A Multipath Degradation Classification and Reinforcement Decision System and Method for Rock Climbing Crane Beams in Pumped Storage Power Stations

By collecting multi-source response data through sensors and processing the data, degradation classification results and reinforcement decision schemes for rock wall crane beams are generated. This solves the problem of difficulty in distinguishing multi-path degradation of rock wall crane beams, and enables accurate reinforcement decisions and reduces the mismatch of measures.

CN123068933APending Publication Date: 2026-11-10中国水利水电建设工程咨询北京有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611548300.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-08
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing technologies struggle to distinguish the multipath causes of degradation in the rock wall crane beams of pumped storage power stations, and lack effective reinforcement decision-making methods, leading to inaccurate and repetitive reinforcement measures.

Method used

Multi-source response data is collected by sensors, and the data processing module performs identification screening and continuous parameter inversion to generate degradation classification results and output reinforcement decision schemes, including degradation indicators for anchorage, contact and rock platform paths. The bridge crane itself forms differentiated detection excitation to reduce multiple solutions and path misjudgment.

Benefits of technology

It enables accurate classification and reinforcement decisions for multi-path degradation of rock wall crane beams, reduces mismatch of measures and over-reinforcement, and improves the accuracy and efficiency of reinforcement.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123068933A_ABST
    Figure CN123068933A_ABST
Patent Text Reader

Abstract

This invention discloses a multi-path degradation classification and reinforcement decision-making system and method for a rock-faced crane beam in a pumped-storage power station. The system includes: a data acquisition interface for collecting multi-source response data on wheel pressure, traction force, position, beam body, anchorage system, interface, and rock platform under combined inspection loads using sensors; a data processing module for performing identifiability screening, state branch identification, and continuous parameter inversion on the multi-source response data to generate degradation classification results and reinforcement decision schemes for anchorage, contact, and rock platform paths; and a result output interface for outputting the degradation classification results and reinforcement decision schemes. This invention utilizes the crane itself to generate differentiated in-service inspection stimuli and first verifies identifiability within a fixed interface state branch before inverting continuous parameters, reducing multiple solutions and path misjudgments. Furthermore, through quantitative mapping of physical parameters to reinforcement measures and phased retesting and updating, it reduces measure mismatch, redundant reinforcement, and over-reinforcement.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of safety monitoring and reinforcement decision-making technology for underground powerhouse structures of pumped storage power stations, specifically relating to a multi-path degradation classification and reinforcement decision-making system and method for rock wall crane beams in pumped storage power stations. Background Technology

[0002] In pumped-storage powerhouses, the crane beams on the rock face of the underground powerhouse transmit lifting loads through the beam body, prestressed anchor cables, the beam-rock interface, and the rock platform. Long-term wheel pressure, traction, surrounding rock creep, and environmental factors can cause anchor cable prestress loss, anchor bond attenuation, interface slippage or opening, and rock platform softening. Degradation along different load-bearing paths may produce similar strain or displacement anomalies, but the required reinforcement measures will differ.

[0003] Existing methods such as visual inspection, beam deformation monitoring, and single anchor cable force measurement are insufficient to distinguish the source of anomalies. Conventional moving load identification also fails to simultaneously address issues such as double-sided wheel pressure, trolley eccentricity, reverse traction, and the identifiability of multiple parameters. Existing maintenance decisions are mostly based on comprehensive scoring, lacking a closed loop that directly maps physical parameters to candidate measures and updates them after reinforcing along different paths. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of difficulty in distinguishing multipath degradation of rock wall crane beams, lack of rigor in the mixed inversion of discrete interface state and continuous parameters, possible mis-involvement of low-sensitivity parameters in the classification, and lack of quantitative comparison and feedback updates for candidate schemes.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A multi-path degradation classification and reinforcement decision system for rock wall crane beams in pumped storage power stations includes: The data acquisition interface is used to collect multi-source response data of wheel pressure, traction force, position, beam, anchoring system, interface and rock platform under combined inspection loads using sensors; The data processing module is used to perform identifiability filtering, state branch identification, and continuous parameter inversion on the multi-source response data to generate degradation classification results and reinforcement decision schemes for anchoring, contact, and rock platform paths. The result output interface is used to output the degradation classification results and reinforcement decision scheme.

[0006] Preferably, the sensors include wheel pressure sensors, longitudinal traction force detection devices, trolley and carriage position encoders, beam strain or displacement sensors, anchor cable axial force or anchor head displacement sensors, beam-rock interface normal opening and tangential slip sensors, and rock platform surface and deep displacement sensors.

[0007] Preferably, the data processing module includes: The benchmark and load configuration unit is used to establish a benchmark parameter field and configure at least two trolley lateral positions, two levels of inspection loads, and forward and reverse controlled longitudinal traction conditions. The multi-source response processing unit is used to correct, spatially align, and normalize the collected multi-source response data based on the reference parameter field and load configuration, thereby forming symmetric and antisymmetric response features. The identifiability analysis unit is used to construct a sensitivity matrix based on the symmetric and antisymmetric response characteristics of each candidate interface state branch under the sampled parameter states of closed bonding, closed micro-slip, complete slip and complete opening. It also performs effective rank, minimum singular value and condition number tests on the selected load-measuring point combination. After the test meets the preset requirements, it enters the parameter inversion of the corresponding branch. The multi-path parameter inversion unit is used to identify the discrete state of the interface and invert the remaining prestress of the anchor cable, the bond stiffness of the anchor section, the interface parameters and the flexibility of the rock platform within the fixed state branch. The degradation classification unit is used to calculate the degradation indices of anchorage, contact and rock platform paths based on the continuous parameters and discrete state of the interface obtained by inversion, and to determine the normal, tracking observation, degradation action or pending review status according to the threshold. The reinforcement decision unit is used to convert candidate reinforcement schemes into coupled model parameters based on the degradation indexes of anchorage, contact and rock platform paths output by the degradation classification unit and the effective parameters output by the multi-path parameter inversion unit, and output a recommended scheme after safety constraint check and comprehensive evaluation. The retest feedback and decision update unit is used to retest the same trajectory and update the reinforcement scheme after path reinforcement. Automatic decision-making stops when there is no feasible scheme or continuous updates are ineffective.

[0008] Preferably, in the reference and load configuration unit: The reference parameter field is either a measured reference parameter field or a reconstructed reference parameter field; the measured reference parameter field is established based on the completion acceptance, anchor cable tensioning, contact quality detection and initial bridge crane operation data; the reconstructed reference parameter field is constrained by at least two independent information sources that cannot be derived from the same original data, of which at least one is on-site measurement or detection information, and the other is design and construction records, response of the opposite crane beam or calibration coupling model information, and satisfies that the normalized reference back substitution error is not greater than a preset upper limit; In the two-level inspection load, the difference between the second inspection load and the first inspection load is not less than the preset minimum load difference. The absolute deviation of the positive and negative traction forces shall not exceed the preset tolerance, and all working conditions shall meet the safety limits for beam stress, anchor cable axial force, interface opening and closing, rock platform displacement and bridge crane operation.

