Shock absorption detection digital twin mapping and dynamic simulation method

Abstract

The application discloses a kind of shock insulation and shock absorption detection digital twin mapping and dynamic simulation method, it is related to shock insulation and shock absorption structure detection and digital twin simulation technical field, under the containment constraint, obtain disturbance event and collect multi-point response data, environmental data and sampling quality label, time alignment is extracted time process feature to generate response signature;Unified mapping state quantity is jointly inversed based on local evidence, and respectively input mechanism twin model and shadow check model, combined with state mapping constraint, perform consistency determination;Through time write back digital twin and carry out scenario dynamic simulation, when not passing, freeze write back and generate supplementary survey scheduling result, and after double-track adjudication, combined with write back result or frozen shunt record forms evidence package, for risk level determination.

Description

Technical Field

[0001] This invention relates to the field of seismic isolation and damping structure inspection and digital twin simulation technology, specifically a digital twin mapping and dynamic simulation method for seismic isolation and damping inspection. Background Technology

[0002] Seismic isolation bearings, dampers, and their connecting components are continuously affected by environmental factors such as earthquakes, wind vibrations, equipment disturbances, temperature, humidity, and corrosion during long-term service. The objects being tested exhibit distinct dynamic response characteristics and are also influenced by changes in operating conditions, sampling quality, and differences in multi-source measurements. For in-service structures, identifying changes in the condition of critical components with minimal or no interruption in operation, and using the test results for subsequent simulation analysis and risk assessment, has become a crucial technical challenge in the operation and maintenance of seismic isolation and damping systems.

[0003] Existing technologies continuously collect displacement, strain, acceleration, and environmental data through multi-sensor monitoring and remote analysis and alarms. After deploying monitoring points at key locations, they conduct monitoring and early warning based on thresholds or trends. However, these technologies still have certain limitations. Such solutions typically focus more on the accumulation of continuous monitoring data and the identification of single-indicator over-limits or alarm output. They do not make sufficient use of time-process information such as the onset timing, transmission path difference, attenuation process, hysteresis closure deviation, and residual recovery trajectory within the same disturbance event window. It is difficult to unify multi-point responses into a response signature that can be directly used for subsequent judgment. When there are large environmental fluctuations or unstable sampling quality, it is also difficult to distinguish the influence between environmental changes and component degradation.

[0004] Existing technologies also incorporate multi-source data into structural models through spatiotemporal alignment using digital twins, model updates, or state estimation. Model updates, sequential Bayesian inference, or full-structure state estimation are employed to support performance evaluation, early warning, or predictive maintenance. However, certain limitations remain. Many solutions still focus on incorporating monitoring data into a single model, online estimation, or single-track early warning. The state benchmarks relied upon for diagnosis, simulation, and risk analysis are not unified, and the write-back conditions, physical boundary constraints, and cross-verification mechanisms for observable evidence after model updates are not clearly defined. Based on currently available publicly available information, there are few methodologies that integrate controlled perturbation event-driven processes, response signature extraction, joint inversion of unified mapping state quantities, dual-track adjudication using mechanistic twin models and shadow verification models, state mapping constraint-gated write-back, and scenario-based dynamic simulation into a single closed loop.

[0005] Therefore, there is an urgent need for a digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection. Under the constraint of the containment, the method takes the disturbance event as the detection entry point, uses response signature to jointly invert the unified mapping state quantity, and forms a continuous processing chain through the mechanism twin model, shadow verification model and state mapping constraints, including consistency judgment, digital twin write-back or freeze diversion, scenario-based dynamic simulation, evidence package generation and risk level determination. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and propose a digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection to solve the above-mentioned problems.

[0007] The objective of this invention is achieved through the following technical solution: a digital twin mapping and dynamic simulation method for seismic isolation and damping detection, comprising: S1, under the premise of satisfying containment constraints, acquiring disturbance events based on current environmental conditions, structural operating status, and historical degradation records, wherein the disturbance events are taken from any event in the window of limited-amplitude micro-disturbance loading trigger events and natural disturbance events during structural operation, and collecting multi-point response data, environmental data, and sampling quality labels corresponding to the disturbance events; S2, performing time alignment on the multi-point response data, environmental data, and sampling quality labels, and extracting data from the aligned event data. S3. Take the oscillation delay, propagation path difference, attenuation envelope, hysteresis closure deviation, residual recovery trajectory, environmental correction deviation, phase lag, and threshold crossing order to generate a response signature for the same key component; S4. Convert the data corresponding to visual displacement, acceleration, strain, temperature and humidity, corrosion, electromagnetic detection, and acoustic non-destructive testing into local evidence, and perform joint inversion of the unified mapping state variables based on the response signature. The unified mapping state variables include at least the equivalent energy dissipation gradient, recovery stiffness attenuation coefficient, threshold drift coefficient, boundary constraint relaxation factor, environmentally sensitive amplification, and response confidence weight; S5. [The text abruptly ends here, likely due to an incomplete sentence or a missing section.] The mapped state variables are simultaneously input into the mechanistic twin model and the shadow verification model. The mechanistic twin model outputs the structural response reconstruction results, parameter update results, and dominant mechanism determination results. The shadow verification model outputs the observable evidence prediction results corresponding to the support loop opening order, damper phase difference change direction, measurement point time delay propagation relationship, and residual error convergence form after environmental switching. S5. The environmental data is compiled into state mapping constraints, and a consistency judgment is performed on whether the deviation between the structural response reconstruction results and the measured response, the deviation between the observable evidence prediction results and the measured evidence, and the parameter update results fall within the physical boundary range. When the consistency judgment passes, the parameter update results are written back to the digital twin. When the dominant mechanism judgment result indicates a switch in the dominant mechanism, the corresponding twin sub-model and simulation boundary conditions are switched according to the state mapping constraints. Based on the written-back digital twin, scenario-based dynamic simulations covering representative seismic wave conditions, temperature change conditions, continuous aftershock conditions, long-term small amplitude fatigue conditions, and component local failure conditions are performed, and the cooperative failure sequence of the isolation components and damping components, the key response amplification range, the safety margin change, and the remaining serviceable time window are output. When the consistency judgment fails, the current round of write-back is frozen and the supplementary measurement scheduling results are generated.

[0008] The containment constraints in S1 are used to limit the disturbance loading amplitude, disturbance duration, disturbance location, and disturbance termination conditions.

[0009] The limited perturbation loading in S1 is selected from at least one of low-amplitude displacement pulse, frequency sweep excitation, short-time additional mass switching, and local electromagnetic excitation.

[0010] When generating a response signature in S2, the oscillation delay, propagation path difference, attenuation envelope, hysteresis closure deviation, residual recovery trajectory, environmental correction deviation, phase lag, and threshold crossing order are associated and encoded according to the same disturbance event window.

[0011] When performing joint inversion of unified mapping state variables in S3, the local evidence of the corresponding unified mapping state variables is fused and inverted according to the same disturbance event window, so that diagnosis, simulation, early warning and lifetime extrapolation all depend on the unified mapping state variables.

[0012] The mechanistic twin model in S4 also outputs at least one of the following: equivalent stiffness, equivalent damping, slip threshold, yield range, and damage evolution parameters. The shadow verification model updates the prediction results of observable evidence synchronously based on the unified mapping state variables.

[0013] In S5, consistency checks are only performed when the comparisons of the deviation between the structural response reconstruction result and the measured response, the deviation between the observable evidence prediction result and the measured evidence, and whether the parameter update result falls within the physical boundary range all pass.

[0014] The environmental effects corresponding to the state mapping constraints include at least temperature, humidity, corrosion, high-frequency background vibration, freeze-thaw conditions, and additional foundation settlement. The criteria for determining the switching of the dominant mechanism are jointly determined by the state mapping constraints and the parameter update results.

[0015] An evidence package is generated based on the scenario-based dynamic simulation results. The evidence package includes at least the source of disturbance, sampling integrity, environmental level, state variable inversion residual, dual-track consistency score, and update confidence interval.

