BIM-fused port infrastructure digital twin operation and maintenance management system and method

Abstract

The invention provides a BIM-fused port infrastructure digital twinborn operation and maintenance management system and method. The method comprises the steps of collecting a port infrastructure BIM full-life-cycle multi-source heterogeneous data set, constructing a port infrastructure real-time three-dimensional digital twinborn body, constructing an equipment health prediction model and generating a fault diagnosis result; simulating tide influence to form a stress distribution prediction result; a three-level linkage optimization model is constructed, a scheduling scheme is generated, a multi-energy collaborative optimization engine optimization scheme is used, a three-dimensional safety matrix is constructed through unmanned aerial vehicle inspection and visual recognition, risks are monitored in real time, and safety early warning is generated. The prediction model is constructed, faults are accurately diagnosed, the tide influence is quantified, and the prediction precision is improved; multi-target collaborative optimization is realized through a linkage optimization model and an engine, and the problem of single-target optimization is solved; and on the basis of a monitoring and early warning system, rapid identification and hierarchical response are realized, a closed-loop link is formed, and the port operation and maintenance adaptability is improved.

Description

Technical Field

[0001] This invention relates to the field of port operation and maintenance technology, and in particular to a digital twin operation and maintenance management system and method for port infrastructure that integrates BIM. Background Technology

[0002] Currently, BIM technology has been gradually introduced into the port infrastructure operation and maintenance field to achieve 3D visualization modeling, and combined with IoT sensors to collect multi-source heterogeneous data such as equipment vibration, temperature, and energy consumption. Traditional methods mostly focus on single-dimensional data monitoring (such as equipment status monitoring or environmental parameter recording). Some advanced ports have attempted to achieve static scene restoration through digital twins, but these applications are mostly limited to the construction phase and lack the ability to deeply integrate and synchronize dynamic data throughout the entire lifecycle. At the operation and maintenance decision-making level, existing technologies mostly rely on manual experience rules or single-objective optimization algorithms (such as minimizing waiting time or energy consumption), and have not yet formed a multi-source data-driven collaborative optimization framework. Furthermore, their ability to integrate and analyze spatiotemporally related variables such as tides and ship dynamics is weak, making it difficult for operation and maintenance solutions to adapt to the complex and ever-changing port operating environment. Summary of the Invention

[0003] This invention aims to at least address the technical problem that existing operation and maintenance solutions are difficult to adapt to the complex and ever-changing port operation environment, and innovatively proposes a digital twin operation and maintenance management system and method for port infrastructure that integrates BIM.

[0004] To achieve the above-mentioned objectives of this invention, this invention provides a digital twin operation and maintenance management method for port infrastructure that integrates BIM, the method comprising: S1. Collect multi-source heterogeneous datasets of the entire lifecycle of port infrastructure BIM, including equipment vibration data, temperature data, equipment energy consumption data, cargo shape data, environmental parameters and ship dynamic data. S2. Based on the aforementioned multi-source heterogeneous dataset covering the entire lifecycle, construct a real-time three-dimensional digital twin of the port infrastructure; S3. Based on the real-time three-dimensional digital twin, construct a health prediction model for port infrastructure and generate diagnostic results for equipment failure type, location, and severity. S4. Simultaneously integrate the OpenFOAM fluid dynamics simulation module into the equipment health prediction model to simulate tidal effects and generate structural stress distribution prediction results; S5. Based on the diagnostic and prediction results, construct a three-level linkage optimization model for ships, berths, and quay cranes, and generate a scheduling scheme. S6. Optimize the scheduling scheme using a multi-energy collaborative optimization engine to obtain an optimized scheduling scheme; S7. Based on the optimized scheduling scheme, a three-dimensional safety matrix is ​​constructed using UAV inspection and visual recognition. Based on the three-dimensional safety matrix, the location of personnel, equipment status and environmental risks are monitored in real time through recognition algorithms, and a safety risk color-coded early warning is generated.

[0005] On the other hand, the present invention also provides a port infrastructure digital twin operation and maintenance management system integrating BIM, the system comprising: processor; Memory used to store processor-executable instructions; The processor is configured to implement the port infrastructure digital twin operation and maintenance management method integrating BIM when executing the executable instructions.

[0006] The beneficial effects of this invention are as follows: This invention constructs a three-level digital twin model at the terminal, equipment, and component levels through deep fusion and real-time synchronization of multi-source heterogeneous data throughout the entire lifecycle. This achieves dynamic mapping of all elements from equipment vibration and cargo morphology to tidal environment, solving the synchronization deviation problem between traditional digital twins and actual scenarios. The equipment health prediction model, constructed by combining the LSTM+XGBoost hybrid algorithm and graph neural network, can accurately diagnose fault types, locations, and severity. Furthermore, it quantifies the impact of tides on structural stress through OpenFOAM fluid dynamics simulation and fluid-structure interaction algorithms, improving prediction accuracy under complex working conditions. The three-level linkage optimization model of ships, berths, and quay cranes, along with the multi-energy collaborative optimization engine, achieves multi-objective collaborative optimization of waiting time, energy consumption cost, and throughput, solving the problem that traditional single-objective optimization cannot adapt to dynamic operating conditions. Finally, relying on the real-time monitoring capabilities of UAV inspection and three-dimensional safety matrix, combined with a color warning system, it realizes rapid identification and graded response of personnel location, equipment status, and environmental risks, forming a closed-loop link of "data fusion - model prediction - collaborative optimization - real-time warning". This significantly improves the port operation and maintenance solution's adaptability to complex and ever-changing operating environments, reduces operation and maintenance costs and safety risks, and enhances the efficiency and reliability of port operations.

