Hoisting digital twin system and management method based on beidou sensing and intelligent fusion

Abstract

This invention discloses a hoisting digital twin system and management method based on BeiDou sensing and intelligent fusion, belonging to the field of intelligent construction technology. The system includes: a BeiDou intelligent positioning device for real-time acquisition of BeiDou positioning data and IMU attitude data of hoisted components; a hierarchical multi-error fusion and suppression unit for receiving the preprocessed data from the BeiDou intelligent positioning device and sequentially performing multi-path error decomposition, data fusion, and dynamic error compensation to output corrected BeiDou positioning and IMU attitude data; and a modular hoisting digital twin collaborative management and control platform for accessing and fusing the corrected BeiDou positioning and IMU attitude data, environmental data, and video data to drive the synchronous display of the BIM model. This invention solves the problems of low positioning accuracy and poor anti-interference capability in traditional hoisting systems, achieving process safety control and improved operational efficiency, and is suitable for hoisting operations of large bridge prefabricated components in complex environments such as sea crossings and mountainous areas.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction technology, specifically to a hoisting digital twin system and management method based on the fusion of BeiDou sensing and intelligence. Background Technology

[0002] The hoisting of large precast bridge components is a critical process that determines the bridge's alignment, structural safety, and construction progress. Its accuracy and efficiency directly affect the overall quality, cost, and risk control of the project. However, the current level of intelligence in hoisting operations still faces severe challenges. Existing technologies suffer from the following prominent bottlenecks: insufficient positioning and sensing capabilities; general-purpose BeiDou equipment exhibits poor signal stability in complex construction environments and lacks built-in preprocessing capabilities, resulting in low data real-time performance and difficulty in meeting the precise control requirements of dynamic hoisting; inefficient and simplistic error processing methods, often relying on traditional filtering techniques that can only handle random noise and are ineffective against complex error sources such as multipath interference, IMU drift, and atmospheric refraction, leading to centimeter-level fluctuations in positioning accuracy and failing to meet millimeter-level installation requirements; weak practicality of the twin management platform, focusing on model visualization and disconnected from the physical process; high costs associated with model solidification and adaptation, and a lack of closed-loop decision-making capabilities linked to error suppression and safety early warning; and poor overall process coordination, with isolated stages from data acquisition and processing to monitoring, resulting in delayed decision-making and execution, making full-process collaborative management difficult to achieve.

[0003] Therefore, there is an urgent need for an integrated hoisting technology solution that is highly resistant to interference and can achieve intelligent collaboration throughout the entire process. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address the shortcomings of the existing technology by providing a hoisting digital twin system and management method based on the fusion of Beidou sensing and intelligence.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] The hoisting digital twin system based on BeiDou sensing and intelligence fusion includes BeiDou intelligent positioning equipment interconnected through standardized data interfaces, a hierarchical multi-error fusion suppression unit, and a modular hoisting digital twin collaborative management and control platform;

[0007] Among them, the Beidou intelligent positioning equipment is used to collect Beidou positioning data and IMU attitude data of hoisted components in real time, and to perform preprocessing and local coordinate transformation;

[0008] The hierarchical multi-error fusion suppression unit is used to receive the preprocessed data from the Beidou intelligent positioning device, and sequentially perform multipath error decomposition, multi-source data fusion, and dynamic error compensation to output corrected Beidou positioning and IMU attitude data. The multipath error decomposition adopts the CEEMDAN-sPCA algorithm, the multi-source data fusion adopts the adaptive Kalman filter algorithm, and the dynamic error compensation integrates the LSTM error prediction model and the NTP clock synchronization mechanism.

[0009] The modular hoisting digital twin collaborative management and control platform is used to access and integrate corrected BeiDou positioning and IMU attitude data, environmental data and video data, drive the BIM model to display synchronously, and trigger closed-loop adjustment instructions or emergency responses according to preset rules. At the same time, it feeds back real-time hoisting condition information to the BeiDou intelligent positioning device for adaptive adjustment of preprocessing parameters.

[0010] Furthermore, the Beidou intelligent positioning device integrates a multi-mode positioning unit, an anti-interference communication unit, an environment adaptation unit, and a data preprocessing unit;

[0011] The multi-mode positioning unit integrates a Beidou RTK chip and an inertial measurement unit (IMU) for synchronously acquiring positioning and IMU attitude data.

[0012] The anti-interference communication unit integrates a dual-link redundant transmission module of 4G / 5G network and dedicated radio, and supports local data caching and breakpoint resume.

[0013] The environmental adaptation unit integrates a temperature and humidity sensor and an adaptive temperature control device. When the temperature or humidity exceeds a preset threshold, it automatically activates a heat dissipation or moisture-proof mechanism.

[0014] The data preprocessing unit integrates an ARM-based processor and is used to remove abrupt outliers using a sliding window algorithm.

[0015] Furthermore, the hierarchical multi-error fusion suppression unit includes a multi-path error decomposition module, a multi-source data fusion module, and a dynamic error compensation module;

[0016] The multipath error decomposition module uses the CEEMDAN-sPCA algorithm to decompose and reconstruct the original BeiDou positioning data in order to suppress multipath interference.

[0017] The multi-source data fusion module uses an adaptive Kalman filter algorithm to fuse the decomposed BeiDou positioning data, IMU attitude data, and environmental wind speed data, and dynamically adjusts the fusion weights of each data source according to the IMU zero bias state.

