A hull structure health monitoring method based on digital twinning

Abstract

The application relates to the technical field of ship body health monitoring, and discloses a ship body structure health monitoring method based on digital twinning. The method establishes a ship body digital twinning model, and arranges monitoring points at key structure positions of the ship body to realize accurate mapping of the physical structure of the ship body and real-time data acquisition of key areas. Based on the structure grid division result of the ship body digital twinning model and the real-time data collected by the monitoring points, dynamic partition division is carried out, so that the partition can dynamically adapt to the state change of the ship body structure under different working conditions. According to the sequence of the dynamic partition, a sequential diagnosis mode is adopted to sequentially execute health state evaluation of each partition until the whole ship partition evaluation is completed. Through the fusion of the digital twinning model and the real-time data, combined with dynamic partition and sequential diagnosis, the method realizes accurate and dynamic monitoring of the health state of the ship body structure, improves the pertinence and reliability of the health evaluation, and provides effective protection for the safe operation of the ship body structure.

Description

Technical Field

[0001] This invention relates to the field of ship hull health monitoring technology, specifically a method for monitoring ship hull structural health based on digital twins. Background Technology

[0002] As vital carriers for water transport and marine development, ships face complex and ever-changing loads and environmental effects on their hull structures during long-term service. Factors such as salt spray corrosion, wave impact, ship vibration, and stress generated by cargo loading in the marine environment continuously cause cumulative damage to the hull structure. Failure to detect and assess these damages in a timely manner can lead to serious consequences such as structural failure and navigational safety accidents. Therefore, effective health monitoring of the hull structure is a crucial step in ensuring safe ship operation and extending service life.

[0003] Traditional methods for monitoring the health of ship structures largely rely on manual inspections, periodic sampling tests, or single-point sensor monitoring. Manual inspections are not only labor-intensive and inefficient, but also limited by the experience and subjective judgment of the inspectors, making it difficult to comprehensively cover all critical areas of the hull and easily overlooking potential structural hazards. While periodic sampling tests can obtain performance data for some structures, they suffer from fixed inspection cycles and discontinuous data, failing to reflect real-time changes in the hull structure's condition during dynamic service. Single-point sensor monitoring often suffers from a lack of systematic planning in sensor placement, resulting in limited data acquisition and making it difficult to accurately assess the overall structural condition of the hull.

[0004] With the development of digital technology, some monitoring methods have begun to incorporate digital modeling techniques. However, in existing technologies, the integration between the digital model of the hull structure and the actual monitoring data is low, making it difficult for the model to dynamically respond to real-time changes in the hull structure. Furthermore, health status assessments often employ a holistic approach, analyzing the entire ship structure. This method not only involves large data processing volumes and low computational efficiency but also struggles to accurately pinpoint specific areas of structural damage, resulting in insufficient relevance and practicality of the assessment results. Because the stress characteristics and damage modes differ across regions of the hull structure, using uniform assessment standards and procedures can easily overlook the unique characteristics of different regions, affecting the accuracy of health status assessments. These issues make traditional monitoring methods insufficient to meet the high-precision, real-time, and comprehensive requirements of modern ship structural health monitoring. Summary of the Invention

[0005] The purpose of this invention is to provide a method for monitoring the health of ship hull structures based on digital twins, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for monitoring the health of ship hull structures based on digital twins, the method comprising:

[0007] Establish a digital twin model of the ship's hull and deploy monitoring points at key structural locations of the hull;

[0008] Dynamic partitioning is performed based on the structural mesh division results of the ship's digital twin model and the real-time data collected from monitoring points;

[0009] Based on the order of dynamic partitions, a sequential diagnostic approach is used to perform health status assessments for each partition in turn, until the entire ship's partition assessment is completed.

