Bridge digital twin data storage and tracing platform based on block chain

Abstract

The invention relates to the technical field of bridge monitoring data storage, in particular to a bridge digital twin data storage and tracing platform based on a block chain. According to the technical scheme, the system comprises a multi-mode tamper-proof acquisition module, a credible data preprocessing module, a component-level block chain evidence storage module, a data-driven twin modeling module and a judicial traceability verification module, wherein the credible data preprocessing module adopts an improved ICP algorithm to realize submillimeter registration of laser point cloud and a BIM model; and the component-level block chain evidence storage module realizes millisecond-level credible evidence storage based on a fragment chain architecture divided by bridge components. According to the method, through hardware fingerprint binding, multi-modal data fusion, fragment chain evidence storage, dynamic twin modeling and judicial-level evidence chain construction, whole-process credible management of bridge monitoring data from collection to tracing is realized, and data authenticity, fusion precision, evidence storage efficiency and model dynamics are remarkably improved; and a real-time and intelligent solution with judicial efficacy is provided for bridge health monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge monitoring data notarization technology, and in particular to a blockchain-based bridge digital twin data notarization and traceability platform. Background Art

[0002] As a vital component of national infrastructure, bridge structural health monitoring and safety assessment are crucial to safeguarding the lifeline of transportation. Traditional bridge monitoring technology primarily relies on discrete placement of single-point sensors, which has the following significant drawbacks:

[0003] Data authenticity risks. Existing sensor data is susceptible to electromagnetic interference or human tampering, lacks hardware-level anti-tampering mechanisms, and is difficult to meet judicial evidence requirements. For example, the raw data from traditional accelerometers and strain gauges is not tied to a unique physical identifier, making it difficult to trace the data source.

[0004] The accuracy of multi-source data fusion is insufficient, the registration error between laser point cloud and BIM model usually exceeds the millimeter level, and the spatiotemporal alignment of different sensors (such as vibration, strain, and environment) relies on manual calibration, which makes it difficult to adapt to dynamic deformation scenarios of bridges.

[0005] Evidence storage efficiency and scalability bottlenecks. Centralized databases present a single point of failure risk, and the need for real-time evidence storage of massive amounts of monitoring data (such as 256Hz vibration signals) poses a challenge to traditional blockchain consensus mechanisms. Existing PoW / PoS algorithms struggle to meet the millisecond-level response requirements for bridge safety warnings.

[0006] The digital twin model lacks dynamics, and the traditional finite element model relies on offline parameter calibration and cannot integrate the latest monitoring data in real time, resulting in the accumulation of deviations between the model prediction and the actual structural behavior over time, affecting the accuracy of the assessment.

[0007] The integrity of the judicial evidence chain is flawed. The existing monitoring system lacks evidence solidification methods that comply with the "Electronic Signature Law". The data collection, transmission, and evidence storage process has not formed a verifiable closed loop, and disputes over the effectiveness of evidence are often faced in legal disputes.

[0008] To summarize, we propose a blockchain-based bridge digital twin data storage and traceability platform. Summary of the Invention

[0009] The purpose of this invention is to address the technical bottlenecks in traditional bridge monitoring in terms of data authenticity, real-time performance, traceability and judicial compliance in the background technology, and to propose a blockchain-based bridge digital twin data storage and traceability platform.

[0010] The technical solution of the present invention is a blockchain-based bridge digital twin data storage and traceability platform, including a multimodal tamper-proof acquisition module, which ensures that the data source cannot be tampered with through hardware fingerprint binding and encrypted sensing technology;

[0011] Trusted data preprocessing module uses an improved ICP algorithm to achieve sub-millimeter registration between laser point clouds and BIM models;

[0012] The component-level blockchain evidence storage module implements millisecond-level trusted evidence storage based on the shard chain architecture divided by bridge components;

[0013] Data-driven twin modeling module, a dynamic model self-verification system that integrates finite element analysis and real-time monitoring data;

[0014] The judicial traceability verification module builds a judicial compliance evidence chain that includes a four-dimensional spatiotemporal index and electronic signature.

[0015] Optionally, the multimodal anti-tampering acquisition module includes:

[0016] Vibration fingerprint unit: STM32L4 series three-axis MEMS accelerometers are placed every 2 meters along the longitudinal direction of the box girder mid-span section. They are configured with an anti-aliasing Butterworth filter (cutoff frequency 100Hz, attenuation slope -40dB / dec). A 24-bit ADS1262 ADC chip is used to collect data at a 256Hz sampling rate. Each sensor has a built-in PUF chip to generate a unique hardware fingerprint.

[0017] Strain sensing unit: An FBG sensor array is deployed at 50cm intervals along the longitudinal direction of the main cable. A dual-wavelength reference method (1540nm / 1550nm) is used to compensate for temperature drift. A Micron Optics si425 interrogator is used to achieve a measurement accuracy of ±0.5με. The data frame header is embedded with spatial metadata containing longitude and latitude coordinates (WGS84) and altitude.

[0018] Crack Tracking Unit: A track-mounted inspection robot equipped with a Basler ace 2 Series line scan camera (5μm resolution) and a 785nm laser scanner moves along the crash barrier to detect surface cracks. An improved U-Net network is used to identify cracks in real time (with an accuracy of ±0.1mm), and the positioning data is bound to the BIM model coordinate system.

