Library body health monitoring and early warning system based on heterogeneous multi-source data fusion

Abstract

The application relates to the technical field of structural health monitoring, and particularly discloses a library body health monitoring and early warning system based on heterogeneous multi-source data fusion. The system realizes real-time acquisition of library wall strain, environmental wind load and three-dimensional deformation data through a multi-source data acquisition module; a spatial gridding layered mapping module is used to divide high-risk, transition and stable areas in combination with the structural characteristics of the library body, and the risk sensitivity is improved through dynamic weight adjustment; a cross-frequency domain correlation feature extraction module constructs a space-time-frequency domain feature matrix, fuses strain spectrum, wind load harmonic and deformation curvature data, and verifies the feature credibility in combination with an adversarial generative network; a resonance risk dynamic identification module identifies resonance hotspots based on energy transmission path analysis and modal cloud image standing wave node line determination; and a closed-loop early warning module realizes hierarchical early warning and strategy self-adaptive adjustment through encrypted scanning, digital twin model iteration and dynamic threshold optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of structural health monitoring, in particular to a library body health monitoring and early warning system based on heterogeneous multi-source data fusion. BACKGROUND

[0002] The fly ash library is a typical large powder library and an important facility and equipment in the thermal power industry, which is used for storing fly ash in the production process of the thermal power plant. The operation and management level of the fly ash library is of great significance to the safe production of the thermal power plant, the protection of the surrounding ecological environment, and the improvement of the solid waste recycling rate. Therefore, the research and development of intelligent equipment for the fly ash library will bring significant ecological, economic and social benefits to the thermal power industry, and will be conducive to the transformation, upgrading and quality improvement of the industry.

[0003] The operation and maintenance of the fly ash library include library body structure health monitoring, fly ash weight and material level monitoring, fly ash hardening removal, fly ash metering and operation management, and the smooth operation of the fly ash library has a great influence on the safe production of the thermal power plant. The production accidents caused by fly ash storage, forced shutdown of the boiler and stop of power generation due to fly ash library failure occur from time to time. Due to the poor internal working environment of the fly ash library (closed space, high dust, strong corrosion), the current operation and maintenance automation level of the fly ash library is low, and manual operation is mainly used, which has the problems of great safety hidden danger, long downtime, high maintenance cost, etc. In addition, the fly ash library also has the safety hidden danger caused by structural instability.

[0004] The prior art has the following disadvantages:

[0005] When the library body is excited by strong wind, the vibration string sensor (10Hz sampling rate) captures the low-frequency strain main frequency (such as 2Hz), and the transient deformation point cloud data obtained by the three-dimensional scanning robot (0.017Hz sampling rate) at the vibration peak moment and the high-frequency fluctuation signal (such as 4Hz) of the environmental wind speed instrument cannot be accurately aligned in the time-frequency domain; when the system forcibly associates the non-synchronous data, due to the phase shift and harmonic aliasing effect, the random vibration harmonic is misjudged as structural resonance, which causes the safety protection mechanism to be triggered. SUMMARY

[0006] The purpose of the present application is to provide a library body health monitoring and early warning system based on heterogeneous multi-source data fusion to solve the problems in the background.

[0007] The purpose of the present application can be achieved by the following technical solutions:

[0008] The library body health monitoring and early warning system based on heterogeneous multi-source data fusion comprises:

[0009] A multi-source data acquisition module, the multi-source data acquisition module is used for acquiring library wall strain data, environmental wind load data and library body three-dimensional deformation data in real time;

[0010] A spatial gridding hierarchical mapping module divides high-risk areas, transition areas and stable areas according to the structural characteristics of the library body, constructs a multi-dimensional spatial grid matrix, and maps the sensor positions to the grid nodes;

[0011] A cross-frequency domain correlation feature extraction module fuses strain spectrum, wind load spectrum and deformation curvature data into a time-space-frequency domain feature matrix by dimensionality elevation, and extracts the coupling features of spatial position and frequency by constraint optimization decomposition;

[0012] A resonance risk dynamic identification module locates the spatial-frequency energy focusing hotspots based on the energy transmission path analysis of the coupling features, and verifies the resonance risk in combination with the weak point parameters of the library body structure;

[0013] A closed-loop early warning and feedback module triggers three-dimensional deformation encryption scanning and finite element model checking when the hotspot area energy intensity exceeds the dynamic threshold, outputs graded early warning and dynamically adjusts the monitoring strategy.

[0014] As a further scheme of the present application, the construction of the multi-dimensional spatial grid matrix specifically includes the following steps:

[0015] The circumferential weld distribution position and historical damage data of the library body welding structure are extracted, and the library wall is divided into high-risk areas, transition areas and stable areas along the height direction in combination with the library wall stress finite element cloud map;

[0016] The high-risk area covers the library bottom ring foundation to the maximum bending moment acting height of the library wall, the transition area covers the bending moment attenuation zone to the structural deformation inflection point height, and the stable area covers the area above the inflection point;

[0017] The high-risk area is equally divided into 12 sector-shaped ring zones, the transition area is equally divided into 8 sector-shaped ring zones, and the stable area is equally divided into 4 sector-shaped ring zones along the circumferential direction with the library body center axis as the reference;

[0018] The high-risk area is divided into 6 layer domains according to the stress gradient, the transition area is divided into 4 layer domains, and the stable area is divided into 2 layer domains along the vertical direction, forming a multi-dimensional spatial grid matrix;

[0019] The chord sensor coordinates are registered with the three-dimensional deformation scanning point cloud, each sensor is associated to the intersection grid node of the corresponding ring zone and layer domain through the nearest neighbor topology mapping, and the high-risk area node is given a 3 times weight coefficient, and the transition area is given a 2 times weight coefficient.