[0016] The risk level of a critical component is increased only when the same critical component continuously exhibits a vulnerability trend in scenario-based dynamic simulation and the corresponding response signature is reproduced in subsequent disturbance events.

[0017] The beneficial effects of this invention are: This scheme integrates controlled micro-disturbance events and natural disturbance events into a single event entry point, triggering the acquisition, judgment, and update process under containment constraints. Compared to long-term passive monitoring methods, this structure synchronously fixes the excitation time, sampling window, and safety interlocking conditions, ensuring that the seismic isolation bearings, dampers, and connecting components respond within the same timeframe. This facilitates earlier detection of signs of weak degradation and local instability, and improves the temporal consistency of detection results.

[0018] This scheme constructs a temporal process fingerprint around the disturbance event window, jointly encoding displacement, velocity, phase lag, decay process, and recovery trajectory. Compared to methods that only extract peak values ​​or single-point thresholds, this structure preserves the dynamic evolution information of components after being subjected to stress, enabling the differentiation between slight performance drifts and changes in dominant mechanisms over time. This improves the ability to identify and distinguish early anomalies and enhances comparability.

[0019] This approach recodes multi-source monitoring results into a unified mapping state quantity, rather than remaining at the level of scattered descriptions of sensor quantities, component quantities, or model parameters. Compared to the method of judging each module separately, the unified mapping state quantity compresses geometric response, energy consumption characteristics, connectivity constraints, and environmental conditions into the same state space, ensuring that field data, historical data, and the objects updated in the twin model maintain the same caliber. This helps reduce cross-module interpretation bias and provides a foundation for subsequent updates and risk comparisons.

[0020] This scheme employs a dual-track evidence adjudication system, combining a mechanistic twin model and a shadow verification model, and overlays physical boundary ranges, consistency thresholds, and freeze / rollback rules, avoiding direct write-back based solely on a single model's fitting result. Compared to single-model updates, this structure incorporates computational convergence, physical realizability, and cross-validation into the admission criteria, intercepting erroneous updates induced by mismatches and accidental perturbations before write-back. This helps maintain the stability of the digital twin version and suppresses the spread of erroneous learning.

[0021] This approach integrates temperature, humidity, corrosion, high-frequency background vibration, freeze-thaw conditions, and additional foundation settlement into state mapping constraints, driving the identification of the dominant mechanism and the switching of sub-models, rather than making empirical corrections after the calculation is complete. Compared to methods where environmental factors are only used as supplementary descriptions, this structure isolates response differences under different boundaries during the state mapping stage, avoiding misjudging environmental changes as component degradation. This improves the robustness of diagnostic conclusions and enhances the comparability of results across different time periods.

[0022] This solution updates the results and writes them back to the digital twin after the judgment is approved, and conducts dynamic simulations around representative earthquakes, temperature abrupt changes, continuous aftershocks, long-term small-amplitude fatigue, and local failure scenarios. Compared to methods that only provide a single diagnostic conclusion, this structure transforms the detection results into predictive conditions for subsequent behavior, allowing the same state quantity to drive risk propagation analysis and maintenance priority ranking. This facilitates the closure of detection and operation and maintenance decisions and provides a basis for maintenance timing and retesting arrangements.

[0023] This solution also establishes mechanisms for evidence retention, dual-condition triggering of risk escalation, retesting after freezing, and collaborative interlocking of execution carriers, creating a closed loop for event initiation, execution, collection, updating, rollback, and recovery. Compared to methods that only output algorithms, this structure synchronously records on-site actions, version status, judgment criteria, and reasons for abnormal termination, and retains the frozen version when conditions are not met. This improves the security, accountability, and operational capabilities of project implementation. Attached Figure Description

[0024] Figure 1 The process of this invention Figure 1 ; Figure 2 The process of this invention Figure 2 ; Figure 3 The process of this invention Figure 3 ; Figure 4 The process of this invention Figure 4 . Detailed Implementation

[0025] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1 like Figure 1 As shown, this embodiment takes an in-service building equipped with seismic isolation bearings, damping components, connecting components, and key measuring points in the superstructure as the object, and constructs a digital twin mapping and dynamic simulation process for seismic isolation and damping detection. Lead-core rubber seismic isolation bearings and friction pendulum bearings are arranged at the bottom of the building, and viscous dampers are arranged at the inter-story locations in the superstructure. Displacement, acceleration, and strain acquisition points are set on the top surface of the bearings, both ends of the dampers, the ends of the transfer layer beams, and the roof layer. Visual displacement acquisition points are set on the corresponding key layers of the building facade, and environmental acquisition points are set on the equipment layer. The entire process is collaboratively completed by an event scheduling unit, a perturbation execution unit, a data acquisition unit, a time alignment unit, a response signature generation unit, a state inversion unit, a mechanism twin model calculation unit, a shadow verification model calculation unit, a consistency determination unit, a write-back control unit, and a scenario-based dynamic simulation unit.

[0027] In the initial deployment phase of this embodiment, unified rules are first established. The event scheduling unit generates a unique event number for each disturbance event, which consists of the detection date, component category, trigger source, and daily sequence number. Disturbance events are divided into two categories: limited-amplitude micro-disturbance loading triggered events and natural disturbance event windows. Limited-amplitude micro-disturbance loading triggered events are triggered by the micro-disturbance execution unit applying a low-amplitude displacement pulse. The engineering default peak value of the displacement pulse is set to 2.0 mm, the duration is set to 0.40 s, and the target is limited to a single seismic isolation bearing or a single damper. Natural disturbance event windows are extracted by the event scheduling unit from wind-induced vibration, equipment start-up and shutdown vibration, or micro-seismic response, and the extraction window length is set to 12 s. The containment constraints are written into the scheduling table by the event scheduling unit during the cold start calibration phase. They include at least the upper limit of the disturbance loading amplitude, the upper limit of the disturbance duration, the limitation of the action position, and the disturbance termination condition. The upper limit of the displacement pulse amplitude is set to 3.0 mm, the upper limit of the sweep frequency excitation frequency is set to 12 Hz, and the disturbance termination condition is set to stop execution immediately when the absolute value of the displacement at any measuring point reaches 80% of the preset safe displacement threshold.

[0028] The data acquisition unit simultaneously collects multi-point response data, environmental data, and sampling quality labels for each disturbance event. Multi-point response data includes the relative displacement between the top and bottom surfaces of the support, the displacement at both ends of the damper, the acceleration of the critical layer, the strain at the beam end, and visual displacement. Environmental data includes temperature, humidity, corrosion potential, and background vibration amplitude. Sampling quality labels include timestamp integrity, sensor online status, visual occlusion status, and abnormal jump markers. The master clock is provided uniformly by the time alignment unit. The sampling frequency for displacement, acceleration, and strain signals is set to 200 Hz, the sampling frequency for visual displacement is set to 50 Hz, and the environmental data refresh cycle is set to 1 s. The time alignment unit uses the event trigger time as the time... Starting from zero, uniform resampling and time-stamp correction are performed on each sampling stream. When samples are missing, segmented linear filling with adjacent valid samples is prioritized. Events with consecutive missing sample lengths exceeding 0.15 s are directly marked as invalid events and sent back to the event scheduling unit for rescheduling. In this embodiment, the timestamp integrity in the sampling quality label is used to form a sampling integrity index. The sampling integrity index is calculated as the ratio of the number of valid samples in the current event window to the number of samples that should be sampled, and together with the sensor online status, visual occlusion status, and abnormal jump markers, it serves as the basis for subsequent event validity determination, state inversion weight shrinking, and evidence packet generation.