[0019] Environmental monitoring unit: A Gill WindMaster Pro three-dimensional ultrasonic anemometer is installed on the top of the bridge tower. It has a measurement range of 0-60m / s (accuracy of ±0.1m / s), and simultaneously collects temperature and humidity (±0.5℃) and atmospheric pressure (±0.1hPa) data. The PTPv2 protocol is used to achieve μs-level time synchronization.

[0020] Optionally, the trusted data preprocessing module includes:

[0021] Hardware trusted unit: implemented on edge computing nodes (NVIDIA Jetson AGX Xavier): Trusted Execution Environment (TEE) based on ARM TrustZone, sensor PUF fingerprint is bound to HMAC-SHA256 of collected data, and data transmission is encrypted using the SM4 national secret algorithm;

[0022] Data cleaning unit: vibration signal processing, three-level noise reduction:

[0023] Sliding average filter (window width 50ms) eliminates high-frequency noise;

[0024] Wavelet threshold denoising (sym5 wavelet, 5-layer decomposition);

[0025] Variational mode decomposition (VMD) extracts the first 10 order eigenmode functions;

[0026] Strain data processing: Moving average method (window width 30s) was used to eliminate transient interference;

[0027] Spatiotemporal alignment unit: Establishes a local coordinate system for the bridge (with the origin at the center of the abutment) and uses an improved ICP algorithm. This includes: registering the laser point cloud data with the BIM model (with an error of <2 mm), applying dynamic time warping (DTW) to align data with different sampling rates, and generating a standard data package with a spatiotemporal stamp (UNIX timestamp + WGS84 coordinates).

[0028] Optionally, the component-level blockchain evidence storage module includes:

[0029] The shard chain architecture unit specifically includes:

[0030] Stay cable shard chain: uses DPoS consensus (3 verification nodes) and stores vibration spectrum characteristics (0-100Hz, resolution 0.1Hz);

[0031] Main beam shard chain: uses PoA consensus (5 CA certified nodes) to store strain and displacement data (sampling interval 10s);

[0032] Pier shard chain: uses PoH consensus (generating timestamps based on sensor data) to store settlement monitoring data;

[0033] Cross-chain verification unit: Develop a cross-shard verification protocol based on threshold signatures: When cross-component data needs to be verified, three shard nodes are randomly selected to generate partial signatures, and the Shamir secret sharing scheme is used to reconstruct the full signature, with verification latency controlled within 500ms.

[0034] Smart contract units, specifically including:

[0035] Data traceability contract: Verify the three-level trust chain of sensor PUF → edge node signature → gateway certificate;

[0036] Feature Evidence Contract: Calculates the wavelet packet energy entropy (128-dimensional feature vector) of the vibration signal and stores it in the Merkle tree;

[0037] Early warning contract: When the main beam strain cumulative value ∑ε i When the value is ≥1500με, the three-level response mechanism is automatically triggered, among which, represents the algebraic sum of the strain measurements of n = 32 fiber Bragg grating sensors arranged on the main beam within the time window T = 1h, με is the microstrain unit, ε i is the i-th strain measurement value.

[0038] Optionally, the data-driven twin modeling module includes:

[0039] Multi-scale modeling unit, specifically including:

[0040] Macro model: BIM-based NURBS surface model (LOD500 accuracy, triangle size ≤ 5cm)

[0041] Mesoscopic model: Finite element mesh established using ABAQUS (element size 10 cm, number of nodes > 500,000) Microscopic model: Concrete aggregate model based on DEM (particle size distribution conforms to the Fuller curve)

[0042] Real-time update unit, development of CUDA accelerated solver: mapping the strain data stored in the blockchain to node loads, and using the Newton-Raphson iteration method to solve the equilibrium equation:

[0043] [K]{u}={F}

[0044] Where [K] is the stiffness matrix, {u} is the node displacement vector, and {F} is the node load vector. The stiffness matrix [K] is dynamically updated according to the real-time elastic modulus, and the full-field stress distribution is output every 10 seconds (accuracy ±0.1MPa).

[0045] Model self-verification unit: Weekly execution:

[0046] Comparison of measured frequency {f 实测} and the model calculated value {f 模型};

[0047] The Bayesian update algorithm is used to correct material parameters:

[0048] E new =E old +K(f 实测 -HEold )

[0049] Among them, E new is the updated elastic modulus, E old is the current elastic modulus, f 实测 is the measured frequency vector, K is the Kalman gain matrix, H is the observation matrix, HE old Predict frequencies for the model.

[0050] Optionally, the judicial retrospective verification module includes:

[0051] Four-dimensional index unit: Constructs a spatiotemporal R* tree index, specifically including:

[0052] Spatial dimension: Divide the bridge into a voxel grid of 1m×1m×0.1m;

[0053] Time dimension: Establish the time axis according to ISO 8601 standard (minimum granularity 1ms);

[0054] The index item contains: voxel coordinates + timestamp + data hash pointer;

[0055] Evidence chain construction units include:

[0056] Original data layer: sensor PUF fingerprint + SM3 hash value;

[0057] Transmission security layer: edge node digital signature + timestamp;

[0058] Evidence layer: Blockchain Merkle tree path proof;

[0059] Model layer: finite element calculation process traceability file;

[0060] The judicial output unit generates an evidence package that complies with the Electronic Signature Law: it is encapsulated in XML format, contains a trusted timestamp signed by the National Time Service Center, generates a timestamp token through the RFC3161 protocol, and is accompanied by a digital certificate chain issued by the CA organization.