[0010] Preferably, the structural mesh division results based on the ship's digital twin model and the real-time data collected from monitoring points are dynamically partitioned, including:

[0011] Based on the structural mesh volume threshold of the ship's digital twin model and the spatial distribution of real-time strain data collected from monitoring points, the structural mesh volume threshold is matched with the spatial distribution of real-time strain data.

[0012] The matching process must satisfy the positive containment relationship between the structural mesh volume threshold and the spatial distribution of real-time strain data.

[0013] Preferably, the dynamic partitioning based on the structural mesh division results of the ship's digital twin model and the real-time data collected from monitoring points further includes:

[0014] Identify the arch foot positions of dynamic stress arches in a ship's digital twin model;

[0015] Based on the geometric relationship between the arch foot position of the dynamic stress arch and the structural mesh division result, the dynamic partition boundary is defined;

[0016] The dynamic stress arch is formed by the characteristics of the ship's navigation load distribution.

[0017] Preferably, identifying the arch foot position of the dynamic stress arch in the ship's digital twin model includes:

[0018] Extract strain gradient data from continuous time series at monitoring points;

[0019] The coordinate position of the arch foot of the dynamic stress arch is determined based on the preset critical value of the strain gradient abrupt change.

[0020] Preferably, before performing the health status assessment of each partition sequentially using a sequential diagnostic approach based on the dynamic partition order, the process includes:

[0021] Obtain the three-dimensional material property parameters of the digital twin model of the ship's hull;

[0022] Calculate the apparent structural deformation parameters for the transverse, longitudinal, and vertical directions respectively;

[0023] The three-dimensional material property parameters include the elastic modulus distribution and the yield strength threshold.

[0024] Preferably, after calculating the apparent structural deformation parameters in the transverse, longitudinal, and vertical directions respectively, the process includes:

[0025] Based on the apparent deformation parameters of the transverse structure, the apparent deformation parameters of the longitudinal structure, and the apparent deformation parameters of the vertical structure, the damage factors of the transverse structure, the damage factors of the longitudinal structure, and the damage factors of the vertical structure are calculated respectively.

[0026] Preferably, after calculating the transverse structural damage factor, longitudinal structural damage factor, and vertical structural damage factor based on the transverse structural apparent deformation parameters, longitudinal structural apparent deformation parameters, and vertical structural apparent deformation parameters, the process includes:

[0027] By integrating the transverse structural damage factor, longitudinal structural damage factor, and vertical structural damage factor, the remaining service strength assessment value of the zone is generated.

[0028] Preferably, the sequential diagnostic approach for evaluating the health status of each partition includes:

[0029] Set a warning trigger offset margin on the partition boundary trajectory line;

[0030] Based on the spatial relationship between real-time strain data and the early warning trigger offset margin, activate the local structural damage diagnosis command.

[0031] Preferably, after generating the remaining service strength assessment value for the partition, the process includes:

[0032] Based on the remaining service strength assessment value of the partition and the design strength threshold of the hull digital twin model, the service safety margin index of the dynamic partition is calculated.

[0033] Preferably, before establishing the digital twin model of the hull, the following steps are included:

[0034] Configure multiphysics coupling rules based on the constitutive characteristics of the hull material;

[0035] Load transfer topology of the ship hull digital twin model is initialized based on multiphysics coupling rules.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] This digital twin-based method for monitoring ship structural health provides a novel technological approach by establishing a digital twin model of the ship and strategically placing monitoring points at key structural locations. The digital twin model accurately maps the physical structural characteristics of the ship, including its geometry, material properties, and connectivity. This allows the monitoring process to move beyond direct observation of the physical entity, enabling a comprehensive understanding of the structural condition through real-time correlation between the digital model and the actual ship. Placing monitoring points at key structural locations allows for targeted real-time data collection in areas of concentrated stress and prone to damage, ensuring the effectiveness and relevance of data acquisition and avoiding data redundancy or missing crucial information issues common in traditional monitoring methods.