[0020] As a further scheme of the present application, the high-risk areas, transition areas and stable areas are divided according to the structural characteristics of the library body, a multi-dimensional spatial grid matrix is constructed, and the sensor positions are mapped to the grid nodes, which at least further include the following steps:

[0021] Real-time micro-deformation data of the weld heat-affected zone is acquired, and when the deformation variance in a single ring belt exceeds a threshold value, the corresponding ring belt is subdivided into 16 sub-ring belts, and a virtual monitoring node is generated at the junction of the sub-ring belts;

[0022] Based on the thickness distribution of the tank wall steel plate, the environmental corrosion rate and the historical load spectrum, the structural risk entropy values of each grid node are calculated;

[0023] The entropy value, real-time strain gradient and weld distance are input into the fuzzy decision maker, and a node weight correction factor is dynamically output, and the correction factor range is set to 1.0-5.0;

[0024] When a certain grid node sensor fails, the adjacent node data is called and the corresponding node state is reconstructed through radial basis interpolation, and a three-dimensional scanning robot is driven to perform intensive scanning on the ring belt where the node is located, so that the missing signal is replaced by laser point cloud data.

[0025] As a further scheme of the present application, the construction process of the space-time-frequency domain feature matrix is:

[0026] A correlation constraint model of strain main frequency components and environmental wind load harmonic components is established, and the strain main frequency amplitude and the wind load harmonic amplitude in the same ring belt position are forced to satisfy a monotonically increasing relationship;

[0027] The strain spectrum, wind load spectrum and deformation curvature data are reorganized into a space-time-frequency domain feature matrix according to time slices, and a non-negative constraint decomposition is used to extract a coupling factor matrix of spatial grid nodes and frequencies;

[0028] Based on the coupling factor matrix, the energy focusing coefficients of each ring belt region are calculated, and when the focusing coefficient of a certain ring belt in a specific frequency band exceeds a reference value and the deformation curvature suddenly changes synchronously, it is marked as a resonance risk hotspot.

[0029] As a further scheme of the present application, the strain spectrum, wind load spectrum and deformation curvature data are reorganized into a space-time-frequency domain feature matrix, and the coupling characteristics of spatial positions and frequencies are extracted through constraint optimization decomposition, at least including the following steps:

[0030] The marginal spectrum entropy of the real-time tank wall vibration signal is analyzed, and when the entropy value is lower than a threshold value, the wind load spectrum analysis frequency band is contracted to within the strain main frequency ±0.5 octave;

[0031] A strain-deformation time sequence generative adversarial network is constructed, a generator generates virtual high-frequency deformation data based on low-frequency deformation curvature, and a discriminator introduces a tank body vibration differential equation to constrain the network output;

[0032] The Hausdorff distance between the generated deformation data and the measured deformation data is calculated, and the singular value decay rate of the coupling factor matrix is combined to output the reliability weight of the space-frequency domain feature;

[0033] Fusion energy focusing coefficient, deformation mutation gradient and feature credibility weight, calculate the resonance risk entropy value through entropy weight method, trigger encryption scanning verification when exceeding dynamic threshold.

[0034] As a further scheme of the present application: the energy transmission path analysis based on the coupling feature, the spatial-frequency energy focusing hot spot is located, comprising the following steps:

[0035] Based on the stiffness attenuation coefficient of the spatial grid node and the continuity parameter of the circumferential weld, a directed topological network of the tank body vibration energy transmission is constructed;

[0036] Wherein the node stiffness attenuation coefficient is calculated by the thickness of the steel plate and the corrosion rate, and the weld continuity is evaluated by fitting the weld displacement gradient through three-dimensional deformation data;

[0037] The resonance energy flow density of each grid node in the target frequency band is calculated, and the energy flow density is the product of the strain spectrum amplitude, the wind load power spectrum density and the node transmission efficiency;

[0038] When the energy flow density of three consecutive nodes in the same ring exceeds 2 times the reference value, the corresponding ring is marked as an energy focusing hot spot;

[0039] The natural frequency of the heat affected zone of the weld in the finite element model is called to verify whether the hot spot frequency band falls within the interval of ±15% of the natural frequency;

[0040] At the same time, detect whether the change rate of the hot spot ring deformation curvature increases to more than 3 times of the previous time.

[0041] As a further scheme of the present application: the resonance risk is verified in combination with the tank structure weak point parameters, comprising the following steps:

[0042] The instantaneous phase of the vibration string sensor in the hot spot ring is extracted, and when the phase coherence of 80% of the sensors in the ring exceeds 85% and lasts for more than 5 seconds, it is determined that the structure is in resonance;

[0043] Fusion energy focusing coefficient, phase coherence and deformation mutation gradient, calculate the correlation entropy value of the three through nonlinear granger causality test;

[0044] When the correlation entropy value is lower than the threshold value and the phase coherence is continuous, output the resonance risk level;

[0045] Drive the three-dimensional scanning robot to perform millisecond-level laser vibration measurement on the hot spot ring, and reconstruct the tank wall vibration mode cloud picture;

[0046] If the cloud picture shows that the standing wave nodal line passes through the circumferential weld, a first-level warning is triggered and a weld fatigue life evaluation report is generated.

[0047] As a further scheme of the present application: the output graded early warning and dynamically adjusts the monitoring strategy, specifically comprising:

[0048] When the energy flow density of the energy focusing hot spot exceeds the dynamic threshold value;

[0049] Primary response: start three-dimensional deformation encryption scanning, and the scanning frequency is increased to 5 times of the conventional frequency;

[0050] Secondary response: call the finite element model to check the resonance frequency matching degree, and if the deviation exceeds 15%, activate the weld damage prediction;

[0051] Tertiary response: when the energy flow density continues for 3 minutes without attenuation, drive the consolidation cleaning robot to remove the material pressure of the hot spot area;

[0052] Calculate the deformation curvature obtained by the encryption scanning and the finite element predicted deformation by Hausdorff distance;

[0053] If the distance value is less than the deformation tolerance threshold, the early warning level is downgraded;

[0054] If the distance value increases and the weld damage probability exceeds 60%, output a secondary early warning;

[0055] Train a hidden Markov model based on historical early warning data;

[0056] When the energy flow density of the hot spot area decreases at a rate lower than the preset value;

[0057] Automatically extend the encryption scanning time to twice the original planned time and shrink the robot operation radius to within 3 meters.

