous directional disturbance vector field is essentially a differential flow with physical constraints, and the technical essence is to realize the vectorization process of the scalar field with direction constraints, while ensuring that: ① the direction dimension strictly inherits the original tensor main axis feature; ② the amplitude dimension completely retains the original projection component information; and ③ the spatial dimension meets the continuous differential flow characteristics. It is particularly suitable for multi-scale numerical simulation scenarios that require to maintain the direction continuity of the physical field.

[0098] Embodiment 8:

[0099] On the basis of embodiment 7, the process of eliminating the connection mutation through the flow smoothing condition in step S303333 provided by the present embodiment comprises the following steps:

[0100] Step S3033331: The directional action quantity is sequentially indexed by the parameter space coordinates, the discrete directional action quantity is connected according to the spatial adjacency topological rule, and the connection mutation is eliminated through the flow smoothing condition; and a continuous directional disturbance vector field covering the whole parameter space is output, and the characteristics meet: the vector length at any point is equal to the directional action value at the point; and the vector direction at any point is parallel to the reference axis.

[0101] Step S3033332: The topologically ordered action quantity chain is extracted, the action quantity values and the reference axis direction of the two adjacent points in the chain are extracted, a linear transition vector segment connecting the two points is generated by taking the energy transmission impedance ratio between the two points as the weight, and a segmented linear vector field covering all adjacent point pairs is output, and the connection direction strictly follows the spatial trend of the energy path;

[0102] Step S3033333: Reconstruct the vector field by applying the curvature continuous boundary condition to the piecewise linear vector field, detecting the direction mutation angle at the turning point of the vector segment, and taking the cotangent value of the mutation angle as the smoothness adjustment coefficient; output the globally continuous directional disturbance vector field that eliminates local mutations, and the characteristics meet: the vector module length at any point is equal to the directional action value at the point, the vector direction at any point is parallel to the reference axis at the point, and the vector field curvature change rate is constrained by the spatial curvature of the energy transmission path.

[0103] In the above embodiments, the present embodiment ensures that the ordering of discrete action points is completely consistent with the energy conduction topology through strict binding of parameter space coordinates and energy transmission paths, forming an ordered action chain with physical meaning. The correlation between adjacent points directly reflects the local impedance characteristics of the energy transmission path, providing a physical constraint basis for subsequent vector field construction. Linear transition vector segments are generated based on energy transmission impedance ratios, so that the connection direction of the piecewise vector field between adjacent points strictly follows the spatial trend of the energy path. This process converts discrete action values into a vector field with path continuity while preserving the physical constraints of the original action module length and reference axis direction. By dynamically adjusting the smoothness coefficient through the cotangent value of the direction mutation angle, the vector field is reconstructed under the curvature continuous boundary condition, eliminating local geometric mutations. The finally output globally continuous vector field meets the three requirements: module length accuracy, direction consistency, and curvature matching. Discrete-continuous conversion under energy path constraints is achieved.

[0104] Embodiment 9:

[0105] As Figure 5 shown, on the basis of Embodiment 1-Embodiment 8, the bridge anti-seismic model online correction system based on parameter identification provided by the embodiments of the application comprises:

[0106] The lossless peeling and reorganization module 1 is used for continuously collecting multi-source mixed response signals of the bridge in real vibration, peeling off environmental noise and known steady-state interference components; the remaining mixed response signal stream is decomposed into a plurality of independent wave units, and adaptive feature slices are automatically generated according to the vibration energy transmission path and the local nonlinear behavior of the structure, each adaptive feature slice comprises three attributes of amplitude, time-varying phase and spatial correlation weight, and the output is a reorganized feature set; the reorganized feature set implicitly contains the spatial and temporal distribution information of the potential damage or parameter drift of the structure, and the reorganization logic is dynamically determined by the instantaneous propagation characteristics of the vibration energy in the structure;