[0029] The response signature generation unit generates response signatures for the same critical component on time-aligned event data. For the first... For each measurement point, the oscillation delay is defined as the difference between the triggering time and the time when the measurement point reaches 1.2 times the baseline root mean square value. The calculation formula is: The transmission path difference is defined as the difference in the oscillation delay between two adjacent measuring points of the same critical component; the attenuation envelope is obtained by the envelope sequence composed of peak values ​​of each period; the hysteresis closure deviation is the normalized difference between the end and start points of a complete response loop; the residual recovery trajectory is the displacement attenuation sequence of the recovery segment after the disturbance stops; the environmental correction deviation is the deviation between the measured characteristic quantity under the current environmental conditions and the same characteristic quantity under the baseline environmental conditions. The response signature generation unit further extracts the phase hysteresis and threshold crossing order, and forms an event-level time process fingerprint. To avoid direct mixing of different dimensions, this embodiment normalizes various characteristic quantities to the same event window using the following formula: In the formula, For the first The start-up delay at each measuring point For the first The moment when each measuring point first reaches the vibration criterion.

[0030] In the formula, For the first The second disturbance event Normalized feature quantity, For the first The second disturbance event Original feature quantity, These are reference characteristic quantities for similar key components under baseline operating conditions. To prevent stable terms with excessively small denominators, this embodiment takes... In this embodiment, the oscillation delay, transmission path difference, attenuation envelope, hysteresis closure deviation, residual recovery trajectory, environmental correction deviation, phase lag, and threshold crossing order are uniformly written into the response signature table. The field order of the response signature table is fixed as event number, component number, feature number, feature value, normalized feature value, generation time, and quality flag.

[0031] The state inversion unit first converts data corresponding to visual displacement, acceleration, strain, temperature and humidity, corrosion, electromagnetic detection, and acoustic non-destructive testing into local evidence, and then performs joint inversion on the unified mapping state variables based on the same disturbance event window. The unified mapping state variable is defined as follows: ,in, Indicates the first The equivalent energy dissipation gradient corresponding to the second perturbation event. Indicates the coefficient of stiffness reduction. Indicates the threshold drift coefficient. This represents the boundary constraint relaxation factor. This indicates a large amount of environmentally sensitive material. This represents the response credibility weight. The initial preset values ​​for the cold start phase are set to... , , , , , The state inversion unit affects the normalized feature vector. Perform a weighted update, and the update formula is written as: In the formula, The state preservation coefficient is set to 0.62 in this embodiment; The calibration matrix is ​​obtained during the cold start calibration phase based on baseline perturbation events; For the first The normalized feature vector of the sub-disturbance event. Each row of the calibration matrix corresponds to a field in the unified mapping state quantity, and each column corresponds to a type of feature quantity in the response signature, thus making the generation source of each field traceable. During the cold start calibration phase, ten to twenty valid disturbance events formed by similar key components under complete installation, without triggering freeze shunting, and in an environment level of level one or two are preferentially selected as baseline samples. The median feature quantity of this batch of samples is used as the baseline. The values ​​are based on the values ​​of the standard matrix, and the calibration matrix is ​​obtained by least-squares fitting. If certain types of local evidence are missing, the state inversion unit will reduce the weight of the corresponding column to 0 based on the sampling quality label, and simultaneously lower the response confidence weight. .

[0032] The mechanistic twin model's computational unit and the shadow verification model's computational unit both receive the same set of unified mapped state variables. The mechanistic twin model is based on... , , and The system updates equivalent stiffness, equivalent damping, slip threshold, yield range, and damage evolution parameters, outputting structural response reconstruction results, parameter update results, and dominant mechanism determination results. The dominant mechanism determination results must distinguish at least four states: elastic shear dominance, slip dominance, viscous energy dissipation dominance, and connection relaxation dominance. The shadow verification model does not output structural parameters; instead, it predicts the support loop opening order, damper phase difference change direction, measurement point time delay propagation relationship, and residual error convergence form after environmental switching based on the same set of unified mapping state variables. The two links are bound with the same event number, and the output results are written to the same event processing record.

[0033] The consistency determination unit performs gating determination on the structural response reconstruction results, observable evidence prediction results, and parameter update results. Therefore, this embodiment defines a structural response consistency score. Consistency score of evidence and boundary validity score The overall consistency score is generated using the following formula: In the formula, For the first The overall consistency score of the disturbance events. , and These are the weights for structural response consistency score, evidence consistency score, and boundary validity score, respectively. In this implementation, we take... , , The sum of the three is 1. Structural response consistency score. The evidence consistency score is calculated from the normalized residuals of the measured response and the reconstructed response. The boundary validity score is obtained by comparing the degree of matching between measured evidence and the evidence predicted by the shadow verification model. The result is calculated based on whether the parameter update result falls within the physical boundary range. In this embodiment, the physical boundary range is defined by both the upper and lower bounds of the parameters and the rate of change limit. Specifically, it includes: the equivalent stiffness update amplitude not exceeding 25% of the baseline value; the equivalent damping update amplitude not exceeding 20% ​​of the baseline value; the slip threshold drift not exceeding 15% of the baseline value; and the parameter change rate for two consecutive events not exceeding 0.10. The write-back control unit performs write-back or freeze shunt according to the following formula: In the formula, For the first The write-back action marker for the second perturbation event. This indicates that a write-back operation will be performed. This indicates that the current write-back round is frozen; To determine the overall consistency score threshold, this implementation uses 0.78. For the first The physical boundary range corresponding to the disturbance event is determined by the consistency determination unit, which outputs the result to the write-back control unit and the event scheduling unit.

[0034] In this embodiment, environmental data is compiled into state mapping constraints, which consist of temperature constraints, humidity constraints, corrosion constraints, high-frequency background vibration constraints, freeze-thaw condition constraints, and additional foundation settlement constraints. Each constraint object contains at least six fields: constraint type, current measurement value, reference measurement value, amplification value, direction of action, and activation flag. When the dominant mechanism changes, the write-back control unit switches the corresponding twin model and simulation boundary conditions according to the state mapping constraints. Taking the change from elastic shear-dominated to slip-dominated support as an example, when the threshold drift coefficient... The boundary constraint relaxation factor is higher than 0.12 twice consecutively. When the value is higher than 0.08, the system switches the support sub-model from shear recovery model to slip recovery model, and synchronously switches the support friction boundary conditions in the scenario-based dynamic simulation to the slip boundary parameter set.

[0035] When writing back the action tag At that time, the write-back control unit writes the parameter update results back to the digital twin, and the scenario-based dynamic simulation unit executes scenario-based dynamic simulations covering representative seismic wave conditions, temperature change conditions, continuous aftershock conditions, long-term small-amplitude fatigue conditions, and component local failure conditions. The scenario-based dynamic simulation outputs the cooperative failure sequence of the isolation components and damping components, the key response amplification range, the safety margin change, and the remaining serviceable time window. For ease of unified recording, this embodiment uses the following formula to calculate the scenario response amplification: In the formula, For the first After the second disturbance event is written back to the scene The response volume is greatly increased. This represents the current maximum value of the key response quantity in this scenario. This serves as the reference maximum value for the baseline digital twin in the same scenario. The scenario-based dynamic simulation unit aggregates the outputs of each scenario into the event processing record and provides the basic fields of the subsequent evidence package to the risk output end.

[0036] When writing back the action tag When writing back to the control unit, the current write-back round is frozen, retaining the parameters of the digital twin from the previous stable version, and a supplementary test scheduling result is generated. The supplementary test scheduling result includes at least the supplementary test object, the reason for supplementary test, priority, suggested disturbance template, and the next scheduling time, and is output to the event scheduling unit and the operation and maintenance terminal. If the reason for freezing is due to abnormal sampling quality labels, resampling is given priority; if the reason for freezing is due to excessive deviation of observable evidence, retesting of the same component is given priority; if the reason for freezing is due to exceeding the physical boundary range, the component is given priority to be transferred to high-frequency monitoring.