[0058] As a further scheme of the present application: the output graded early warning and dynamically adjusts the monitoring strategy, comprising the following steps:

[0059] Input the encryption scanning data, robot operation record and energy flow density change rate into the digital twin, and dynamically correct the boundary constraint conditions and material constitutive parameters of the finite element model;

[0060] When the secondary early warning continues for 10 minutes without elimination;

[0061] Synchronously start the three-dimensional topography acquisition system to reconstruct the library wall topology model;

[0062] And call the consolidation cleaning robot to perform preventive vibration on the hot spot ring;

[0063] Statistical structure relaxation coefficient after each early warning is eliminated, which is defined as the ratio of the energy flow density decay rate to the deformation recovery rate;

[0064] When the relaxation coefficient is greater than 1.5 times of the reference value for 3 consecutive times;

[0065] The dynamic threshold is raised by 20%, and a structure stiffness reinforcement scheme is generated.

[0066] Advantages of the present application:

[0067] (1) The present application fuses heterogeneous multi-source data (library wall strain, environmental wind load, three-dimensional deformation, etc.), combines dynamic spatial gridding modeling and cross-frequency domain feature extraction technology, constructs a space-time-frequency domain feature matrix, and accurately identifies resonance risk hotspots. At the same time, the closed-loop early warning module realizes real-time response and strategy adjustment to high-risk areas through encryption scanning, digital twin model iteration and dynamic threshold optimization, reduces false alarm rate and improves early warning timeliness, and ensures efficient management and control of the safety of the library structure.

[0068] (2) Through dynamic weight correction of spatial grid nodes (based on fuzzy decision maker and structural risk entropy), sensor fault tolerance (radial basis interpolation and laser scanning compensation), and strategy optimization driven by hidden Markov model, the system can maintain stable operation under complex working conditions. In addition, the collaborative updating mechanism of digital twin and finite element model enables the system to have continuous learning ability, adapt to environmental changes and structural degradation trends, prolong equipment life and reduce maintenance costs. BRIEF DESCRIPTION OF DRAWINGS

[0069] The present application will be further described below with reference to the accompanying drawings.

[0070] Figure 1 is a block diagram of the system of the present application. DETAILED DESCRIPTION

[0071] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0072] Please refer to Figure 1 The present application is a library health monitoring and early warning system based on heterogeneous multi-source data fusion, which comprises:

[0073] A multi-source data acquisition module is used to acquire library wall strain data, environmental wind load data and library three-dimensional deformation data in real time.

[0074] A spatial gridding hierarchical mapping module is used to divide high-risk areas, transition areas and stable areas according to the structural characteristics of the library, construct a multi-dimensional spatial grid matrix, and map the sensor positions to the grid nodes.

[0075] A cross-frequency domain correlation feature extraction module, which fuses strain spectrum, wind load spectrum and deformation curvature data into a space-time-frequency domain feature matrix, and extracts spatial position and frequency coupling features through constrained optimization decomposition;

[0076] A resonance risk dynamic identification module, which locates the space-frequency energy focusing hotspots based on energy transmission path analysis of coupling features, and verifies the resonance risk in combination with library structure weak point parameters;

[0077] A closed-loop early warning and feedback module, which triggers three-dimensional deformation intensive scanning and finite element model checking when the hotspot area energy intensity exceeds the dynamic threshold, outputs graded early warning and dynamically adjusts the monitoring strategy.

[0078] In the multi-source data acquisition module, 172 vibrating wire strain sensors are installed on the surface of the library wall, including 112 in the horizontal direction and 60 in the vertical direction. The sensors are arranged according to the gradient distribution principle: 6 rings are divided in the 0-7.2 meter height range at the bottom of the library wall, 12 sensors are installed at equal intervals in each ring; 4 rings are divided in the 7.2-15 meter height range in the middle, 8 sensors are installed in each ring; 2 rings are divided in the 15 meters or more height range at the top, 4 sensors are installed in each ring. All sensors are connected to 6 32-channel vibrating wire acquisition boxes through armored cables, the acquisition boxes are placed in waterproof protective shells and fixed on the welded buckles 1.5 meters high from the ground on the library wall. The data sampling rate is 10 Hz, the sensor temperature value is recorded synchronously during data collection for strain temperature compensation, and the compensation coefficient is determined by laboratory calibration as 0.5 micro-strain per Celsius degree.

[0079] A three-dimensional ultrasonic wind speed instrument is installed at the center of the library roof, measuring a height of 3 meters above the highest point of the library roof. The wind speed instrument has a sampling rate of 20 Hz, a measurement range of 0-60 m / s, and an accuracy of ±0.1 m / s. To eliminate the interference of turbulence on the library roof, the wind speed instrument is equipped with a fairing and a dynamic yaw correction mechanism: when the wind direction changes by more than 30 degrees within 10 seconds, the motor automatically drives the wind speed instrument to rotate to the wind direction. The wind load spectrum analysis uses the Hanning window weighted fast Fourier transform, with a window length of 1024 sampling points and an overlap rate of 50%.

[0080] A high-precision track system is laid along the wall of the library, carrying a laser three-dimensional scanning robot. The total length of the track is 94.2 meters (based on the diameter of the library body of 30 meters), and the positioning accuracy of the robot is ±0.1 mm. The scanning uses a line laser triangulation method to form a dot matrix on the surface of the library wall with a spacing of 0.5 mm, and a single full-library scan takes 60 seconds. The deformation curvature calculation process is: first extract the local surface of the library wall from the point cloud data, and use the moving least squares method to fit the surface equation; then calculate the maximum eigenvalue of the curvature tensor as the deformation curvature index. Under the conventional monitoring mode, scanning is performed once every hour, and under the early warning mode, it is increased to once every minute.

[0081] A three-dimensional coordinate system is established with the center of the library floor as the origin, and the Z axis is vertically upward. The clocks of all sensors are synchronized through the GPS time module, with a time error of less than 1 millisecond. The space registration method is: 12 reference target balls are pre-set on the library wall, and the three-dimensional scanning robot scans the target ball coordinates every time it starts, mapping the position of the vibrating string sensor to the target ball coordinate system, with a position mapping residual error of within 2 mm. The environmental wind load data is aligned with the strain data through time stamp.

[0082] Strain data preprocessing includes third-order Butterworth low-pass filtering (cutoff frequency 5 Hz) and temperature drift compensation. Wind load data preprocessing uses median filtering to eliminate pulse noise and converts to wind pressure distribution in the library coordinate system (based on the pre-calibrated wind pressure coefficient matrix of computational fluid dynamics). Three-dimensional deformation data is registered with historical point clouds through the iterative closest point algorithm to extract the deformation displacement field. The processed data is stored in a time slice structure, each slice containing timestamp, spatial grid code, strain frequency spectrum amplitude, wind load power spectral density, and deformation curvature.

