Bridge monitoring data storage system based on cloud computing

Abstract

The application discloses a kind of bridge monitoring data storage systems based on cloud computing, specifically related to bridge monitoring technical field, including optical fiber sensing network acquisition unit, edge computing preprocessing unit, multi-modal data fusion storage module, dynamic hierarchical compression engine, fault-data correlation analysis unit and distributed cloud storage cluster.The application constructs full-link optimization system, front-end accurately captures key response, middle-end filters abnormal data to reduce invalid transmission, back-end is based on health rating dynamic compression and combined with cloud storage mechanism, so that storage efficiency is improved by more than 40%, solve storage cost and bandwidth pressure;By comparing and marking key data with knowledge graph, combined with fault-tolerant strategy, a triple protection system is constructed;Fusion index is formed by using space-time correlation, supporting multidimensional analysis, security and operation and maintenance are considered in authority control, cold and hot stratification optimizes storage cost, to provide high availability, scalable data management platform for bridge monitoring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bridge monitoring, in particular to a bridge monitoring data storage system based on cloud computing. BACKGROUND

[0002] In the field of bridge health monitoring, with the rapid development of sensor technology and the Internet of Things, the multi-modal data generated by bridge monitoring (such as strain, vibration, image, etc.) presents an explosive growth, and the traditional storage system faces problems such as high cost of mass data storage, low transmission efficiency, easy loss of feature data, and insufficient security protection. The existing scheme usually adopts fixed sampling frequency and unified compression strategy, which cannot accurately capture the key response of the bridge structure under different stress states, resulting in redundant storage of low-value data and possible loss of fault feature data due to excessive compression. At the same time, multi-source data (such as BIM model, sensor data, crack image) lack an efficient spatio-temporal correlation mechanism, making it difficult to support multi-dimensional joint analysis of structural damage.

[0003] In addition, the traditional storage architecture does not dynamically adjust the storage strategy for the health status of the bridge, and the mixed storage of hot and cold data leads to high response delay of high-frequency access data, and lacks fine-grained permission control and fault tolerance mechanism, making data availability and security difficult to meet long-term monitoring needs. Therefore, an intelligent storage system based on cloud computing is urgently needed to realize efficient management and deep mining of bridge monitoring data through dynamic sampling, hierarchical compression, spatio-temporal data fusion, and cloud distributed storage, and to improve the accuracy and timeliness of structural safety evaluation. SUMMARY

[0004] The main purpose of the present application is to provide a bridge monitoring data storage system based on cloud computing, which can effectively solve the problems in the background art.

[0005] To achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0006] A bridge monitoring data storage system based on cloud computing, which is configured to:

[0007] An optical fiber sensing network acquisition unit extracts bridge structure surface strain gradient distribution data, dynamically adjusts FBG sensor sampling frequency according to strain gradient, and generates original sensing signal stream;

[0008] An edge computing preprocessing unit calls the original sensing signal stream, performs vibration signal mutation feature extraction and timestamp alignment, and generates a preprocessed signal set;

[0009] A multi-modal data fusion storage module integrates the preprocessed signal set, bridge BIM model spatial coordinates, and crack image features, establishes a spatio-temporal correlation mapping, and generates a fusion storage index;

[0010] Dynamic hierarchical compression engine: monitor structural health state rating, dynamically select compression algorithm and data retention frequency band, output hierarchical compression data stream;

[0011] Fault-data correlation analysis unit: compare real-time spectral features with historical fault knowledge graph, mark key data storage nodes, and generate compression strategy optimization instructions;

[0012] Distributed cloud storage cluster: implement network delay-based erasure code sharding storage, perform hot and cold data tiered migration, and output monitoring data access interface.

[0013] Preferably, the optical fiber sensing network acquisition unit comprises:

[0014] The FBG sensor array is deployed according to the stress distribution topology of the bridge, and the sensor spacing is dynamically set to 0.5-1.2 meters according to the high strain gradient area of the finite element model;

[0015] Dual-wavelength demodulation module: contains a first channel with a center wavelength of 1520nm±2nm and a second channel with a center wavelength of 1550nm±2nm, and compensates for temperature drift error in real time through dual-channel differential calculation;

[0016] The flexible packaging structure has a thermal expansion coefficient matched with the solidification shrinkage coefficient of concrete, which is:

[0017] 1.0±0.2×10 -6 / ℃

[0018] Preferably, the edge computing preprocessing unit comprises:

[0019] The Daubechies-4 wavelet packet entropy value calculation algorithm is used to extract the mutation characteristics of the frequency of 50-300Hz in the vibration signal;

[0020] The abnormal data filter triggers full-data upload when the sampling frequency is 200-500Hz and the duration is 0.1-0.5s;

[0021] The timestamp synchronization module aligns multi-sensor data, and the synchronization accuracy is 10±2μs.