[0009] Preferably, in the multi-source response processing unit, the process of forming symmetric and antisymmetric response features includes: For the same load level and trolley lateral position under the controlled traction conditions in both the forward and reverse directions, forward and reverse response data were collected after corrections for zero point, temperature, slow variable terms and bridge torsion. Based on the measured wheel pressure resultant force in the forward and reverse directions of the test section, the wheel pressure normalization process is performed on the forward and reverse responses respectively to obtain the forward and reverse wheel pressure normalized responses; Based on the measured forward and reverse traction forces of the test section, the forward and reverse responses are normalized to obtain the forward and reverse traction normalized responses. The average of the normalized response of the positive wheel pressure and the normalized response of the negative wheel pressure at the same measuring point under the same working condition is taken as the symmetrical response feature of the measuring point. The symmetrical response feature characterizes the overall support flexibility and the change of anchoring force. The difference between the normalized response of the forward traction and the normalized response of the reverse traction at the same measuring point under the same working condition is taken as the antisymmetric response feature of the measuring point. The antisymmetric response feature characterizes the tangential stick-slip response of the beam-rock interface.

[0010] Preferably, the identifiability analysis unit includes: The prior parameter space sampling unit is used to form a finite set of sampling parameters within the prior parameter range of each candidate interface state branch. The sensitivity matrix construction subunit is used to construct a continuous parameter sensitivity matrix and perform singular value decomposition for each sampling parameter state. The identifiability determination sub-unit is used to verify whether the selected load-measuring point combination simultaneously meets the requirements of effective rank, minimum singular value and condition number under all sampling conditions. The working condition supplement subunit is used to supplement the trolley position, load level, traction working condition or measuring point according to the minimum singular value gain when the judgment is not met. The iteration termination determination subunit is used to output a pending review status when the preset number of adjustments is not met.

[0011] Preferably, the multi-path parameter inversion unit includes: The interface state recognition subunit is used to determine the current discrete interface state of each detection segment according to the priority order of fully open, fully slip, closed micro-slip, and closed adhesion. The historical state marking subunit is used to record the historical markings of the interface after it is opened and then re-contacted, based on the discrete state of the interface and monitoring data, and to correct the prior parameter interval of the corresponding segment according to the historical markings. The fixed-state branch parameter inversion sub-unit is used to invert the remaining prestress of the anchor cable, the equivalent bond stiffness of the anchoring section, the interface tangential parameters, the interface normal parameters, and the incremental flexibility of the rock platform within the fixed-state branch that has passed the identifiability test, based on the current interface discrete state, the corrected parameter prior interval, and the load-measuring point combination that has passed the test. The confidence screening subunit is used to remove parameters that fail the confidence test based on the inverted parameters and confidence intervals, and to limit parameters belonging to the preset low sensitivity category to participate in degradation classification only when the overall confidence interval is in the abnormal region, so as to obtain the set of effective parameters after screening and the corresponding state branch information. The model selection subunit is used to eliminate branches with excessive back-substitution errors based on the effective parameter set, state branch information, and back-substitution error of each branch. Among the remaining branches, the branch with the smallest Bayesian information criterion is selected as the target state branch, and the target state branch and effective parameters are output to the degeneracy classification unit. The pending verification marking subunit is used to mark the current segment as pending verification and output the pending verification status when it is determined that the current segment does not meet any state criterion or the monitoring criteria are contradictory.

[0012] Preferably, the degradation classification unit includes: The effective parameter discrimination subunit is used to receive continuous parameters and confidence intervals from the multi-path parameter inversion unit and identify the effective parameter set that passes the tests of identifiability, back substitution error and confidence. The degradation component calculation sub-unit is used to calculate the degradation components of anchor cable prestress, anchor bond degradation, interface tangential degradation, interface normal degradation, residual slip degradation, rock platform flexibility degradation, and deep displacement based on the non-empty effective parameter set, residual slip, rock platform flexibility degradation, and deep displacement, respectively; wherein the interface tangential degradation component is 1 when there is complete slip, and the interface normal degradation component is 1 when there is complete opening; The bearing path index aggregation sub-unit is used to classify each degradation component with valid data support into the anchorage path component set, contact path component set, and rock platform path component set, respectively. Each path index takes the maximum value of the corresponding set; when the degradation component set of a certain path is empty, the path is marked as pending review. The threshold determination subunit is used to determine whether each path is in a normal, tracking and observation, or degenerate action state based on preset warning thresholds and action thresholds. The comprehensive status output subunit is used to summarize the judgment results of each path, output single-path degradation, compound degradation, pending review or normal status, and output them to the hardening decision unit.

[0013] Preferably, the hardening decision unit includes: The candidate scheme generation sub-unit is used to map the degradation state of anchorage, contact and rock platform paths to candidate schemes such as tensioning, anchorage section reinforcement, new anchorage, contact grouting, consolidation grouting and deep anchorage based on the degradation classification results. The scheme parameter conversion sub-unit is used to convert each candidate scheme into initial stress increment, cable element, bond-slip element, interface constitutive parameter, surrounding rock material parameter or structural surface parameter in the coupled model; The reinforcement effect prediction sub-unit is used to predict the recovery degree and safety margin of each candidate scheme based on the updated coupling model; The safety constraint check subunit is used to check whether the predicted response of each candidate solution meets the upper limit type and lower limit type safety index limit, and to list the solutions that pass the check as feasible solutions. The multi-index comprehensive evaluation subunit is used to conduct a weighted comprehensive evaluation of feasible solutions based on the degree of restoration, safety margin, amount of work, construction disturbance and uncertainty. The weight template is determined and locked before calculation based on whether there is complete slippage or complete opening before reinforcement, the safety margin before reinforcement and the remaining maintenance window. The recommended solution output sub-unit is used to output the feasible solution with the highest comprehensive evaluation value. When no feasible solution exists, it outputs bridge crane operation restriction, temporary stability and special design instructions.

[0014] This invention also provides a multi-path degradation classification and reinforcement decision-making method for rock wall crane beams in pumped storage power stations, and the system includes: Under combined inspection loads, sensors are used to collect multi-source response data of wheel pressure, traction force, position, beam, anchoring system, interface and rock platform. The multi-source response data is subjected to identifiability screening, state branch identification, and continuous parameter inversion to generate degradation classification results and reinforcement decision schemes for anchorage, contact, and rock platform paths.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention utilizes the bridge crane itself to form differentiated in-service detection excitation, and first checks the identifiability within the fixed interface state branch, and then inverts the continuous parameters, which can reduce multiple solutions and path misjudgment; furthermore, through the quantitative mapping of physical parameters to reinforcement measures and the phased retesting and updating, it reduces measure mismatch, repeated reinforcement and over-reinforcement. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1This is a schematic diagram of the multi-path degradation classification and reinforcement decision-making system for rock wall crane beams in a pumped storage power station, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the bridge crane combined load and multi-source monitoring arrangement according to an embodiment of the present invention; wherein, (a) is a cross section: combined inspection load and multi-source monitoring arrangement; (b) is a plane: schematic diagram of the detection section, trolley lateral position and forward / reverse traction; Figure 3 Flowchart for interface state recognition, identifiability analysis, and multi-path parameter inversion; Figure 4 A flowchart for three-path degradation classification, candidate scheme evaluation, and retesting and updating. Figure 5 This is a flowchart illustrating the overall closed-loop process for multi-path degradation classification and reinforcement decision-making of rock wall crane beams. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] Example 1 This implementation method uses the crane beam on the rock wall of one side of the underground powerhouse of a pumped storage power station as the inspection object, and establishes a detection system along the axis of the crane beam. The coordinates are used to divide the detection zones based on anchor cable arrangement, construction joints, surrounding rock type, cross-influence of tunnel entrances, and existing anomaly records. It should be noted that the following zone lengths, number of monitoring points, load levels, and thresholds are for illustrative purposes only and do not constitute limitations.