[0083] Automatic diagnosis is performed every 24 hours: the vibrating string sensor measures the string damping coefficient through the excitation coil, and if the damping changes more than 15% of the initial value, it is marked as faulty; the anemometer triggers the calibration program when the deviation is greater than 5% through sound wave time of flight cross verification; the scanning robot verifies the accuracy through standard ball diameter measurement, and automatically adjusts the optical calibration parameters when the error exceeds 0.2 mm. The diagnosis results are updated in real time to the monitoring database.

[0084] When the wind speed exceeds 20 m / s, the vibrating string sampling rate is increased to 20 Hz, and the scanning robot starts the wind-resistant image stabilization mode (hydraulic leg locks the track). When rain and snow weather is detected, the vibrating string sensor protection cover starts electric heating and dehumidification, and the anemometer enables ultrasonic self-cleaning function. When the environmental temperature is lower than -10℃, the temperature control module in the acquisition box maintains the temperature in the box at 5±2℃.

[0085] In the spatial grid layer mapping module, according to the circumferential weld distribution position and historical damage detection report of the silo body welding structure, combined with the silo wall stress finite element analysis results, the silo wall is divided into high-risk area, transition area and stable area along the height direction. The division range of the high-risk area is from the silo bottom ring foundation surface to the maximum bending moment acting height of the silo wall, which is determined by finite element analysis, and the typical value is 25%-30% of the total height of the silo body. The transition area covers the maximum bending moment acting height to the structural deformation inflection point height, which is determined by long-term deformation monitoring data, which is characterized by a significant decrease in curvature change rate. The stable area is above the deformation inflection point to the top of the silo. In actual engineering, for a silo with a height of 30 meters, the high-risk area is usually set to 0-7.2 meters, the transition area is 7.2-15 meters, and the stable area is above 15 meters.

[0086] With the center axis of the silo body as the reference, the high-risk area is equally divided into 12 sector rings along the circumferential direction, each corresponding to a 30-degree central angle; the transition area is equally divided into 8 sector rings, each corresponding to a 45-degree central angle; and the stable area is equally divided into 4 sector rings, each corresponding to a 90-degree central angle. When divided vertically, the high-risk area is divided into 6 layers according to the stress gradient, with each layer being about 1.2 meters high; the transition area is divided into 4 layers, each layer being about 1.95 meters high; and the stable area is divided into 2 layers, each layer being about 7.5 meters high. Thus, a multi-dimensional spatial grid matrix composed of ring bands and layers is formed, with the high-risk area containing 72 grid nodes (12 ring bands x 6 layers), the transition area containing 32 nodes (8 ring bands x 4 layers), and the stable area containing 8 nodes (4 ring bands x 2 layers).

[0087] The installation coordinates of the vibrating wire sensors are registered with the silo wall point cloud data obtained by three-dimensional laser scanning, and each sensor is positioned to the corresponding grid node through the nearest neighbor topology mapping algorithm. During the mapping process, for sensors near the boundaries of ring bands or layers, they are attributed according to the distance ratio between their actual positions and the centers of the nodes. After the mapping is completed, weight coefficients are assigned according to the risk levels of the areas: the grid nodes in the high-risk area are set to a weight of 3.0, the nodes in the transition area are set to a weight of 2.0, and the nodes in the stable area remain the default weight of 1.0. The weight coefficients will directly affect the contribution calculation in subsequent data analysis.

[0088] Real-time monitoring of micro-deformation data in the weld heat-affected zone, when the system detects that the deformation variance in a certain ring band exceeds the set threshold (usually 2 times the standard deviation of the average deformation) for 3 consecutive times, the ring band is automatically subdivided into 16 sub-ring bands, each corresponding to a 22.5-degree central angle. Virtual monitoring nodes are generated at the junctions of the sub-ring bands, and their initial data are generated by bilinear interpolation from the data of adjacent actual sensors. The subdivided grid will improve the monitoring resolution of the local area, especially for abnormal conditions such as weld cracking or corrosion intensification.

[0089] For each grid node, the structural risk entropy value is calculated by integrating the library wall steel plate thickness measurement, environmental corrosion rate monitoring data and historical load spectrum analysis results. The steel plate thickness data comes from the construction acceptance report, the corrosion rate is back calculated through the library wall surface potential detection, and the historical load spectrum comes from long-term strain monitoring statistics. The entropy value is calculated using the information entropy theory, reflecting the uncertainty of the structural state of the node location, and the higher the entropy value, the greater the risk. The calculation results are normalized to relative values in the range of 0-1, which are used for subsequent dynamic weight adjustment.

[0090] The structural risk entropy value, real-time strain gradient (the ratio of the strain difference of adjacent nodes to the distance) and the distance to the nearest weld are input into the fuzzy decision system. The system has five levels of fuzzy rules, such as "if the entropy value is high and the strain gradient is large, then the weight is greatly increased". After fuzzy reasoning and defuzzification, the dynamic weight correction factor of each node is output, with the correction range limited to 1.0-5.0. This process is executed every 10 minutes to ensure that the weight distribution responds to the changes in the structure in a timely manner.

[0091] When a sensor corresponding to a certain grid node fails, the real-time monitoring data of the adjacent 8 nodes is first called, and the radial basis function interpolation algorithm is used to reconstruct the state data of the failed node. In the interpolation process, the interpolation weight is allocated according to the spatial distance and weight coefficient of each adjacent node and the failed node. At the same time, the three-dimensional scanning robot is controlled to perform intensive scanning on the ring belt where the failed node is located, and the scanning frequency is increased to 3 times that of the regular mode. After curvature calculation on the obtained high-density point cloud data, the output of the failed sensor is replaced, forming a fault-tolerant mechanism with multiple source data complementation.

[0092] The real-time state of the spatial grid matrix is displayed through a three-dimensional visualization interface, where different ring belts and layers are rendered with gradient colors, and the color depth corresponds to the risk level. Manual interaction is supported to adjust the grid parameters, including temporarily adding virtual nodes and manually modifying weight coefficients. All adjustments are recorded in the database for subsequent analysis of the impact of grid optimization on monitoring effectiveness.