[0038] Dynamic zoning is performed based on the structural mesh generation results from the ship's digital twin model and real-time data from monitoring points. This allows the zoning method to dynamically adapt to changes in the actual state of the ship's structure. Under different navigation conditions and loading conditions, the stress state and deformation of different regions of the hull will change. Dynamic zoning can adjust the zoning range and boundaries according to these real-time changes, ensuring that each zoning accurately reflects the local characteristics of the structure under the current state. This dynamically adjusted zoning method overcomes the shortcomings of traditional fixed zoning methods, which cannot adapt to dynamic structural changes, making subsequent health status assessments more closely aligned with the actual service condition of the hull.

[0039] By sequentially assessing the health status of each dynamic zoning zone using a sequential diagnostic approach, the complex task of overall ship structural health assessment can be broken down into ordered, localized assessment processes. Each zone's assessment can focus on its structural characteristics, monitoring data patterns, and potential damage modes. Detailed analysis of individual zones allows for more precise identification of structural anomalies or signs of damage. This sequential diagnostic approach avoids the excessive data processing and computational complexity of overall ship assessments, reducing the computational load and facilitating the timely identification and focused analysis of problem areas. This local-to-global assessment logic gradually accumulates assessment information, resulting in a more comprehensive and reliable overall ship zoning assessment, providing more valuable references for hull structure maintenance decisions.

[0040] The deep integration of digital twin models and real-time monitoring data enables health status assessments to move beyond reliance on single-factor empirical judgments or static data. Instead, they are based on dynamically updated digital models and real-time collected structural response data, achieving dynamic tracking and precise characterization of the ship's structural health status. This method can promptly capture early signs of structural damage, buying time for ship maintenance, reducing operational interruptions and safety risks caused by structural failures, and improving the stability and economy of ship operations. Attached Figure Description

[0041] Figure 1 This is a schematic diagram illustrating the working principle of the ship hull structure health monitoring method based on digital twins as described in this invention.

[0042] Figure 2 A diagram illustrating the working principle of dynamic partitioning.

[0043] Figure 3 This is a schematic diagram illustrating the working principle of identifying the arch foot position of a dynamic stress arch.

[0044] Figure 4 This diagram illustrates the working principle of calculating the structural damage factor based on apparent structural deformation parameters. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Please see Figure 1 This invention provides a method for monitoring the health of ship hull structures based on digital twins. Its core lies in constructing a dynamic interaction between a virtual model and a physical entity to achieve real-time assessment and early warning of the ship's structural condition. The following detailed description, in conjunction with specific implementation methods, illustrates this method.

[0047] A digital twin model of the ship's hull structure is established, accurately reflecting the actual hull's geometry, material properties, and connection relationships. Sensors are deployed as monitoring points at key structural locations, such as stress concentration areas, weld areas, and major supporting components, to collect real-time data on physical quantities such as strain and vibration. Based on the structural mesh generation results of the constructed digital twin model and combined with the real-time data (mainly strain data) collected from the monitoring points, the hull structure is dynamically partitioned. This partitioning is not fixed but dynamically adjusted according to the structural response state reflected by the real-time data. Following the order of the partitions determined after dynamic partitioning—which can be based on the partition's importance, stress level, or spatial location—a sequential diagnostic approach is used to perform health status assessments on each partition in turn. The assessment process utilizes the corresponding subset of the digital twin model for each partition and the real-time data from the monitoring points within that partition for analysis and calculation. Once the assessment of one partition is completed, the assessment process automatically proceeds to the next partition until the health status assessment of all partitions of the entire ship is completed.

[0048] Example 1: See Figure 2Based on the structural mesh generation results of the ship's digital twin model and the real-time data collected by the monitoring point network, a dynamic partitioning operation is performed. This process includes two interrelated core steps: matching the structural mesh volume threshold with the spatial distribution of real-time strain data, and partitioning the boundaries based on the location of the arch feet of the dynamic stress arch.