[0021] like Figure 1 As shown, a multi-path degradation classification and reinforcement decision system for rock wall crane beams in pumped storage power stations includes: a data acquisition interface, a data processing module, and a result output interface.

[0022] Specifically, such as Figure 2As shown, (a) is the cross-section: combined inspection load and multi-source monitoring layout; (b) is the plane: schematic diagram of the detection section, trolley lateral position and forward / reverse traction; beam strain or displacement sensors are arranged on the representative section of the crane beam on the rock wall to be detected, anchor cable axial force or anchor head displacement sensors are arranged at the corresponding positions of the prestressed anchor cable anchor head or anchorage section, directional opening and closing and tangential sliding sensors are arranged at the beam-rock contact interface, and displacement sensors are arranged on the surface of the rock platform and deep in the rock mass, with at least one deep displacement measuring point located outside the expected relaxation zone; wheel pressure sensors are arranged on both sides of the bridge crane wheel set, longitudinal traction force detection devices are arranged at the longitudinal drive or traction parts of the bridge crane, and position encoders are arranged on the trolley running mechanism and the trolley running mechanism respectively; wheel pressure verification and structural response measuring points are arranged on the crane beam on the opposite side of the rock wall. The data of each measuring point, wheel pressure, traction force and position are collected synchronously, and the time series data are converted into a spatial sequence according to the position of the trolley.

[0023] Specifically, such as Figure 2 As shown, strain or displacement sensors are installed on representative sections of the crane beam on the side rock wall to be tested. Anchor cable axial force or anchor head displacement sensors are installed at corresponding positions of the prestressed anchor cable anchor heads or anchorage sections. Directional opening and closing and tangential slip sensors are installed at the beam-rock contact interface. Displacement sensors are also installed on the rock platform surface and deep within the rock mass. At least one deep displacement measurement point is located outside the expected relaxation zone. Wheel pressure sensors are installed on both sides of the bridge crane's wheel sets. Longitudinal traction force detection devices are installed on the longitudinal drive or traction parts of the bridge crane. Position encoders are installed on the trolley traveling mechanism and the main trolley traveling mechanism, respectively. Wheel pressure verification and structural response measurement points are installed on the crane beam on the opposite side rock wall. The testing section is divided along the crane beam axis according to the anchor cable arrangement, construction joints, and surrounding rock zoning. Data on each measurement point, wheel pressure, traction force, and position are collected synchronously. Time-series data are converted into spatial sequences based on the main trolley position. Sampling frequency, synchronization error, missing rate, and repeatability are preset based on the on-site monitoring accuracy.

[0024] The data processing module is used to perform identifiable filtering, state branch identification, and continuous parameter inversion on multi-source response data to generate degradation classification results and reinforcement decision schemes for anchoring, contact, and rock platform paths. A further implementation method includes a data processing module comprising: a baseline and load configuration unit, a multi-source response processing unit, an identifiability analysis unit, a multi-path parameter inversion unit, a degradation classification unit, a hardening decision unit, and a retest feedback and decision update unit.

[0025] A reference and load configuration unit is used to establish a reference parameter field and configure at least two trolley lateral positions, two levels of inspection loads, and controlled longitudinal traction conditions in both directions; a further implementation method is that, in the reference and load configuration unit: The reference parameter field is either a measured reference parameter field or a reconstructed reference parameter field. The measured reference parameter field is established based on the completion acceptance, anchor cable tensioning, contact quality inspection, and initial bridge crane operation data. The reconstructed reference parameter field is constrained by at least two independent information sources that cannot be derived from the same original data. At least one of these sources is on-site measurement or inspection information, and the other is design and construction records, the response of the opposite crane beam, or calibration coupled model information. The reconstructed reference parameter field must satisfy a normalized reference back-substitution error not exceeding a preset upper limit. In the two-level inspection load, the second inspection load... With the first inspection load The difference is not less than the preset minimum load difference, that is In this embodiment, the first inspection load The second inspection load can be 10% to 20% of the rated lifting capacity. 25% to 40% can be taken; the forward and reverse traction forces respectively meet the requirements. and The absolute deviation of the positive and negative traction forces shall not exceed the preset tolerance, and all working conditions shall meet the safety limits for beam stress, anchor cable axial force, interface opening and closing, rock platform displacement and bridge crane operation.

[0026] Specifically, a calibrated measured reference parameter field or a reconstructed reference parameter field is established, and at least two levels of inspection loads with at least two trolley lateral positions and load differences not less than the preset minimum value are configured, as well as positive and negative controlled longitudinal traction conditions with opposite directions and comparable amplitudes within the detection section; among them, the lateral position change of the trolley is used to change the wheel pressure distribution on both sides of the bridge crane, the two levels of inspection loads are used to excite the normal opening and closing of the beam-rock interface, anchorage force transmission and incremental deformation of the rock platform, and the positive and negative controlled longitudinal traction is used to excite the tangential stick-slip of the beam-rock interface.

[0027] A multi-source response processing unit is used to correct, spatially align, and normalize the collected multi-source response data based on a reference parameter field and load configuration, forming symmetrical and anti-symmetrical response characteristics. A further implementation involves the following steps in the multi-source response processing unit: for the same load level and trolley lateral position under controlled forward and reverse traction conditions, positive and negative response data are collected after corrections for zero point, temperature, slow variable terms, and bridge torsion; using the measured wheel pressure resultant force of the positive and negative directions in the tested section as a reference, wheel pressure normalization is performed on the positive and negative responses respectively. The normalized responses of the positive and negative wheel pressures were obtained through a process of normalization. Based on the measured positive and negative traction forces of the test section, the positive and negative responses were normalized to obtain the normalized responses of the positive and negative traction forces. The normalized responses of the positive and negative wheel pressures at the same test point under the same working condition were averaged to obtain the symmetrical response characteristics of the test point. The symmetrical response characteristics characterize the overall support flexibility and anchorage force variation. The normalized responses of the positive and negative traction forces at the same test point under the same working condition were half-differenced to obtain the antisymmetric response characteristics of the test point. The antisymmetric response characteristics characterize the tangential stick-slip response of the beam-rock interface.

[0028] Specifically, for the same load level and lateral position of the trolley, let , For positive and negative correction responses, , To correspond to the zero-point response, , To measure the combined wheel pressure in both forward and reverse directions in the participating section, , Given the measured traction forces in both forward and reverse directions, the normalized wheel pressure response and traction normalized response in both directions are as follows: In the formula, , These are the normalized responses to forward and reverse wheel pressure, respectively. , These are the normalized responses for positive and negative traction, respectively. and These are the minimum effective wheel pressure threshold and the minimum effective traction force threshold, respectively, used to avoid distortion of the normalization results when the load amplitude is too small. Symmetrical response characteristics Antisymmetric response characteristics ; It mainly characterizes the overall support flexibility and anchorage stress changes. The main characteristic is the tangential stick-slip response at the beam-rock interface. Class characteristics are further standardized according to the baseline standard deviation or the allowable variation in the design.