[0037] The working process of this embodiment is as follows: After the detection cycle arrives, the event scheduling unit reads the historical degradation archive and the real-time environmental state, generates a disturbance event, and calls the micro-disturbance execution unit or captures the natural disturbance event window; the data acquisition unit completes the acquisition of multi-point response data, environmental data, and sampling quality labels, and the time alignment unit forms a unified event data packet; the response signature generation unit generates a time process fingerprint, and the state inversion unit obtains a unified mapped state quantity; the mechanism twin model and the shadow verification model respectively output the structural response reconstruction result, parameter update result, dominant mechanism judgment result, and observable evidence prediction result; the consistency judgment unit forms a comprehensive consistency score and triggers the write-back control unit; the write-back control unit executes write-back or freeze diversion; the scenario-based dynamic simulation unit outputs five types of scenario results after write-back, and the event scheduling unit then enters the next detection cycle based on the results of this round.

[0038] By unifying the limited perturbation loading trigger event and the natural perturbation event window into a perturbation event, the detection entry point is changed from long-term passive monitoring to event-based detection with identification purpose; by uniformly encoding the oscillation delay, phase lag, attenuation envelope, and threshold crossing order into a response signature, the detection basis is expanded from static final value to time process characteristics; by uniformly mapping state quantities, local evidence is compressed into a master state that can be directly used for diagnosis, write-back, and simulation; by using a dual-track parallel mechanism twin model and shadow verification model, the write-back action no longer depends solely on the fitting result, but is jointly determined by three types of criteria: structural response, observable evidence, and physical boundary; by compiling environmental data into state mapping constraints and participating in the dominant mechanism judgment, sub-model switching and scenario-based dynamic simulation have a unified entry point, thus forming a closed-loop implementation path of "event acquisition—state mapping—dual-track judgment—write-back control—scenario deduction—supplementary test scheduling".

[0039] In another implementation, the event scheduling unit detects that the background vibration of the roof layer is in a low-interference range during the 37th detection cycle, and the historical degradation file shows that the threshold drift coefficient of friction pendulum support No. 2 has been continuously increasing in the last three detection cycles. It then generates event number E20260415-FP2-037 and calls the perturbation execution unit to apply a low-amplitude displacement pulse with a peak value of 2.0 mm and a duration of 0.40 s to friction pendulum support No. 2. During this event, the data acquisition unit collects relative displacement of the upper and lower plates of the support, acceleration of adjacent layers, beam end strain, visual displacement, and environmental data. The time alignment unit completes timescale unification. The response signature generation unit finds that the vibration start-up delay, phase lag, and threshold crossing order have all shifted relative to the baseline event. Based on this, the state inversion unit updates the threshold drift coefficient to 0.14 and the boundary constraint relaxation factor to 0.09. The mechanistic twin model determines that the dominant mechanism of this support has shifted from elastic shear dominance to slip dominance. The shadow verification model synchronously predicts that the lap opening order has shifted forward and the residual error convergence time has increased. The consistency determination unit calculated a comprehensive consistency score of 0.82, and the parameter update result fell within the physical boundary range. The write-back control unit executed the write-back and switched the support sub-model. Subsequently, the scenario-based dynamic simulation unit output the critical response amplification corresponding to this support as 1.18 and 1.11 under continuous aftershock conditions and long-term small amplitude fatigue conditions, respectively. The safety margin decreased but was still higher than the preset lower limit. The parameter update result, scenario output result, and supplementary record of this event were written into the event processing record, providing a basis for subsequent evidence package generation and risk level determination.

[0040] Example 2 like Figure 1 and Figure 2As shown, this embodiment enhances the settings for disturbance scheduling boundaries, perturbation template configurations, and response signature association encoding methods for different key components. The detection objects still include seismic isolation bearings, viscous dampers, connecting components, and their corresponding measuring points. The event scheduling unit, perturbation execution unit, data acquisition unit, time alignment unit, response signature generation unit, state inversion unit, mechanistic twin model calculation unit, shadow verification model calculation unit, consistency determination unit, and write-back control unit continue to participate in the entire processing chain. However, this embodiment adjusts the containment constraints from a uniform threshold to a constraint table configured hierarchically according to component categories, and incorporates the perturbation template and response signature encoding rules into the scheduling and determination process.

[0041] The event scheduling unit pre-establishes a correspondence table between component categories and disturbance templates. For lead-core rubber seismic isolation bearings, the low-amplitude displacement pulse template is prioritized, with the default peak value of the displacement pulse set to 1.8 mm, an allowable adjustment range of 1.2 mm to 2.4 mm, and a duration of 0.35 s to 0.55 s. For friction pendulum bearings, the combined template of low-amplitude displacement pulse and frequency sweep excitation is prioritized, with the displacement pulse peak value set to 2.0 mm to 3.0 mm and the frequency sweep excitation frequency range set to 2 Hz to 8 Hz. For viscous dampers, the frequency sweep excitation template is prioritized, with a sweep duration of 6 s and a frequency change slope of 1 Hz / s. For connected components with locally achievable additional load conditions, the short-time additional mass switching template is prioritized, with the additional mass set to 1.5% to 2.5% of the equivalent mass of the corresponding component. The event scheduling unit selects the initial template based on the component category, the response confidence weight of the three most recent events, and the background vibration level, and writes the selected template, target component, preset peak value, preset duration, and termination conditions into the current round of scheduling records.

[0042] In this embodiment, the containment constraint uses a normalized risk index for unified assessment. For the first... For each disturbance event, the event scheduling unit calculates the displacement percentage, duration percentage, and local response percentage, and takes the maximum value among the three as the containment risk indicator. In the formula, For the first Safety risk indicators for secondary disturbance events. This refers to the target displacement or target excitation amplitude during this round of disturbance. This represents the maximum allowable value for this component category. For the duration of this round of disturbance, This is the maximum duration. This represents the peak value of the real-time local response. This represents the maximum allowed response level for this component category. When, the perturbation execution unit executes directly according to the preset template; when When this happens, the event scheduling unit automatically lowers the template amplitude by one level before execution; when When the current template is determined to be outside the safe shell constraint, the perturbation execution unit stops loading and reassigns the event to the natural perturbation event window monitoring.

[0043] To ensure a unified basis for switching perturbation templates across different components, the event scheduling unit further calculates the template matching value. Let the first... The baseline frequency corresponding to the target component category of the secondary disturbance event is The dominant frequency of the background vibration is The credibility weight of the response to the most recent valid event is . Then the template matching value is written as: In the formula, A smaller value indicates that the current template is more suitable for the key component. When, keep the original template; when When switching to an adjacent template within the same component category; When the active perturbation stops, the monitoring is switched to natural perturbation event window. With this setting, lead-core rubber isolation bearings will preferentially accept displacement pulse templates when the background vibration is stable, friction pendulum bearings will preferentially accept displacement pulse and frequency sweep combined templates when the background vibration is close to its slip identification frequency band, and viscous dampers will preferentially accept frequency sweep excitation templates when low displacement is achievable but frequency characteristics are obvious.

[0044] In this embodiment, the response signature generation unit no longer outputs only a single feature, but performs fixed-order association encoding on the time process quantities within the same disturbance event window. The inputs for association encoding are, in order: oscillation delay, propagation path difference, attenuation envelope coefficient, hysteresis closure error, residual recovery slope, environmental correction error, phase lag, and threshold crossing order. The response signature generation unit first arranges the input quantities according to a unified clock, then binds them according to the component number and event number, and writes them into the same signature record. To facilitate subsequent reproduction and determination, this embodiment uses a weighted association method to form an enhanced response signature value: In the formula, For the first The enhanced response signature value of the critical component corresponding to the secondary disturbance event. For the first Class-normalized time process quantity, The corresponding weights are as follows. In this embodiment, the weight for oscillation delay is 0.16, the weight for propagation path difference is 0.12, the weight for attenuation envelope coefficient is 0.10, the weight for hysteresis closure deviation is 0.14, the weight for residual recovery slope is 0.14, the weight for environmental correction deviation is 0.10, the weight for phase lag is 0.12, and the weight for threshold crossing order is 0.12. The sum of all weights is 1. In addition to recording the enhanced response signature value, the response signature table still retains all individual time process quantities. Subsequent reproduction determination will preferentially use the enhanced response signature value combined with the phase lag and threshold crossing order.