[0049] The structural mesh volume threshold is an upper limit constraint on the volume set when dividing the digital twin model into structural units. This threshold is predetermined based on the dimensions, material properties, and expected computational accuracy of the hull components. The spatial distribution of real-time strain data is a spatial representation of the strain field formed by a network of sensors continuously collected and mapped onto the spatial coordinates of the digital twin model from key structural locations on the hull (such as bulkhead joints, main longitudinal beams, and deck beam nodes). The core of the matching operation lies in corresponding the volume range of each structural mesh to the real-time strain distribution area reflected by its internal or associated monitoring points. This correspondence must satisfy the positive containment principle, meaning that the volume range of the structural mesh must be able to completely contain the physical spatial range represented by the strain data of its associated monitoring points. For example, a specific structural mesh unit has a spatial boundary that defines a three-dimensional volume region. The strain data collected by several monitoring points within this region and the strain change trends they reflect should be confined within the volume boundary of that mesh unit. If real-time strain data shows that the strain distribution in a certain region exceeds the volume boundary of its current mesh, or if changes in the strain gradient suggest mechanical behavior across the mesh boundary, it indicates that the current meshing may not effectively characterize the real-time structural response state of that region. In this case, the system will trigger a mesh adjustment or re-meshing mechanism based on the matching results, ensuring that the adjusted mesh volume can once again satisfy the positive containment relationship of its internal strain data spatial distribution. This dynamic matching ensures that the structural discretization units of the digital twin model can always effectively capture and carry the real-time mechanical information of their corresponding physical regions, providing an accurate geometric and physical basis for subsequent partitioning.

[0050] Identifying the arch foot location of dynamic stress arches in a ship's digital twin model is crucial for defining dynamic zoning boundaries. A dynamic stress arch is a temporary, high-stress path or region formed within the ship's structure under specific navigation load conditions, such as encountering specific waves, making large-angle turns, or uneven cargo distribution. Its shape and location dynamically change with load distribution. The arch foot location is the stress concentration point or the start / end point of load transfer at the structural boundary or connection point with other structural components. Identifying the arch foot location relies on analyzing continuous time-series strain gradient data from monitoring points. The system continuously collects and processes strain data from each monitoring point, calculating its spatial gradient change (i.e., the amount of strain change per unit distance). By analyzing continuous time-series strain gradient data, the system tracks the evolution of gradient values ​​in space and time. A preset critical value for strain gradient abrupt change is a threshold set based on hull material properties, structural form, and historical data or simulation analysis. When the strain gradient value at a certain location or its adjacent area consistently exceeds or reaches this preset critical value in a continuous time series, that location is identified as a potential dynamic stress arch arch foot location. The coordinates of this location were precisely extracted and recorded.

[0051] After successfully identifying the arch foot locations of the dynamic stress arch, the system dynamically delineates the boundaries of the partitions by combining the geometric features contained in the structural mesh generation results of the ship's digital twin model. The structural mesh generation results include rich geometric features such as the spatial coordinates of mesh nodes, the shape of mesh elements (e.g., tetrahedrons, hexahedrons), and the connection topology between elements. The partitioning process first maps the identified arch foot coordinate points onto the mesh nodes or elements of the digital twin model. Then, the correlation between these arch foot points and the surrounding mesh geometric features is analyzed. This correlation may manifest as: arch foot points located at the vertices or edges of specific mesh elements; lines connecting arch foot points traversing specific mesh element sequences; and mesh elements surrounding the arch foot points exhibiting specific deformation modes or stress concentration trends. Based on these correlations, the system dynamically determines the boundary trajectories of the partitions. For example, the partition boundary might be defined as a virtual line connecting two critical arch points, extending to both sides along this line within a certain range (this range may be determined by the mesh element size or stress influence range); alternatively, the partition boundary might be defined along the envelope of the high stress gradient region around the arch point, which is composed of mesh element boundaries. The goal of partitioning is to ensure that the resulting dynamic partitions effectively encompass the critical structural response regions dominated by dynamic stress arches under the current load distribution characteristics, making the structural behavior within the partition relatively homogeneous or correlated, facilitating centralized and efficient local health status assessment. Therefore, the partition boundary is not fixed but adjusts in real time with changes in sailing loads, the formation and dissipation of dynamic stress arches, and the migration of arch positions, always reflecting the most critical mechanical state region of the hull structure. The entire dynamic partitioning process is a closed loop: real-time data drives arch identification and mesh matching, the matching results and arch positions jointly guide the partition boundary division, and the partitioned partitions provide target areas for subsequent targeted health assessments. The assessment results may also be fed back to optimize mesh division or arch identification parameters.