[0029] The identifiability analysis unit is used to construct a sensitivity matrix based on the sampled parameter states of each candidate interface state branch, including closed bonding, closed microslip, complete slip, and complete opening, using symmetric and antisymmetric response characteristics. It also performs effective rank, minimum singular value, and condition number checks on the selected load-measurement point combination. After verifying that the checks meet preset requirements, it proceeds to parameter inversion for the corresponding branch. Figure 3 As shown. A further implementation includes the following: the identifiability analysis unit comprises: a priori parameter space sampling unit, used to form a finite set of sampling parameters within the priori parameter interval of each candidate interface state branch; a sensitivity matrix construction subunit, used to construct a continuous parameter sensitivity matrix and perform singular value decomposition for each sampling parameter state; an identifiability judgment subunit, used to verify whether the selected load-measuring point combination simultaneously satisfies the requirements of effective rank, minimum singular value, and condition number under all sampling states; a working condition supplement subunit, used to supplement the trolley position, load level, traction working condition, or measuring point according to the minimum singular value gain when the judgment is not satisfied; and an iteration termination judgment subunit, used to output a state to be reviewed when the preset number of adjustments is still not satisfied.

[0030] Specifically, regarding the first Each candidate current interface state branch forms a finite set of sampled parameters within the prior parameter range. and for each Constructing a continuous parameter sensitivity matrix The effective rank is defined as: The selected load-measuring point combination satisfies the following under all sampling conditions: In the formula, For the first In the nth state branch Class-standardized response to the first Normalized sensitivity of a continuous parameter, For the first A continuous parameter, For the first The baseline standard deviation or allowable variation of the response; , and The sensitivity matrix is ​​respectively the first The number, maximum, and minimum singular values, For indicator functions, For condition numbers, For the first The number of consecutive parameters in each state branch; For effective singular value relative thresholds, This is the lower bound for the minimum singular value. The upper limit of the condition number is determined in advance based on the measurement noise level, the prior interval of the parameters, and the cross-validation results of the benchmark operating condition. If the conditions are not met, the trolley position, load level, traction condition, or measuring point is supplemented according to the minimum singular gain under the most unfavorable sampling state. If the adjustment still does not meet the requirements, the output will show a status pending review.

[0031] For each state branch that passes the identifiability test, the inversion objective is jointly constituted by the normalized response residual, spatial continuity, and parameter priors. Branches with excessive back-substitution errors or violations of physical boundaries are eliminated, and the branch with the smallest Bayesian information criterion is selected from the remaining branches. Low-sensitivity parameters only participate in degradation classification when their confidence intervals are entirely in the outlier region.

[0032] A multi-path parameter inversion unit is used to identify the discrete state of the interface and invert the remaining prestress of the anchor cable, the bond stiffness of the anchorage section, interface parameters, and the flexibility of the rock platform within a fixed-state branch. A further implementation of this unit includes: an interface state identification subunit, used to determine the current discrete state of the interface in each detection section according to the priority order of fully open, fully slipped, closed with slight slip, and closed with bond; a historical state marking subunit, used to record historical markings of interfaces that have re-contacted after opening based on the discrete state of the interface and monitoring data, and to correct the prior parameter intervals of the corresponding sections based on the historical markings; and a fixed-state branch parameter inversion subunit, used to invert the remaining prestress of the anchor cable, the bond stiffness of the anchorage section, etc., within a fixed-state branch that has passed the identifiability test, based on the current discrete state of the interface, the corrected prior parameter intervals, and the load-measuring point combinations that have passed the test. The system includes: effective bond stiffness, interface tangential parameters, interface normal parameters, and incremental compliance of the rock platform; a confidence screening subunit, used to eliminate parameters that fail the confidence test based on the inverted parameters and confidence intervals, and to limit parameters belonging to the preset low-sensitivity category to participate in degradation classification only when the overall confidence interval is in the anomalous area, thus obtaining the set of effective parameters and the corresponding state branch information; a model selection subunit, used to eliminate branches with excessive back-substitution errors based on the set of effective parameters, state branch information, and back-substitution errors of each branch, and to select the branch with the smallest Bayesian information criterion as the target state branch from the remaining branches, and output the target state branch and effective parameters to the degradation classification unit; and a verification mark subunit, used to mark the current segment as verification pending and output the verification pending state when it is determined that the current segment does not meet any state criterion or the monitoring criteria are contradictory.

[0033] In this embodiment, a coupled model is established, consisting of the bridge crane, track, double-sided crane beams, anchorage system, beam-rock interface, and surrounding rock platform. The bridge crane and crane beams are represented by spatial beam, shell, or solid elements; the wheels and track are represented by compression-only contact elements; the anchor cables are represented by cable elements and bond-slip elements; and the surrounding rock is represented by a continuous or discontinuous medium model capable of characterizing the main structural surfaces. The torsional stiffness of the bridge frame, the track elevation difference, and the actual wheel pressure on both sides are used as model inputs.

[0034] For anchor cables equipped with independent axial force monitoring, the calibration measurement at the zero-moving load stage is used as the residual prestress. Observations or strong priors; for anchor cables that are not independently monitored but are in the same row, have the same structure, and are subjected to similar forces, As a grouping parameter inversion, if the sensitivity and confidence interval requirements are not met, the prestress of each anchor cable is not output. The equivalent bond stiffness of the anchorage section is: In the formula, For the first Section No. The equivalent bond stiffness of the anchorage section of the anchor cable. This represents the increment of anchor cable axial force between the two levels of inspection loads. For the average relative slippage of the corresponding anchorage section, To prevent excessively small positive denominators, the average relative slip of the anchorage section can be obtained by integration using distributed optical fiber or multi-point strain monitoring, or by subtracting the elastic elongation of the free section, anchor deformation, and rock mass reference displacement from the anchor head displacement; local pull-out tests can be used to calibrate the bond-slip relationship. The incremental compliance of the rock platform section is defined by the increment between two load levels: In the formula, For the first Sectional rock platform incremental flexibility, and The first The displacement of each rock platform response measuring point under the second and first inspection loads. and The first The measured wheel pressure of each participating wheel set under two levels of inspection load. For the first The normalized weights of the measurement points are summed as follows: The multipath parameter inversion unit is based on the first... Normal contact force at the interface of the section Normal opening and closing displacement tangential stress and residual slip after unloading The current interface state is determined according to the following top-down priority order: (1) The fully open state satisfies: (2) The complete slip state satisfies: (3) The closed, slightly sliding state first satisfies: And meet at least one of the following conditions: (4) The closed bonding state satisfies: In the formula, The threshold for determining zero normal contact force. To set the opening displacement threshold, This is the threshold value for the absolute value of tangential stress that separates closed adhesion from closed microslip. The threshold value of the absolute value of the tangential stress for complete slip. The residual slip measurement threshold, For complete slip residual slip threshold, and satisfying and The threshold is determined by sensor resolution, interface shearing or opening / closing tests, and the design model. If any state criterion is not met or the monitoring criteria contradict each other, a state pending review is output. Historical markers indicating re-contact after opening are used to select prior intervals for the tangential and normal parameters of the re-contact interface and candidate contact recovery schemes; they are not used as the current state alone. Continuous parameters are inverted within each fixed current state model, and the model with the smallest Bayesian information criterion among those whose normalized back-substitution error meets the upper limit is selected. Full slip and full opening are directly characterized by discrete states; low-sensitivity parameters only participate in degradation classification when their confidence intervals are entirely in the anomalous region.