[0093] In the cross-frequency domain correlation feature extraction module, based on the dynamic characteristics of the library structure, a physical correlation model of the strain main frequency component and the environmental wind load harmonic component is established. For the ring belt position of each spatial grid node, the strain main frequency amplitude and the wind load harmonic amplitude are forced to satisfy a monotonically increasing relationship. In specific implementation, the strain main frequency (usually the peak frequency in the range of 0.5-5 Hz) of each ring belt is first obtained through the vibrating wire sensor, and then the corresponding wind load harmonic component (usually an integer multiple of the strain main frequency) is extracted from the wind speed instrument record. The system automatically checks the amplitude trend of the two, and if a ring belt appears an abnormal situation of increasing strain amplitude and decreasing wind load amplitude, it is determined that the ring belt data is abnormal and the review program is started. This constraint model effectively avoids false correlation of non-wind-induced vibration.

[0094] The pretreated strain spectrum, wind load spectrum and deformation curvature data are reorganized according to 5-minute time slices. Within each slice, the strain spectrum takes the amplitude spectrum of the 0.1-10 Hz frequency band at each spatial grid node, the wind load spectrum takes the power spectral density of the 0.5-20 Hz frequency band, and the deformation curvature takes the rate of change in this time period. The constructed three-dimensional feature matrix corresponds to the spatial grid node number in the first dimension, the frequency component in the second dimension, and the third dimension contains three types of data channels. When filling the matrix, cubic spline interpolation smoothing is used for the data of the frequency points not directly measured to ensure the integrity of the matrix.

[0095] Non-negative constraint decomposition is performed on the space-time-frequency domain feature matrix to extract the spatial grid node factor matrix and the frequency factor matrix. Special constraint conditions are set during the decomposition process: nodes in the same ring share the fundamental frequency component in the frequency factor matrix. After solving by an iterative optimization algorithm, the core tensor reflecting the spatial-frequency coupling characteristics is obtained. Based on this tensor, the energy focusing coefficient of each ring is calculated. The calculation method is: the coupling strength in a certain frequency band (such as 2-4 Hz) is spatially integrated in the ring, and then divided by the total area of the ring. When the energy focusing coefficient of a certain ring exceeds 1.8 times the historical statistical benchmark value, and the deformation curvature of the ring changes more than 3 times the average value in the same period, it is marked as a potential resonance risk hotspot.

[0096] The marginal spectrum entropy of the library wall vibration signal is calculated in real time to quantify the concentration degree of vibration energy distribution. When the spectrum entropy value is lower than the set threshold value (indicating that the energy is concentrated in a few frequency bands), the wind load spectrum analysis bandwidth is automatically contracted: taking the current strain main frequency as the center, only the wind load components within ±0.5 octave range are retained to participate in subsequent analysis. For example, when the strain main frequency is detected as 2 Hz, the wind load analysis frequency band is limited to 1.5-3 Hz. This mechanism effectively suppresses wideband noise interference and improves the signal-to-noise ratio of feature extraction. The degree of frequency band contraction is dynamically adjusted according to the spectrum entropy value, forming an adaptive analysis window.

[0097] An adversarial neural network containing a generator and a discriminator is constructed. The generator takes low-frequency deformation curvature (from regular scans once an hour) and strain spectrum as input, and outputs predicted high-frequency deformation field (equivalent to virtual scan data every 6 minutes). The discriminator not only judges the authenticity of the data, but also introduces the library vibration control equation as a physical constraint: it checks whether the generated data satisfies the vibration modal relationship determined by the structural stiffness matrix and mass matrix. During network training, a transfer learning strategy is adopted: first pre-train on small-scale calibration test data, and then gradually adapt to the actual ash bunker characteristics through online learning. The generated virtual data is labeled with a confidence index for subsequent fusion.

[0098] The Hausdorff distance between the calculated deformation data and the measured deformation data is calculated, which reflects the maximum difference between the two in spatial distribution. At the same time, the singular value decay rate of the coupling factor matrix is analyzed to evaluate the stability of feature extraction. The two indicators are input into the fuzzy reasoning system, and the credibility weight of the spatial-frequency domain feature is output, with a weight range of 0-1. For feature components with a credibility less than 0.6, the system automatically triggers data reacquisition or decomposition parameter adjustment process to ensure the reliability of the analysis results.

[0099] The entropy weight method is used to fuse the energy focusing coefficient, deformation mutation gradient and feature credibility weight. First, normalize each index, then determine the objective weight coefficient according to the information entropy. The calculated resonance risk entropy value is quantized to 0-100, and when it exceeds the dynamic threshold (initially set to 75), the system performs a three-level response: the first level is encrypted scanning verification (increased to 1 times per minute), the second level is real-time checking of the finite element model, and the third level is early warning information push. The dynamic threshold is automatically adjusted according to the historical false alarm rate, and if there is no false alarm for 3 consecutive times, the threshold is lowered by 5 points to improve sensitivity.

[0100] For the annulus marked as a resonance hotspot, the system automatically retrieves the deformation data of the last 3 encrypted scans, and verifies the risk persistence through the curvature change trend. At the same time, feature extraction parameters (such as the number of non-negative decomposition layers, frequency band contraction ratio, etc.) are recorded to the knowledge base, and when a real resonance is verified, the parameter settings are optimized for subsequent analysis. All feature extraction cases are periodically reviewed and evaluated every month, and the baseline and threshold parameters are updated to form a self-optimizing closed-loop system.

[0101] In the resonance risk dynamic identification module, a directed topological network is constructed based on the mechanical properties of the spatial grid nodes. The node stiffness decay coefficient is calculated by subtracting the corrosion loss from the measured steel plate thickness, and the corrosion rate is updated monthly based on the potential detection data. The weld continuity parameter is extracted from three-dimensional deformation data: 5 monitoring points are taken on both sides of the weld, the angle between the displacement gradient vectors is calculated, and an angle less than 15 degrees is determined as good continuity. The network edge weight is set as the energy transfer efficiency between adjacent nodes, which depends on the node spacing and the integrity rating of the connecting weld. After the network is constructed, reachability analysis is performed to identify the main path of energy transfer and potential blockage points.