[0061] Optionally, the wavelet threshold denoising in the data cleaning unit is specifically implemented as follows:

[0062] Use the sym5 wavelet basis to perform 5-layer decomposition and set the adaptive threshold:

[0063]

[0064] Among them, λ j is the threshold parameter of the j-th layer wavelet decomposition, σ j is the estimated value of the noise standard deviation of the detail coefficient of the jth layer, N is the number of data points of the detail coefficient of the jth layer, σ j Estimated based on the median of detail coefficients of each layer:

[0065]

[0066] Among them, cD j is the detail coefficient of the j-th layer wavelet decomposition, 0.6745 is the conversion coefficient between the median and standard deviation of the Gaussian distribution, and hard threshold processing is performed on the high-frequency coefficients above the third layer to retain the inter-layer correlation;

[0067] The variational mode decomposition (VMD) parameters are set as follows: eigenmode number K = 10; penalty factor α = 2000; convergence criterion ε = 1×10 -6 ; After decomposition, the energy proportion of the first five IMF components is extracted as the eigenvector.

[0068] Optionally, the shard chain architecture unit further includes a shard chain management unit, which adopts a Hybrid-BFT consensus algorithm, specifically including:

[0069] View switching protocol: Set a 200ms timeout threshold to trigger a view switch when the primary node fails;

[0070] Node reputation evaluation: Dynamically adjust voting weights based on historical verification success rates:

[0071]

[0072] in, is the reputation weight of node i at time t, N correct N is the number of transactions that have been successfully verified by node i in history. total is the total number of transactions that node i participates in verifying;

[0073] Random verification group selection: 21 verification nodes are selected from 50 candidate nodes using the VRF algorithm;

[0074] Byzantine fault tolerance: tolerates ≤ 7 malicious nodes (f = 7, n = 3f + 1 = 22);

[0075] The cross-shard verification protocol uses a (t,n)=(3,5) threshold signature. When a cross-shard transaction needs to be verified, the main chain coordinator generates a verification request hash H and randomly selects three shard nodes to generate a partial signature σ i =Sign(H,sk i ), reconstruct the complete signature by Lagrange interpolation Complete cross-shard transaction verification within 800ms, where σ i The partial signature generated for the i-th shard node, Sign is the digital signature algorithm, H is the hash value of the verification request, sk i is the private key of the i-th shard node, λ iis the Lagrange interpolation coefficient, satisfying ∑λ i =1.

[0076] Optionally, the CUDA acceleration solver in the real-time update unit is configured as follows:

[0077] Thread block partitioning: The finite element mesh is divided into 1024 thread blocks, with 256 threads per block;

[0078] Stiffness matrix compression storage: using CSR format, compression rate ≥ 85%;

[0079] Iterative convergence condition: Set the residual norm threshold ||R||2<1×10 -4 , ||R||2 is the second norm of the residual force vector;

[0080] The Newton-Raphson iteration implementation steps are:

[0081] Initial displacement field {u (0)} Load the latest state from the blockchain, calculate the residual force vector {R}={F}-[K]{u}, solve the linear system [K]{Δu}={R}, and update the displacement field {u (k+1)}={u (k)}+{Δu}, repeat until convergence, the maximum number of iterations is set to 20, each iteration takes ≤50ms, and the full solution completion time is ≤1s.

[0082] Optionally, the method for constructing the spatiotemporal R* tree of the four-dimensional index unit includes:

[0083] The node capacity is set to 50 entries, and the minimum bounding rectangle (MBR) expansion strategy is adopted;

[0084] Insertion algorithm optimization: prioritize the branch with the smallest MBR area increment and adopt a forced reinsertion mechanism;

[0085] Splitting strategy: Implement axial splitting, giving priority to splitting along the time dimension;

[0086] Query optimization: Create auxiliary hash indexes for common query patterns;

[0087] The finite element calculation process traceability file in the evidence chain construction unit includes:

[0088] Input data fingerprint: blockchain transaction ID + feature hash;

[0089] Solver configuration parameters: time step Δt = 0.1s, number of iterations k max =20, convergence threshold∈=1×10 -4 ;

[0090] Intermediate result verification: record the SHA3-256 summary of the residual norm every 10 iterations;

[0091] Output verification: The correlation coefficient ρ between the displacement field and the blockchain-stored strain data is ≥ 0.98.

[0092] Compared with the prior art, the present invention has the following beneficial technical effects:

[0093] Through physical fingerprint binding and encrypted sensing technology, monitoring data can be guaranteed to be tamper-proof from the source, meeting judicial evidence requirements.

[0094] Achieve sub-millimeter registration between laser point clouds and BIM models, eliminate multi-sensor sampling differences through spatiotemporal alignment technology, and build an accurate digital map of the bridge structure.

[0095] The millisecond-level evidence storage capability and cross-chain verification protocol based on the shard chain architecture support real-time on-chain uploading and rapid tracing of massive data, meeting the timeliness requirements of bridge safety warnings.