[0035] The degradation classification unit is used to calculate the degradation indices of anchorage, contact, and rock platform paths based on the continuous parameters and discrete interface states obtained from the inversion, and to determine the status as normal, under observation, in deterioration action, or pending review based on threshold values; such as Figure 4As shown. A further implementation involves a degradation classification unit comprising: an effective parameter discrimination subunit, used to receive continuous parameters and confidence intervals from a multi-path parameter inversion unit, and identify the effective parameter set that passes the tests of identifiability, back-substitution error, and confidence level; and a degradation component calculation subunit, used to calculate, based on the non-empty effective parameter set, residual slip, rock platform incremental flexibility, and deep displacement, the anchor cable prestress degradation component, anchor bond degradation component, interface tangential degradation component, interface normal degradation component, residual slip degradation component, rock platform flexibility degradation component, and deep displacement degradation component, respectively; wherein, when there is complete slip, the interface tangential degradation component is 1, and when fully open, it is 1. The interface normal degradation component is 1; the bearing path index aggregation subunit is used to classify each degradation component with valid data support into the anchoring path component set, contact path component set, and rock platform path component set respectively, and the maximum value of each path index is taken from the corresponding set; when the degradation component set of a certain path is empty, the path is marked as pending review; the threshold determination subunit is used to determine whether each path is in a normal, tracking and observation, or degradation action state according to the preset warning threshold and action threshold; the comprehensive status output subunit is used to summarize the determination results of each path, output the single path degradation, compound degradation, pending review, or normal state, and output them to the reinforcement decision unit.

[0036] Specifically, the first Effective parameter set of the section , , and The remaining prestress of the anchor cable, the equivalent bond stiffness of the anchorage section, the interface tangential parameters, and the interface normal parameters are obtained through tests of identifiability, back-substitution error, and confidence level. Based on the non-empty effective parameter set, residual slip, rock platform incremental compliance, and deep displacement, the prestress degradation components of the anchor cable are obtained. Anchorage bond degradation component Interface tangential degradation component Interface normal degradation component Residual slip degradation component Components of Rock Platform Flexibility Degradation and deep displacement degradation components Fully slide command Fully open season .

[0037] Degraded components with valid data support are respectively grouped into anchorage path component sets. Contact path component set and rock platform path component set When the set is empty, the path is marked as pending review, and the anchoring, contact, and rock platform path indices are taken as the maximum value of their respective effective components: For the The first section Each carrying path is assigned a warning threshold. and action threshold ,in These represent the anchoring path, contact path, and rock platform path, respectively; the threshold values ​​are calculated based on normal inspection statistics and design maintenance limits, and all values ​​are greater than 0 and not greater than 1. , , In the formula, and The first Section No. The mean and standard deviation of the indicators for each path under normal inspection conditions. This is the statistical control coefficient. greater than The minimum warning threshold, These are action threshold benchmarks determined based on design maintenance limits, specific safety assessments, or historical engineering data. The minimum interval between the warning threshold and the action threshold, and satisfying... .when When the path is normal, the corresponding path will be considered normal; when When, the corresponding path is determined to be for tracking and observation; when When the corresponding path is in a degraded action state, it will be determined. If the set of degraded components for a path is empty, the path index will not be calculated and it will be marked as pending review; the contact path will be directly determined to enter the degraded action state in the fully slipped or fully open state.

[0038] When local cracks in the beam, loose track pressure plates, uneven track, or sensor malfunctions can be independently explained as abnormalities, special treatment should be carried out on the beam, track, or monitoring system first, and they should not be forcibly classified into the three load-bearing paths.

[0039] The reinforcement decision unit is used to convert candidate reinforcement schemes into coupled model parameters based on the degradation indices of anchorage, contact, and rock platform paths output by the degradation classification unit and the effective parameters output by the multi-path parameter inversion unit. After safety constraint checks and comprehensive evaluation, a recommended scheme is output. A further implementation method includes a candidate scheme generation subunit, used to map the degradation state of anchorage, contact, and rock platform paths to candidate schemes for tensioning, anchorage section reinforcement, new anchorage, contact grouting, consolidation grouting, and deep anchorage, based on the degradation classification results. Each degradation path and candidate scheme is shown in Table 1. Table 1 The scheme parameter conversion subunit is used to convert each candidate scheme into initial stress increment, cable element, bond-slip element, interface constitutive parameter, surrounding rock material parameter, or structural surface parameter in the coupled model; the reinforcement effect prediction subunit is used to predict the recovery degree and safety margin of each candidate scheme based on the updated coupled model; the safety constraint check subunit is used to check whether the predicted response of each candidate scheme meets the upper limit and lower limit safety index limits, and list the schemes that pass the check as feasible schemes; the multi-index comprehensive evaluation subunit is used to perform a weighted comprehensive evaluation of feasible schemes according to recovery degree, safety margin, engineering quantity, construction disturbance, and uncertainty. The weight template is determined and locked before calculation based on whether there is complete slip or complete opening before reinforcement, the safety margin before reinforcement, and the remaining maintenance window; the recommended scheme output subunit is used to output the feasible scheme with the largest comprehensive evaluation value, and output bridge crane operation restriction, temporary stability, and special design instructions when no feasible scheme exists.

[0040] Specifically, candidate measures are generated according to Table 1. When the remaining prestress of the anchor cable decreases but the equivalent bond stiffness of the anchorage section still meets the requirements, a candidate scheme for additional tensioning is generated first. When the equivalent bond stiffness of the anchorage section decreases, or the existing anchor cable does not have effective conditions for additional tensioning, a candidate scheme for strengthening the anchorage section or adding new anchorages is generated. The reinforcement parameters are primarily determined by the results of trial grouting, grout material testing, anchor cable pull-out testing, construction records, and re-measurement in this project. When on-site evidence is insufficient, a calibrated conservative range that does not overestimate the reinforcement effect is adopted. The tensioning scheme is converted into the initial stress increment of the specified anchor cable element; the anchorage section reinforcement scheme is converted into the equivalent bond stiffness and corresponding bond-slip parameter update of the corresponding anchorage section bond-slip element; the new anchorage scheme is converted into a new cable element and bond-slip element; the contact grouting scheme is converted into an update of the interface voids, effective contact area, and tangential and normal constitutive parameters of the abnormal zone; the consolidation grouting scheme is converted into an update of the material parameters or structural surface parameters of the target rock mass area; and the deep anchorage scheme is converted into a new anchor cable element that crosses the abnormal zone and enters the stable surrounding rock. Section No. The additional tension of the anchor cable satisfies: In the formula, To compensate for the tension, For target prestress, To reinforce the remaining prestress, To account for the redistribution of anchor cable internal forces, staged tensioning re-measurement, and safety constraints, correction factors are applied, and The deep anchorage range should extend beyond the identified anomalous zone and into the stable surrounding rock. Its length is determined based on the boundary of the anomalous zone, the anchor cable inclination angle, and the minimum anchorage length within the stable surrounding rock. The interface, anchorage, and surrounding rock parameters after reinforcement should be determined primarily based on on-site grouting tests, grout material tests, anchor cable pull-out tests, construction quality inspections, and local verification tests. If on-site test data for this project is lacking, initial values ​​for candidate parameters should be determined using a rule base from similar projects or a calibrated empirical interval, and calibrated during post-reinforcement re-measurement along the same trajectory.