[0052] Example 2: See Figure 3Identifying the arch foot locations of dynamic stress arches in a ship's digital twin model relies on systematic analysis of continuous time-series strain gradient data from monitoring points. A network of monitoring points continuously collects strain data from key structural locations on the hull, forming a data stream that varies over time. The system preprocesses this raw strain data, including filtering to eliminate noise interference, data alignment to unify timestamps, and outlier detection and handling. The preprocessed strain data is used to calculate the spatial strain gradient. The strain gradient characterizes the rate of change of strain within a unit distance of the structure, reflecting the intensity and directionality of local strain. Calculations are typically performed within the grid framework of the digital twin model, using the spatial distance and strain difference between adjacent monitoring points to obtain the strain gradient values ​​at each monitoring point location or its vicinity through finite difference or other numerical methods. These calculated strain gradient values ​​and their spatial distribution constitute a continuous time-series strain gradient dataset.

[0053] The preset strain gradient abrupt change threshold is the core criterion for identifying the arch foot location. This threshold is not a fixed single value, but is comprehensively set based on the specific material properties of the hull structure, structural form, historical operational data, and simulation analysis results from a digital twin model. It may be an absolute threshold, a relative rate of change threshold relative to the background strain gradient level, or a composite criterion combining spatial and temporal gradient changes. The system continuously monitors the strain gradient values ​​calculated at each location and compares them with the preset abrupt change threshold in real time. When the strain gradient value at a specific location or its immediate vicinity consistently reaches or exceeds the preset abrupt change threshold in continuous time-series data, that location is identified as a potential dynamic stress arch foot location. The determination process needs to consider the continuity over time to eliminate misjudgments caused by transient interference. Once confirmed, the system accurately records the three-dimensional coordinate information of the arch foot location in the digital twin model. The arch foot locations of dynamic stress arches typically appear at structural connections, boundary constraint points, or geometric abrupt change areas. These locations are prone to stress concentration under specific load conditions and are key areas of focus for structural health monitoring.

[0054] Before performing a sequential diagnostic assessment of the health status of each partition according to the dynamic partitioning order, it is essential to acquire and utilize the three-dimensional material property parameters from the ship's digital twin model. These parameters are crucial inputs during the construction of the digital twin model, accurately describing the material mechanical behavior at different locations within the ship's structure. The three-dimensional material property parameters primarily consist of two core components: the elastic modulus distribution and the yield strength threshold. The elastic modulus distribution describes the stiffness characteristics of the ship's structural materials at various points in three-dimensional space, i.e., the material's ability to resist elastic deformation; its value may vary at different locations due to material type, processing technology, or usage history. The yield strength threshold defines the critical stress level at which the material begins to undergo irreversible plastic deformation or failure at various points in three-dimensional space. These parameters are typically stored in the digital twin model as field variables, associated with structural mesh nodes or elements.