[0096] By integrating blockchain evidence data with finite element analysis, real-time updating and self-verification of model parameters can be achieved, continuously improving the prediction accuracy of bridge health status.

[0097] Build a complete chain of evidence including four-dimensional spatiotemporal index and electronic signature to ensure that the entire process of data collection, transmission and evidence storage is traceable and meets judicial compliance requirements.

[0098] Through hardware fingerprint binding, multimodal data fusion, shard chain evidence storage, dynamic twin modeling and judicial-grade evidence chain construction, the present invention realizes the trusted management of the entire process of bridge monitoring data from collection to traceability, significantly improving data authenticity, fusion accuracy, evidence storage efficiency and model dynamics, and providing a real-time and intelligent solution with judicial effectiveness for bridge health monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] Figure 1 The principle block diagram of the blockchain-based bridge digital twin data storage and traceability platform of the present invention is given. DETAILED DESCRIPTION

[0100] The technical solutions of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. It will be apparent that the described embodiments are some, but not all, embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the accompanying drawings herein may be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but rather merely represents selected embodiments of the present invention.

[0101] Example

[0102] The blockchain-based bridge digital twin data storage and traceability platform includes a multimodal tamper-proof acquisition module, a trusted data preprocessing module, a component-level blockchain storage module, a data-driven twin modeling module, and a judicial traceability verification module. Each module is described in detail below.

[0103] 1. In this embodiment, the multimodal tamper-proof acquisition module ensures that the data source cannot be tampered with through hardware fingerprint binding and encrypted sensing technology. The multimodal tamper-proof acquisition module includes:

[0104] Vibration fingerprint unit: STM32L4 series triaxial MEMS accelerometers are placed every 2 meters longitudinally along the mid-span section of the box girder. They are equipped with an anti-aliasing Butterworth filter (cutoff frequency 100Hz, attenuation slope -40dB / dec) and use a 24-bit ADS1262 ADC chip to acquire data at a 256Hz sampling rate. Each sensor has a built-in PUF chip to generate a unique hardware fingerprint. The sensor housing encapsulates an anti-tamper trigger circuit that immediately locks and generates an alarm in the event of unauthorized disassembly. Each frame of data is embedded with a PUF fingerprint, a UNIX timestamp (μs level), and GPS coordinates (WGS84) to ensure data traceability. Temperature compensation and a dual-wavelength reference method reduce temperature drift errors to ±2με, improving long-term monitoring stability. Through spatial correlation, strain data is bound to BIM model vertices, supporting three-dimensional stress field visualization.

[0105] Strain sensing unit: An FBG sensor array is deployed at 50cm intervals along the longitudinal direction of the main cable. A dual-wavelength reference method (1540nm / 1550nm) is used to compensate for temperature drift. A Micron Optics si425 interrogator is used to achieve a measurement accuracy of ±0.5με. The data frame header is embedded with spatial metadata containing longitude and latitude coordinates (WGS84) and altitude.

[0106] Crack Tracking Unit: A track-mounted inspection robot equipped with a Basler ace 2 Series line scan camera (5μm resolution) and a 785nm laser scanner moves along the crash barrier to detect surface cracks. An improved U-Net network is used to identify cracks in real time (with an accuracy of ±0.1mm), and the positioning data is bound to the BIM model coordinate system.

[0107] Environmental monitoring unit: A Gill WindMaster Pro three-dimensional ultrasonic anemometer is installed on the top of the bridge tower. It has a measurement range of 0-60m / s (accuracy of ±0.1m / s), and simultaneously collects temperature and humidity (±0.5℃) and atmospheric pressure (±0.1hPa) data. The PTPv2 protocol is used to achieve μs-level time synchronization.

[0108] The multimodal tamper-proof acquisition module ensures that data collection cannot be tampered with from the source through hardware fingerprint binding and encrypted sensing technology; multiple types of sensors are deployed in a coordinated manner (vibration, strain, cracks, and environment) to achieve all-round perception of the bridge structure, providing real and reliable raw data for subsequent analysis.

[0109] 2. In this embodiment, the trusted data preprocessing module uses an improved ICP algorithm to achieve submillimeter-level registration between the laser point cloud and the BIM model; the trusted data preprocessing module includes:

[0110] Hardware trusted unit: implemented on edge computing nodes (NVIDIA Jetson AGX Xavier): Trusted Execution Environment (TEE) based on ARM TrustZone, sensor PUF fingerprint is bound to HMAC-SHA256 of collected data, and data transmission is encrypted using the SM4 national secret algorithm;

[0111] Data cleaning unit: vibration signal processing, three-level noise reduction:

[0112] Sliding average filter (window width 50ms) eliminates high-frequency noise;

[0113] Wavelet threshold denoising (sym5 wavelet, 5-layer decomposition);

[0114] Variational mode decomposition (VMD) extracts the first 10 order eigenmode functions;

[0115] Strain data processing: The moving average method (window width 30s) is used to eliminate transient interference; the wavelet threshold denoising in the data cleaning unit is specifically implemented as follows:

[0116] Use the sym5 wavelet basis to perform 5-layer decomposition and set the adaptive threshold:

[0117]

[0118] Among them, λ j is the threshold parameter of the j-th layer wavelet decomposition, σ j is the noise standard deviation estimate of the j-th layer detail coefficient, N j is the number of data points of the detail coefficient of the jth layer, σ j Estimated based on the median of detail coefficients of each layer:

[0119]

[0120] Among them, cD j is the detail coefficient of the j-th layer wavelet decomposition, 0.6745 is the conversion coefficient between the median and standard deviation of the Gaussian distribution, and hard threshold processing is performed on the high-frequency coefficients above the third layer to retain the inter-layer correlation;

[0121] The variational mode decomposition (VMD) parameters are set as follows: the number of eigenmodes K = 10; the penalty factor α = 2000; the convergence criterion ε = 1×10 -6 ; After decomposition, the energy proportion of the first five IMF components is extracted as the eigenvector.