[0018] The dynamic error compensation module integrates an LSTM error prediction model and an NTP clock synchronization mechanism to predict and compensate for satellite clock errors and atmospheric refraction errors.

[0019] Furthermore, the modular hoisting digital twin collaborative management and control platform adopts a four-layer architecture, including:

[0020] The data access layer is used to access multi-source sensor data in real time via the MQTT protocol, and supports data compression, encryption and offline caching;

[0021] The model computation layer has a built-in lightweight BIM model engine and LOD technology to drive the synchronization between the virtual model and physical components, and realize the virtual-real mapping.

[0022] The interactive display layer adopts a modular card layout to provide real-time visual display of the entire hoisting process status, deviation data, and monitoring videos.

[0023] The emergency response layer has a built-in three-level early warning mechanism and emergency plan library, which can automatically trigger early warnings, generate adjustment instructions, or lock device permissions based on the deviation level.

[0024] The hoisting digital twin management method based on BeiDou sensing and intelligent fusion, implemented based on any one of the hoisting digital twin systems based on BeiDou sensing and intelligent fusion, includes:

[0025] Step S1: Deploy Beidou intelligent positioning equipment on the hoisting component and initialize the coordinate system and equipment;

[0026] Step S2: The Beidou intelligent positioning device collects Beidou positioning data and IMU attitude data in real time and preprocesses them, and transmits them to the hierarchical multi-error fusion suppression unit through an anti-interference communication link;

[0027] Step S3: The hierarchical multi-error fusion suppression unit performs hierarchical error suppression processing on the received BeiDou positioning data and IMU attitude data to obtain the corrected BeiDou positioning and IMU attitude data.

[0028] Step S4: The modular hoisting digital twin collaborative management and control platform accesses the corrected Beidou positioning and IMU attitude data, driving the digital twin model to perform synchronous display and status monitoring;

[0029] Step S5: When the deviation between the monitoring data and the design value exceeds the threshold, an adjustment instruction is automatically generated and sent to the actuator, forming a closed loop of perception-processing-control-execution-feedback.

[0030] Furthermore, in step S2, the preprocessing specifically includes the following steps:

[0031] Based on real-time hoisting condition information from the modular hoisting digital twin collaborative management and control platform, a dynamic threshold for data jump judgment is dynamically calculated; the condition information includes hoisting stage identifier, hoisting rope inclination angle, hoisting rope length, and the historical steady movement speed of the component;

[0032] Based on the dynamic threshold, within a unified sliding time window, collaborative analysis and filtering are performed on BeiDou positioning data and IMU attitude data. Specifically, this includes: when the instantaneous change in BeiDou positioning data exceeds the dynamic threshold, the attitude data output by the IMU at the corresponding time is simultaneously verified; if the motion state reflected by the IMU attitude data does not exceed the preset state change threshold, the BeiDou positioning data is determined to have jumped to an abnormal value and is removed.

[0033] Furthermore, in step S3, the hierarchical error suppression processing specifically includes:

[0034] The CEEMDAN-sPCA algorithm is used to decompose and suppress multipath interference components in BeiDou positioning data.

[0035] Adaptive Kalman filtering is used to dynamically fuse BeiDou positioning, IMU attitude and wind speed data, and data weights are dynamically allocated according to the hoisting stage;

[0036] The LSTM error prediction model is used to predict systematic errors, and dynamic compensation is performed by combining the NTP synchronization time.

[0037] Furthermore, the error suppression process in step S3 also includes a thermoelastic deformation compensation step, specifically including:

[0038] Based on the component material properties, structural parameters, hoisting constraints, and real-time temperature gradient, the thermoelastic deformation compensation is calculated using a dedicated physical model. The specific formula is as follows:

[0039] ;

[0040] in, This represents the amount of thermoelastic deformation compensation of the component along the positioning direction, and α represents the linear expansion coefficient of the bridge component material, which is preset based on the component material and updated in real time through a cloud database. This indicates the design length between component positioning reference points, which is retrieved in real-time from the hoisting digital twin model. This represents the real-time temperature gradient of the component, i.e., the difference between the component surface temperature collected by environmental sensors and the design reference temperature. This represents the elastic modulus of the component material. Representing temperature stress, through Calculate, where k represents the constraint coefficient and μ represents the Poisson's ratio of the component material. Indicates the length of the constraint feature.

[0041] Furthermore, in the hierarchical error suppression processing, the LSTM error prediction model is an adaptive HA-LSTM model for hoisting conditions, specifically including:

[0042] Based on K-means clustering of historical hoisting data, the corresponding initial weight matrix of LSTM network is generated according to the hoisting, translation and docking conditions;

[0043] Introduce an adjustment coefficient related to the operating condition indicator signal into the forget gate mechanism;

[0044] An attention mechanism unit is added to the hidden layer to assign higher attention weights to the input data corresponding to sudden changes in wind speed or sudden changes in boom swing angle.

[0045] Furthermore, in the hierarchical error suppression process, the adaptive Kalman filter employs a stage-environment coupled adjustment strategy, specifically including:

[0046] The weighting ratio of the fusion of BeiDou positioning data, IMU attitude data, and environmental data is dynamically allocated according to the hoisting stage.

[0047] The adjustment step size of the filter gain is coupled with the current hoisting stage and the real-time wind speed level.