[0041] For the first to pass the safety constraint check Each of the candidate schemes has a value range of [value range]. ~ Predicted recovery level Safety margin Normalized engineering quantities Normalized construction disturbance and normalized uncertainty For upper limit safety indicators where an increase in response value leads to an increase in risk, the safety margin is calculated based on the remaining distance from the predicted response to the upper limit. For lower-limit safety indicators where a decrease in response value leads to an increase in risk, the safety margin is calculated based on the remaining distance from which the predicted response exceeds the lower limit. Specifically, for the first [item] that passes the safety constraint check... The candidate solutions are predicted to have the following recovery levels and safety margins: In the formula, For the first The predicted recovery degree of the target degradation parameters by each candidate scheme. To include the key parameter set in the evaluation of this scheme, , and These are the predicted value, the value before reinforcement, and the baseline value, respectively. To prevent positive numbers with excessively small denominators.

[0042] For the first Safety margin of each safety indicator ; To predict the response, and These are upper limit and lower limit safety limits, respectively. To reinforce the pre-response, To prevent positive numbers with excessively small denominators.

[0043] The quantities of work, construction disturbances, and uncertainties are calculated based on the additional tension, anchor cable length, drilling and grouting volume, bridge crane downtime and construction impact range, and the degree of variation of key prediction parameters, respectively. These are then normalized to 0–1 using a pre-fixed reference interval for the same project that does not change with the candidate set. A comprehensive evaluation is then performed using the following formula: In the formula, For the first The comprehensive evaluation value of each candidate solution, , , , and These are weights for the degree of restoration, safety margin, workload, construction disturbance, and uncertainty, respectively. Each weight is a non-negative integer, and the sum of the weights is [value missing]. Before calculating candidate solutions, the weight template is used based on whether there is complete slippage or complete opening before reinforcement, and the safety margin before reinforcement. And the remaining maintenance window is determined and locked; safety margin of candidate solutions. It is only used for evaluating the corresponding scheme and not for adjusting the weight template in reverse.

[0044] When there is complete slip, complete opening or When using a security-first template, the following is true: The preset safety margin threshold is used; when the safety priority condition is not met and the remaining maintenance window is less than the preset short-term window, the fast recovery template is used; in other cases, the balanced decision template is used. The most feasible solution is the recommended solution. If no feasible solution exists, output instructions for limiting the operation of the bridge crane, temporary stabilization, and special design.

[0045] The retest feedback and decision update unit is used to retest and update the reinforcement scheme along the same trajectory after path reinforcement. Automatic decision-making stops when no feasible scheme is found or continuous updates are ineffective. Specifically, the retest feedback and decision update unit uses the same bridge crane running direction, inspection load level, trolley lateral position, running speed, traction force, spatial sampling interval, and data processing method as before reinforcement for the retest. After the first reinforcement update The recovery index of each continuous parameter is: In the formula, Parameters before reinforcement For the first The parameters after the second reinforcement update For the corresponding reference parameters, To prevent positive numbers with excessively small denominators, Restricted to ~ Inside, the closer the value is... This indicates that the parameters are closer to the baseline state.

[0046] Key parameters that have not yet met the corresponding acceptance criteria should be grouped into a set of substandard parameters; after the contact path is reinforced, it should also be confirmed that the interface has exited the fully sliding or fully opening state and meets the requirements. , Residual slippage and safety response thresholds. Anchorage paths should be inspected. , and the redistribution of internal forces in adjacent anchor cables; the rock platform path should be checked. and deep displacement. When the set of non-compliant parameters is empty, the current path is determined to meet the recovery requirements; when the minimum recovery index increment of the non-compliant parameter is lower than the preset limit for two consecutive times, the automatic decision-making of the current path is stopped and a special review instruction is output; Indicates the first The absolute deviation of the parameters relative to the benchmark after the next update; the absolute deviation of any non-compliant key parameter from the benchmark parameter. When the gain is greater than the previous update and the increment exceeds the preset deterioration limit, the automatic decision-making will stop immediately and a special review instruction will be output without waiting for two consecutive low gains.

[0047] This embodiment verifies the effectiveness of combined inspection loads, state branch identification, and identifiability screening in differentiating multipath degradation. A three-dimensional solid-contact-structural surface coupled model is used to generate the structural response, and a beam-interface spring-equivalent surrounding rock model is used for parameter inversion. This ensures that the generated response model and the inverted model maintain a different modeling approach, reducing the beneficial deviations caused by verification using the same model. In this simulation embodiment, the ratio of the remaining prestress of the anchor cable to the reference value is 0.55–1.00, the ratio of the equivalent bond stiffness of the anchorage section to the reference value is 0.40–1.00, the ratio of the interface tangential and normal parameters to their respective reference values ​​is 0.30–1.00, and the ratio of the incremental compliance of the rock platform to the reference value is 1.00–2.00. These parameter ranges are only used to illustrate the verification process and do not constitute a limitation on the specific engineering parameter values.

[0048] In the parameter state settings, different degrees of single-parameter degradation states were set for the remaining prestress of the anchor cable, the bond stiffness of the anchorage section, the tangential and normal parameters of the interface, and the flexibility of the rock platform. Simultaneously, dual-path composite degradation (anchorage-contact, anchorage-rock platform, and contact-rock platform) and three-path composite degradation were set, resulting in a total of 45 parameter states. Each parameter state was superimposed with 20 sets of independent random measurement noise, yielding a total of 900 test samples. The actual degradation type of each test sample was determined based on a pre-defined action threshold and interface state criteria such as complete opening and complete slippage. The previous full version used a verification design with 45 parameter states and 900 test samples.

[0049] To examine the role of each core technical feature, three comparison methods were set up: the first method used only the beam strain and displacement response for inversion; the second method used all multi-source responses but did not implement identifiability screening; the third method used the complete process of combined load, multi-source response, and identifiability screening proposed in this invention. The results are shown in Table 2.

[0050] Table 2 Table 2 shows that when only beam strain and displacement are used, the effective rank of the continuous parameter sensitivity matrix is ​​insufficient, with the average parameter error and misclassification rate being 24.8% and 31.7%, respectively. When all multi-source responses are introduced but identifiability screening is not implemented, the effective rank reaches 5, but the condition number remains at 420, the average parameter error is 18.6%, and the misclassification rate is 22.0%. After adopting the complete process of this invention, the condition number decreases to 86, the average parameter error decreases to 7.4%, and the misclassification rate decreases to 4.0%. The results indicate that the synergistic effect of lateral position change of the trolley, two-stage inspection load, controlled forward and reverse traction, and identifiability screening within the fixed state branch can improve the stability of anchorage, contact, and rock platform path parameter separation and degradation classification. The above results are calculation results under the conditions set in this simulation embodiment and are used to illustrate the technical effect of this invention. They do not constitute a general guarantee of identification accuracy for different engineering projects.