[0102] For each grid node, the resonant energy flow density is calculated within the target frequency band (usually 1-10 Hz). The strain spectrum amplitude is taken from the temperature-compensated data of the vibrating wire sensor, the wind power spectrum density is obtained from the results of the ultrasonic anemometer after turbulence correction, and the node transmission efficiency is obtained from the topological network analysis. During the calculation, the parameters are first normalized, and then the geometric mean of the three is taken as the energy flow density index. The system continuously tracks the energy flow density distribution of each ring belt, and when it detects that the energy flow density of three consecutive nodes in the same ring belt exceeds twice the historical baseline value of the ring belt and the duration exceeds 3 sampling periods, the ring belt is marked as an energy focusing hotspot.

[0103] The natural frequency parameters of the hotspot ring belt corresponding position are extracted from the finite element model database, which has been corrected by modal test. Compare the real-time monitored hotspot frequency band with the model natural frequency to verify whether it falls within the ±15% interval. For the heat-affected zone of the weld, additionally consider the material performance degradation coefficient, and extend the lower limit of the natural frequency interval to -20%. After the matching verification is passed, the system automatically retrieves the fatigue cumulative damage history data of the weld as an auxiliary basis for risk level assessment.

[0104] In the ring belt marked as a hotspot, extract the instantaneous phase data of all vibrating wire sensors. Calculate the instantaneous phase angle of each sensor signal by Hilbert transform, and then evaluate the phase consistency of the sensors in the ring belt using circular statistics. When the phase difference of 80% or more of the sensors remains within ±15 degrees (corresponding to 85% coherence), and this state lasts for more than 5 seconds, it is determined that the structure is in resonance. To improve robustness, the system will exclude abnormal phase data caused by sensor failure.

[0105] A three-element time series containing the energy focusing coefficient, phase coherence index, and deformation mutation gradient is constructed. An improved Granger causality test method is used to analyze the non-linear correlation between parameters: the time series is divided into 10-second sliding windows, and the conditional entropy and transfer entropy are calculated in each window. Finally, the correlation entropy value is synthesized by information geometry method, and the smaller the value, the stronger the deterministic relationship between parameters. The system sets a dynamic entropy threshold, and when the measured entropy value is lower than the threshold and the phase coherence continuously meets the conditions, the resonance risk level is determined to be "confirmed".

[0106] Control the three-dimensional scanning robot to perform millisecond-level laser vibration measurement on the hotspot ring belt, with a sampling frequency of 1 kHz and a duration of 10 seconds. The obtained vibration displacement data is reconstructed into a library wall vibration cloud map through modal decomposition algorithm. Analyze the standing wave characteristics in the cloud map, and focus on checking the intersection position of the standing wave nodal line (vibration amplitude zero point connecting line) and the circumferential weld. When the nodal line is detected to pass through the weld and the strain gradient at the intersection exceeds the allowed value, it is determined to be a high-risk state. The system automatically generates a special evaluation report containing the vibration modal animation and the weld stress concentration coefficient.

[0107] According to the continuous monitoring results, the risk level is dynamically adjusted: if the energy flow density continues to rise while the phase coherence remains, the risk level is raised by one level every 5 minutes; if the deformation rate slows down or the entropy value rebounds, the downgrade evaluation process is started. The risk level is divided into four levels: observation (blue), warning (yellow), emergency (orange) and danger (red), each corresponding to different response plans. All level changes are time-stamped and the decision basis is recorded to form a complete risk evolution archive.

[0108] For cases confirmed as resonance, the system calls the weld fatigue analysis module. Based on the vibration amplitude, frequency and material S-N curve, the fatigue damage increment caused by this resonance event is calculated. Combined with historical cumulative damage data, the remaining fatigue life is predicted. When the predicted life is less than 3 months, a special inspection work order is triggered and pushed to the maintenance management system. The evaluation report includes key parameters such as crack initiation probability and critical crack size suggestion.

[0109] In the closed-loop warning and feedback module, the system monitors the energy flow density indicators of each ring in real time. When the energy flow density of the hot spot area exceeds the dynamic threshold (initial value set to 2.5 times the historical average of the ring), the hierarchical response mechanism is started. The first level response is to increase the three-dimensional deformation scanning frequency: in the normal mode, the scanning is once every hour, which is increased to once every 12 minutes (5 times frequency), and the scanning range is focused on the hot ring and its adjacent two rings. The second level response calls the finite element model for real-time verification. When the deviation between the measured resonance frequency and the model predicted value exceeds 15%, the weld damage probability prediction algorithm based on fracture mechanics is automatically activated. The third level response is triggered when the energy flow density continues to increase for 3 minutes without attenuation trend. The control board cleaning robot moves to the hot ring directly below and performs directional vibration work to loosen the accumulated material, and the vibration intensity is adjusted according to the energy flow density value.

[0110] The deformation curvature data obtained by encrypted scanning is spatially matched and verified with the finite element predicted value. The improved Hausdorff distance algorithm is used to calculate the difference between the two: first, the scanning point cloud and the predicted surface are meshed, and the maximum value of the curvature difference in each mesh element is calculated. The system presets the deformation tolerance threshold to be 0.15 radians / meter. When the measured distance is less than this value, it is determined as a false alarm and the warning level is downgraded; when the distance value continues to increase and the weld damage probability model output value exceeds 60%, it is upgraded to the second warning state. During the verification process, the system automatically marks the areas with significant differences, which are used to guide the local modification of the finite element model.

[0111] The Hidden Markov Model is trained based on historical early warning cases, which contains three hidden states (risk rising, risk stable, risk falling) and six observation variables (energy flow density, deformation curvature, etc.). When the decline rate of energy flow density in the hotspot area is lower than the preset value (usually 15% per hour), the model outputs a strategy adjustment suggestion: extend the duration of the encryption scan from the default 30 minutes to 60 minutes, and at the same time, shrink the working radius of the cleaning robot from 5 meters to 3 meters, forming a more intensive monitoring-disposal closed loop. The strategy adjustment parameters are automatically updated every week to ensure adaptation to the time-varying characteristics of the library structure performance.