[0022] Preferably, the multi-modal data fusion storage module adopts a hierarchical storage architecture, wherein:

[0023] Real-time vibration data is stored in InfluxDB partition in time series format, and the sampling interval is 5-10ms;

[0024] After SIFT feature extraction, the structural crack image is stored in the Neo4j graph database to store the crack topology relationship;

[0025] The spatial mapping of the bridge BIM model and the sensing data is established, and the binding accuracy is ± 5cm three-dimensional coordinate data.

[0026] Preferably, the dynamic hierarchical compression engine performs compression according to the health state rating:

[0027] Health state (rating I): retain 1-5Hz characteristic frequency data;

[0028] Early warning state (rating II): retain 50-100Hz full-band data;

[0029] Fault state (rating III): original data lossless storage;

[0030] An improved LZW compression algorithm is adopted, and when the real-time calculated compression ratio reaches 10:1-15:1, the byte-level repetitive pattern detection module is started.

[0031] Preferably, the fault-data correlation analysis unit is built-in historical fault knowledge graph, containing 10-200Hz vibration frequency spectrum and crack propagation mode mapping relationship library;

[0032] When the fundamental frequency offset reaches 5-8%, the corresponding position 200±10Hz high-frequency vibration data is automatically associated and stored, and a compression strategy instruction is generated and fed back to the dynamic hierarchical compression engine, and the instruction update period is 5-30s.

[0033] Preferably, the distributed cloud storage cluster adopts erasure code fragmentation strategy (EC6+3), and the fragmentation size is dynamically adjusted to 64-256KB according to network delay;

[0034] Cold and hot data layering rules:

[0035] Hot data layer: SSD stores near 72±12 hours high-frequency monitoring data;

[0036] Cold data layer: HDD stores long-term trend data after feature extraction;

[0037] The data access interface integrates RBAC permission control, and supports single sensor node granularity authorization.

[0038] Compared with the prior art, the present application has the following beneficial effects:

[0039] 1、The application constructs a full-link optimization system from data acquisition to cloud storage, the front-end optical fiber sensing network dynamically adjusts the sampling frequency according to the bridge strain gradient, accurately captures the key structure response; the middle-end edge computing node filters abnormal vibration data in real time through the wavelet packet entropy algorithm, significantly reduces invalid transmission; the rear-end dynamic hierarchical compression engine adaptively switches the compression strategy based on the structure health rating, while retaining the fault feature frequency band, realizing high ratio compression. Combined with the hot and cold data layering mechanism of the distributed cloud storage cluster and the network-aware erasure code fragmentation technology, the overall storage efficiency of the system is improved by more than 40%, effectively solving the storage cost and bandwidth pressure brought by massive monitoring data.

[0040] 2、The application compares the fundamental frequency offset and high-frequency component in real time through the historical fault knowledge graph, actively marks the key data nodes of the high-risk period, and generates compression optimization instructions to dynamically control the compression engine, ensuring lossless storage of fault feature data; combined with the redundant sensor voting mechanism and EC6+3 erasure code fault tolerance strategy, a "fault identification-data protection-error repair" triple protection system is constructed. This design makes the key data availability reach 99.7%, greatly improving the timeliness and accuracy of structure safety diagnosis.

[0041] 3、The application uses spatial hashing algorithm to associate BIM spatial coordinates, vibration time series and crack topology relationship in space and time, forming a traceable "structure-response-damage" fusion index to support multi-dimensional joint analysis; the permission control system based on RBAC and OAuth2.0 realizes 6-level access authorization of single sensor node granularity, considering data security and flexible operation and maintenance; the cold and hot data layering uses a quantitative weight model to trigger intelligent migration, ensuring the real-time of high-frequency data while optimizing the long-term storage cost, and the system provides a high-availability, scalable integrated data management platform for bridge health monitoring. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is the overall flowchart of the application;

[0043] Figure 2 It is the flowchart of the dynamic hierarchical compression engine in the application. DETAILED DESCRIPTION

[0044] In order to make the technical means, creative features, purposes and effects of the application easy to understand, the application is further described below in conjunction with specific embodiments.