[0055] After acquiring the three-dimensional material property parameters, and combining them with real-time strain data collected from monitoring points, the system calculates the apparent structural deformation parameters of the hull structure in three orthogonal directions. These three directions are defined as: transverse (breadth direction, perpendicular to the ship's longitudinal axis and parallel to the waterline); longitudinal (length direction, along the ship's longitudinal axis); and vertical (depth direction, perpendicular to the waterline). The calculation process is performed for each dynamic zone or evaluation unit. For the calculation of the transverse structural apparent deformation parameters, the system extracts the real-time strain component data of all monitoring points within the zone in the transverse direction, and simultaneously acquires the corresponding three-dimensional elastic modulus distribution data for that zone. Using the fundamental relationships of mechanics of materials, the system combines the monitored transverse strain components with the corresponding local elastic modulus, and calculates the parameter value reflecting the overall macroscopic deformation degree of the zone in the transverse direction through integration or weighted averaging methods. This parameter may be expressed as average strain, maximum strain, or some form of equivalent deformation. The calculation of the longitudinal structural apparent deformation parameters follows the same logic, but processes the longitudinal strain component data, reflecting the overall deformation state of the zone in the ship's length direction. The calculation of the apparent deformation parameters of the vertical structure is based on the vertical strain component data, characterizing the deformation of the partitions along the vertical height. The calculations for each direction are independent, and the results quantify the instantaneous deformation response of the structure in that direction. These apparent deformation parameters serve as the foundational input data for subsequent more in-depth damage assessment and strength analysis, directly reflecting the deformation state of the structure in different dimensions under the current loading conditions. The calculation process fully utilizes the spatial material property information provided by the digital twin model, making the assessment of deformation parameters more consistent with the actual physical characteristics of the structure.

[0056] Example 3: See Figure 4After acquiring the apparent deformation parameters of the transverse, longitudinal, and vertical structures respectively, the system performs calculations of the transverse, longitudinal, and vertical structural damage factors. The structural damage factor in each direction aims to quantify the degree of degradation of the structural material properties or the potential failure risk in that specific direction. The calculation process closely integrates the three-dimensional material property parameters provided by the ship's digital twin model with the measured apparent deformation parameters.

[0057] The calculation process is performed independently for each dynamic partition. For each partition, the system first acquires pre-calculated apparent deformation parameters in three directions within that partition. These parameters reflect the macroscopic deformation state of the partition's structure in the transverse (breadth direction), longitudinal (length direction), and vertical (depth direction) directions during the current monitoring period. Simultaneously, the system extracts the corresponding three-dimensional material property parameters for that partition from the ship's digital twin model, primarily including elastic modulus distribution data and yield strength threshold distribution data. The elastic modulus distribution describes the material's ability to resist elastic deformation at different locations within the partition, while the yield strength threshold defines the critical stress level at which irreversible plastic deformation or failure occurs at different locations within the partition.

[0058] The calculation of the transverse structural damage factor focuses on the structural response in the beam direction. The system utilizes the transverse structural apparent deformation parameters of this zone (e.g., a scalar value reflecting the transverse mean strain or equivalent deformation), combined with representative elastic modulus values ​​within the zone (potentially spatially averaged or strain-weighted effective values) and a transverse yield strength threshold (which may differ from the global yield strength due to material anisotropy or stress state). The computational model considers the constitutive behavior of the material in the transverse direction and the relationship between the current deformation state and the material's ultimate limit state. The calculation of the longitudinal structural damage factor follows a similar logic but is entirely based on the longitudinal structural apparent deformation parameters and longitudinally relevant material properties. Longitudinal stress modes typically involve tension / compression and bending, and the computational model considers the impact of these load modes on damage accumulation. The calculation of the vertical structural damage factor relies on the vertical structural apparent deformation parameters and vertically relevant material properties. The vertical response is often related to gravity, buoyancy, and wave impact loads, and the calculation must consider ballast conditions and local hydrostatic pressure effects.