[0045] This embodiment further establishes an event validity determination chain to prevent events with insufficient sampling quality or insufficient disturbance strength from being mistakenly sent into the unified mapping state inversion. After the event ends, the data acquisition unit extracts the sampling integrity index, synchronization deviation, and minimum identification amplitude ratio from the sampling quality label. The time alignment unit sets the sampling integrity index as... Synchronization deviation is denoted as The minimum recognition amplitude percentage is denoted as The consistency determination unit then generates an event validity score. In the formula, For the first The event validity score of the secondary disturbance event. To allow for an upper limit on synchronization deviation, this implementation uses 5 ms. When , or If any one of the above conditions is met, the current event is directly determined to be an invalid event; if all of the above individual conditions are met and When the current event is determined to be a valid event and sent to the state inversion unit; when When the current event is marked as a boundary event, the event scheduling unit will perform an additional retest of the same key component using the same template in the next detection cycle.

[0046] To ensure the stability of subsequent risk escalation assessments, this embodiment fixes the fields directly involved in the reproduction assessment as the enhanced response signature value, phase hysteresis, and threshold crossing order. Only when the difference between the enhanced response signature value of a subsequent disturbance event and the enhanced response signature value of the most recent stable write-back event does not exceed a preset tolerance, and the phase hysteresis changes in the same direction and the threshold crossing order maintains the same sequential relationship, will the response signature generation unit mark the event as a valid reproduction event suitable for reproduction assessment. The preset tolerance is provided by the event scheduling unit during the cold start calibration phase; in this embodiment, it is taken as 1.15 times the standard deviation of the enhanced response signature values ​​of the most recent ten valid events.

[0047] The following describes an enhanced scheduling process. The event scheduling unit detected that the response confidence weight of viscous damper No. 3 was below 0.90 in the two most recent valid events, while the background vibration dominant frequency remained around 4.5 Hz. Therefore, a frequency sweep excitation template was selected first, with a preset sweep range of 2 Hz to 6 Hz and a sweep duration of 6 s. Before the perturbation was executed, the event scheduling unit calculated the containment risk index. Template matching value The conditions for direct execution are met. After the event ends, the data acquisition unit outputs a sampling integrity index of 0.98, a synchronization deviation of 2 ms, a minimum identification amplitude ratio of 0.91, and an event validity score of 0.93. Therefore, the event is sent to the state inversion unit as a valid event. The response signature generation unit simultaneously obtains three reproduction fields: enhanced response signature value, phase lag, and threshold crossing order, and writes them into the response signature table. This setting ensures that subsequent unified mapping state quantity inversion and reproduction determination no longer rely on a single peak quantity, but rather on the associated time process quantity under the same event window, thereby improving the consistency of disturbance event scheduling and signature reproduction under different component categories.

[0048] Example 3 like Figures 1 to 3 As shown, this embodiment, based on Embodiments 1 and 2, enhances the fusion and inversion of the unified mapping state quantity, the switching of the dominant mechanism driven by environmental constraints, the dual-track consistency adjudication, the evidence package generation, and the risk level enhancement closed loop. The event scheduling unit, data acquisition unit, time alignment unit, response signature generation unit, state inversion unit, mechanism twin model operation unit, shadow verification model operation unit, consistency judgment unit, write-back control unit, and scenario-based dynamic simulation unit continue to use the role divisions of Embodiments 1 and 2. The unified mapping state quantity is still defined as... , respectively, represent the equivalent energy dissipation gradient, recovery stiffness attenuation coefficient, threshold drift coefficient, boundary constraint relaxation factor, environmental sensitivity amplification, and response confidence weight. This embodiment does not change the aforementioned object names and event numbering rules, but only further tightens the local evidence, environmental level, update confidence interval, evidence package field, and risk escalation triggering conditions for entering the unified mapping state quantity.

[0049] To ensure the reproducibility of the fusion inversion, environmental levels, and updated confidence intervals, this embodiment employs a unified anchoring rule. Structural response prediction mapping. The mechanistic twin model calibrated in Example 1 was obtained by least-squares fitting on baseline and primary environmental samples; observable evidence predicts mapping. The shadow calibration model is synchronously calibrated on the same batch of samples according to the corresponding event numbers. Environmental normalization quantity. , , and The deviation of the current measured value from the median reference value of valid event samples that did not trigger freezing diversion during the past thirty days of stable operation is normalized to the respective allowable upper limit of engineering fluctuation, and truncated to between 0 and 1; freeze-thaw or additional foundation settlement markers. The value is either 0 or 1 depending on whether a preset trigger condition is met. Sensitivity matrix. This is achieved by applying differential perturbations of no more than 1% of the reference value to each of the six state fields in the current stable version, and recording the increment of each local evidence response. The termination condition for the three rounds of iterations is that the relative changes of the candidate update values ​​in the six fields in two adjacent rounds are no higher than 3%; if this condition is not met by the third round, convergence is determined according to the 8% criterion described later. The reasons for freezing and diverting are recorded in the order of high residual, low determinism, physical boundary violation, and failure of dual-track consistency, for subsequent supplementary testing scheduling and evidence package archiving.

[0050] After each valid disturbance event, the state inversion unit first organizes visual displacement, acceleration, strain, temperature and humidity, corrosion potential, electromagnetic detection characteristics, and acoustic non-destructive testing characteristics into a local evidence set within the same event window. Among them, visual displacement, acceleration, and strain constitute the structural response evidence vector. The observable evidence vector consists of temperature and humidity corrected displacement deviation, corrosion potential deviation, electromagnetic detection characteristic quantity, acoustic echo characteristic quantity, phase hysteresis, and threshold crossing order. The state inversion unit generates a diagonal weight matrix based on the sampling integrity index, response confidence weight, and environmental level in the sampling quality label. , and And calculate the first one according to the following formula. The fusion inversion results corresponding to the perturbation event: In the formula, For the first Candidate update values ​​for the unified mapping state variables of the second perturbation event. This represents the structural response prediction mapping given by the mechanistic twin model. This represents the predictive mapping of observable evidence given by the shadow verification model. This is the unified mapping state quantity corresponding to the previous stable version. The state inversion unit adopts a three-round iterative solution. The first round uses only the structural response evidence vector, the second round introduces the observable evidence vector, and the third round introduces the prior constraint vector. If the difference between the candidate update values ​​obtained in the second and third rounds in the six fields does not exceed 8% of the reference upper limit of the corresponding field, the result of the third round is determined as the result of the fusion inversion in this round. Otherwise, the event in this round is marked as an inversion non-convergence event and directly enters the freeze split.

[0051] To ensure that the fused inversion results have an executable threshold, the state inversion unit further generates state variable inversion residuals. The residual is composed of the structural response evidence residual, the observable evidence residual, and the state transitions between two adjacent events. The calculation formula is as follows: In the formula, A smaller value indicates a greater consistency between the fusion inversion result and the current evidence and the previous stable version. In this embodiment, When the value is not higher than 0.26, the fusion inversion results enter the dual-track consistency judgment. When the value is higher than 0.26 but not higher than 0.34, the state inversion unit will respond with the confidence weight. Downgraded to The event will be reclassified as a boundary event and then proceed to the consensus adjudication process. When the value is higher than 0.34, the current event is directly identified as a high residual event, the write-back control unit freezes the current write-back, and at the same time generates the supplementary test scheduling result for the target component.