[0107] The time evolution trajectory modeling module 2 is configured to construct a non-uniform evolution trajectory of each independent wave unit in the recombined feature set on a continuous time axis; the non-uniform evolution trajectory quantifies the change rule of the feature unit of each independent wave unit with the duration of vibration by establishing a phase coupling strength matrix and an energy attenuation gradient field between adjacent slices; and the abnormal coupling nodes in the phase coupling strength matrix and the mutation regions in the gradient field are automatically marked as parameter sensitive domains;

[0108] The parameter space directional disturbance module 3 is configured to generate a directional disturbance vector field in the parameter space of the bridge seismic model based on the spatial positioning and intensity distribution of the parameter sensitive domains; the directional disturbance vector field implements asymmetric fine tuning on the original bridge seismic model parameters with the sensitive domain intensity as the disturbance weight and the phase coupling direction in the evolution trajectory as the disturbance axis; the transient response verification is performed on the current vibration input adaptive feature after each disturbance, the parameter sensitivity feedback loop is formed by comparing the convergence speed change of the feature slices in the energy attenuation gradient field before and after the disturbance, and the model parameter set after online correction is output, wherein the correction direction is autonomously guided by the feature evolution trajectory and the correction amplitude is dynamically constrained by the sensitive domain intensity.

[0109] In the above embodiment, the bridge seismic model online correction system based on parameter identification is realized by the collaborative operation of the lossless stripping and recombination module, the time evolution trajectory modeling module and the parameter space directional disturbance module, and the following is achieved: dynamic decomposition and feature extraction of multi-source signals, decomposition of mixed response signals into independent wave units by stripping noise and interference components, generation of adaptive feature slices reflecting the local nonlinear behavior and energy transmission characteristics of the structure, and provision of high-fidelity input for subsequent analysis. The non-uniform evolution trajectory is constructed, the evolution rule of the feature unit with time is quantified by the phase coupling strength matrix and the energy attenuation gradient field, and the abnormal coupling nodes and the gradient mutation regions are marked as parameter sensitive domains to realize accurate positioning of the abnormal regions of the structure. The directional disturbance vector field is generated based on the spatial distribution of the sensitive domains, asymmetric fine tuning is performed on the model parameters, the correction direction is guided by the feature evolution trajectory, and the correction amplitude is constrained by the sensitive domain intensity; the transient response verification and feedback loop are used to ensure that the correction process responds to local nonlinear changes and maintains overall stability.

[0110] In summary, the bridge seismic model online autonomous correction is realized, which can dynamically adapt to the structural parameter drift and potential damage, improve the response prediction accuracy under strong earthquake action, and avoid relying on preset damage modes or manual intervention.

[0111] Figure 6 A block diagram of an exemplary electronic device suitable for implementing an embodiment of the present application is shown.

[0112] The electronic device can include a central processor / microprocessor / master control chip, etc. 4; a storage medium 5 coupled to the central processor / microprocessor / master control chip, etc. 4 and storing computer executable instructions therein for performing steps of various methods of embodiments of the present application when executed by the processor.

[0113] The central processor / microprocessor / master control chip, etc. 4 can include, but not limited to, for example, one or more processors or microprocessors, etc.

[0114] The storage medium 5 can include, but not limited to, for example, random access memory (RAM), read only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (such as hard disks, floppy disks, solid state drives, removable disks, CD-ROMs, DVD-ROMs, Blu-ray discs, etc.).

[0115] In addition, the electronic device can also include, but not limited to, a data bus 6, an input / output bus / external bus / device bus, etc. 7, a display 8 and an input / output device 9 (such as a keyboard, a mouse, a speaker, etc.), etc.

[0116] The central processor / microprocessor / master control chip, etc. 4 can communicate with external devices (8, 9, etc.) via wired or wireless networks (not shown) through the I / O bus 7.

[0117] The storage medium 5 can also store at least one computer executable instruction for performing steps of various functions and / or methods in embodiments described in the present technology when executed by the central processor / microprocessor / master control chip, etc. 4.

[0118] In one embodiment, the at least one computer executable instruction can also be compiled or composed into a software product, wherein one or more computer executable instructions are executed by the processor to perform steps of various functions and / or methods in embodiments described in the present technology.