[0051] A 36-m long rock wall crane beam was selected as the example inspection object. The beam was divided into six inspection sections, numbered Z1-Z6, from one end to the other, with each section being 6 m long. The trolley was positioned at three locations: near the wall, mid-span, and far from the wall. Two levels of inspection loads were applied: 15% and 30% of the rated lifting weight, respectively. Controlled forward and reverse operation was implemented at each trolley position and load level. The operating speed within each inspection section was 0.20 m / s, and the absolute deviation of the controlled forward and reverse traction forces was controlled within 10%. After data correction, spatial alignment, load normalization, state branch identification, and parameter inversion, section Z2 was determined to have anchorage path degradation, section Z4 to have contact path degradation, and section Z5 to have rock platform path degradation. The remaining sections were not described as typical degradation sections in this embodiment.

[0052] Based on the aforementioned degradation results, the system generates corresponding candidate reinforcement schemes: For section Z2, evaluation includes tensioning, anchorage section reinforcement, and new anchorage measures; for section Z4, evaluation includes overall contact grouting, zoned contact grouting, and additional interface anchorage measures; and for section Z5, evaluation includes deep anchorage, consolidation grouting, and their combinations. The recommended scheme is determined after a comprehensive evaluation of the candidate schemes based on restoration effect, safety margin, workload, construction disturbance, and uncertainties. After reinforcement is completed and the materials reach the specified strength and the anchor cable axial force stabilizes, a repeat measurement of the same trajectory is performed using the same trolley position, inspection load level, running direction, running speed, traction force, spatial sampling interval, and data processing method as before reinforcement. The updated interface state, continuous parameters, and restoration index are used to determine whether the current path meets the restoration requirements. If the restoration requirements are not met, subsequent candidate schemes are regenerated. Once the requirements are met, other degradation paths are processed until all degradation paths meet the final inspection requirements or the automatic update termination condition is met.

[0053] The result output interface is used to output degradation classification results and reinforcement decision schemes. Specifically, in this embodiment, a graphical software interface installed on the monitoring workstation or server displays the detection section, inspection status, degradation classification, path indicators, recommended or alternative schemes, restricted operation, and special review instructions. It can also generate electronic reports, export structured data, or provide the results to the power plant monitoring system, maintenance management system, or mobile terminal via a communication interface. The implementation form of the result output interface is not limited to a single software interface. The engineering basic database, benchmark parameter library, and reinforcement rule library are shared by all modules. On-site construction parameters and retest deviations are used to correct the range of prior parameters and effect prediction coefficients, but do not automatically change safety control values, identifiability criteria, and final inspection conditions.

[0054] Example 2 This invention also provides a multi-path degradation classification and reinforcement decision-making method for rock wall crane beams in pumped storage power stations, applying the system of Embodiment 1, including: Under combined inspection loads, sensors are used to collect multi-source response data of wheel pressure, traction force, position, beam, anchoring system, interface and rock platform. The multi-source response data is subjected to identifiability screening, state branch identification, and continuous parameter inversion to generate degradation classification results and reinforcement decision schemes for anchorage, contact, and rock platform paths.

[0055] like Figure 5 As shown, the detailed steps of the present invention include: Step S101: Divide the detection area, configure the multi-source monitoring system, and establish or reconstruct the calibrated measured reference parameter field; Step S102: Based on the bridge crane structure, rated lifting capacity and special safety limits, configure at least two trolley lateral positions, at least two levels of non-overlapping inspection loads, and forward and reverse controlled longitudinal traction conditions; Step S103: Synchronously collect actual wheel pressure, traction force, bridge crane position, and the response of beam, anchorage system, beam-rock interface and surrounding rock of the platform on both sides. After correction, spatial alignment and load normalization, a symmetric response feature matrix and an antisymmetric response feature matrix are formed. Step S104: Based on the interface normal contact force, opening and closing displacement and residual slip, establish current state branches of closed adhesion, closed micro-slip, complete slip and complete opening, set the re-contact history mark after opening, and perform identification test on the load-measuring point combination in each fixed current state branch; Step S105: Invert the remaining prestress of the anchor cable, the equivalent bond stiffness of the anchorage section, the interface tangential parameters, the interface normal parameters, and the incremental compliance of the rock platform within the state branch that has passed the identifiability test, and determine the target state branch based on the back-substitution error and the Bayesian information criterion. Step S106: Calculate the degradation indices of anchorage, contact, and rock platform paths, and determine the status as normal, under observation, single-path degradation, composite degradation, or pending review. Step S107: Map the degradation type to candidate solutions such as tensioning, new anchorage, contact grouting, consolidation grouting, or deep anchorage, and output recommended solutions through coupled model prediction, safety constraint check, and multi-index evaluation. Step S108: After each load-bearing path reinforcement is completed, the same trajectory retest is carried out. The subsequent plan is adjusted according to the updated interface status and continuous parameters until the final inspection requirements are met or the automatic update termination conditions are met.

[0056] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A multi-path degradation classification and reinforcement decision-making system for rock wall crane beams in pumped storage power stations, characterized in that, include: The data acquisition interface is used to collect multi-source response data of wheel pressure, traction force, position, beam, anchoring system, interface and rock platform under combined inspection loads using sensors; The data processing module is used to perform identifiability filtering, state branch identification, and continuous parameter inversion on the multi-source response data to generate degradation classification results and reinforcement decision schemes for anchoring, contact, and rock platform paths. The result output interface is used to output the degradation classification results and reinforcement decision scheme.

2. The system according to claim 1, characterized in that, The sensors include wheel pressure sensors, longitudinal traction force detection devices, trolley and carriage position encoders, beam strain or displacement sensors, anchor cable axial force or anchor head displacement sensors, beam-rock interface normal opening and tangential slip sensors, and rock platform surface and deep displacement sensors.

3. The system according to claim 1, characterized in that, The data processing module includes: The benchmark and load configuration unit is used to establish a benchmark parameter field and configure at least two trolley lateral positions, two levels of inspection loads, and forward and reverse controlled longitudinal traction conditions. The multi-source response processing unit is used to correct, spatially align, and normalize the collected multi-source response data based on the reference parameter field and load configuration, thereby forming symmetric and antisymmetric response features. The identifiability analysis unit is used to construct a sensitivity matrix based on the symmetric and antisymmetric response characteristics of each candidate interface state branch under the sampled parameter states of closed bonding, closed micro-slip, complete slip and complete opening. It also performs effective rank, minimum singular value and condition number tests on the selected load-measuring point combination. After the test meets the preset requirements, it enters the parameter inversion of the corresponding branch. The multi-path parameter inversion unit is used to identify the discrete state of the interface and invert the remaining prestress of the anchor cable, the bond stiffness of the anchor section, the interface parameters and the flexibility of the rock platform within the fixed state branch. The degradation classification unit is used to calculate the degradation indices of anchorage, contact and rock platform paths based on the continuous parameters and discrete state of the interface obtained by inversion, and to determine the normal, tracking observation, degradation action or pending review status according to the threshold. The reinforcement decision unit is used to convert candidate reinforcement schemes into coupled model parameters based on the degradation indexes of anchorage, contact and rock platform paths output by the degradation classification unit and the effective parameters output by the multi-path parameter inversion unit, and output a recommended scheme after safety constraint check and comprehensive evaluation. The retest feedback and decision update unit is used to retest the same trajectory and update the reinforcement scheme after path reinforcement. Automatic decision-making stops when there is no feasible scheme or continuous updates are ineffective.