[0048] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0049] 1. This invention employs an innovative decomposition-fusion-compensation hierarchical error suppression system, comprehensively utilizes the CEEMDAN-sPCA algorithm to suppress multipath interference, adopts a stage-environment coupled adaptive Kalman filter to dynamically optimize data fusion weights, introduces an LSTM model to predict and compensate for system errors, and combines a thermoelastic deformation compensation model to effectively offset the influence of temperature stress. Ultimately, it stabilizes and improves the overall positioning accuracy to the millimeter level, solving the problem of accuracy fluctuation.

[0050] 2. The Beidou intelligent positioning device of the present invention adopts a scenario-based deep integration design, integrates multi-mode sensors, realizes data spatiotemporal alignment through a built-in preprocessing unit, and combines IP67 protection, shockproof buffering and adaptive temperature control mechanism, as well as dual-link redundant communication of 4G / 5G and dedicated radio, to ensure continuous and reliable data transmission in harsh environments such as high salt spray at sea and strong electromagnetic interference in mountainous areas. The overall environmental adaptability of the device is more than 10 times better than that of general equipment.

[0051] 3. The modular hoisting digital twin collaborative management and control platform of the present invention is based on a four-layer architecture. It achieves rapid loading and smooth rendering of large BIM models through a lightweight engine and LOD technology. The platform deeply integrates error suppression units and a three-level emergency response mechanism. It can automatically generate adjustment instructions based on real-time deviations to drive the actuators, realizing a leap from visualization to closed-loop management and control of decision-making and execution.

[0052] 4. This invention eliminates the safety risks of high-altitude operations through remote and precise control with human-machine isolation. At the same time, the intelligent collaboration of the entire process improves the hoisting efficiency of single components and reduces the rework rate. All process data is automatically archived and connected to the full life cycle management platform to form a complete digital archive, realizing full traceability of the operation process and continuous optimization of the process. Attached Figure Description

[0053] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0054] Figure 1 This is a system schematic diagram according to an embodiment of the present invention;

[0055] Figure 2 This is a flowchart illustrating an embodiment of the present invention;

[0056] Figure 3 This is a flowchart of the hierarchical multi-error fusion suppression unit in an embodiment of the present invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] To facilitate understanding of this invention, the meanings of the English abbreviations used herein are explained as follows:

[0059] IMU: Inertial Measurement Unit, used to measure the three-axis attitude angles and acceleration of an object.

[0060] BIM: Building Information Modeling, is a system that integrates information related to building projects based on three-dimensional digital technology.

[0061] RTK: Real-Time Kinematic, a real-time dynamic differential positioning technology based on carrier phase observations.

[0062] LSTM: Long Short-Term Memory, a special type of recurrent neural network suitable for processing and predicting time series data.

[0063] CEEMDAN: Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, used for the decomposition of nonlinear and non-stationary signals.

[0064] sPCA: Sparse Principal Component Analysis, is a feature extraction method that introduces sparse constraints on top of PCA.

[0065] LOD: Levels of Detail, is a rendering optimization technique that dynamically adjusts the model's detail based on the distance of objects from the viewpoint.

[0066] NTP: Network Time Protocol, a protocol used to synchronize time in computer networks.

[0067] MQTT: Message Queuing Telemetry Transport, a lightweight publish / subscribe message transport protocol.

[0068] ARM: Advanced RISC Machines, an advanced reduced instruction set computer processor architecture.

[0069] IP67: Ingress Protection Rating 67, a protection rating code indicating complete dust protection (level 6) and the ability to withstand immersion in 1 meter of water for 30 minutes without damage (level 7).

[0070] GNSS: Global Navigation Satellite System, including BeiDou, GPS, GLONASS, Galileo, etc.

[0071] HA-LSTM: Hierarchical Attention-based Long Short-Term Memory, is an improved model that introduces attention mechanisms and hierarchical structures on the basis of LSTM.

[0072] K-means: K-means clustering is a partition-based clustering method.

[0073] LZ4: LZ4 compression algorithm, a lossless data compression algorithm known for its fast compression and decompression speeds.

[0074] AES: Advanced Encryption Standard, a symmetric-key encryption algorithm.

[0075] CGCS2000: China Geodetic Coordinate System 2000, the national geodetic coordinate system currently used in China.

[0076] BDS-3: BeiDou Navigation Satellite System Phase 3.

[0077] GPS: Global Positioning System, the U.S. satellite navigation system.

[0078] GLONASS: Global Navigation Satellite System, the Russian satellite navigation system.

[0079] Galileo: Galileo Navigation Satellite System, the European Union's satellite navigation system.

[0080] like Figure 1 As shown, the hoisting digital twin system based on BeiDou sensing and intelligent fusion includes BeiDou intelligent positioning equipment interconnected through standardized data interfaces, a hierarchical multi-error fusion suppression unit, and a modular hoisting digital twin collaborative management and control platform.

[0081] Among them, the Beidou intelligent positioning equipment is used to collect Beidou positioning data and IMU attitude data of hoisted components in real time, and to perform preprocessing and local coordinate transformation;

[0082] The hierarchical multi-error fusion suppression unit is used to receive the preprocessed data from the Beidou intelligent positioning device, and sequentially perform multipath error decomposition, multi-source data fusion, and dynamic error compensation to output corrected Beidou positioning and IMU attitude data. The multipath error decomposition adopts the CEEMDAN-sPCA algorithm, the multi-source data fusion adopts the adaptive Kalman filter algorithm, and the dynamic error compensation integrates the LSTM error prediction model and the NTP clock synchronization mechanism.