[0119] Figure 7 A schematic diagram of a computer readable storage medium according to embodiments of the present application is shown.

[0120] As Figure 7As shown, the non-transitory computer-readable storage medium 11 stores instructions, for example, computer-readable instructions 10. When the computer-readable instructions 10 are run by the processor, the methods described above can be performed. The non-transitory computer-readable storage medium includes, but is not limited to, for example, a volatile memory and / or a non-volatile memory. The volatile memory, for example, can include a random access memory (RAM), a cache, and / or the like. The non-volatile memory, for example, can include a read-only memory (ROM), a hard disk, a flash memory, and / or the like. For example, the non-transitory computer-readable storage medium 11 can be connected to a computing device such as a computer, and then, when the computing device runs the computer-readable instructions 10 stored on the non-transitory computer-readable storage medium 11, the methods described above can be performed.

[0121] In several embodiments provided by the present application, it should be understood that the disclosed apparatus and method can be implemented in other manners. For example, the described apparatus embodiments are merely schematic. The division of the units is merely a logical function division. There can be another division manner for the actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0122] The units described as separated components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Some or all of the units can be selected according to the actual needs to achieve the purposes of the embodiments of the present application.

[0123] In addition, each functional unit in the embodiments of the present application can be integrated in one processing unit, or each unit can exist physically as separated, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware, or in the form of software functional units.

[0124] If the integrated unit is implemented in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for executing all or part of the steps of the embodiments of the present application by a computer device (which can be a personal computer, a server, or a network device, etc.). The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (English full name: Read-Only Memory, English abbreviation: ROM), a random access memory (English full name: Random Access Memory, English abbreviation: RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0125] The above embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An online correction method for bridge seismic model based on parameter identification, characterized in that: The method includes: decomposing the bridge vibration response signal into independent wave units and generating adaptive feature slices containing amplitude, time-varying phase and spatial weight; Based on the phase coupling strength matrix and energy attenuation gradient field analysis, the characteristic evolution trajectory is automatically identified as abnormal coupling nodes and gradient mutation areas as parameter sensitive domains; Based on the spatial location and strength distribution of parameter sensitive regions, a directional disturbance vector field is generated in the parameter space of the bridge seismic model. The directional perturbation vector field uses the sensitive domain intensity as the perturbation weight and the phase coupling direction in the evolution trajectory as the perturbation axis to implement asymmetric fine-tuning of the original bridge seismic model parameters.

2. The method for online correction of bridge seismic model based on parameter identification according to claim 1, characterized in that: The process of asymmetric fine-tuning of the original bridge seismic model parameters includes the following steps: The spatial location and strength distribution of the parameter sensitive domain are input into the space-parameter mapper. Based on the spatial relationship between the boundary coordinates of the parameter sensitive domain and the parameters of the bridge seismic model, a parameter influence weight matrix is ​​established. The strength values ​​of the sensitive domain are assigned to the corresponding parameter groups according to the weight matrix, and the output is a parameter sensitivity distribution map with strength annotations. The phase coupling direction in the non-uniform evolution trajectory is input into the direction-parameter converter to extract the main axis of the phase coupling direction in the structural space. According to the structural dynamic constitutive relation, the spatial axis is mapped to the parameter-related direction vector in the parameter space, and the output is the parameter coupling axis set that defines the parameter perturbation direction. The parameter sensitivity distribution map and the parameter coupling axial set are input into the field synthesis operator. The disturbance amplitude scalar field is generated based on the intensity value of the parameter sensitivity distribution map. The disturbance direction tensor field is constructed based on the direction vector of the parameter coupling axial set. The Kronecker product operation is performed on the disturbance amplitude scalar field and the disturbance direction tensor field, and a directional disturbance vector field with spatial heterogeneity is output. Its vector distribution satisfies both intensity non-uniformity and directional asymmetry.