[0007] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0008] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of a digital twin operation and maintenance management method for port infrastructure that integrates BIM, according to the present invention. Detailed Implementation

[0009] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0010] Example 1 like Figure 1 As shown, a digital twin operation and maintenance management method for port infrastructure integrating BIM is proposed, the method comprising: S1. Collect multi-source heterogeneous datasets of the entire lifecycle of port infrastructure BIM, including equipment vibration data, temperature data, equipment energy consumption data, cargo shape data, environmental parameters and ship dynamic data. Step S1 requires detailed explanation of the deployment of a BIM-IoT three-core sensing network within the port infrastructure to collect multi-source heterogeneous datasets. These datasets undergo preprocessing to ensure their quality and consistency. Specifically, firstly, equipment vibration and temperature data are denoised using wavelet transform or Kalman filtering algorithms to remove outliers, and missing data is filled in using interpolation methods. Secondly, equipment energy consumption data is normalized using time series analysis methods to facilitate subsequent integration with other data sources. For cargo morphology data, a high-precision cargo model is generated using 3D point cloud reconstruction technology and matched and calibrated with the geometric information in the BIM model. Environmental parameters are collected in real-time through a sensor network and compared with historical meteorological data to verify their accuracy. Finally, ship dynamic data is acquired through the AIS system and spatiotemporally aligned with the port scheduling plan to ensure data consistency in both time and space. All preprocessed data will be integrated into a unified multi-source heterogeneous dataset.

[0011] S2. Based on the aforementioned multi-source heterogeneous dataset covering the entire lifecycle, construct a real-time three-dimensional digital twin of the port infrastructure; S3. Based on the real-time three-dimensional digital twin, construct a health prediction model for port infrastructure and generate diagnostic results for equipment failure type, location, and severity. S4. Simultaneously integrate the OpenFOAM fluid dynamics simulation module into the equipment health prediction model to simulate tidal effects and generate structural stress distribution prediction results; S5. Based on the diagnostic and prediction results, construct a three-level linkage optimization model for ships, berths, and quay cranes, and generate a scheduling scheme. S6. Optimize the scheduling scheme using a multi-energy collaborative optimization engine to obtain an optimized scheduling scheme; S7. Based on the optimized scheduling scheme, a three-dimensional safety matrix is ​​constructed using UAV inspection and visual recognition. Based on the three-dimensional safety matrix, the location of personnel, equipment status and environmental risks are monitored in real time through recognition algorithms, and a safety risk color-coded early warning is generated.

[0012] The technical principle of a BIM-integrated digital twin operation and maintenance management method for port infrastructure in this embodiment is as follows: First, through the deep integration of BIM technology and IoT sensing networks, efficient collection and processing of data throughout the entire lifecycle of port infrastructure is achieved. In the process of constructing a real-time 3D digital twin, corresponding sensors and data fusion methods are used to ensure consistency and synchronization between the virtual model and the physical entity. The establishment of the equipment health prediction model combines machine learning algorithms and physical simulation technology, accurately capturing the changing trends of equipment operating status and achieving early warning of faults through multi-dimensional feature extraction. In terms of structural stress analysis, by introducing the OpenFOAM fluid dynamics simulation module, not only the influence of static loads is considered, but also the long-term impact of dynamic factors such as tides on the structural safety of the wharf is fully assessed. The entire methodology, through the synergistic effect of multi-level optimization models and intelligent algorithms, ultimately forms a comprehensive operation and maintenance management framework integrating prediction, diagnosis, optimization, and early warning functions.

[0013] As an optional embodiment of the present invention, optionally, constructing a real-time three-dimensional digital twin of the port infrastructure based on the full lifecycle multi-source heterogeneous dataset in step S2 includes: S201. Preprocess the full lifecycle multi-source heterogeneous dataset to obtain a full lifecycle data asset map; In step S201, it needs to be explained in detail that the method for obtaining the full lifecycle data asset graph in this embodiment is as follows: The full lifecycle multi-source heterogeneous dataset is cleaned, classified, and labeled, and a data asset graph is constructed using knowledge graph technology. Specifically, firstly, natural language processing technology is used to parse unstructured data, extract key entities and relationships, and transform them into a structured form. Secondly, for structured data, existing association rule mining algorithms are used to identify implicit relationships between data, forming a multi-layered data node network. Finally, combined with the business logic of port infrastructure, semantic rules for the data asset graph are defined to ensure that it accurately reflects the physical and logical relationships in the actual scenario.

[0014] S202. Based on the full lifecycle data asset map, the port infrastructure is reconstructed in three dimensions using Revit+Dynamo parametric modeling and laser scanning point cloud fusion method to obtain three-dimensional models of the wharf structure, quay cranes, and storage yard. In step S202, it is necessary to explain in detail that, in this embodiment, the 3D reconstruction first utilizes a high-precision laser scanning device to perform a comprehensive scan of the port infrastructure, generating dense point cloud data. Subsequently, point cloud processing software is used to denoise, register, and segment the original data, extracting key geometric features such as the wharf outline, quay crane structure, and yard layout. Next, a parametric model is constructed based on the Revit platform, precisely matching the processed point cloud data with the geometric information in the BIM model to ensure consistency in the model's dimensions and position. Simultaneously, the Dynamo visual programming tool is used to automate the modeling process, dynamically adjusting model parameters according to preset rules to adapt to different scenario requirements. The final output 3D model not only includes static geometric information but also integrates real-time sensor data, thus dynamically reflecting the actual state changes of the port infrastructure.

[0015] S203. Based on the real-time acquisition of the full life cycle multi-source heterogeneous dataset, the data is synchronized with the three-dimensional model to obtain a real-time synchronized digital twin. In step S203, the real-time synchronization process involves several steps. First, based on a unified time base, various data streams from the multi-source heterogeneous dataset across the entire lifecycle are timestamped to ensure data from different sensors and systems can be integrated on the same time dimension. Second, existing spatial mapping algorithms (such as Laplacian eigenmaps) are used to associate the collected real-time data with corresponding entities in the 3D model. For example, equipment vibration data is mapped to specific quay crane components, or environmental parameters are bound to specific areas of the yard. During this process, Kafka message queue technology is used to achieve efficient data transmission and low-latency updates. Edge computing nodes are also used to preprocess key data to reduce the computational burden on the main server. Finally, a real-time rendering engine visualizes the synchronized data, creating a dynamically updated digital twin.