[0045] Example one, as shown in Figure 1 and Figure 2 , a bridge monitoring data storage system based on cloud computing, the bridge monitoring data storage system based on cloud computing is configured to:

[0046] The fiber sensing network acquisition unit extracts the bridge structure surface strain gradient distribution data, adjusts the FBG sensor sampling frequency based on the strain gradient, and generates the original sensing signal stream;

[0047] The edge computing preprocessing unit calls the original sensing signal stream, performs vibration signal mutation feature extraction and timestamp alignment, and generates a preprocessed signal set;

[0048] The multi-modal data fusion storage module integrates the preprocessed signal set, bridge BIM model spatial coordinates, and crack image features, establishes a spatio-temporal correlation mapping, and generates a fusion storage index;

[0049] The dynamic hierarchical compression engine monitors the structure health state rating, dynamically selects the compression algorithm and data retention frequency band, and outputs the hierarchical compression data stream;

[0050] The fault-data correlation analysis unit compares the real-time frequency spectrum features with the historical fault knowledge graph, marks the key data storage nodes, and generates compression strategy optimization instructions;

[0051] The distributed cloud storage cluster implements network delay-based erasure code sharding storage, performs hot and cold data hierarchical migration, and outputs the monitoring data access interface.

[0052] First, deploy the fiber sensing network, use a 48-channel FBG demodulator (model SM125) to build a distributed acquisition node. The sensor array is arranged at the bridge main beam web with a spacing of 0.8 meters. After verifying the strain gradient distribution through the finite element model, the sensor density in the high stress area is dynamically adjusted to 0.5 meters. The edge computing node uses the NVIDIA Jetson AGXXavier platform, and the Daubechies-4 wavelet packet algorithm is deployed. When the vibration signal frequency exceeds 50Hz is detected, the 500Hz high-speed sampling mode is started. When multi-modal data fusion is performed, the spatial hash algorithm is used to establish a mapping relationship between the BIM model coordinates (X=1024.35, Y=358.76, Z=+25.4) and the vibration data, with an error control within ±3cm. The fault analysis unit contains a SQLite database containing 200 groups of historical fault spectra. When a 6% base frequency shift is detected, the high-frequency data full retention instruction is triggered.

[0053] The fiber sensing network acquisition unit includes:

[0054] The FBG sensor array is deployed according to the bridge stress distribution topology, and the sensor spacing is dynamically set to 0.5-1.2 meters according to the high strain gradient area of the finite element model;

[0055] Dual-wavelength demodulation module: including a first channel with a center wavelength of 1520nm±2nm and a second channel with a center wavelength of 1550nm±2nm, and real-time compensation for temperature drift error is realized by dual-channel differential calculation;

[0056] The flexible packaging structure has a thermal expansion coefficient matched with a concrete solidification shrinkage coefficient, and is characterized in that:

[0057] 1.0±0.2×10 -6 / ℃

[0058] Further, the sensor package adopts a polyimide-silica gel composite structure, and has a thermal expansion coefficient α=1.05×10 -6 / ℃ (test standard GB / T20624.2);

[0059] In the dual-wavelength demodulation module, the 1520nm channel is used for strain measurement, and the 1550nm channel is used for temperature compensation, and the light intensity difference ΔP of the two channels is controlled in the range of 5-8dBm. When deployed, a redundant sensor group is arranged at the root of the pier, each group including three sensor nodes distributed at 120°, and single-point error is eliminated through a voting mechanism. The temperature compensation algorithm adopts:

[0060] Δε=k1·Δλ1-k2·Δλ2

[0061] Wherein, k1=0.78με / pm (strain coefficient), k2=0.05με / ℃ (temperature compensation coefficient).

[0062] The temperature drift compensation effect of the dual-wavelength demodulation module is verified by the actual bridge environment:

[0063] In the temperature difference range of-10℃ to 45℃, the strain measurement error is stabilized within ±2.5με (the error is ±15με without compensation), and the voting mechanism of the redundant sensor group adopts a two-out-of-three majority decision algorithm, so that the false alarm rate caused by single-point failure is reduced to 0.3%. After 2000 times of thermal cycle test , the sensor and the concrete peeling rate are less than 1%, which significantly improves the long-term monitoring stability. The measured data shows that the dynamic adjustment of the sampling frequency strategy improves the completeness of data capture in the high stress area by 35%, and reduces the redundant data volume in the low stress area by 30%.