[0083] The modular hoisting digital twin collaborative management and control platform is used to access and integrate corrected BeiDou positioning and IMU attitude data, environmental data and video data, drive the BIM model to display synchronously, and trigger closed-loop adjustment instructions or emergency responses according to preset rules. At the same time, it feeds back real-time hoisting condition information to the BeiDou intelligent positioning device for adaptive adjustment of preprocessing parameters.

[0084] The Beidou intelligent positioning device integrates a multi-mode positioning unit, an anti-interference communication unit, an environment adaptation unit, and a data preprocessing unit.

[0085] The multi-mode positioning unit integrates a Beidou RTK chip and an inertial measurement unit (IMU) for synchronously acquiring positioning and IMU attitude data.

[0086] The anti-interference communication unit integrates a dual-link redundant transmission module of 4G / 5G network and dedicated radio, and supports local data caching and breakpoint resume.

[0087] The environmental adaptation unit integrates a temperature and humidity sensor and an adaptive temperature control device. When the temperature or humidity exceeds a preset threshold, it automatically activates a heat dissipation or moisture-proof mechanism.

[0088] The data preprocessing unit integrates an ARM-based processor and is used to remove abrupt outliers using a sliding window algorithm.

[0089] The hierarchical multi-error fusion suppression unit includes a multi-path error decomposition module, a multi-source data fusion module, and a dynamic error compensation module;

[0090] The multipath error decomposition module uses the CEEMDAN-sPCA algorithm to decompose and reconstruct the original BeiDou positioning data in order to suppress multipath interference.

[0091] The multi-source data fusion module uses an adaptive Kalman filter algorithm to fuse the decomposed BeiDou positioning data, IMU attitude data, and environmental wind speed data, and dynamically adjusts the fusion weights of each data source according to the IMU zero bias state.

[0092] The dynamic error compensation module integrates an LSTM error prediction model and an NTP clock synchronization mechanism to predict and compensate for satellite clock errors and atmospheric refraction errors.

[0093] The modular hoisting digital twin collaborative management and control platform adopts a four-layer architecture, including:

[0094] The data access layer is used to access multi-source sensor data in real time via the MQTT protocol, and supports data compression, encryption and offline caching;

[0095] The model computation layer has a built-in lightweight BIM model engine and LOD technology to drive the synchronization between the virtual model and physical components, and realize the virtual-real mapping.

[0096] The interactive display layer adopts a modular card layout to provide real-time visual display of the entire hoisting process status, deviation data, and monitoring videos.

[0097] The emergency response layer has a built-in three-level early warning mechanism and emergency plan library, which can automatically trigger early warnings, generate adjustment instructions, or lock device permissions based on the deviation level.

[0098] Example 1: Modular Implementation of BeiDou High-Precision Intelligent Positioning Equipment

[0099] This equipment adopts a modular integrated design, with overall dimensions of 135mm × 135mm × 70mm, an IP67 protection rating, an operating temperature range of -30℃ to +60℃, and resistance to a 2m drop on concrete. The specific functional unit design is as follows:

[0100] Multi-mode positioning unit: Equipped with a Beidou-3 GNSS RTK chip, it supports multi-frequency signal reception from the entire system, including BDS-3 (B1I / B2I / B3I bands), GPS (L1 / L2 / L5 bands), GLONASS, and Galileo. The positioning accuracy reaches ±2mm+1ppm in the horizontal plane and ±8mm+1ppm in the vertical plane. The data update frequency is adjustable from 5 to 20Hz, meeting the real-time attitude capture requirements of dynamic hoisting.

[0101] Anti-interference communication unit: It integrates a dual-channel redundant transmission module of 4G / 5G wireless network and 410-470MHZ dedicated radio, supports 8GB local data caching and breakpoint resume function - when the main link (4G / 5G) signal is interrupted, it automatically switches to the backup radio link, and the data transmission interruption time is ≤0.5 seconds, ensuring communication stability in complex environments.

[0102] Environmental Adaptation Unit: It adopts a high-strength aluminum alloy shell, with a built-in shock-absorbing structure (which can withstand vibrations below 1000HZ) and dustproof and waterproof sealing strips. It integrates temperature and humidity sensors. When the temperature is >60℃ or the humidity is >90%RH, it automatically starts the cooling fan / moisture-proof module, making it suitable for high salt spray at sea and high temperature difference in mountainous areas.

[0103] Data preprocessing unit: Built-in low-power ARM Cortex-M4 microprocessor, using sliding window algorithm to remove outliers (outlier identification accuracy ≥98%), converting geodetic coordinates (CGCS2000) into construction local coordinate system data, with data preprocessing latency ≤10ms, reducing the processing pressure on the backend system.

[0104] Example 2: Design of a hierarchical multi-error fusion suppression unit:

[0105] Multipath error decomposition module: The CEEMDAN-sPCA (Adaptive Noise Complete Set Empirical Mode Decomposition-Sparse Principal Component Analysis) algorithm is used to decompose BeiDou positioning data into "effective signal component - multipath interference component - noise component". The effective signal is extracted through sparse principal component analysis. The multipath interference suppression rate is ≥92%, which solves the positioning deviation problem caused by component occlusion and ground reflection.