[0112] A digital twin of the grey library is constructed, with a parameterized finite element model at its core. After each encryption scan is completed, the deformation field data, robot vibration operation records (including vibration position, intensity, duration), and energy flow density change curve are input into the twin. The system adjusts the model parameters through the back propagation algorithm: the boundary constraint conditions are corrected according to the measured support reaction force data, and the material constitutive parameters are optimized according to the stress-strain relationship curve. The corrected model is immediately used for the next round of early warning analysis, forming a continuously evolving simulation environment. A historical copy is kept every time the model version is updated, facilitating retrospective comparative analysis.

[0113] When the secondary early warning state lasts for 10 minutes without being resolved, the cross-system collaborative disposal process is started. First, the three-dimensional topography acquisition system is called to perform a full circumferential scan of the hotspot ring, reconstructing a topological model with millimeter-level precision, and paying special attention to surface micro-cracks in the weld area. At the same time, the board consolidation cleaning robot is instructed to switch to the preventive vibration mode: arrange vibration points at intervals of 2 meters below the hotspot ring, each point vibrates for 30 seconds, and the vibration frequency is set to 80% of the ring's natural frequency to avoid resonance. The operation data of the two systems is shared in real time, the topography data is used to verify the vibration effect, and the vibration record is used to optimize the operation parameters.

[0114] After each early warning is resolved, the system calculates the structure relaxation coefficient as a performance evaluation indicator. This coefficient is defined as the ratio of the energy flow density decay rate to the deformation recovery rate, reflecting the structure damping characteristics. When calculating, take the monitoring data within 30 minutes after the early warning is resolved, and obtain the change rate through linear regression. The baseline value of the coefficient is obtained by statistical analysis of the stable operation data in the first 100 days (typical value is 1.2-1.8). When the relaxation coefficient of 3 consecutive early warning events exceeds 1.5 times the baseline value, it is determined that the structure stiffness has degenerated, and the system automatically adjusts the dynamic threshold permanently by 20% to enhance the conservatism, and generates a reinforcement ring installation position and steel plate reinforcement scheme.

[0115] The early warning system is comprehensively evaluated every month: the false alarm rate, the missed alarm rate, the average response time, etc. are counted, and the correlation between the early warning level and the subsequent damage development is analyzed. The evaluation results are used to optimize three aspects of parameters: adjusting the trigger threshold of each level of response, correcting the material parameter updating strategy of the finite element model, and improving the vibration operation procedure of the cleaning robot. All optimization schemes need to be verified by simulation of the digital twin before deployment to ensure system stability.

[0116] The working principle of the present application: The present application realizes high-precision health monitoring and dynamic risk control of the reservoir body structure through multi-source data acquisition (reservoir wall strain, environmental wind load, three-dimensional deformation), dynamic spatial gridding modeling (high-risk / transition / stable area division and node weight optimization), cross-frequency domain feature extraction (spatio-temporal-frequency domain matrix construction and adversarial generation verification), multi-dimensional resonance risk identification (energy focusing coefficient, phase coherence, modal cloud chart standing wave node line) and closed-loop strategy optimization (digital twin iterative model, hierarchical early warning and robot response). The system innovatively integrates physical modeling and data-driven methods, and through dynamic weight adjustment, energy transmission path analysis and adaptive threshold optimization, improves the accuracy of resonance risk identification and the efficiency of early warning response.

[0117] The above describes one embodiment of the present application in detail, but the content described is only the preferred embodiment of the present application, and cannot be considered as limiting the scope of the present application. Any equivalent changes and improvements made in the scope of the present application should still belong to the scope of the present application.

Claims

1. A library body health monitoring and early warning system based on heterogeneous multi-source data fusion, characterized in that, The application relates to a multi-source data acquisition module for acquiring library wall strain data, environmental wind load data and library body three-dimensional deformation data in real time. A spatial gridding layered mapping module is used for dividing high-risk areas, transition areas and stable areas according to the structure characteristics of the library body, constructing a multi-dimensional spatial grid matrix, and mapping the sensor positions to the grid nodes. The multi-dimensional spatial grid matrix is constructed, and the construction specifically comprises the following steps: The circumferential weld distribution position and historical damage data of the library body welding structure are extracted, the library wall stress finite element cloud chart is combined, and the library wall is divided into high-risk areas, transition areas and stable areas along the height direction. The high-risk area is divided from the library bottom annular foundation surface to the maximum bending moment acting height of the library wall, the transition area covers the maximum bending moment acting height to the structure deformation inflection point height, and the stable area is the part above the deformation inflection point to the library top. The library body center axis is taken as the reference, the high-risk area is equally divided into 12 fan-shaped ring zones along the circumferential direction, the transition area is equally divided into 8 fan-shaped ring zones, and the stable area is equally divided into 4 fan-shaped ring zones. The high-risk area is divided into 6 layer domains according to the stress gradient along the vertical direction, the transition area is divided into 4 layer domains, and the stable area is divided into 2 layer domains, thereby forming a multi-dimensional spatial grid matrix. The chord vibration sensor coordinates are matched with the three-dimensional deformation scanning point cloud, each sensor is associated with the intersection grid node of the corresponding ring zone and layer domain through the nearest neighbor topological mapping, and the high-risk area node is given a 3 times weight coefficient, and the transition area is given a 2 times weight coefficient. A cross-frequency domain associated feature extraction module is used for upgrading and fusing strain frequency spectrum, wind load frequency spectrum and deformation curvature data into a time-space-frequency domain feature matrix, and coupling features of spatial positions and frequencies are extracted through constraint optimization decomposition. A resonance risk dynamic identification module is used for positioning the space-frequency energy focusing hot spot based on the energy transmission path analysis of the coupling features, verifying the resonance risk in combination with the weak point parameters of the library body structure. A closed-loop early warning and feedback module is used for triggering three-dimensional deformation encryption scanning and finite element model checking when the hotspot area energy intensity exceeds the dynamic threshold value, outputting a graded early warning and dynamically adjusting the monitoring strategy. The multi-dimensional spatial grid matrix is constructed, and the construction specifically comprises the following steps:

2. The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, Real-time micro-deformation data of the weld heat affected zone are acquired, when the deformation variance in a single ring zone exceeds a threshold value, the ring zone is subdivided into 16 sub-ring zones, and virtual monitoring nodes are generated at the junctions of the sub-ring zones; Based on the library wall steel plate thickness distribution, the environmental corrosion rate and the historical load spectrum, the structure risk entropy values of the grid nodes are calculated; The entropy values, real-time strain gradients and weld distances are input into a fuzzy decision maker, a node weight correction factor is dynamically output, and the correction factor range is set to 1.0-5.