[0016] S204. Perform multi-scale model fusion and lightweight processing on the real-time synchronized digital twin to construct a three-level digital twin model at the dock level, equipment level, and component level.

[0017] In step S204, it is necessary to explain in detail that, in this embodiment, the method for multi-scale model fusion and lightweight processing is as follows: First, based on the real-time synchronized digital twin, existing multi-resolution analysis techniques are used to perform hierarchical analysis on the three-dimensional models of the wharf structure, quay crane, and yard. By extracting geometric features and physical properties at different scales, a wharf-level macro model, an equipment-level meso model, and a component-level micro model are constructed respectively. Second, during the model fusion process, existing graph neural networks (such as GraphSAGE or GCN) are used to model the relationships between the three-level models, ensuring consistency in data transmission and state updates at each level. Finally, to meet the lightweight model requirements, existing mesh simplification algorithms are used to optimize complex geometric structures, while LOD (Level of Detail) technology is combined to dynamically adjust the model's level of detail to adapt to the performance requirements of different application scenarios. After the above processing, the final output three-level digital twin model not only has high-precision expressive capabilities but also enables efficient loading and real-time interaction in the actual operating environment.

[0018] As an optional embodiment of the present invention, optionally, in step S3, based on the real-time three-dimensional digital twin, constructing an equipment health prediction model for port infrastructure and generating diagnostic results of equipment failure type, location, and severity includes: S301. Based on the real-time three-dimensional digital twin, the equipment vibration time-frequency features are extracted using wavelet packet transform, and the cargo shape 3D features are extracted using a convolutional neural network. Based on the equipment vibration time-frequency features and cargo shape 3D features, the equipment topology relationship data is fused using a graph neural network to generate a holographic feature matrix containing spatiotemporal correlation features. In step S301, the specific method for generating the holographic feature matrix in this embodiment is as follows: First, wavelet packet transform is used to perform multi-scale decomposition of the equipment vibration signal, extracting the energy distribution and temporal statistical features of different frequency bands to form a time-frequency feature vector of equipment vibration. Second, a convolutional neural network is used to extract features from the three-dimensional point cloud data of the cargo shape, capturing key information about its geometric shape and spatial distribution to generate a 3D feature vector of the cargo shape. Subsequently, a graph neural network is used to fuse the above two feature vectors with equipment topology relationship data, where the equipment topology relationship data includes information such as physical connections, logical dependencies, and spatial layout between equipment. Through the message passing mechanism of the graph neural network, the features are propagated and aggregated in the equipment topology structure, thereby generating a holographic feature matrix containing spatiotemporal correlation features. The final output holographic feature matrix can comprehensively reflect the dynamic correlation between the equipment operating status and the changes in cargo shape.

[0019] S302. Establish an equipment health prediction model based on the LSTM+XGBoost hybrid algorithm, train the equipment health prediction model using the holographic feature matrix and historical fault samples, fine-tune the equipment health prediction model using transfer learning, and automatically tune the parameters through Bayesian optimization to generate an equipment health prediction model containing a classifier, locator and regressor. In step S302, the construction process of the equipment health prediction model in this embodiment is as follows: First, the time-series data of the equipment is modeled based on LSTM (Long Short-Term Memory) to capture the time-dependent features of the equipment's operating status. Simultaneously, the existing XGBoost (Extreme Gradient Boosting) algorithm is used to efficiently process nonlinear features and extract key fault patterns. The two algorithms are combined to form a hybrid model, achieving unified classification, localization, and regression functions through a multi-task learning framework. Second, during training, a holographic feature matrix is ​​used as input, combined with historical fault sample data, to optimize model parameters through supervised learning. To improve the model's generalization ability, transfer learning technology is employed to transfer the knowledge of the pre-trained model to the current task, and Bayesian optimization is used to automatically adjust hyperparameters, ensuring the model's adaptability in different scenarios. The final generated equipment health prediction model includes a classifier to determine the fault type, a locator to determine the fault location, and a regressor to assess the fault severity.

[0020] S303. Input the real-time collected multi-source heterogeneous dataset of the entire life cycle into the completed equipment health prediction model. The classifier determines the fault type, the locator determines the equipment component-level location, and the regressor determines the severity, generating diagnostic results of equipment fault type, location, and severity.

[0021] In step S303, it needs to be explained in detail that, in this embodiment, the specific method for generating the diagnostic result is as follows: First, the real-time collected multi-source heterogeneous dataset covering the entire lifecycle is input into the trained equipment health prediction model. The classifier, based on the feature distribution of the input data, calculates the probability distribution of fault types through a multi-layer neural network and outputs the most likely fault category. The locator, combining the equipment topology and the spatial features of the input data, utilizes the message passing mechanism of a graph neural network to accurately locate the specific component-level location where the fault occurs. The regressor models the continuous values ​​of fault-related features, assesses the severity of the fault, and outputs the results in the form of quantitative indicators. The final diagnostic result includes detailed information on the fault type, specific location, and severity.

[0022] The expression for the equipment health prediction model is: LSTM temporal feature extractor: XGBoost classifier (fault types): XGBoost locator (fault location): XGBoost Regressor (Severity): in, Represents the temporal hidden state matrix. Indicates an LSTM network. This represents the holographic feature matrix, which includes the time-frequency features of equipment vibration, the 3D features of cargo shape, and topological relationship features. This represents the LSTM parameter set, including the input gate, forget gate, output gate weight matrices, and bias vectors. This represents the transpose of a vector. Represents the predicted fault type, taking values ​​from a discrete set. , The function representing the fault type classifier, Represents the set of classifier parameters. Represents a set of fault types. Indicates the fault type The weight vector, Indicates the fault type The bias term, Indicates the fault type The weight vector, Indicates the fault type The bias term, Indicates the predicted fault location, with values ​​taken from continuous space or discrete component identifiers. This represents the fault location prediction function. This represents the set of locator parameters, including the weak classifier weight coefficients and the weak classifier tree function, used to construct the ensemble learning model. The total number of weak classifier trees is controlled by the hyperparameters of the XGBoost algorithm, and the optimal value is determined through cross-validation (e.g., ...). =100), balancing model complexity and generalization ability. Indicates the number of weak classifier trees. Indicates the first The weight coefficients of the weak classifier trees are automatically learned through the XGBoost algorithm, reflecting the contribution of each tree to the final prediction result (e.g., >0 indicates that the tree makes a positive contribution to the prediction of fault location. Indicates the first A weak classifier tree function, typically a regression tree or a classification tree, takes LSTM hidden states as input. The output is the tree's prediction of the fault location (such as leaf node weights or class probabilities). This indicates the predicted severity level, with values ​​ranging from [0,1] (0 for no fault, 1 for severe fault). Classification functions representing severity, This represents the total set of parameters of the regressor, including Regression coefficients of trees and global bias terms , Indicates the first The regression coefficients of a weak classifier (decision tree) are used to adjust the weight of the tree's contribution to the final prediction result. This indicates a global bias term.