[0122] The spatiotemporal alignment unit establishes a local bridge coordinate system (with the abutment center as the origin) and employs an improved ICP algorithm. This includes registering the laser point cloud data with the BIM model (to an accuracy of <2 mm), applying dynamic time warping (DTW) to align data of varying sampling rates, and generating a standardized data package with spatiotemporal stamps (UNIX timestamps + WGS84 coordinates). Cross-modal data is fused in a unified spatiotemporal coordinate system, supporting four-dimensional (3D + time) state reconstruction.

[0123] The trusted data preprocessing module ensures data transmission security based on a trusted execution environment and national secret algorithms. It combines three-level noise reduction and spatiotemporal alignment technology to eliminate noise interference and unify the spatiotemporal benchmarks of multi-source data, significantly improving data quality and availability.

[0124] 3. Among them, the component-level blockchain evidence storage module realizes millisecond-level trusted evidence storage based on the shard chain architecture divided by bridge components; the component-level blockchain evidence storage module includes:

[0125] The shard chain architecture unit specifically includes:

[0126] Stay cable shard chain: uses DPoS consensus (3 verification nodes) and stores vibration spectrum characteristics (0-100Hz, resolution 0.1Hz);

[0127] Main beam shard chain: uses PoA consensus (5 CA certified nodes) to store strain and displacement data (sampling interval 10s);

[0128] Pier shard chain: uses PoH consensus (generating timestamps based on sensor data) to store settlement monitoring data;

[0129] The shard chain architecture unit also includes a shard chain management unit, which uses the Hybrid-BFT consensus algorithm, specifically including:

[0130] View switching protocol: Set a 200ms timeout threshold to trigger a view switch when the primary node fails;

[0131] Node reputation evaluation: Dynamically adjust voting weights based on historical verification success rates:

[0132]

[0133] in, is the reputation weight of node i at time t, N correctN is the number of transactions that have been successfully verified by node i in history. total is the total number of transactions that node i participates in verifying;

[0134] Random verification group selection: 21 verification nodes are selected from 50 candidate nodes using the VRF algorithm;

[0135] Byzantine fault tolerance: tolerates ≤ 7 malicious nodes (f = 7, n = 3f + 1 = 22);

[0136] The cross-shard verification protocol uses a (t,n)=(3,5) threshold signature. When a cross-shard transaction needs to be verified, the main chain coordinator generates a verification request hash H and randomly selects three shard nodes to generate a partial signature σ. i =Sign(H,sk i ), reconstruct the complete signature by Lagrange interpolation Complete cross-shard transaction verification within 800ms, where σ i The partial signature generated for the i-th shard node, Sign is the digital signature algorithm, H is the hash value of the verification request, sk i is the private key of the i-th shard node, λ i is the Lagrange interpolation coefficient, satisfying ∑λ i =1.

[0137] Cross-chain verification unit: Develop a cross-shard verification protocol based on threshold signatures: When cross-component data needs to be verified, three shard nodes are randomly selected to generate partial signatures, and the Shamir secret sharing scheme is used to reconstruct the full signature, with verification latency controlled within 500ms.

[0138] Smart contract units, specifically including:

[0139] Data traceability contract: Verify the three-level trust chain of sensor PUF → edge node signature → gateway certificate;

[0140] Feature Evidence Contract: Calculates the wavelet packet energy entropy (128-dimensional feature vector) of the vibration signal and stores it in the Merkle tree;

[0141] Early warning contract: When the main beam strain cumulative value ∑ε i When the value is ≥1500με, the three-level response mechanism is automatically triggered, among which, represents the algebraic sum of the strain measurements of n = 32 fiber Bragg grating sensors arranged on the main beam within the time window T = 1h, με is the microstrain unit, ε i is the i-th strain measurement value.

[0142] The component-level blockchain evidence storage module supports millisecond-level evidence storage of single components through a shard chain architecture, and the cross-chain verification protocol ensures the credibility of multi-component data associations; smart contracts automatically execute data traceability and early warning triggering to achieve efficient evidence storage and intelligent application of monitoring data.

[0143] 4. In addition, the data-driven twin modeling module integrates finite element analysis and a dynamic model self-verification system with real-time monitoring data; the data-driven twin modeling module includes:

[0144] Multi-scale modeling unit, specifically including:

[0145] Macro model: BIM-based NURBS surface model (LOD500 accuracy, triangle size ≤ 5cm)

[0146] Mesoscopic model: Finite element mesh established using ABAQUS (element size 10 cm, number of nodes > 500,000) Microscopic model: Concrete aggregate model based on DEM (particle size distribution conforms to the Fuller curve)

[0147] Real-time update unit, development of CUDA accelerated solver: mapping the strain data stored in the blockchain to node loads, and using the Newton-Raphson iteration method to solve the equilibrium equation:

[0148] [K]{u}={F}

[0149] Where [K] is the stiffness matrix, {u} is the node displacement vector, and {F} is the node load vector. The stiffness matrix [K] is dynamically updated according to the real-time elastic modulus, and the full-field stress distribution is output every 10 seconds (accuracy ±0.1MPa).