[0106] Multi-source data fusion module: Employs an adaptive Kalman filter algorithm to fuse pre-processed BeiDou positioning data, IMU attitude data, and wind speed sensor data—dynamically adjusting the filter gain to ensure the smoothness and continuity of the fused position-attitude data;

[0107] Dynamic error compensation module: Integrates NTP (Network Time Protocol) clock synchronization mechanism (clock synchronization accuracy ≤1ms) and LSTM (Long Short-Term Memory) error prediction model. The model is trained using historical data (hoisting data from the past month) to predict satellite clock error, ionospheric / tropospheric refraction error and equipment installation deviation in real time, and generates dynamic compensation values ​​(compensation accuracy ≤0.3mm) to further correct the positioning data and ensure that the final accuracy is stable at the millimeter level.

[0108] Example 3: Construction of a Modular Lifting Digital Twin Collaborative Management and Control Platform:

[0109] It adopts a four-layer modular architecture, which supports flexible configuration and rapid adaptation to different scenarios. The specific layer design is as follows:

[0110] Data access layer: Supports access from multiple sources such as Beidou positioning devices, IMU, anemometers, and high-definition cameras; uses the MQTT protocol to achieve real-time data transmission (transmission latency ≤500ms); integrates LZ4 data compression (compression rate ≥50%) and AES-128 encryption; supports offline data caching (cache capacity ≥10GB); automatically caches data when the network is interrupted and resumes transmission later.

[0111] Model Calculation Layer: Built-in lightweight BIM model engine (supports Revit / 3ds Max model import, model compression rate ≥70%), adopts LOD (Level of Detail) technology—for long-distance (>50m) display, it calls low-precision models with ≤100,000 faces, and for short-distance (≤10m) display, it loads high-precision models with ≥1 million faces, with a rendering frame rate ≥30FPS; based on the seven-parameter Bursa coordinate transformation model and 4×4 rigid body transformation matrix, it drives the virtual model and physical components to move synchronously, with a virtual-real synchronization delay ≤1 second;

[0112] Interactive Display Layer: Adopts a modular card layout, including: "Time and Weather Module (displays real-time date, temperature, and wind speed) - Video Monitoring Module (supports real-time viewing / playback) - Structure View Module (supports top / front view for viewing hoisting deviations) - Status Monitoring Module (displays yaw angle, roll angle, pitch angle, and hoisting rope tension) - Positioning Data Module (displays design values / measured values / differences, with difference accuracy ≤ 0.1mm) - Process Display Module (displays the current process)"; supports mouse drag and zoom operations, with a visualization response latency ≤ 200ms;

[0113] Emergency Response Layer: Establish a three-level early warning mechanism—Mild warning (e.g., weak signal at a single measuring point): pop-up notification and automatic switching of data source; Moderate warning (e.g., deviation 1-10mm or wind speed 8-12m / s): trigger audible and visual alarms with a sound level ≥80dB and generate adjustment suggestions (e.g., "Raise A1 lifting point by 2mm"); Severe warning (e.g., deviation ≥10mm or wind speed ≥15m / s): automatically lock the operation permission of the hoisting equipment, call the emergency plan library (built-in 10+ types of fault handling procedures, such as "total station assisted positioning switching procedure"), and guide the operator to handle the situation.

[0114] like Figure 2 As shown, the hoisting digital twin management method based on the fusion of BeiDou sensing and intelligence is implemented based on any one of the hoisting digital twin systems based on the fusion of BeiDou sensing and intelligence, and includes:

[0115] Step S1: Deploy Beidou intelligent positioning equipment on the hoisting component and initialize the coordinate system and equipment;

[0116] Step S2: The Beidou intelligent positioning device collects Beidou positioning data and IMU attitude data in real time and preprocesses them, and transmits them to the hierarchical multi-error fusion suppression unit through an anti-interference communication link;

[0117] Step S3: The hierarchical multi-error fusion suppression unit performs hierarchical error suppression processing on the received BeiDou positioning data and IMU attitude data to obtain the corrected BeiDou positioning and IMU attitude data.

[0118] Step S4: The modular hoisting digital twin collaborative management and control platform accesses the corrected Beidou positioning and IMU attitude data, driving the digital twin model to perform synchronous display and status monitoring;

[0119] Step S5: When the deviation between the monitoring data and the design value exceeds the threshold, an adjustment instruction is automatically generated and sent to the actuator, forming a closed loop of perception-processing-control-execution-feedback.

[0120] In step S2, the preprocessing specifically includes the following steps:

[0121] Based on real-time hoisting condition information from the modular hoisting digital twin collaborative management and control platform, a dynamic threshold for data jump judgment is dynamically calculated; the condition information includes hoisting stage identifier, hoisting rope inclination angle, hoisting rope length, and the historical steady movement speed of the component;

[0122] Based on the dynamic threshold, within a unified sliding time window, collaborative analysis and filtering are performed on BeiDou positioning data and IMU attitude data. Specifically, this includes: when the instantaneous change in BeiDou positioning data exceeds the dynamic threshold, the attitude data output by the IMU at the corresponding time is simultaneously verified; if the motion state reflected by the IMU attitude data does not exceed the preset state change threshold, the BeiDou positioning data is determined to have jumped to an abnormal value and is removed.