[0052] After obtaining the fused inversion results, the state inversion unit simultaneously generates update confidence intervals for each of the six state fields. To this end, the state inversion unit forms a sensitivity matrix based on the sensitivity of the unified mapping state quantities to the local evidence from this round. The updated covariance matrix is ​​obtained by the following formula. and update confidence interval : In the formula, The stabilization coefficient is set to 0.08 in this embodiment. Take 1.96. The update confidence interval half-width corresponds to the six state fields. Write back to the control unit and perform upper limit determination on each of the six fields: equivalent energy consumption gradient. The confidence interval half-width is no higher than 0.18, and the recovery stiffness attenuation coefficient is... The confidence interval half-width is no higher than 0.15, and the threshold drift coefficient is... The confidence interval half-width is no higher than 0.05, and the boundary constraint relaxation factor... The confidence interval half-width is no higher than 0.05, and the environmentally sensitive amplification is significant. The confidence interval half-width is no higher than 0.12, and the response confidence weight is... The confidence interval half-width is no higher than 0.10. If any field exceeds the corresponding upper limit, the event in this round is reclassified as a low-determinism event, the write-back control unit is frozen, and the supplementary test scheduling results are output.

[0053] In this embodiment, environmental data is no longer used solely as a correction factor in feature extraction; instead, environmental levels are generated by the environmental constraint compilation unit. This information is then written into the state mapping constraint table. The environmental level is calculated jointly by temperature deviation, humidity deviation, corrosion potential deviation, high-frequency background vibration, and freeze-thaw or additional foundation settlement markers. In the formula, This is the normalized temperature deviation. This is the normalized humidity deviation. This is the normalized corrosion potential deviation. To normalize the high-frequency background vibration, Additional settlement markers for freeze-thaw cycles or foundations. Environmental Class Environments with a value below 0.30 are classified as Level 1, 0.30 to 0.55 as Level 2, 0.55 to 0.75 as Level 3, and above 0.75 as Level 4. The environment constraint compilation unit maps Level 1 environments to the baseline constraint set and Level 2 environments to the relaxed environment sensitivity amplification. The constraint set for updating the step size corresponds to simultaneously tightening the threshold drift coefficient across the three-level environments. and boundary constraint relaxation factor The upper limit constraint set corresponds to the Level 4 environment as a constraint set that prohibits active perturbation and only allows natural perturbation event windows to enter the processing chain.

[0054] In this embodiment, the mechanism twin model computation unit, in addition to outputting the structural response reconstruction results, parameter update results, and dominant mechanism determination results, also explicitly outputs equivalent stiffness, equivalent damping, slip threshold, yield range, and damage evolution parameters. The shadow verification model computation unit explicitly outputs four types of observable evidence prediction results: support loop opening order, damper phase difference change direction, measurement point time delay propagation relationship, and residual error convergence pattern. The consistency determination unit still generates a dual-track consistency score based on the structural response consistency score, evidence consistency score, and boundary validity score from Embodiment 1. In this embodiment, a dominant mechanism switching determination is added. If two consecutive valid events simultaneously satisfy... Not less than 0.12 The environmental level is not less than 0.08, the opening sequence of the support loop precedes the change in the phase difference of the damper, and the environmental level is... If the coefficient is not less than 0.55, the dominant mechanism switches from elastic shear to slip; if two consecutive valid events simultaneously satisfy the recovery stiffness attenuation coefficient... Not higher than 0.88, the phase difference of the damper is reversed, the strain threshold crossing sequence of the connecting components occurs before the displacement threshold crossing sequence of the damper, and the environmental level is... If the value is not lower than 0.55, the dominant mechanism will switch from viscous energy dissipation to connection relaxation. Once the dominant mechanism switches, the write-back control unit will immediately and synchronously switch the corresponding twin model and the set of boundary conditions for the scenario-based dynamic simulation.

[0055] The evidence package generation unit generates an evidence package for the current event after each round of scenario-based dynamic simulation. The evidence package is generated after the dual-track consistency adjudication ends and the scenario-based dynamic simulation is completed; both write-back and freeze routing generate evidence packages. The evidence package header is fixed and includes the disturbance source, sampling integrity, environment level, state variable inversion residual, dual-track consistency score, update confidence interval, dominant mechanism, write-back flag, scenario-based dynamic simulation output summary, and current risk level. The sampling integrity is derived from the sampling integrity index in the sampling quality label. The evidence package is written to the event archive table according to the event number and to the version traceability table according to the digital twin version number, for subsequent risk level determination and operation and maintenance terminal access. If the current event enters the freeze routing, the write-back flag in the evidence package is marked as frozen, and the scenario-based dynamic simulation output summary is rewritten as the simulation output of the stable version before freezing under the current environment constraints.

[0056] In this embodiment, the risk level upgrade is triggered only when both the "persistent vulnerability trend" and the "response signature corresponding to the recurrence of subsequent disturbance events" are met simultaneously. To this end, the risk assessment unit arranges the scenario-based dynamic simulation results that have been written back in the most recent three rounds by event number; when any of the five scenarios experiences at least two instances within three consecutive rounds where the safety margin decreases by no less than 12% and the remaining serviceable time window shrinks by no less than 15%, a persistent vulnerability trend assessment is recorded. The value is 1 if the event is not specified, otherwise it is 0. The determination of subsequent disturbance event recurrence follows the enhanced response signature criterion in Example 2: when the difference in the enhanced response signature of subsequent events falls within the preset tolerance, the direction of phase hysteresis changes is consistent, and the threshold crossing order remains consistent, the recurrence determination value is recorded. The value is 1 if it is not specified, otherwise it is 0. The rule for increasing the risk level is written as follows: In the formula, This indicates that the risk level of the target component has been raised by one level. This indicates that the current risk level will be maintained. If the risk level is raised, the risk output unit will simultaneously perform three actions: push maintenance suggestions for the target component to the maintenance terminal, issue a high-frequency detection plan to the event scheduling unit, and issue restriction instructions for the corresponding operating conditions to the operation control terminal; if If so, only the risk profile will be updated, and no restriction instructions will be triggered.

[0057] The following is a high-level closed-loop operation simulation. During the 11th detection cycle, the event scheduling unit detects that the No. 2 friction pendulum support has experienced a sudden change in diurnal temperature variation. The environmental constraint compilation unit then generates an environmental level. The component was then incorporated into a Level 3 environmental constraint set. The perturbation execution unit subsequently applied a low-amplitude displacement pulse with a peak value of 2.2 mm and a duration of 0.42 s to the friction pendulum support No. 2. The data acquisition unit completed the acquisition of visual displacement, acceleration, strain, temperature and humidity, corrosion potential, electromagnetic detection features, and acoustic echo features, while the time alignment unit formed a unified event window. The state inversion unit obtained candidate update values ​​based on the fusion inversion objective function, where the threshold drift coefficient... The boundary constraint relaxation factor increased to 0.13. The residual value increased to 0.09, indicating that the state variable was inverted. The confidence interval half-width for all six fields is 0.24, and the half-width of the update confidence interval falls within the corresponding upper limit. The mechanistic twin model operation unit outputs the determination result of the shift in the dominant mechanism from elastic shear to slip, while the shadow verification model operation unit simultaneously outputs the observable evidence prediction results of the forward shift in the support loop opening order and the slower convergence of residual errors. The consistency determination unit calculates the dual-track consistency score. The value is 0.81, and the write-back control unit executes the write-back and switches the support sub-model accordingly. The scenario-based dynamic simulation unit then outputs the safety margin changes and remaining serviceable time window under representative seismic wave conditions, temperature abrupt change conditions, continuous aftershock conditions, long-term small-amplitude fatigue conditions, and component local failure conditions. The evidence package generation unit generates an evidence package according to the event number and writes it into the event archive table. The scenario-based dynamic simulation results of the subsequent two rounds of write-back events continue to show that the support maintains a decreasing safety margin and a shrinking remaining serviceable time window under continuous aftershock conditions and long-term small-amplitude fatigue conditions. Therefore, the risk assessment unit will continuously determine the vulnerability trend. Set to 1. In the next round of natural disturbance event window, the enhanced response signature difference falls within the preset tolerance, the phase lag direction is consistent, and the threshold crossing order remains consistent, thus reproducing the judgment quantity. When the risk level is set to 1, the risk level increase rule is triggered. The risk output unit raises the risk level of friction pendulum support No. 2 by one level and simultaneously pushes maintenance suggestions, high-frequency testing plans and operating condition limit instructions.