[0023] As an optional embodiment of the present invention, optionally, in step S4, the OpenFOAM fluid dynamics simulation module is synchronously integrated into the equipment health prediction model to simulate the tidal effects and form a structural stress distribution prediction result, including: S401. Collect tidal environment parameters, unify the spatial reference of the tidal environment parameters with the real-time three-dimensional digital twin, and obtain a tidal-geometric coupling dataset that is consistent with the spatiotemporal scale of the real-time three-dimensional digital twin. In step S401, it is necessary to explain in detail that a network of tidal monitoring sensors deployed in the port area acquires key parameters such as tidal height, current velocity, and direction in real time. Simultaneously, by combining astronomical tidal forecast data and historical statistical information, short-term and long-term tidal change trends are predicted. Secondly, to ensure the spatiotemporal consistency of tidal environmental parameters with the real-time 3D digital twin, existing Geographic Information System (GIS) technology is used to perform spatial reference transformation on the tidal data, mapping it to the same coordinate system as the 3D model. During this process, timestamp technology is used to precisely calibrate the tidal data in the time dimension, thereby forming a tidal-geometric coupled dataset that is completely synchronized with the real-time 3D digital twin. The final output dataset not only contains dynamic change information of the tidal environment but can also be seamlessly integrated into the digital twin.

[0024] S402. Based on the tidal-geometric coupling dataset, mesh generation is performed using the OpenFOAM fluid dynamics simulation module, boundary conditions are set, and a dynamic simulation model of tidal wave propagation-flow velocity field-pressure field is constructed using the PISO solver. The numerical simulation results containing the velocity vector field and pressure gradient field are output. In step S402, it is necessary to explain in detail that, in this embodiment, when performing mesh generation using the OpenFOAM fluid dynamics simulation module, firstly, based on the geometric features and physical properties in the tidal-geometric coupling dataset, an adaptive meshing technique is used to finely divide the complex structural region, while a coarse-grained mesh is used for open water to improve computational efficiency. Secondly, when setting boundary conditions, combined with actual tidal environment parameters, the inlet boundary is defined as dynamically changing tidal height and velocity distribution, the outlet boundary is defined as a free outflow condition, and a no-slip boundary condition is applied to the solid wall. Subsequently, the pressure-velocity coupling calculation is achieved through the PISO (Pressure-Implicit with Splitting of Operators) solver to capture the nonlinear effects during tidal wave propagation and the dynamic interaction between the flow velocity field and the pressure field. The final output numerical simulation results include high-resolution velocity vector fields and pressure gradient fields, which can accurately reflect the fluid dynamic effects on port infrastructure under tidal action.

[0025] S403. The numerical simulation results are mapped to the grid nodes of the real-time three-dimensional digital twin using the Kriging interpolation algorithm, and a dynamic stress distribution cloud map of the wharf structure is generated using data mapping rules. In step S403, the application process of the Kriging interpolation algorithm in this embodiment is as follows: First, based on the velocity vector field and pressure gradient field data in the numerical simulation results, the Kriging interpolation method is used to perform spatial interpolation processing on these physical field information. This method constructs a semi-variogram model to evaluate the spatial correlation between different grid nodes and generates a continuous physical field distribution map based on this. Second, when mapping the interpolation results to the grid nodes of the real-time three-dimensional digital twin, strict data mapping rules are adopted to ensure that the interpolated data completely matches the spatial topology of the digital twin. During this process, combined with finite element analysis technology, the dynamic stress distribution of the wharf structure under tidal action is calculated, and a high-precision stress distribution cloud map is generated. The final output cloud map can intuitively reflect the stress change trend of the wharf structure under different tidal conditions.

[0026] S404. Based on the dynamic stress distribution cloud of the wharf structure, the stress concentration factor of the key stress points of the wharf is calculated using the ANSYS Mechanical finite element analysis module. Combined with historical tidal environment parameters, the fatigue region of the structure is predicted by the fluid-structure interaction algorithm. Finally, the predicted stress distribution of the structure, including static stress and dynamic stress, is output.

[0027] The expression for calculating the stress concentration factor is: , ; in, express Stress concentration factor at a point Indicates the first Key stress points of each dock This indicates the calculation performed using the ANSYS Mechanical finite element analysis module. Maximum local stress at point express Point nominal stress, Indicates static load. Indicates geometric dimensions, express Point cross-sectional area; In step S404, it is necessary to explain in detail that, in this embodiment, the calculation process of the stress concentration factor is further refined into the following steps: First, the key stress points of the wharf are meshed using the ANSYS Mechanical finite element analysis module to ensure that the mesh density can accurately capture local stress changes. Second, under static and dynamic loads, the maximum local stress and nominal stress of each key stress point are calculated respectively. The maximum local stress is obtained through high-precision finite element simulation, while the nominal stress is theoretically derived based on the structural geometry and load distribution. Subsequently, combined with historical tidal environment parameters, the fatigue region of the wharf structure is predicted using a fluid-structure interaction algorithm. In this process, a cumulative effect analysis of the time dimension is introduced to evaluate the impact of long-term tidal action on the structural fatigue performance. The final generated structural stress distribution prediction result not only includes detailed distribution information of static and dynamic stresses but also identifies potential fatigue weak areas.