[0150] Model self-verification unit: Weekly execution:

[0151] Comparison of measured frequency {f 实测} and the model calculated value {f 模型};

[0152] The Bayesian update algorithm is used to correct material parameters:

[0153] E new =E old +K(f 实测 -HE old )

[0154] Among them, E new is the updated elastic modulus, E old is the current elastic modulus, f 实测 is the measured frequency vector, K is the Kalman gain matrix, H is the observation matrix, HE old Predict frequencies for the model.

[0155] The CUDA accelerated solver configuration in the real-time update unit is:

[0156] Thread block partitioning: The finite element mesh is divided into 1024 thread blocks, with 256 threads per block;

[0157] Stiffness matrix compression storage: using CSR format, compression rate ≥ 85%;

[0158] Iterative convergence condition: Set the residual norm threshold ||R||2<1×10 -4 , ||R||2 is the second norm of the residual force vector;

[0159] Newton-Raphson iteration implementation steps:

[0160] Initial displacement field {u (0)} Load the latest state from the blockchain, calculate the residual force vector {R}={F}-[K]{u}, solve the linear system [K]{Δu}={R}, and update the displacement field {u (k+1)}={u (k)} + {Δu}, repeating until convergence, with a maximum number of iterations set to 20. Each iteration takes ≤50ms, and the full solution takes ≤1s. By integrating blockchain-stored data with the finite element model, weekly parameter self-correction (Bayesian algorithm) is achieved, keeping the model prediction error within ±3%, significantly improving assessment accuracy. The data-driven twin modeling module integrates blockchain-stored data with finite element analysis, dynamically updating model parameters and self-verifying them, mapping bridge structural behavior in real time. Multi-scale modeling capabilities span the macro to micro levels, enhancing the accuracy and adaptability of digital twins.

[0161] 5. In this embodiment, the judicial retrospective verification module constructs a judicial compliance evidence chain including a four-dimensional spatiotemporal index and an electronic signature. The judicial retrospective verification module includes:

[0162] Four-dimensional index unit: Constructs a spatiotemporal R* tree index, specifically including:

[0163] Spatial dimension: Divide the bridge into a voxel grid of 1m×1m×0.1m;

[0164] Time dimension: Establish the time axis according to ISO 8601 standard (minimum granularity 1ms);

[0165] The index item contains: voxel coordinates + timestamp + data hash pointer;

[0166] Evidence chain construction units include:

[0167] Original data layer: sensor PUF fingerprint + SM3 hash value;

[0168] Transmission security layer: edge node digital signature + timestamp;

[0169] Evidence layer: Blockchain Merkle tree path proof;

[0170] Model layer: finite element calculation process traceability file;

[0171] The judicial output unit generates an evidence package that complies with the Electronic Signature Law: it is encapsulated in XML format, contains a trusted timestamp signed by the National Time Service Center, generates a timestamp token through the RFC3161 protocol, and is accompanied by a digital certificate chain issued by the CA organization.

[0172] It is worth noting that the spatiotemporal R* tree construction method of the four-dimensional index unit includes:

[0173] The node capacity is set to 50 entries, and the minimum bounding rectangle (MBR) expansion strategy is adopted;

[0174] Insertion algorithm optimization: prioritize the branch with the smallest MBR area increment and adopt a forced reinsertion mechanism;

[0175] Splitting strategy: Implement axial splitting, giving priority to splitting along the time dimension;

[0176] Query optimization: Create auxiliary hash indexes for common query patterns;

[0177] The finite element calculation process traceability file in the evidence chain construction unit includes:

[0178] Input data fingerprint: blockchain transaction ID + feature hash;

[0179] Solver configuration parameters: time step Δt = 0.1s, number of iterations k max =20, convergence threshold∈=1×10 -4 ;

[0180] Intermediate result verification: record the SHA3-256 summary of the residual norm every 10 iterations;

[0181] Output Verification: The correlation coefficient ρ between the displacement field and the strain data stored on the blockchain is ≥ 0.98. The four-dimensional spatiotemporal R*-tree index supports microsecond-level timestamp positioning. The XML evidence package complies with the requirements of the Electronic Signature Law and includes a trusted timestamp and CA certificate chain certified by the National Time Service Center. The judicial traceability verification module uses the four-dimensional spatiotemporal index to accurately locate data. The complete evidence chain (raw data → transmission → storage → model) ensures full traceability. The XML evidence package complies with judicial standards and includes a trusted timestamp and digital certificate, strengthening the legal validity of the data.

[0182] The above specific embodiments are merely several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art may make various alternative improvements and combinations to the above specific embodiments.