[0123] The specific formula for the dynamic threshold is as follows:

[0124] ;

[0125] in, Indicates a dynamic threshold. Indicates the basic security threshold. This represents the environmental attenuation factor, which is dynamically adjusted based on the current environmental signal quality. When the signal quality is poor, it is set to 0.5-0.8, lowering the threshold for more stringent screening. , , The weights corresponding to the hoisting stage, real-time motion features, and historical motion features are respectively, and their sum is 1. This represents the influence function during the hoisting phase. Represents the real-time motion feature function. Represents the characteristic function of historical motion;

[0126] The specific formulas for the influence function of the hoisting stage, the real-time motion characteristic function, and the historical motion characteristic function are as follows:

[0127] ;

[0128] ;

[0129] ;

[0130] in, Indicates the angle between the suspension rope and the plumb line. Indicates the speed at which the hoisting rope is retracted or extended. Indicates the nominal or current length of the suspension rope. , This represents the adjustment factor, which can be 0.3 or 0.2 respectively. This represents the standard deviation of the component's motion velocity calculated from the most recent N valid data points. N can be fine-tuned according to the current hoisting stage. During the stable docking stage, a larger N, such as 100, can be used to correspond to a smoother historical benchmark. During the violent translation stage, a smaller N, such as 30, can be used to respond more quickly to changes in motion. This represents a reference speed used for normalization, taking the typical safe speed allowed during the current lifting phase.

[0131] The value of the influence function during the hoisting stage is determined by comprehensive calibration based on the kinematic characteristics, allowable range of motion, and accuracy requirements of each stage of the hoisting operation. The specific determination criteria are as follows:

[0132] During the lifting phase (0.15), the component has just left the ground and experiences initial swaying and acceleration impact, but its range of motion is relatively limited. This phase requires a high level of real-time dynamic response, allowing for a certain degree of normal motion variation while remaining sensitive to abnormal jumps.

[0133] During the translation phase (0.40), the component moves horizontally in the air and is affected by wind load, inertia, and boom sway. The amplitude and speed of the movement are the greatest, and the allowable range of data variation is the widest.

[0134] During the docking phase, the component is about to be in place, requiring millimeter-level precision control. The movement should be extremely slow and smooth, with a minimum value of 0.03, to ensure high fidelity of the positioning data and provide reliable data support for accurate docking.

[0135] In practical applications, the coefficients can be fine-tuned based on specific engineering experience.

[0136] The specific formula for the motion state reflected by the IMU attitude data is as follows:

[0137] ;

[0138] in, Indicators of exercise intensity The characteristic radius is typically taken as the distance from the IMU installation location to the lifting point or the centroid of the component. This represents the three-axis angular velocity vector at time t. This represents the three-axis acceleration vector at time t. Represents the gravitational acceleration vector. Represents the magnitude of the vector;

[0139] The preset state change threshold is dynamically set based on the IMU's background noise and the current hoisting stage, and the specific formula is as follows:

[0140] ;

[0141] in, Indicates the threshold for state change. The baseline noise threshold is determined experimentally by collecting IMU data under absolutely static conditions and calculating the statistical distribution of the motion intensity index. The 99th percentile of this distribution is taken as the baseline noise threshold, representing the noise level of the IMU itself. This represents the stage relaxation coefficient, typically 0 ≤ ≤0.5, This represents the stage relaxation factor, and its value is related to the hoisting stage: 0.3 for the lifting stage, 0.6 for the translation stage, and 0.0 for the docking stage.

[0142] like Figure 3 As shown, in step S3, the hierarchical error suppression processing specifically includes:

[0143] The CEEMDAN-sPCA algorithm is used to decompose and suppress multipath interference components in BeiDou positioning data.

[0144] Adaptive Kalman filtering is used to dynamically fuse BeiDou positioning, IMU attitude and wind speed data, and data weights are dynamically allocated according to the hoisting stage;

[0145] The LSTM error prediction model is used to predict systematic errors, and dynamic compensation is performed by combining the NTP synchronization time.

[0146] The error suppression process in step S3 also includes a thermoelastic deformation compensation step, which specifically includes:

[0147] Based on the component material properties, structural parameters, hoisting constraints, and real-time temperature gradient, the thermoelastic deformation compensation is calculated using a dedicated physical model. The specific formula is as follows:

[0148] ;

[0149] in, This represents the amount of thermoelastic deformation compensation of the component along the positioning direction, and α represents the linear expansion coefficient of the bridge component material, which is preset based on the component material and updated in real time through a cloud database. This indicates the design length between component positioning reference points, which is retrieved in real-time from the hoisting digital twin model. This represents the real-time temperature gradient of the component, i.e., the difference between the component surface temperature collected by environmental sensors and the design reference temperature. This represents the elastic modulus of the component material. Representing temperature stress, through Calculation, where Represents the constraint coefficient. The Poisson's ratio represents the material of the component. It represents the length of the constraint feature. For simply supported or two-point hoisting, it represents the distance between the two hoisting points. For multi-point hoisting, it represents the average distance between adjacent hoisting points, or the length of the maximum deformation area.

[0150] In the hierarchical error suppression process, the LSTM error prediction model is an adaptive HA-LSTM model for hoisting conditions, specifically including:

[0151] Based on K-means clustering of historical hoisting data, the corresponding initial weight matrix of LSTM network is generated according to the hoisting, translation and docking conditions;

[0152] Introduce an adjustment coefficient related to the operating condition indicator signal into the forget gate mechanism;

[0153] An attention mechanism unit is added to the hidden layer to assign higher attention weights to the input data corresponding to sudden changes in wind speed or sudden changes in boom swing angle.