[0028] As an optional embodiment of the present invention, optionally, in step S5, a three-level linkage optimization model of ships, berths, and quay cranes is constructed based on the diagnostic results and prediction results to generate a scheduling scheme, including: S501. Based on the diagnostic results, prediction results and ship dynamic data, use the constraint rule engine to generate a set of equipment availability constraints, a set of structural bearing constraints and a set of ship-berth matching constraints, and output a multi-source dataset and a set of constraint conditions. In step S501, it needs to be explained in detail that the application process of the constraint rule engine in this embodiment is as follows: First, based on the diagnostic and prediction results, and combined with ship dynamic data, key information such as equipment health status, structural stress distribution, and tidal influence are extracted. This information is used to generate an equipment availability constraint set to ensure that the scheduling scheme can fully consider the actual operating capacity and potential failure risks of the equipment. Second, by analyzing the stress distribution prediction results of the wharf structure, a structural bearing constraint set is constructed to prevent structural damage caused by overloading or fatigue accumulation. In addition, using parameters such as ship tonnage, berthing time, and cargo loading and unloading requirements from the ship dynamic data, a ship-berth matching constraint set is generated to optimize the allocation efficiency of berth resources. The final output multi-source dataset and constraint set not only contain real-time equipment status and structural performance information.

[0029] S502. Based on the multi-source dataset and constraint set, construct a three-level linkage optimization model for ships, berths and quay cranes, and set an objective function to minimize the total waiting time, minimize energy consumption cost and maximize throughput. Solve the model using a genetic algorithm and output the initial scheduling scheme matrix. In step S502, the following steps require detailed explanation: First, based on the multi-source dataset and constraint set, the collaborative relationships between ships, berths, and quay cranes are clarified, and these relationships are transformed into mathematical expressions. Second, when setting the objective function, core indicators of port operations are comprehensively considered, including total waiting time, energy consumption cost, and throughput. Total waiting time is quantified by statistically analyzing the time interval from ship arrival to the start of loading / unloading operations; energy consumption cost is calculated by combining equipment operating power and scheduling path length; and throughput is evaluated based on the amount of cargo loaded / unloaded per unit time. Subsequently, a genetic algorithm is used to solve the model. In this process, the initial population consists of a randomly generated scheduling scheme matrix, with each individual representing a possible scheduling strategy. Through selection, crossover, and mutation operations, the individuals in the population are gradually optimized until the convergence condition is met or the preset number of iterations is reached. The final output initial scheduling scheme matrix not only reflects the optimal matching relationship between ships, berths, and quay cranes but also significantly improves the overall operational efficiency of the port in practical applications.

[0030] S503. Based on the initial scheduling scheme matrix, perform real-time scheduling and conflict resolution to obtain the resolved scheduling scheme; In step S503, it is necessary to explain in detail that, in this embodiment, the real-time scheduling and conflict resolution process is further refined into the following steps: First, based on the initial scheduling scheme matrix, scheduling instructions are issued to the control systems of ships, berths, and quay cranes to achieve preliminary task allocation. Second, during actual execution, potential scheduling conflicts are dynamically detected by real-time monitoring of equipment status, cargo loading and unloading progress, and changes in the external environment. For example, when a quay crane malfunctions and stops operating, or a ship arrives late, the system automatically triggers the conflict resolution mechanism. Subsequently, priority strategies and dynamic adjustment algorithms in the rule base are used to quickly respond to conflicts. Specifically, for high-priority tasks, such as emergency cargo loading and unloading or the berthing of large ships, the system will reallocate resources and adjust the work sequence; for low-priority tasks, resource competition is alleviated by extending waiting time or optimizing path planning. The final output of the resolved scheduling scheme can not only effectively cope with emergencies, but also minimize the impact on the overall scheduling plan while ensuring port operational efficiency.

[0031] S504. The resolved scheduling scheme is verified and dynamically adjusted to obtain the final verified scheduling scheme.

[0032] In step S504, the verification and dynamic adjustment process of the scheduling scheme includes the following key steps: First, based on the resolved scheduling scheme, the feasibility and effectiveness of the scheme are comprehensively evaluated using a simulation platform. By simulating ship berthing, cargo loading and unloading, and quay crane operations, potential bottlenecks or resource conflicts are detected. Second, during the verification process, the scheme is dynamically corrected using real-time data streams. For example, when external environmental parameters (such as tidal changes or weather conditions) fluctuate significantly, the system automatically triggers a dynamic adjustment mechanism to recalculate the optimal scheduling path and time allocation. Furthermore, to ensure the robustness of the scheduling scheme, a multi-objective optimization algorithm (such as a genetic algorithm) is introduced to perform secondary optimization of the adjusted scheme, focusing on balancing the relationship between efficiency, cost, and safety. The final output scheduling scheme undergoes rigorous verification and can adapt to the complex and ever-changing port operating environment, thereby comprehensively improving the overall operational level of the port.

[0033] As an optional embodiment of the present invention, the expression of the objective function may be: in, This represents the set of optimization variables (including berthing time, number of quay cranes allocated, and berth selection). Represents a ship index set, , and Represents the normalized weighting coefficients. Indicates the maximum allowed waiting time. Indicates the first Total waiting time for the ships This indicates the port's total energy consumption capacity. Indicates the first Energy consumption cost per ship This indicates the target throughput (set by the port's annual operations plan). This indicates the total throughput of the port.