[0058] Example 4 like Figures 1 to 4As shown, this embodiment specifically explains the execution carrier that initiates the methods and actions in Embodiments 1 to 3, the detection objects they affect, the acquisition chain through which they enter the computation processing chain, and the carrier that completes write-back, freeze, sub-model switching, and risk output. The detection objects remain in-service buildings equipped with seismic isolation bearings, damping components, connecting components, and key measuring points in the superstructure. The field equipment layer includes an event scheduling unit, a perturbation execution unit, a data acquisition unit, and field safety interlocking components; the computer room side includes a computation processing unit and a twin management unit; and the operation and maintenance side includes a risk output unit and an operation and maintenance terminal. All the above units work collaboratively according to event numbers, without changing the definitions of perturbation events, response signatures, unified mapping state quantities, state mapping constraints, evidence packages, and risk levels in Embodiments 1 to 3.

[0059] The event scheduling unit is located in the field control cabinet and consists of an industrial controller, an isolated power supply, and a hard-wired shutdown loop. The industrial controller pre-stores a containment constraint table, a disturbance template table, a component number table, and the archive of the last ten rounds of stable events, and maintains bidirectional communication with the twin management unit. Before each round of testing begins, the event scheduling unit first reads the testing plan issued by the maintenance terminal, the previous stable version number returned by the twin management unit, the current risk level returned by the risk output unit, and the environmental constraint compilation results. When the environmental level is level four, the current component risk level is at the highest monitoring level, the field shutdown loop closure is abnormal, any actuator self-test fails, or any key measurement point has an abnormal online status, the event scheduling unit prohibits active perturbation and only allows natural disturbance event windows to enter the processing chain; only when none of the above prohibition conditions are triggered does the event scheduling unit issue the target component, disturbance template, preset peak value, duration, stop threshold, and event number for this round to the perturbation execution unit.

[0060] The perturbation actuators are respectively configured for lead-core rubber seismic isolation bearings, friction pendulum bearings, viscous dampers, and locally accessible connecting components. For seismic isolation bearings, the perturbation actuators use electro-hydraulic displacement actuators installed near the maintenance position of the bearing. The displacement actuator contacts the maintenance connection of the target bearing through a detachable mechanical clamp. The maximum thrust is set to 15 kN according to the engineering default value, and the upper limit of displacement stroke is set to 5 mm. Before execution, the physical limit is limited to within 1.25 times the scheduled value by a mechanical limit block. For viscous dampers, the perturbation actuators use electromagnetic exciters and dedicated clamping seats to perform frequency sweep excitation on the outside of the damper lugs. The upper limit of the exciter's rated output is set to 2 kN, and the upper limit of continuous working time is set to 8 s. For locally accessible connecting components, the perturbation actuators use short-time additional mass switching components. The additional mass block is released and retrieved through an electromagnetic lock. The upper limit of the additional mass is configured at 2.5% of the equivalent mass of the corresponding component. After receiving the instruction from the event scheduling unit, the perturbation execution unit first completes the zero-position correction, fixture lock confirmation, and actuator no-load test operation before entering the formal loading. If the zero-position deviation exceeds 0.2 mm, the fixture lock signal is missing, or the no-load current exceeds 15% of the rated value, the active perturbation of this round is terminated and the monitoring of the natural disturbance event window is switched.

[0061] The data acquisition unit, located within the field acquisition box, comprises a displacement acquisition card, an acceleration acquisition card, a strain acquisition card, a visual displacement acquisition terminal, an environmental acquisition terminal, a corrosion potential acquisition terminal, an electromagnetic detection front-end, and an acoustic non-destructive testing front-end. Each acquisition front-end receives the same master clock pulse via a unified clock synchronizer. The synchronous sampling frequency for displacement, acceleration, and strain signals is maintained at 200 Hz, the visual displacement sampling frequency at 30 Hz, and the environmental quantity refresh cycle at 1 s. The data acquisition unit enters a pre-triggered recording state 1 s before the event begins and continues recording with an 8-slow delay after the event ends to cover any remaining recovery trajectories. The data acquisition unit generates real-time online status, timestamp integrity, signal saturation markers, visual occlusion markers, and abnormal jump markers for each sampling channel, and packages the original sampling records and quality labels together before sending them to the computational processing unit. If any critical channel fails continuously for more than 0.15 s, the data acquisition unit immediately sends a quality anomaly signal to the event scheduling unit and the computational processing unit, and the current event is automatically reclassified as a retest event.

[0062] The computational processing unit (CFD) is located within the industrial computer in the data center and is responsible for time alignment, response signature generation, unified mapping state variable inversion, dual-track consistency judgment, and scenario-based dynamic simulation. After receiving a data packet with an event number from the data acquisition unit, the CCD first forms a unified event window according to the time alignment rules of Examples 1 and 2, and then completes the calculation of response signature, unified mapping state variable, state variable inversion residual, environmental level, update confidence interval, and dual-track consistency score according to the criteria of Examples 1 to 3. If the CCD determines that the current event is a high residual event, a low deterministic event, or a physical boundary violation event, the CCD only sends a freeze command to the twin management unit and a supplementary test scheduling suggestion to the event scheduling unit; if the CCD determines that the current event meets the write-back conditions, it sends the parameter update result, the dominant mechanism judgment result, the state mapping constraints, and the scenario-based dynamic simulation input conditions to the twin management unit.

[0063] The twin management unit is located on a dedicated database server and is used to store the baseline version, the current stable version, candidate update versions, the event archive table, the evidence package table, and the version traceability table. Upon receiving a write-back command from the computational processing unit, the twin management unit does not directly overwrite the current stable version. Instead, it first generates a candidate update version, writes the parameter update results, dominant mechanism, environment level, update confidence interval, and event number into the candidate update record, and then performs a version consistency check. The version consistency check covers at least three items: first, the event number referenced by the candidate update version must be consistent with the event number in the current round's response signature, evidence package, and scenario-based dynamic simulation results; second, the timestamp of the candidate update version must be later than the timestamp of the current stable version; third, the dominant mechanism corresponding to the candidate update version must be consistent with the switched twin sub-model type. If all three checks pass, the twin management unit upgrades the candidate update version to the current stable version; if any check fails, the twin management unit deletes the candidate update version while retaining the original stable version, and simultaneously writes the rollback record into the version traceability table.

[0064] The risk output unit, located within the operations and maintenance server, is responsible for receiving evidence packages, risk levels, scenario-based dynamic simulation summaries, and supplementary testing scheduling results. It then pushes corresponding actions to the operations and maintenance terminal, the event scheduling unit, and the operation control terminal. When the current event enters a freeze-through phase, the risk output unit pushes the supplementary testing reason, target component, suggested supplementary testing time, and suggested disturbance template to the operations and maintenance terminal; it pushes the supplementary testing plan to the event scheduling unit; and it only pushes monitoring prompts to the operation control terminal without changing the operating conditions. When the current event triggers a risk level upgrade, the risk output unit pushes maintenance suggestions and a checklist to the operations and maintenance terminal; it issues a high-frequency detection plan to the event scheduling unit; and it issues a restriction command for the corresponding component's operating condition to the operation control terminal. Upon receiving the restriction command, the operation control terminal adds the restricted operating condition to the prohibited list outside the field control whitelist and returns an execution confirmation signal to the risk output unit and the event scheduling unit.