Claims

1. The blockchain-based bridge digital twin data storage and traceability platform is characterized by: include: Multimodal tamper-proof acquisition module ensures that the data source cannot be tampered with through hardware fingerprint binding and encrypted sensing technology; Trusted data preprocessing module uses an improved ICP algorithm to achieve sub-millimeter registration between laser point clouds and BIM models; The component-level blockchain evidence storage module implements millisecond-level trusted evidence storage based on the shard chain architecture divided by bridge components; Data-driven twin modeling module, a dynamic model self-verification system that integrates finite element analysis and real-time monitoring data; The judicial traceability verification module builds a judicial compliance evidence chain that includes a four-dimensional spatiotemporal index and electronic signature.

2. The blockchain-based bridge digital twin data storage and traceability platform according to claim 1 is characterized in that: The multimodal anti-tampering acquisition module includes: Vibration fingerprint unit: STM32L4 series three-axis MEMS accelerometers are placed every 2 meters along the longitudinal direction of the box girder mid-span section. They are configured with an anti-aliasing Butterworth filter and use a 24-bit ADS1262 ADC chip to collect data at a 256Hz sampling rate. Each sensor has a built-in PUF chip to generate a unique hardware fingerprint. Strain sensing unit: An FBG sensor array is deployed at 50cm intervals along the longitudinal direction of the main cable. A dual-wavelength reference method is used to compensate for temperature drift. A Micron Optics si425 demodulator is used to achieve a measurement accuracy of ±0.5με. The data frame header is embedded with spatial metadata containing latitude, longitude, and altitude coordinates. Crack Tracking Unit: A track-mounted inspection robot equipped with a Basler ace 2 Series line scan camera and a 785nm laser scanner moves along the crash barrier to detect surface cracks. An improved U-Net network is used to identify cracks in real time, and the positioning data is bound to the BIM model coordinate system. Environmental monitoring unit: A Gill WindMaster Pro three-dimensional ultrasonic anemometer is installed on the top of the bridge tower with a measurement range of 0-60m / s. It simultaneously collects temperature, humidity, and atmospheric pressure data and uses the PTPv2 protocol to achieve μs-level time synchronization.

3. The blockchain-based bridge digital twin data storage and traceability platform according to claim 2 is characterized in that: The trusted data preprocessing module includes: Hardware trusted unit: implemented in edge computing nodes: based on the ARM TrustZone trusted execution environment, the sensor PUF fingerprint is bound to the HMAC-SHA256 of the collected data, and data transmission is encrypted using the SM4 national secret algorithm; Data cleaning unit: vibration signal processing, three-level noise reduction: Sliding average filtering eliminates high-frequency noise; Wavelet threshold denoising; Variational mode decomposition is used to extract the first 10 order intrinsic mode functions; Strain data processing: using moving average method to eliminate transient interference; Spatiotemporal alignment unit: Establishes a local coordinate system for the bridge with the origin at the center of the abutment, and adopts an improved ICP algorithm. Specifically, it aligns the laser point cloud data with the BIM model, applies dynamic time warping to align data with different sampling rates, and generates a standard data package with time and space stamps.

4. The blockchain-based bridge digital twin data storage and traceability platform according to claim 3 is characterized in that: The component-level blockchain evidence storage module includes: The shard chain architecture unit specifically includes: Stay cable shard chain: uses DPoS consensus to store vibration spectrum characteristics; Main beam shard chain: uses PoA consensus to store strain and displacement data; Bridge pier shard chain: uses PoH consensus to store settlement monitoring data; Cross-chain verification unit: Develop a cross-shard verification protocol based on threshold signatures: When cross-component data needs to be verified, three shard nodes are randomly selected to generate partial signatures, and the Shamir secret sharing scheme is used to reconstruct the full signature, with verification latency controlled within 500ms. Smart contract units, specifically including: Data traceability contract: Verify the three-level trust chain of sensor PUF → edge node signature → gateway certificate; Feature Evidence Contract: Calculates the wavelet packet energy entropy of the vibration signal and stores it in the Merkle tree; Early warning contract: When the main beam strain cumulative value ∑ε i When the value is ≥1500με, the three-level response mechanism is automatically triggered, among which, It represents the algebraic sum of the strain measurements of n = 32 fiber Bragg grating sensors arranged on the main beam within the time window T = 1h, με is the microstrain unit, ε i is the i-th strain measurement value.

5. The blockchain-based bridge digital twin data storage and traceability platform according to claim 4 is characterized in that: The data-driven twin modeling module includes: Multi-scale modeling unit, specifically including: Macro model: NURBS surface model based on BIM; Mesoscopic model: finite element mesh was established using ABAQUS; Micro model: concrete aggregate model based on DEM; Real-time update unit, development of CUDA accelerated solver: mapping the strain data stored in the blockchain to node loads, and using the Newton-Raphson iteration method to solve the equilibrium equation: [K]{u}={F} Where [K] is the stiffness matrix, {u} is the node displacement vector, {F} is the node load vector, and the stiffness matrix [K] is dynamically updated according to the real-time elastic modulus, outputting the full-field stress distribution every 10 seconds. Model self-verification unit: Weekly execution: Comparison of measured frequency {f 实测 } and the model calculated value {f 模型 }; The Bayesian update algorithm is used to correct material parameters: HAVE BEEN new =E old +K(f 实测 -HE old ) Among them, E new is the updated elastic modulus, E old is the current elastic modulus, f 实测 is the measured frequency vector, K is the Kalman gain matrix, H is the observation matrix, HE old Predict frequencies for the model.