[0034] As an optional embodiment of the present invention, optionally, in step S6, the scheduling scheme is optimized using a multi-energy collaborative optimization engine to obtain an optimized scheduling scheme, including: S601. Based on the aforementioned scheduling scheme, the real-time electricity price of the port power grid, the SOC of the energy storage system, and the energy consumption data stream of the equipment are spatiotemporally aligned with the scheduling scheme, and a synchronized multi-energy time-series data cube is output. In step S601, it is necessary to explain in detail that the construction process of the multi-energy time-series data cube in this embodiment is as follows: First, based on the time axis and spatial distribution information in the scheduling scheme, the real-time electricity price of the port power grid, the State of Charge (SOC) of the energy storage system, and the energy consumption data stream of the equipment are precisely aligned. Through the alignment operation, it is ensured that the energy supply and demand relationship at each time node can match the specific scheduling task. Second, in the data processing process, an interpolation algorithm is used to fill any possible data gaps, and a sliding window technique is used to smooth the time-series data to reduce noise interference and improve data quality. Subsequently, the synchronized multi-energy data is organized into a three-dimensional data cube structure according to the time dimension, spatial dimension, and attribute dimension. This structure not only facilitates subsequent analysis and optimization but also intuitively reflects the dynamic relationship between energy consumption and scheduling tasks. The final multi-energy time-series data cube is generated.

[0035] S602. Based on the multi-energy time series data cube, the LSTM+attention mechanism model is used to fuse historical energy data, weather parameters and scheduling variables to output the predicted energy demand value for the next n hours. Combined with the real-time status of the real-time three-dimensional digital twin, an energy-equipment-structure coupling model is constructed to quantify the threshold of the impact of high energy consumption on equipment life and structural safety. In step S602, it is necessary to explain in detail that, in this embodiment, firstly, based on a multi-energy time-series data cube, an LSTM+attention mechanism model is used to perform deep fusion analysis on historical energy data, weather parameters, and scheduling variables. This model generates a predicted energy demand value for the next n (24) hours by capturing long-term dependencies in the time series and combining the attention mechanism to highlight the impact of key features. Secondly, based on the prediction results, an energy-equipment-structure coupling model is constructed by combining the state information of a real-time three-dimensional digital twin. This model can quantify the wear and tear on equipment lifespan and the potential impact on the safety of the terminal structure caused by high-energy-consumption operation. Specifically, by introducing fatigue accumulation theory and material stress-life curves, the remaining service life of the equipment under different energy consumption modes is evaluated; at the same time, the finite element analysis method is used to simulate the changes in structural stress distribution under high-energy-consumption conditions to determine its safety threshold. The final output analysis results can significantly improve the safety and economy of port operations in practical applications.

[0036] S603. Based on the energy demand forecast and the energy-equipment-structure coupling model, a multi-objective function is constructed, and the optimal weights are automatically learned through a genetic algorithm to output the optimized scheduling scheme.

[0037] The expression for the multi-objective function is: in, This represents the set of optimization variable vectors, including decision variables such as quay crane power allocation, energy storage charging and discharging power, and ship berthing time periods, which directly drive the direction of energy synergistic optimization. , and This represents the normalized weighting coefficients, which are automatically learned to their optimal values ​​using a genetic algorithm, balancing the priorities of energy cost, structural safety, and equipment lifespan. This indicates the maximum permissible energy cost, set by the port's annual operating budget. Indicates total energy cost, Indicates the structural stress safety threshold. This represents the structural stress index, output from an energy-equipment-structure coupling model, quantifying the impact of high energy consumption on the structural safety of the wharf. Indicates the design life of the equipment. The equipment lifespan impact index is calculated from the equipment's real-time load rate and historical failure data, quantifying the accelerated decay effect of high energy consumption on equipment lifespan. In step S603, the construction and optimization process of the multi-objective function is further refined into the following steps: First, based on the predicted energy demand and the energy-equipment-structure coupling model, the specific composition of the optimization variable vector set is clarified. This set includes key decision variables such as quay crane power allocation, energy storage system charging and discharging power, and ship berthing time periods. These variables directly determine the direction and effect of energy synergistic optimization. Second, when setting the objective function, the core needs of port operation are comprehensively considered, including three dimensions: energy cost, structural safety, and equipment lifespan. By introducing normalized weight coefficients, objectives of different dimensions are unified into the same evaluation system, and the optimal weights are automatically learned using a genetic algorithm to achieve a dynamic balance among the three. Subsequently, during the optimization process, the actual value of the total energy cost is calculated by combining the maximum allowable energy cost set in the port's annual operating budget; simultaneously, the structural stress index is obtained through finite element analysis and compared with the safety threshold to assess the impact of high energy consumption on the structural safety of the wharf. In addition, based on real-time equipment load rate and historical fault data, the accelerated decay effect of high energy consumption operation on equipment lifespan is quantified, generating an equipment lifespan impact index. The final optimized scheduling scheme can not only significantly reduce energy costs, but also achieve good results in ensuring structural safety and extending equipment life, thereby comprehensively improving the overall operational level of the port.

[0038] As an optional embodiment of the present invention, optionally, in step S7, based on the optimized scheduling scheme, a three-dimensional safety matrix is ​​constructed using UAV inspection and visual recognition. Based on the three-dimensional safety matrix, a recognition algorithm is used to monitor personnel location, equipment status, and environmental risks in real time, generating a safety risk color-coded early warning, including: S701. Based on the optimized scheduling scheme, deploy drone inspection tasks, plan drone flight paths, and cover key areas of the port, including berths, quay cranes and surrounding environment. In step S701, the deployment process of the UAV inspection task in this embodiment is as follows: First, based on the optimized scheduling scheme, the distribution of key port areas, including berths, quay cranes, and their surrounding environment, is determined. The selection of these areas is based on their importance in port operations and their potential safety risk levels. Second, considering the UAV's endurance, sensor performance, and inspection requirements, the UAV's flight path is planned to ensure maximum coverage without omissions. During path planning, a suitable path planning method (such as Dijkstra's algorithm) is used, which comprehensively considers obstacle distribution, flight altitude limitations, and weather conditions to generate the optimal path. Subsequently, the path information is integrated with the scheduling system to achieve automated allocation and real-time monitoring of UAV tasks. The final output UAV inspection path not only efficiently completes safety monitoring tasks but also significantly improves the precision of port safety management in practical applications.