[0065] The field safety interlock component runs through the event scheduling unit, the perturbation execution unit, and the data acquisition unit, and is used to realize the stop, freeze, and restore action chain. The field safety interlock component receives at least the actuator overcurrent signal, displacement over-limit signal, clamp detachment signal, critical measuring point disconnection signal, and control cabinet emergency stop signal. After any signal is triggered, the field safety interlock component performs the following actions: immediately cuts off the power output of the perturbation execution unit; retracts the actuator to the mechanical zero position and locks it; marks the current event as a forced termination event; sends the forced termination signal to the event scheduling unit, the data acquisition unit, the computing and processing unit, and the twin management unit; and retains the sampling records for 2 seconds before termination and 8 seconds after termination. Only when the event scheduling unit completes fault reset confirmation, the perturbation execution unit completes zero position return confirmation, the data acquisition unit completes critical channel online confirmation, and the twin management unit confirms that the current stable version has not been rewritten, can the next round of active perturbation be reactivated.

[0066] This embodiment also clarifies the evidence package invocation chain and the risk level invocation chain. After the core fields of the evidence package are generated by the calculation and processing unit, they are written to the twin management unit. The twin management unit then adds the version number, write-back flag, and archiving time before pushing it to the risk output unit. The risk output unit is not allowed to modify the evidence package fields; it is only permitted to add sending records, reading records, and processing status records. The risk level is adjusted upwards or maintained by the risk output unit according to the criteria in Embodiment 3. The event scheduling unit only reads the risk level result and does not recalculate the risk level. The operation control terminal only executes the restriction instructions issued by the risk output unit and does not participate in the risk level determination. Through the above settings, the risk level determination subject, execution subject, and feedback subject are fixedly separated, thereby avoiding the repeated triggering of the same action on both the field and the computer room sides.

[0067] In one implementation, at the start of the 13th detection cycle, the event scheduling unit reads that the current risk level of the No. 2 friction pendulum support is medium monitoring level, the environmental level is level 3, the previous stable version number is V017, and the on-site shutdown loop, actuator self-test, and key measuring point online status are all normal. Therefore, it sends event number E20260415-FP2-052, a low-amplitude displacement pulse template, a peak value of 2.0 mm, a duration of 0.40 s, and a stop threshold to the perturbation execution unit. After completing zero-position correction and fixture locking confirmation, the perturbation execution unit executes the perturbation. The data acquisition unit simultaneously acquires the structural response, environmental quantities, and mass tags, and sends the event data packet to the computation processing unit. After completing time alignment, response signature generation, unified mapping state quantity inversion, and dual-track consistency judgment, the computation processing unit determines that the current event meets the write-back conditions and generates a result showing that the dominant mechanism remains unchanged due to slip dominance, as well as a scenario-based dynamic simulation summary. The twin management unit generates candidate update versions based on this information. After completing checks on event number consistency, timestamp order, and sub-model type consistency, it upgrades the candidate update versions to the current stable version V018 and generates a corresponding evidence package. Upon receiving the evidence package, the risk output unit confirms that the risk level for this round remains unchanged and only pushes the update completion record to the operations and maintenance terminal. If a clamp detachment signal occurs in the perturbation execution unit during the same round, the field safety interlock component will immediately cut off power output, reclaim the actuator, freeze the current event, and retain sampling records before and after termination. The calculation and processing unit only outputs the freeze result and retesting suggestions. The twin management unit maintains the stable version V017, and the risk output unit pushes the forced termination record and retesting plan to the operations and maintenance terminal, thus forming a complete stop, freeze, rollback, and recovery closed loop.

[0068] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be modified within the scope of the concept described herein by means of the above teachings or the technology or knowledge in related fields.

Claims

1. A digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection, characterized in that, include: S1. Under the premise of satisfying the containment constraints, based on the current environmental conditions, structural operating status and historical degradation records, the disturbance events are obtained. The disturbance events are taken from any event in the window of the limited micro-perturbation loading trigger event and the natural disturbance event during the structural operation process, and multi-point response data, environmental data and sampling quality labels corresponding to the disturbance events are collected. S2. Time-align the multi-point response data, the environmental data, and the sampling quality labels, and extract the oscillation delay, propagation path difference, attenuation envelope, hysteresis closure deviation, residual recovery trajectory, environmental correction deviation, phase lag, and threshold crossing order from the aligned event data to generate a response signature for the same key component; S3. Convert the data corresponding to visual displacement, acceleration, strain, temperature and humidity, corrosion, electromagnetic detection, and acoustic non-destructive testing into local evidence, and perform joint inversion of the unified mapping state variables based on the response signature. The unified mapping state variables include at least the equivalent energy dissipation gradient, recovery stiffness attenuation coefficient, threshold drift coefficient, boundary constraint relaxation factor, environmental sensitivity amplification, and response confidence weight. S4. Simultaneously input the unified mapping state variables into the mechanistic twin model and the shadow verification model. The mechanistic twin model outputs the structural response reconstruction result, parameter update result, and dominant mechanism determination result. The shadow verification model outputs the observable evidence prediction results corresponding to the support loop opening order, damper phase difference change direction, measurement point time delay propagation relationship, and residual error convergence form after environmental switching. S5. Compile the environmental data into state mapping constraints, and perform consistency checks on whether the deviation between the structural response reconstruction result and the measured response, the deviation between the observable evidence prediction result and the measured evidence, and whether the parameter update result falls within the physical boundary range. The consistency determination process involves writing the parameter update results back to the digital twin when the consistency determination passes. When the dominant mechanism determination result indicates a switch in the dominant mechanism, the corresponding twin sub-model and simulation boundary conditions are switched according to the state mapping constraints. Based on the written-back digital twin, scenario-based dynamic simulations covering representative seismic wave conditions, temperature change conditions, continuous aftershock conditions, long-term small amplitude fatigue conditions, and component local failure conditions are performed, and the collaborative failure sequence of isolation components and damping components, key response amplification range, safety margin changes, and remaining serviceable time windows are output. If the consistency determination fails, the current write-back is frozen and supplementary measurement scheduling results are generated.

2. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, The containment constraints in S1 are used to limit the disturbance loading amplitude, disturbance duration, disturbance location, and disturbance termination conditions.

3. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, The limited perturbation loading in S1 is selected from at least one of low-amplitude displacement pulse, frequency sweep excitation, short-time additional mass switching, and local electromagnetic excitation.

4. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, When generating the response signature in S2, the oscillation delay, the transmission path difference, the attenuation envelope, the hysteresis closure deviation, the residual recovery trajectory, the environmental correction deviation, the phase hysteresis, and the threshold crossing order are associated and encoded according to the same disturbance event window.

5. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, When performing joint inversion on the unified mapping state quantity in S3, the local evidence corresponding to the unified mapping state quantity is fused and inverted according to the same disturbance event window, so that diagnosis, simulation, early warning and lifetime extrapolation all depend on the unified mapping state quantity.

6. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, The mechanistic twin model in S4 also outputs at least one of the following: equivalent stiffness, equivalent damping, slip threshold, yield range, and damage evolution parameters. The shadow verification model synchronously updates the prediction results of the observable evidence based on the unified mapping state quantity.

7. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, The consistency check in S5 is performed only when the comparisons of the deviation between the reconstructed structural response and the measured response, the deviation between the predicted observable evidence and the measured evidence, and whether the parameter update result falls within the physical boundary range all pass.

8. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, The environmental effects corresponding to the state mapping constraints include at least temperature, humidity, corrosion, high-frequency background vibration, freeze-thaw conditions, and additional foundation settlement. The criteria for determining the switching of the dominant mechanism are jointly determined by the state mapping constraints and the parameter update results.

9. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 1, characterized in that, An evidence package is generated based on the scenario-based dynamic simulation results. The evidence package includes at least the source of disturbance, sampling integrity, environmental level, state variable inversion residual, dual-track consistency score, and update confidence interval.

10. The digital twin mapping and dynamic simulation method for seismic isolation and vibration reduction detection according to claim 9, characterized in that, The risk level of a critical component is increased only when the same critical component continuously exhibits a vulnerability trend in scenario-based dynamic simulation and the corresponding response signature is reproduced in subsequent disturbance events.

Images