[0064] The edge computing preprocessing unit comprises:

[0065] The Daubechies-4 wavelet packet entropy value calculation algorithm is adopted to extract the mutation characteristics of the frequency of 50-300Hz in the vibration signal;

[0066] The abnormal data filter triggers full-quantity data uploading when the sampling frequency is 200-500Hz and the duration is 0.1-0.5s;

[0067] The timestamp synchronization module aligns multi-sensor data with a synchronization accuracy of 10±2μs.

[0068] Further, the edge node configures a 512MB dedicated cache area, and when more than 3 vibration pulses above 300Hz are continuously detected, a full-data uploading mode is started.

[0069] The time synchronization module adopts PTPv2 protocol, and 6 time nodes are deployed in the optical fiber network to ensure that the cross-section data synchronization error is less than 8μs.

[0070] The anomaly filter sets two levels of threshold values: a primary threshold value triggers local data saving (vibration amplitude > 2.5g), and a secondary threshold value triggers cloud alarm (continuous overrun for 0.3s). The preprocessed data packet is added with a four-tuple label (timestamp | position code | data type | confidence), such as “202505291512 | B2-3-7 | VIB | 0.92”.

[0071] In the anomaly data filtering link, the wavelet packet entropy value calculation adopts:

[0072]

[0073] H is the entropy value (unit: bit), p i is the energy proportion of the i-th frequency band, N=8 is the decomposition layer number, and when H>4.2, it is determined as an abnormal signal, triggering full uploading.

[0074] Further, the 512MB dedicated cache area adopts a ring queue design, supporting a burst data peak throughput of 120MB / s (for 5 seconds), effectively avoiding data loss caused by network jitter. After the deployment of 6 PTPv2 time nodes in the time synchronization module, the synchronization error across a 200-meter bridge span is measured to be 6.8μs (better than the design index). The two-level threshold mechanism of the anomaly filter is verified by a vehicle flow impact test: when a heavy truck passes through the bridge, the primary threshold value is triggered to save the key waveform, and in the earthquake simulation shaking table test, the 0.3s continuous overrun triggers the cloud alarm, with a false trigger rate of <1 time / month. The detection rate of the entropy value calculation anomaly determination model (H>4.2) for crack propagation events is 92%, which is 27% higher than that of the traditional threshold method.

[0075] The multi-modal data fusion storage module adopts a hierarchical storage architecture, in which:

[0076] The real-time vibration data is stored in the InfluxDB partition in time series format, with a sampling interval of 5-10ms;

[0077] After the SIFT feature extraction of the structural crack image, the Neo4j graph database is used to store the crack topological relationship;

[0078] The spatial mapping between the bridge BIM model and the sensor data is established, and the binding accuracy is ±5 cm.

[0079] Specifically, the InfluxDB partition adopts a bridge span sharding strategy, and each bridge span is configured with an independent time series database instance.

[0080] During crack image processing, the SIFT feature extraction sets the key point threshold to 0.03, and at least 50 feature vectors are extracted for each crack. The spatial mapping service is deployed with a PostGIS extension module, and the bridge origin is set as the center point of the 0# pier (WGS84 coordinates: 40.7682°N, 114.8865°E) when establishing the three-dimensional coordinate system. The binding accuracy is optimized by the RANSAC algorithm. When constructing the crack topology relationship in the graph database, the "adjacent" relationship is defined as the edge distance <15 cm and the strike angle <30°.

[0081] The dynamic hierarchical compression engine performs compression according to the health status rating:

[0082] Health status (rating I): retain 1-5 Hz characteristic frequency data;

[0083] Warning state (rating II): retain 50-100 Hz full-band data;

[0084] Fault state (rating III): original data lossless storage;

[0085] An improved LZW compression algorithm is used, and when the real-time calculated compression ratio reaches 10:1-15:1, the byte-level repeated pattern detection module is started.

[0086] The compression engine is configured with a three-level strategy:

[0087] Level I state enables the improved LZW algorithm, sets the sliding window to 8 KB, and the dictionary size to 4K items;

[0088] Level II state switches to Zstd compression, and sets the compression level to 7;

[0089] Level III state enables lossless FLAC encoding.