4. The system according to claim 3, characterized in that, In the reference and load configuration unit: The reference parameter field is either a measured reference parameter field or a reconstructed reference parameter field; the measured reference parameter field is established based on the completion acceptance, anchor cable tensioning, contact quality detection and initial bridge crane operation data; the reconstructed reference parameter field is constrained by at least two independent information sources that cannot be derived from the same original data, of which at least one is on-site measurement or detection information, and the other is design and construction records, response of the opposite crane beam or calibration coupling model information, and satisfies that the normalized reference back substitution error is not greater than a preset upper limit; In the two-level inspection load, the difference between the second inspection load and the first inspection load is not less than the preset minimum load difference. The absolute deviation of the positive and negative traction forces shall not exceed the preset tolerance, and all working conditions shall meet the safety limits for beam stress, anchor cable axial force, interface opening and closing, rock platform displacement and bridge crane operation.

5. The system according to claim 3, characterized in that, In the multi-source response processing unit, the process of forming symmetric and antisymmetric response features includes: For the same load level and trolley lateral position under the controlled traction conditions in both the forward and reverse directions, forward and reverse response data were collected after corrections for zero point, temperature, slow variable terms and bridge torsion. Based on the measured wheel pressure resultant force in the forward and reverse directions of the test section, the wheel pressure normalization process is performed on the forward and reverse responses respectively to obtain the forward and reverse wheel pressure normalized responses; Based on the measured forward and reverse traction forces of the test section, the forward and reverse responses are normalized to obtain the forward and reverse traction normalized responses. The average of the normalized response of the positive wheel pressure and the normalized response of the negative wheel pressure at the same measuring point under the same working condition is taken as the symmetrical response feature of the measuring point. The symmetrical response feature characterizes the overall support flexibility and the change of anchoring force. The difference between the normalized response of the forward traction and the normalized response of the reverse traction at the same measuring point under the same working condition is taken as the antisymmetric response feature of the measuring point. The antisymmetric response feature characterizes the tangential stick-slip response of the beam-rock interface.

6. The system according to claim 3, characterized in that, The identifiability analysis unit includes: The prior parameter space sampling unit is used to form a finite set of sampling parameters within the prior parameter range of each candidate interface state branch. The sensitivity matrix construction subunit is used to construct a continuous parameter sensitivity matrix and perform singular value decomposition for each sampling parameter state. The identifiability determination sub-unit is used to verify whether the selected load-measuring point combination simultaneously meets the requirements of effective rank, minimum singular value and condition number under all sampling conditions. The working condition supplement subunit is used to supplement the trolley position, load level, traction working condition or measuring point according to the minimum singular value gain when the judgment is not met. The iteration termination determination subunit is used to output a pending review status when the preset number of adjustments is not met.

7. The system according to claim 3, characterized in that, The multipath parameter inversion unit includes: The interface state recognition subunit is used to determine the current discrete interface state of each detection segment according to the priority order of fully open, fully slip, closed micro-slip, and closed adhesion. The historical state marking subunit is used to record the historical markings of the interface after it is opened and then re-contacted, based on the discrete state of the interface and monitoring data, and to correct the prior parameter interval of the corresponding segment according to the historical markings. The fixed-state branch parameter inversion sub-unit is used to invert the remaining prestress of the anchor cable, the equivalent bond stiffness of the anchoring section, the interface tangential parameters, the interface normal parameters, and the incremental flexibility of the rock platform within the fixed-state branch that has passed the identifiability test, based on the current interface discrete state, the corrected parameter prior interval, and the load-measuring point combination that has passed the test. The confidence screening subunit is used to remove parameters that fail the confidence test based on the inverted parameters and confidence intervals, and to limit parameters belonging to the preset low sensitivity category to participate in degradation classification only when the overall confidence interval is in the abnormal region, so as to obtain the set of effective parameters after screening and the corresponding state branch information. The model selection subunit is used to eliminate branches with excessive back-substitution errors based on the effective parameter set, state branch information, and back-substitution error of each branch. Among the remaining branches, the branch with the smallest Bayesian information criterion is selected as the target state branch, and the target state branch and effective parameters are output to the degeneracy classification unit. The pending verification marking subunit is used to mark the current segment as pending verification and output the pending verification status when it is determined that the current segment does not meet any state criterion or the monitoring criteria are contradictory.

8. The system according to claim 7, characterized in that, The degradation classification unit includes: The effective parameter discrimination subunit is used to receive continuous parameters and confidence intervals from the multi-path parameter inversion unit and identify the effective parameter set that passes the tests of identifiability, back substitution error and confidence. The degradation component calculation sub-unit is used to calculate the degradation components of anchor cable prestress, anchor bond degradation, interface tangential degradation, interface normal degradation, residual slip degradation, rock platform flexibility degradation, and deep displacement based on the non-empty effective parameter set, residual slip, rock platform flexibility degradation, and deep displacement, respectively; wherein the interface tangential degradation component is 1 when there is complete slip, and the interface normal degradation component is 1 when there is complete opening; The bearing path index aggregation sub-unit is used to classify each degradation component with valid data support into the anchorage path component set, contact path component set, and rock platform path component set, respectively. Each path index takes the maximum value of the corresponding set; when the degradation component set of a certain path is empty, the path is marked as pending review. The threshold determination subunit is used to determine whether each path is in a normal, tracking and observation, or degenerate action state based on preset warning thresholds and action thresholds. The comprehensive status output subunit is used to summarize the judgment results of each path, output single-path degradation, compound degradation, pending review or normal status, and output them to the hardening decision unit.

9. The system according to claim 3, characterized in that, The reinforcement decision-making unit includes: The candidate scheme generation sub-unit is used to map the degradation state of anchorage, contact and rock platform paths to candidate schemes such as tensioning, anchorage section reinforcement, new anchorage, contact grouting, consolidation grouting and deep anchorage based on the degradation classification results. The scheme parameter conversion sub-unit is used to convert each candidate scheme into initial stress increment, cable element, bond-slip element, interface constitutive parameter, surrounding rock material parameter or structural surface parameter in the coupled model; The reinforcement effect prediction sub-unit is used to predict the recovery degree and safety margin of each candidate scheme based on the updated coupling model; The safety constraint check subunit is used to check whether the predicted response of each candidate solution meets the upper limit type and lower limit type safety index limit, and to list the solutions that pass the check as feasible solutions. The multi-index comprehensive evaluation subunit is used to conduct a weighted comprehensive evaluation of feasible solutions based on the degree of restoration, safety margin, amount of work, construction disturbance and uncertainty. The weight template is determined and locked before calculation based on whether there is complete slippage or complete opening before reinforcement, the safety margin before reinforcement and the remaining maintenance window. The recommended solution output sub-unit is used to output the feasible solution with the highest comprehensive evaluation value. When no feasible solution exists, it outputs bridge crane operation restriction, temporary stability and special design instructions.

10. A multi-path degradation classification and reinforcement decision-making method for rock wall crane beams in pumped storage power stations, using the system described in any one of claims 1-9, characterized in that, include: Under combined inspection loads, sensors are used to collect multi-source response data of wheel pressure, traction force, position, beam, anchoring system, interface and rock platform. The multi-source response data is subjected to identifiability screening, state branch identification, and continuous parameter inversion to generate degradation classification results and reinforcement decision schemes for anchorage, contact, and rock platform paths.