[0039] S702. Utilize drones to collect real-time image data and environmental parameters, and extract personnel location, equipment status, and environmental risk information through visual recognition algorithms (YOLOv7) to generate a multi-dimensional safety status dataset. In step S702, it is important to explain in detail that, in this embodiment, the high-definition camera and sensor devices mounted on the drone can efficiently collect real-time image data and environmental parameters. This data is processed using visual recognition algorithms (such as YOLOv7) to extract key information, including personnel location, equipment operating status, and potential environmental risk factors. Specifically, the algorithm first classifies and locates target objects in the image, such as identifying whether workers are in dangerous areas or whether equipment is operating abnormally. Subsequently, combined with environmental parameters (such as temperature, humidity, and wind speed), it further analyzes risk factors that may affect safety. The resulting multi-dimensional safety status dataset not only contains spatial distribution information but also covers dynamic changes over time.

[0040] S703. Map the safety status dataset to a real-time 3D digital twin to construct a 3D safety matrix. Use a deep learning model to analyze the 3D safety matrix, identify potential safety hazards, quantify risk levels, and provide color-coded warnings for safety risks.

[0041] In step S703, it is necessary to explain in detail that, in this embodiment, the construction and analysis process of the three-dimensional safety matrix is ​​further refined into the following steps: First, the multi-dimensional safety status dataset is mapped onto a real-time three-dimensional digital twin to form a three-dimensional safety matrix covering key areas of the port. This matrix not only includes personnel location, equipment status, and environmental risk information, but can also be dynamically updated to reflect real-time changes in port operations. Second, a deep learning model is used to perform multi-level analysis on the three-dimensional safety matrix to identify potential safety hazards. For example, spatial features are extracted through convolutional neural networks, and dynamic risk trends are captured by combining time series analysis, thereby achieving accurate judgment of complex scenarios. Subsequently, based on the analysis results, the levels of various risks are quantified and transformed into an intuitive color-coded warning system. Specifically, according to the severity of the risk, the warning levels are divided into three levels: low, medium, and high, corresponding to green, yellow, and red indicators, respectively, to facilitate rapid response by management personnel. The final output of the three-dimensional safety matrix and its color-coded warning system can significantly improve the efficiency and accuracy of port safety management in practical applications.

[0042] Example 2 A digital twin operation and maintenance management system for port infrastructure integrating BIM, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to implement a BIM-integrated digital twin operation and maintenance management method for port infrastructure when executing executable instructions.

[0043] It should be noted that the computer device includes a processor, a memory, and may also include one or more of a multimedia component, an input / output (I / O) interface, and a communication component.

[0044] The processor controls the overall operation of the computer device to complete all or part of the steps in the aforementioned BIM-integrated digital twin operation and maintenance management method for port infrastructure.

[0045] Memory is used to store various types of data to support the operation of the computer device. This data may include, for example, instructions for any application or method used to operate on the computer device, as well as application-related data. Memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0046] The multimedia component may include a screen and an audio component, wherein the screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals; for example, the audio component may include a microphone for receiving external audio signals, the received audio signals may be further stored in memory or transmitted via a communication component; the audio component may also include at least one speaker for outputting audio signals.

[0047] I / O interfaces provide interfaces between the processor and other interface modules, such as keyboards, mice, buttons, etc.; these buttons can be virtual buttons or physical buttons.

[0048] The communication component is used for wired or wireless communication between the computer device and other devices; wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G or 5G, or one or more combinations thereof, and the corresponding communication component may include: Wi-Fi module, Bluetooth module, NFC module, mobile communication module.

[0049] As a preferred embodiment, the computer device may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the aforementioned port infrastructure digital twin operation and maintenance management method integrating BIM.

[0050] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A digital twin operation and maintenance management method for port infrastructure integrating BIM, characterized in that, The method includes: Collect multi-source heterogeneous datasets of the entire lifecycle of port infrastructure BIM, including equipment vibration data, temperature data, equipment energy consumption data, cargo shape data, environmental parameters and ship dynamic data; Based on the aforementioned multi-source heterogeneous dataset covering the entire lifecycle, a real-time three-dimensional digital twin of the port infrastructure is constructed. Based on the real-time 3D digital twin, a health prediction model for port infrastructure is constructed, and diagnostic results of equipment failure type, location, and severity are generated. The OpenFOAM fluid dynamics simulation module is synchronously integrated into the equipment health prediction model to simulate the tidal effects and generate structural stress distribution prediction results. Based on the diagnostic and prediction results, a three-level linkage optimization model for ships, berths, and quay cranes is constructed to generate a scheduling scheme. The scheduling scheme is optimized using a multi-energy collaborative optimization engine to obtain an optimized scheduling scheme; Based on the optimized scheduling scheme, a three-dimensional safety matrix is ​​constructed using UAV inspection and visual recognition. Based on the three-dimensional safety matrix, the location of personnel, equipment status and environmental risks are monitored in real time through recognition algorithms, and a safety risk color-coded early warning is generated.

2. The method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 1, characterized in that, Building a real-time 3D digital twin of port infrastructure includes: The full lifecycle multi-source heterogeneous dataset is preprocessed to obtain a full lifecycle data asset map; Based on the full lifecycle data asset map, the port infrastructure is reconstructed in three dimensions using Revit+Dynamo parametric modeling and laser scanning point cloud fusion method to obtain three-dimensional models of the wharf structure, quay cranes, and storage yard. The three-dimensional model is synchronized with the real-time acquired multi-source heterogeneous dataset of the entire life cycle to obtain a real-time synchronized digital twin. The real-time synchronized digital twin is subjected to multi-scale model fusion and lightweight processing to construct a three-level digital twin model at the dock level, equipment level, and component level.