[0090] The dynamic switching threshold sets a hysteresis interval, and the health status needs to be rated I for 5 minutes before enabling compression, while the warning state is switched immediately. The byte-level repeated detection module sets the minimum pattern length to 16 bytes, and when >3 repetitions are detected, the pattern replacement encoding is started.

[0091] Dynamic efficiency evaluation formula of improved LZW compression algorithm:

[0092]

[0093] R cReal-time compression rate(%) F raw Original data size(bytes) D size Dictionary item space W dict =4096 Maximum dictionary capacity. When R c Automatic switching of compression algorithm when below 12% for 10 seconds.

[0094] The fault-data correlation analysis unit is built-in historical fault knowledge graph, containing 10-200Hz vibration spectrum and crack propagation mode mapping relationship database;

[0095] When the fundamental frequency offset reaches 5%-8%, automatically associate 200±10Hz high-frequency vibration data at the corresponding position, generate compression strategy instructions and feedback to the dynamic hierarchical compression engine, instruction update period 5-30s.

[0096] Further, the knowledge graph is constructed by Neo4j graph database, and the nodes include: fault type (such as crack propagation), spectral feature (center frequency, harmonic component), and treatment measure three kinds of entities.

[0097] Correlation rule setting: when detecting 180-220Hz component and Q factor>12, automatically associate shear crack mode.

[0098] Optimization instruction generation period is dynamically adjusted, network delay<50ms, update every 5 seconds, delay>100ms, extend to 30 seconds.

[0099] Instruction encoding adopts TLV format, type field identifies instruction type (0xA1=data retention, 0xA2=compression adjustment).

[0100] Spectral feature correlation degree calculation model:

[0101]

[0102] Where, A f Real-time spectral amplitude H f Historical fault feature amplitude f is frequency(Hz). When S corr >0.85, generate optimization instruction, correlation confidence

[0103] Further, the knowledge graph is built into 300 groups of actual bridge fault cases, forming the association relationship of 182 spectral feature nodes and 47 crack modes. The spectral correlation degree model(Scorr>0.85) successfully marks the key data in the broken wire event of the cable-stayed cable, and the required data amount for subsequent analysis is reduced by 82%.

[0104] Instruction update period dynamic adjustment algorithm: based on the historical network delay median to predict the next period length, keep the instruction transmission success rate above 99.6%. TLV instruction field adds "urgency" identification (0-255), when detecting that the fundamental frequency offset is >7%, send 0xA1|0xFF highest priority instruction, ensure 200±10Hz data zero compression storage.

[0105] The distributed cloud storage cluster adopts erasure code fragmentation strategy (EC6+3), and the fragmentation size is dynamically adjusted to 64-256KB according to network delay;

[0106] Cold and hot data layering rules:

[0107] Hot data layer: SSD stores high-frequency monitoring data within 72±12 hours;

[0108] Cold data layer: HDD stores long-term trend data after feature extraction;

[0109] Data access interface integrates RBAC permission control, supports single sensor node granularity authorization.

[0110] Further, the erasure code fragmentation adopts Reed-Solomon (6,3) coding, and the fragmentation size is dynamically adjusted according to network quality: 256KB large fragmentation is used in good network (packet loss rate <0.1%), and 64KB fragmentation is used in poor network (packet loss rate >1%).

[0111] Hot data layer adopts RAID10 array, configures 4 pieces of NVMeSSD (Samsung PM1733), and cold data layer adopts 8-disk NAS (Seagate Exos 16TB).

[0112] Trigger condition quantification of hot and cold data migration:

[0113] W d =0.6×access frequency+0.4×fault correlation

[0114] W d >0.8: retained in SSD layer (high-frequency access within 72 hours);

[0115] W d <0.3: migrated to HDD layer (weekly aggregated storage);

[0116] Migration decision is executed every 2 hours (avoid I / O jitter)

[0117] Access interface realizes permission control based on OAuth2.0, defines 6-level access permissions:

[0118] Level 1 can only read vibration data, level 4 can access raw images, and level 6 has full data download permission.

[0119] Further, the erasure code fragmentation strategy combines network awareness: when the packet loss rate > 1%, automatically switch to EC4+2 mode and reduce the fragmentation to 64KB, reduce the transmission failure rate from 5.7% to 0.9%. The cold and hot data migration weight model (Wd) introduces a time decay factor: the access frequency is calculated by exponential decay (half-life period = 24h), avoiding long-term occupation of SSD by old data. RBAC permission control realizes field-level security: 4-level permission users can access the original image but hide the coordinate metadata, and 6-level users unlock all fields.