[0154] To address the characteristics of hoisting time-series data—strong dependence on operational conditions and frequent sudden interference—a triple-customized improvement to the LSTM network is implemented to overcome the prediction accuracy bottleneck of general-purpose LSTM in hoisting scenarios. Specifically, this includes:

[0155] 1. Weight initialization driven by working condition recognition: Based on historical data of hoisting under the same working condition (lifting, translation, docking) using K-means clustering (cluster number K=3, accurately matching the three stages of hoisting), a dedicated LSTM initial weight matrix for this working condition is generated to avoid the slow convergence problem caused by general initialization and improve the model's adaptation speed to hoisting time series data.

[0156] 2. Improve the forget gate mechanism, the formula is as follows: ,in, This is a working condition identification signal, a one-bit valid encoded vector, such as during the lifting phase: =[1, 0, 0], This indicates the corresponding adjustment coefficient: n=0.8 for the lifting stage, n=0.5 for the translation stage, and n=0.2 for the docking stage. For the weight of the forget gate, For bias terms, The sigmoid activation function is used to strengthen the weight of recent data and improve positioning accuracy.

[0157] 3. Sudden Error Capture Unit: An attention mechanism is added to the LSTM hidden layer to capture sudden wind speed changes (rate of change ≥ 2 m / s). s) The input data corresponding to abrupt changes in the boom swing angle (rate of change ≥ 0.5° / s) are assigned attention weights. , This represents the sudden change in wind speed. This method accurately captures and suppresses sudden errors in swing angle values, addressing the issue of insufficient response of general-purpose LSTM to sudden interference in hoisting scenarios.

[0158] In the hierarchical error suppression process, the adaptive Kalman filter employs a stage-environment coupled adjustment strategy, specifically including:

[0159] The weighting of the fusion of BeiDou positioning data, IMU attitude data, and environmental data is dynamically allocated during the hoisting phase, specifically including:

[0160] Lifting phase: Positioning weight = 0.3, Attitude weight = 0.5, Environmental weight = 0.2;

[0161] During the hoisting phase: Positioning weight = 0.4, Attitude weight = 0.4, Environmental weight = 0.2;

[0162] Docking phase: Positioning weight = 0.6, Attitude weight = 0.3, Environment weight = 0.1;

[0163] Simultaneously, environmental factors are dynamically adjusted: when the wind speed v≥5m / s, the environmental weight is increased by 0.1 and the attitude weight is decreased by 0.1; when the temperature gradient |ΔT|≥8℃, the positioning weight is decreased by 0.05 and the environmental weight is increased by 0.05.

[0164] The adjustment step size of the filter gain is coupled with the current hoisting stage and the real-time wind speed level, and the calculation formula is as follows:

[0165] ;

[0166] in, This indicates the adjustment step size of the filter gain. This indicates that the step size reference value is being adjusted. The coefficients represent the stages: 1 for the lifting stage, 2 for the translation stage, and 3 for the docking stage. This represents the integer operation. Indicates real-time wind speed. Indicates the reference unit for wind speed. and This represents the coupling coefficient.

[0167] Any combination of one or more computer-readable media may be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0168] The examples described herein are merely preferred embodiments of the invention and are not intended to limit the concept and scope of the invention. Any modifications and improvements made by those skilled in the art to the technical solutions of the invention without departing from the design concept of the invention should fall within the protection scope of the invention.

Claims

1. A hoisting digital twin management method based on BeiDou sensing and intelligent fusion, characterized in that, include: Step S1: Deploy Beidou intelligent positioning equipment on the hoisting component and initialize the coordinate system and equipment. Step S2: The Beidou intelligent positioning device collects Beidou positioning data and IMU attitude data in real time and preprocesses them, and transmits them to the hierarchical multi-error fusion suppression unit through an anti-interference communication link; Step S3: The hierarchical multi-error fusion suppression unit performs hierarchical error suppression processing on the received BeiDou positioning data and IMU attitude data to obtain the corrected BeiDou positioning and IMU attitude data. Step S4: The modular hoisting digital twin collaborative management and control platform accesses the corrected Beidou positioning and IMU attitude data, driving the digital twin model to perform synchronous display and status monitoring; Step S5: When the deviation between the monitoring data and the design value exceeds the threshold, an adjustment instruction is automatically generated and sent to the actuator, forming a closed loop of perception-processing-control-execution-feedback. The process of suppressing delamination error in step S3 also includes a thermoelastic deformation compensation step, specifically including: Based on the component material properties, structural parameters, hoisting constraints, and real-time temperature gradient, the thermoelastic deformation compensation is calculated using a dedicated physical model. The specific formula is as follows: in, This represents the amount of thermoelastic deformation compensation of the component along the positioning direction, and α represents the linear expansion coefficient of the bridge component material, which is preset based on the component material and updated in real time through a cloud database. This indicates the design length between component positioning reference points, which is retrieved in real-time from the hoisting digital twin model. This represents the real-time temperature gradient of the component, i.e., the difference between the component surface temperature collected by environmental sensors and the design reference temperature. This represents the elastic modulus of the component material. Representing temperature stress, through Calculate, where k represents the constraint coefficient and μ represents the Poisson's ratio of the component material. Indicates the length of the constraint feature.

2. The method of claim 1, wherein, In step S2, the preprocessing specifically includes the following steps: Based on real-time hoisting condition information from the modular hoisting digital twin collaborative management and control platform, a dynamic threshold for data jump judgment is dynamically calculated; the condition information includes hoisting stage identifier, hoisting rope inclination angle, hoisting rope length, and the historical steady movement speed of the component; Based on the dynamic threshold, within a unified sliding time window, collaborative analysis and filtering are performed on BeiDou positioning data and IMU attitude data. Specifically, this includes: when the instantaneous change in BeiDou positioning data exceeds the dynamic threshold, the attitude data output by the IMU at the corresponding time is simultaneously verified; if the motion state reflected by the IMU attitude data does not exceed the preset state change threshold, the BeiDou positioning data is determined to have jumped to an abnormal value and is removed.