3. The method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 1, characterized in that, The diagnostic results generated include the type, location, and severity of the equipment fault. Based on the real-time three-dimensional digital twin, wavelet packet transform is used to extract the time-frequency features of equipment vibration, and convolutional neural network is used to extract the 3D features of cargo shape. Based on the time-frequency features of equipment vibration and the 3D features of cargo shape, graph neural network is used to fuse equipment topology data to generate a holographic feature matrix containing spatiotemporal correlation features. An equipment health prediction model is established based on the LSTM+XGBoost hybrid algorithm. The model is trained using the holographic feature matrix and historical fault samples. Transfer learning is used to fine-tune the model, and Bayesian optimization is used to automatically tune the parameters, generating an equipment health prediction model that includes a classifier, a locator, and a regressor. The real-time collected multi-source heterogeneous datasets throughout the entire lifecycle are input into the completed equipment health prediction model. The classifier determines the fault type, the locator determines the equipment component-level location, and the regressor determines the severity, generating diagnostic results for the equipment fault type, location, and severity.

4. The port infrastructure digital twin operation and maintenance management method integrating BIM as described in claim 3, characterized in that, The expression for the equipment health prediction model is: in, Represents the temporal hidden state matrix. Indicates an LSTM network. Represents the holographic feature matrix. Represents the LSTM parameter set, Represents the predicted fault type, taking values ​​from a discrete set. , The function representing the fault type classifier, Represents the set of classifier parameters. Represents a set of fault types. Indicates transpose. Indicates the fault type The weight vector, Indicates the fault type The bias term, Indicates the fault type The weight vector, Indicates the fault type The bias term, Indicates the predicted fault location. This represents the fault location prediction function. This represents the set of locator parameters. Indicates the number of weak classifier trees. Indicates the first The weight coefficients of the weak classifier trees. Indicates the first A weak classifier tree function, Indicates the predicted severity. Classification functions representing severity, This represents the total set of parameters of the regressor. Indicates the first Regression coefficients of weak classifiers This represents the global bias term.

5. A method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 1, characterized in that, The predicted results of structural stress distribution include: Collect tidal environment parameters, unify the spatial reference of the tidal environment parameters with the real-time three-dimensional digital twin, and obtain a tidal-geometric coupling dataset that is consistent with the spatiotemporal scale of the real-time three-dimensional digital twin. Based on the tidal-geometry coupled dataset, the OpenFOAM fluid dynamics simulation module was used to perform mesh generation, set boundary conditions, and the PISO solver was used to construct a dynamic simulation model of tidal wave propagation-flow velocity field-pressure field, outputting numerical simulation results including the flow velocity vector field and pressure gradient field. The numerical simulation results are mapped to the grid nodes of the real-time three-dimensional digital twin using the Kriging interpolation algorithm, and a dynamic stress distribution cloud map of the wharf structure is generated using data mapping rules. Based on the dynamic stress distribution cloud of the wharf structure, the stress concentration factor of the key stress points of the wharf is calculated using the ANSYS Mechanical finite element analysis module. Combined with historical tidal environment parameters, the fatigue region of the structure is predicted by the fluid-structure interaction algorithm. Finally, the predicted stress distribution of the structure, including static stress and dynamic stress, is output.

6. The method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 1, characterized in that, The generated scheduling scheme includes: Based on the diagnostic results, prediction results and ship dynamic data, the constraint rule engine is used to generate a set of equipment availability constraints, a set of structural load-bearing constraints and a set of ship-berth matching constraints, and outputs a multi-source dataset and a set of constraint conditions. Based on the multi-source dataset and constraint set, a three-level linkage optimization model of ships, berths and quay cranes is constructed, and an objective function is set to minimize the total waiting time, minimize energy consumption cost and maximize throughput. The model is solved by a genetic algorithm and the initial scheduling scheme matrix is ​​output. Based on the initial scheduling scheme matrix, real-time scheduling and conflict resolution are performed to obtain the resolved scheduling scheme. The resolved scheduling scheme is verified and dynamically adjusted to obtain the final verified scheduling scheme.

7. A method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 6, characterized in that, The expression for the objective function is: in, This represents the set of optimization variables (including berthing time, number of quay cranes allocated, and berth selection). Represents a ship index set, , and Represents the normalized weighting coefficients. Indicates the maximum allowed waiting time. Indicates the first Total waiting time for the ships This indicates the port's total energy consumption capacity. Indicates the first Energy consumption cost per ship Indicates the target throughput. This indicates the total throughput of the port.

8. A method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 1, characterized in that, The optimized scheduling scheme includes: Based on the aforementioned scheduling scheme, the real-time electricity price of the port power grid, the SOC of the energy storage system, and the energy consumption data stream of the equipment are spatiotemporally aligned with the scheduling scheme, and a synchronized multi-energy time-series data cube is output. Based on the multi-energy time-series data cube, an LSTM+attention mechanism model is used to fuse historical energy data, weather parameters and scheduling variables to output the predicted energy demand for the next n hours. Combined with the real-time status of the real-time 3D digital twin, an energy-equipment-structure coupling model is constructed to quantify the threshold of the impact of high energy consumption on equipment life and structural safety. Based on the energy demand forecast and the energy-equipment-structure coupling model, a multi-objective function is constructed, and the optimal weights are automatically learned through a genetic algorithm to output an optimized scheduling scheme.

9. A method for digital twin operation and maintenance management of port infrastructure integrating BIM as described in claim 1, characterized in that, The generation of color-coded security risk warnings includes: Based on the optimized scheduling scheme, drone inspection tasks are deployed, drone flight paths are planned, and key areas of the port are covered, including berths, quay cranes and the surrounding environment. By using drones to collect real-time image data and environmental parameters, and by using visual recognition algorithms to extract personnel location, equipment status and environmental risk information, a multi-dimensional safety status dataset is generated. The safety status dataset is mapped to a real-time 3D digital twin to construct a 3D safety matrix. A deep learning model is used to analyze the 3D safety matrix to identify potential safety hazards, quantify risk levels, and provide color-coded warnings for safety risks.

10. A digital twin operation and maintenance management system for port infrastructure integrating BIM, characterized in that, The system includes: processor; Memory used to store processor-executable instructions; The processor is configured to implement the port infrastructure digital twin operation and maintenance management method integrating BIM as described in any one of claims 1 to 9 when executing the executable instructions.

Images