[0120] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A cloud computing based bridge monitoring data storage system characterized in that, The cloud computing-based bridge monitoring data storage system is configured to: The optical fiber sensing network acquisition unit extracts bridge structure surface strain gradient distribution data, dynamically adjusts the FBG sensor sampling frequency according to the strain gradient, and generates an original sensing signal stream; The edge computing preprocessing unit calls the original sensing signal stream, performs vibration signal mutation feature extraction and timestamp alignment, and generates a preprocessed signal set; The multi-modal data fusion storage module integrates the preprocessed signal set, bridge BIM model spatial coordinates, and crack image features, establishes a spatiotemporal correlation mapping, and generates a fusion storage index; The dynamic hierarchical compression engine monitors the structure health state rating, dynamically selects the compression algorithm and data retention frequency band, and outputs the hierarchical compression data stream; The fault-data correlation analysis unit compares real-time spectral features with historical fault knowledge graphs, marks key data storage nodes, and generates compression strategy optimization instructions; The distributed cloud storage cluster implements network delay-based erasure coding sharding storage, performs hot and cold data tiered migration, and outputs a monitoring data access interface; The multi-modal data fusion storage module adopts a hierarchical storage architecture, in which: Real-time vibration data is stored in InfluxDB partitions in time series format, with a sampling interval of 5-10 ms; After SIFT feature extraction, structural crack images are stored in a Neo4j graph database to store crack topological relationships; A spatial mapping of the bridge BIM model and sensing data is established, and three-dimensional coordinate data with an accuracy of ±5 cm are bound; The dynamic hierarchical compression engine performs compression according to the health state rating: Health state (rating I): retain 1-5 Hz characteristic frequency data; Warning state (rating II): retain 50-100 Hz full-band data; Fault state (rating III): original data is stored losslessly; An improved LZW compression algorithm is used, and when the real-time compression ratio reaches 10:1-15:1, a byte-level repetition pattern detection module is started.

2. The bridge monitoring data storage system based on cloud computing according to claim 1, wherein, The optical fiber sensing network acquisition unit includes: An FBG sensor array deployed according to the bridge stress distribution topology, with a sensor spacing dynamically set to 0.5-1.2 meters according to the high strain gradient region of the finite element model; A dual-wavelength demodulation module includes a first channel with a center wavelength of 1520 nm±2 nm and a second channel with a center wavelength of 1550 nm±2 nm, which compensates for temperature drift errors in real time through dual-channel differential calculation; A flexible packaging structure with a thermal expansion coefficient matched to the concrete solidification shrinkage coefficient: 1.0±0.2×10 -6 / ℃。 3. The bridge monitoring data storage system based on cloud computing according to claim 1, characterized in that: The edge computing preprocessing unit includes: A Daubechies-4 wavelet packet entropy value calculation algorithm is used to extract mutation features in the vibration signal with a frequency of 50-300 Hz; An abnormal data filter triggers full-quantity data upload when the sampling frequency is 200-500 Hz and the duration is 0.1-0.5 s; A timestamp synchronization module aligns multi-sensor data with a synchronization accuracy of 10±2 μs.

4. The cloud computing based bridge monitoring data storage system as claimed in claim 1, wherein: The fault-data correlation analysis unit has a built-in historical fault knowledge graph, which includes a mapping relationship library of 10-200 Hz vibration spectrum and crack propagation mode; When the base frequency offset reaches 5% to 8%, the automatic association storage stores the corresponding position 200±10Hz high-frequency vibration data, generates a compression strategy instruction feedback to the dynamic hierarchical compression engine, and the instruction update period is 5-30s.

5. The bridge monitoring data storage system based on cloud computing according to claim 1, characterized in that: The distributed cloud storage cluster adopts an erasure code fragmentation strategy (EC6+3), and the fragmentation size is dynamically adjusted to 64-256KB according to network delay; Cold and hot data layering rules: Hot data layer: SSD stores near 72±12 hours of high-frequency monitoring data; Cold data layer: HDD stores long-term trend data after feature extraction; The data access interface integrates RBAC permission control and supports single sensor node granularity authorization.

Images