3. The method of claim 2, wherein, In step S3, the hierarchical error suppression process specifically includes: The CEEMDAN-sPCA algorithm is used to decompose and suppress multipath interference components in BeiDou positioning data. Adaptive Kalman filtering is used to dynamically fuse BeiDou positioning, IMU attitude and wind speed data, and data weights are dynamically allocated according to the hoisting stage; The LSTM error prediction model is used to predict systematic errors, and dynamic compensation is performed by combining the NTP synchronization time.

4. The method of claim 3, wherein, In the hierarchical error suppression process, the LSTM error prediction model is an adaptive HA-LSTM model for hoisting conditions, specifically including: Based on K-means clustering of historical hoisting data, the corresponding initial weight matrix of LSTM network is generated according to the hoisting, translation and docking conditions; Introduce an adjustment coefficient related to the operating condition indicator signal into the forget gate mechanism; An attention mechanism unit is added to the hidden layer to assign higher attention weights to the input data corresponding to sudden changes in wind speed or sudden changes in boom swing angle.

5. The method of claim 4, wherein, In the hierarchical error suppression process, the adaptive Kalman filter employs a stage-environment coupled adjustment strategy, specifically including: The weighting ratio of the fusion of BeiDou positioning data, IMU attitude data, and environmental data is dynamically allocated according to the hoisting stage. The adjustment step size of the filter gain is coupled with the current hoisting stage and the real-time wind speed level.

6. A hoisting digital twin system based on BeiDou sensing and intelligent fusion, used to implement the hoisting digital twin management method based on BeiDou sensing and intelligent fusion as described in any one of claims 1-5, characterized in that, This includes BeiDou intelligent positioning devices interconnected through standardized data interfaces, hierarchical multi-error fusion suppression units, and modular hoisting digital twin collaborative management and control platforms; Among them, the Beidou intelligent positioning equipment is used to collect Beidou positioning data and IMU attitude data of hoisted components in real time, and to perform preprocessing and local coordinate transformation; The hierarchical multi-error fusion suppression unit is used to receive the preprocessed data from the Beidou intelligent positioning device, and sequentially perform multipath error decomposition, multi-source data fusion, and dynamic error compensation to output corrected Beidou positioning and IMU attitude data. The multipath error decomposition adopts the CEEMDAN-sPCA algorithm, the multi-source data fusion adopts the adaptive Kalman filter algorithm, and the dynamic error compensation integrates the LSTM error prediction model and the NTP clock synchronization mechanism. The modular hoisting digital twin collaborative management and control platform is used to access and integrate corrected BeiDou positioning and IMU attitude data, environmental data and video data, drive the BIM model to display synchronously, and trigger closed-loop adjustment instructions or emergency responses according to preset rules. At the same time, it feeds back real-time hoisting condition information to the BeiDou intelligent positioning device for adaptive adjustment of preprocessing parameters.

7. The system of claim 6, wherein, The Beidou intelligent positioning device integrates a multi-mode positioning unit, an anti-interference communication unit, an environment adaptation unit, and a data preprocessing unit. The multi-mode positioning unit integrates a Beidou RTK chip and an inertial measurement unit (IMU) for synchronously acquiring positioning and IMU attitude data. The anti-interference communication unit integrates a dual-link redundant transmission module of 4G / 5G network and dedicated radio, and supports local data caching and breakpoint resume. The environmental adaptation unit integrates a temperature and humidity sensor and an adaptive temperature control device. When the temperature or humidity exceeds a preset threshold, it automatically activates a heat dissipation or moisture-proof mechanism. The data preprocessing unit integrates an ARM-based processor and is used to remove abrupt outliers using a sliding window algorithm.

8. The system of claim 7, wherein, The hierarchical multi-error fusion suppression unit includes a multi-path error decomposition module, a multi-source data fusion module, and a dynamic error compensation module; The multipath error decomposition module uses the CEEMDAN-sPCA algorithm to decompose and reconstruct the original BeiDou positioning data in order to suppress multipath interference. The multi-source data fusion module uses an adaptive Kalman filter algorithm to fuse the decomposed BeiDou positioning data, IMU attitude data, and environmental wind speed data, and dynamically adjusts the fusion weights of each data source according to the IMU zero bias state. The dynamic error compensation module integrates an LSTM error prediction model and an NTP clock synchronization mechanism to predict and compensate for satellite clock errors and atmospheric refraction errors.

9. The system of claim 8, wherein, The modular hoisting digital twin collaborative management and control platform adopts a four-layer architecture, including: The data access layer is used to access multi-source sensor data in real time via the MQTT protocol, and supports data compression, encryption and offline caching; The model computation layer has a built-in lightweight BIM model engine and LOD technology to drive the synchronization between the virtual model and physical components, and realize the virtual-real mapping. The interactive display layer adopts a modular card layout to provide real-time visual display of the entire hoisting process status, deviation data, and monitoring videos. The emergency response layer has a built-in three-level early warning mechanism and emergency plan library, which can automatically trigger early warnings, generate adjustment instructions, or lock device permissions based on the deviation level.

Images