6. The blockchain-based bridge digital twin data storage and traceability platform according to claim 5 is characterized in that: The judicial retrospective verification module includes: Four-dimensional index unit: Constructs a spatiotemporal R* tree index, specifically including: Spatial dimension: Divide the bridge into a voxel grid of 1m×1m×0.1m; Time dimension: Establish a time axis according to ISO 8601 standard; The index item contains: voxel coordinates + timestamp + data hash pointer; Evidence chain construction units include: Original data layer: sensor PUF fingerprint + SM3 hash value; Transmission security layer: edge node digital signature + timestamp; Evidence layer: Blockchain Merkle tree path proof; Model layer: finite element calculation process traceability file; The judicial output unit generates an evidence package that complies with the Electronic Signature Law: it is encapsulated in XML format, contains a trusted timestamp signed by the National Time Service Center, generates a timestamp token through the RFC3161 protocol, and is accompanied by a digital certificate chain issued by the CA organization.

7. The blockchain-based bridge digital twin data storage and traceability platform according to claim 3 is characterized in that: The wavelet threshold denoising in the data cleaning unit is specifically implemented as follows: Use the sym5 wavelet basis to perform 5-layer decomposition and set the adaptive threshold: Among them, λ j is the threshold parameter of the j-th layer wavelet decomposition, σ j is the noise standard deviation estimate of the j-th layer detail coefficient, N j is the number of data points of the detail coefficient of the jth layer, σ j Estimated based on the median of detail coefficients of each layer: Among them, cD j is the detail coefficient of the j-th layer wavelet decomposition, 0.6745 is the conversion coefficient between the median and standard deviation of the Gaussian distribution, and hard threshold processing is performed on the high-frequency coefficients above the third layer to retain the inter-layer correlation; The variational mode decomposition parameters are set as follows: eigenmode number K = 10; penalty factor α = 2000; convergence criterion ε = 1×10 -6 ; After decomposition, the energy proportion of the first five IMF components is extracted as the eigenvector.

8. The blockchain-based bridge digital twin data storage and traceability platform according to claim 4 is characterized in that: The shard chain architecture unit also includes a shard chain management unit. The shard chain management unit adopts the Hybrid-BFT consensus algorithm, which specifically includes: View switching protocol: Set a 200ms timeout threshold to trigger a view switch when the primary node fails; Node reputation evaluation: Dynamically adjust voting weights based on historical verification success rates: in, is the reputation weight of node i at time t, N correct N is the number of transactions that have been successfully verified by node i in history. total is the total number of transactions that node i participates in verifying; Random verification group selection: 21 verification nodes are selected from 50 candidate nodes using the VRF algorithm; Byzantine fault tolerance: tolerates ≤7 malicious nodes; The cross-shard verification protocol uses a (t,n)=(3,5) threshold signature. When a cross-shard transaction needs to be verified, the main chain coordinator generates a verification request hash H and randomly selects three shard nodes to generate a partial signature σ i =Sign(H,sk i ), where the complete signature is reconstructed by Lagrange interpolation Complete cross-shard transaction verification within 800ms, where σ i The partial signature generated for the i-th shard node, Sign is the digital signature algorithm, H is the hash value of the verification request, sk i is the private key of the i-th shard node, λ i is the Lagrange interpolation coefficient, satisfying ∑λ i =1.

9. The blockchain-based bridge digital twin data storage and traceability platform according to claim 5 is characterized in that: The CUDA accelerated solver configuration in the real-time update unit is: Thread block partitioning: The finite element mesh is divided into 1024 thread blocks, with 256 threads per block; Stiffness matrix compression storage: using CSR format, compression rate ≥ 85%; Iterative convergence condition: Set the residual norm threshold ||R||2<1×10 -4 , ||R||2 is the second norm of the residual force vector; The Newton-Raphson iteration implementation steps are: Initial displacement field {u (0) } Load the latest state from the blockchain, calculate the residual force vector {R}={F}-[K]{u}, solve the linear system [K]{Δu}={R}, and update the displacement field {u (k+1) }={u (k) }+{Δu}, repeat until convergence, the maximum number of iterations is set to 20, each iteration takes ≤50ms, and the full-field solution completion time is ≤1s.

10. The blockchain-based bridge digital twin data storage and traceability platform according to claim 6 is characterized in that: The method for constructing a spatiotemporal R* tree of the four-dimensional index unit includes: The node capacity is set to 50 entries, and the minimum bounding rectangle (MBR) expansion strategy is adopted; Insertion algorithm optimization: prioritize the branch with the smallest MBR area increment and adopt a forced reinsertion mechanism; Splitting strategy: Implement axial splitting, giving priority to splitting along the time dimension; Query optimization: Create auxiliary hash indexes for common query patterns; The finite element calculation process traceability file in the evidence chain construction unit includes: Input data fingerprint: blockchain transaction ID + feature hash; Solver configuration parameters: time step Δt = 0.1s, number of iterations k max =20, convergence threshold∈=1×10 -4 ; Intermediate result verification: record the SHA3-256 summary of the residual norm every 10 iterations; Output verification: The correlation coefficient ρ between the displacement field and the blockchain-stored strain data is ≥ 0.98.

Images