3. The method for online correction of bridge seismic model based on parameter identification according to claim 2, characterized in that: The process of outputting a spatially heterogeneous directional perturbation vector field includes the following steps: The intensity values ​​of the parameter sensitivity distribution map are input into the amplitude field generator. An intensity interpolation grid is established based on the parameter space coordinates. The discrete intensity values ​​are converted into a continuous scalar distribution through the radial basis function. The disturbance amplitude scalar field covering the entire parameter space is output. The field intensity gradient is consistent with the spatial energy anomaly distribution of the sensitive domain. The direction vectors of the parameter-coupled axial set are input into the direction field generator. The action anchor points of each axial vector are marked in the parameter space. The direction vector diffusion flow is generated based on the anchor point position. The conflict-free tensor field is formed through the divergence-curl constraint equation. The perturbation direction tensor field with smooth continuity is output. Its streamline topology inherits the spatial characteristics of the phase transfer path fracture mode. The disturbance amplitude variable field and the disturbance direction tensor field are input into the field coupler, and the dual fields are aligned in the parameter space homogeneous coordinate system. The orthogonal projection multiplication is performed on the disturbance amplitude scalar and the disturbance direction tensor of each spatial point, and the action dimension of the amplitude scalar is expanded along the main axis of the direction tensor. The output is a full parameter space directional disturbance vector field that integrates the amplitude and direction properties.

4. The method for online correction of bridge seismic model based on parameter identification according to claim 3, characterized in that: The process of inputting the disturbance amplitude variable field and the disturbance direction tensor field into the field coupler includes the following steps: The disturbance amplitude scalar field and the disturbance direction tensor field are combined to establish a unified measurement grid based on the parameter space coordinates, and the field intensity sampling points of the disturbance amplitude scalar field are topologically matched with the streamline control points of the disturbance direction tensor field. Output isomorphic reference field pairs with spatial correspondence, whose grid nodes carry both amplitude values ​​and direction tensors; Extract the principal action axis of the direction tensor of each spatial point from the isomorphic reference field pair, and establish an orthogonal projection coordinate system along the principal action axis; decompose the amplitude scalar into axial action components; Output the projection amplitude component set with direction constraints; The projected amplitude component set is used to construct a vector generation framework based on the principal axis of the direction tensor; The amplitude components are transformed from scalars to vectors along the frame, and the transformation results of all spatial points are aggregated to output the full-space directional perturbation vector field that satisfies the asymmetric constraints.

5. The method for online correction of bridge seismic model based on parameter identification according to claim 4, characterized in that: The process of aggregating the transformation results of all spatial points includes the following steps: The principal axis of the direction tensor is used as the reference axis to construct an orthogonal zeroing auxiliary surface in the vertical plane; a three-dimensional non-commutative frame containing principal direction constraints is output, and its reference axis direction inherits the spatial directivity of the phase transfer path fracture feature; Assign the amplitude component value of each spatial point of the projected amplitude component set to the reference axis of the three-dimensional non-commutative frame; impose zero value constraints on the orthogonal zeroing auxiliary surface; Outputs the directional action acting only along the reference axis direction; Directed actions are sequentially indexed with parameter space coordinates, discrete directed actions are connected according to spatial adjacency topological rules, and connection mutations are eliminated through manifold smoothing conditions; a continuous directional perturbation vector field covering the entire parameter space is output.

6. The method for online correction of bridge seismic model based on parameter identification according to claim 5, characterized in that: The process of eliminating connection mutations through manifold smoothing conditions includes the following steps: The directional action is arranged in a natural sequence with the coordinates in the parameter space, and the discrete points are rearranged according to the spatial ordering rules of the energy transfer path. The topologically ordered action chain with path dependence is output, and the correlation between its adjacent points comes from the topology of the energy transfer path. From a topologically ordered action chain, the action values ​​and reference axis directions of two adjacent points in the chain are extracted, and a linear transition vector segment connecting the two points is generated using the energy transfer impedance ratio between the two points as a weight. The output is a piecewise linear vector field covering all pairs of adjacent points, and its connection direction strictly follows the spatial direction of the energy path. The piecewise linear vector field is divided into two parts, and the directional mutation angle at the turning point of the vector segment is detected. The cotangent value of the mutation angle is used as the smoothness adjustment coefficient, and the curvature continuous boundary condition is applied to reconstruct the vector field; the global continuous directional perturbation vector field that eliminates local mutations is output.