0. ​ When a certain grid node sensor fails, the data of adjacent nodes is called and the state of the node is reconstructed by radial basis interpolation, while driving the three-dimensional scanning robot to perform an encrypted scan on the ring where the node is located to replace the missing signal with laser point cloud data. 3.The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, The construction process of the space-time-frequency domain feature matrix is: A correlation constraint model of the strain main frequency component and the environmental wind load harmonic component is established to force the strain main frequency amplitude and the wind load harmonic amplitude in the same ring position to satisfy a monotonic increasing relationship; The strain spectrum, wind load spectrum and deformation curvature data are reorganized into a space-time-frequency domain feature matrix according to time slicing, and a non-negative constraint decomposition is used to extract the coupling factor matrix of the spatial grid node and the frequency; Based on the coupling factor matrix, the energy focusing coefficient of each ring zone is calculated, and when the focusing coefficient of a certain ring zone in a specific frequency band exceeds a reference value and the deformation curvature suddenly changes, the ring zone is marked as a resonance risk hotspot.

4. The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, The strain spectrum, wind load spectrum and deformation curvature data are dimensionally fused into a space-time-frequency domain feature matrix, and the coupling characteristics of the spatial position and the frequency are extracted by constraint optimization decomposition, which at least includes the following steps: The marginal spectrum entropy of the real-time analysis of the library wall vibration signal is analyzed, and when the entropy value is lower than the threshold value, the wind load spectrum analysis frequency band is contracted to within the strain main frequency ± 0.5 octave; A strain-deformation time sequence generative adversarial network is constructed, the generator generates virtual high-frequency deformation data under the condition of low-frequency deformation curvature, and the discriminator introduces the library body vibration differential equation constraint network output; The Hausdorff distance between the generated deformation data and the measured deformation data is calculated, and the singular value decay rate of the coupling factor matrix is combined to output the credibility weight of the space-frequency domain feature; The energy focusing coefficient, deformation mutation gradient and feature credibility weight are fused, the resonance risk entropy value is calculated by entropy weight method, and when the value exceeds the dynamic threshold value, the encrypted scanning verification is triggered.

5. The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, The energy transfer path analysis based on the coupling characteristics locates the space-frequency energy focusing hotspot, including the following steps: Based on the stiffness attenuation coefficient of the spatial grid node and the ring weld continuity parameter, a directed topological network of the library body vibration energy transfer is constructed; The node stiffness attenuation coefficient is calculated from the steel plate thickness and the corrosion rate, and the weld continuity is evaluated by fitting the weld displacement gradient based on the three-dimensional deformation data; The resonance energy flow density of each grid node in the target frequency band is calculated, and the energy flow density is the product of the strain spectrum amplitude, the wind load power spectrum density and the node transfer efficiency; When the energy flow density of three consecutive nodes in the same ring exceeds 2 times the reference value, the ring is marked as an energy focusing hotspot; The natural frequency of the weld heat affected zone in the finite element model is called to verify whether the hotspot frequency band falls within the interval of ± 15% of the natural frequency; At the same time, it is detected whether the deformation curvature change rate of the hotspot ring increases by more than 3 times the value at the previous time.

6. The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, The resonance risk is verified in combination with the library structure weak point parameters, including the following steps: The instantaneous phase of the vibrating string sensor in the hotspot ring is extracted, and when the phase coherence of 80% of the sensors in the ring exceeds 85% and lasts for more than 5 seconds, it is determined that the structure is resonating; The energy focusing coefficient, phase coherence and deformation mutation gradient are fused, and the correlation entropy value of the three is calculated by nonlinear Granger causality test; When the correlation entropy value is lower than the threshold value and the phase coherence persists, output the resonance risk level; Drive the three-dimensional scanning robot to perform millisecond laser vibration measurement on the hot spot ring, and reconstruct the library wall vibration mode cloud picture; If the cloud picture shows that the standing wave node line passes through the circumferential weld, trigger a first-level warning and generate a weld fatigue life evaluation report.

7. The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, The output hierarchical warning and dynamic adjustment of the monitoring strategy, specifically includes: When the energy flow density of the energy focusing hot spot exceeds the dynamic threshold value; First-level response: start three-dimensional deformation encryption scanning, and the scanning frequency is increased to 5 times of the regular frequency; Second-level response: call the finite element model to check the resonance frequency matching degree, and if the deviation exceeds 15%, activate the weld damage prediction; Third-level response: when the energy flow density does not decay for 3 minutes, drive the consolidation cleaning robot to relieve the material pressure in the hot spot area; Calculate the deformation curvature obtained by encryption scanning and the finite element predicted deformation by Hausdorff distance; If the distance value is less than the deformation tolerance threshold, the warning level is downgraded; If the distance value increases and the weld damage probability exceeds 60%, output a second-level warning; Train the hidden Markov model based on historical warning data; When the energy flow density of the hot spot area decreases at a rate lower than the preset value; Automatically extend the encryption scanning time to twice the original plan and shrink the robot operation radius to within 3 meters. 8.The library health monitoring and early warning system based on heterogeneous multi-source data fusion according to claim 1, characterized in that, The output hierarchical warning and dynamic adjustment of the monitoring strategy, including the following steps: Input the encryption scanning data, robot operation record and energy flow density change rate into the digital twin, and dynamically correct the boundary constraint conditions and material constitutive parameters of the finite element model; When the second-level warning lasts for 10 minutes without being eliminated; Synchronously start the three-dimensional topography acquisition system to reconstruct the library wall topology model; And call the consolidation cleaning robot to perform preventive vibration on the hot spot ring; Statistical the structure relaxation coefficient after each warning is eliminated, which is defined as the ratio of the energy flow density decay rate to the deformation recovery rate; When the relaxation coefficient is greater than 1.5 times of the reference value for 3 consecutive times; Increase the dynamic threshold value by 20% and generate a structure stiffness strengthening scheme.

Images