7. The method for online correction of bridge seismic model based on parameter identification according to claim 1, characterized in that: After each disturbance, the current vibration input adaptive features are used to verify the transient response. By comparing the changes in the convergence speed of the feature slices in the energy attenuation gradient field before and after the disturbance, a parameter sensitivity feedback loop is formed. The model parameter set that completes online correction is output, and its correction direction is autonomously guided by the feature evolution trajectory, and the correction amplitude is dynamically constrained by the sensitive domain strength.

8. The method for online correction of bridge seismic model based on parameter identification according to claim 1, characterized in that: The multi-source mixed response signals of the bridge under real vibration are continuously collected to remove environmental noise and known steady-state interference components. The remaining mixed response signal stream is decomposed into several independent fluctuation units, and adaptive feature slices are automatically generated based on the vibration energy transfer path and the local nonlinear behavior of the structure. Each adaptive feature slice contains the triple attributes of amplitude, time-varying phase and spatial correlation weight, and the output is a reconstructed feature set. The reconstructed feature set already implies the spatiotemporal distribution information of potential structural damage or parameter drift, and its reconstructing logic is dynamically determined by the instantaneous propagation characteristics of vibration energy in the structure.

9. The method for online correction of bridge seismic model based on parameter identification according to claim 8, characterized in that: For each independent fluctuation unit in the reorganized feature set, a non-uniform evolution trajectory is constructed on the continuous time axis. The non-uniform evolution trajectory quantifies the change pattern of the characteristic unit of each independent fluctuation unit with the vibration duration by establishing the phase coupling intensity matrix and energy attenuation gradient field between adjacent slices. The abnormal coupling nodes in the phase coupling intensity matrix and the mutation areas in the gradient field will be automatically marked as parameter sensitive domains.

10. A bridge seismic model online correction system based on parameter identification according to the bridge seismic model online correction method based on parameter identification according to any one of claims 1 to 9, characterized in that: The online correction of the bridge seismic model based on parameter identification includes: The non-destructive stripping and reassembly module is used to continuously collect the multi-source mixed response signals of the bridge under real vibration, stripping away the environmental noise and known steady-state interference components. The remaining mixed response signal stream is decomposed into several independent wave units, and adaptive feature slices are automatically generated based on the vibration energy transfer path and the local nonlinear behavior of the structure, outputting the reassembled feature set. The time evolution trajectory modeling module is used to construct a non-uniform evolution trajectory for each independent fluctuation unit in the recombined feature set on a continuous time axis. The non-uniform evolution trajectory quantifies the change of the characteristic unit of each independent fluctuation unit with the vibration duration by establishing a phase coupling strength matrix and energy attenuation gradient field between adjacent slices. Abnormal coupling nodes in the phase coupling strength matrix and mutation areas in the gradient field are automatically marked as parameter sensitive domains. The parameter space directional perturbation module is used to generate a directional perturbation vector field in the parameter space of the bridge seismic model based on the spatial positioning and intensity distribution of the parameter sensitive domain. The directional perturbation vector field uses the sensitive domain intensity as the perturbation weight and the phase coupling direction in the evolution trajectory as the perturbation axis to perform asymmetric fine-tuning on the original bridge seismic model parameters. After each disturbance, the transient response is verified using the current vibration input adaptive characteristics. By comparing the changes in the convergence speed of the feature slices in the energy attenuation gradient field before and after the disturbance, a parameter sensitivity feedback loop is formed. The model parameter set that has completed online correction is output, and its correction direction is autonomously guided by the feature evolution trajectory, and the correction amplitude is dynamically constrained by the sensitive domain intensity.

Images