[0059] The damage factor for each direction is calculated using a dimensionless mathematical expression, mapping the current apparent structural deformation state to a numerical range characterizing the degree of damage; for example, 0 represents intact, and 1 represents complete failure. This expression integrates three key elements: deformation parameters, material stiffness (elastic modulus), and material strength (yield threshold). Taking the transverse structural damage factor as an example... Taking the calculation of as an example, its core relationship can be expressed as:

[0060]

[0061] In this expression: The calculated transverse structural damage factor is a dimensionless quantity with a value range typically in the [0,1] interval. A larger value indicates a higher degree of damage in that direction. The apparent deformation parameter representing the lateral structure of the partition is a scalar value that was previously calculated and reflects the overall lateral deformation of the partition. The elastic modulus value representing the lateral aspect of the partition is derived from the three-dimensional elastic modulus distribution data of the digital twin model and has undergone spatial or equivalent processing for the lateral response being evaluated. The threshold value representing the representative yield strength of the partition in the lateral direction is derived from the three-dimensional yield strength distribution data of the digital twin model, and is also processed for the lateral response. This represents a specific functional relationship or computational model. The specific form of this model is determined based on the principles of materials mechanics, damage mechanics theory, or empirical correlation rules. Its core idea is: given a material stiffness... Under the conditions, the current deformation state The corresponding stress level (or derived quantities such as strain energy density) and the material's yield strength Compare the functions and consider possible nonlinearities, cyclic loading effects, or existing damage history (if the model includes memory functionality). The output is the damage factor. .

[0062] Longitudinal structural damage factor and vertical structural damage factors The calculation adopts the same as Similar principles and functional forms However, the input parameters are replaced with the structural apparent deformation parameters in their respective directions ( , ) and representative material property parameters in the corresponding directions ( , ; , The calculations were performed independently, yielding damage assessment results in three orthogonal directions. , , These damage factor values ​​provide relatively independent health status indicators of the partition structure in three main directions, revealing the directional characteristics of potential damage or intensity degradation. As intermediate results, they provide necessary input for subsequent fusion to generate overall partition status indicators and to conduct safety margin assessments. The entire calculation process is completed automatically within the computation engine of the digital twin system, executed sequentially for each dynamic partition.

[0063] Example 4: After calculating the transverse structural damage factor, longitudinal structural damage factor, and vertical structural damage factor, the system performs a fusion operation to generate a residual service strength assessment value for the partition. This value aims to comprehensively reflect the degree to which the overall load-bearing capacity of the dynamic partition is preserved in its current state. An example is taken of a cargo hold section located amidships. Assume that the calculated damage factors for this partition in the three directions are: transverse structural damage factor... This indicates a certain degree of performance degradation in the beam direction; longitudinal structural damage factor This indicates that the ship's condition is relatively good in the longitudinal direction; vertical structural damage factor This indicates the optimal condition in the depth direction. The higher the value of these damage factors, the more severe the degradation of structural integrity or load-bearing capacity in that direction.

[0064] The fusion process is not a simple arithmetic average, but rather considers the differences in the importance of different directions in the overall load-bearing capacity of the hull structure, as well as potential damage coupling effects. The system assigns a weighting coefficient to the damage factor for each direction based on the hull structure type (e.g., bulk carrier, tanker, container ship), the location of the section within the hull (e.g., the midships section bears a larger total longitudinal bending moment, while the bottom section bears water pressure), and historical load characteristics. This weighting reflects the contribution or sensitivity of different directions to the overall residual strength of the section. For example, for the midships cargo hold section, the total longitudinal bending moment is the primary load; therefore, the longitudinal structural damage factor... The transverse structure, crucial for maintaining the compartment shape and local strength, may be assigned the highest weight; the vertical structure, primarily bearing uniformly distributed pressure in the midship section, may have the lowest weight. Furthermore, the fusion model may introduce nonlinear terms to account for the mutual influence of multiaxial damage; for example, increased transverse damage may weaken longitudinal bearing capacity. The system applies a pre-defined fusion algorithm (such as weighted geometric mean, weighted mean based on the maximum value of damage in the main direction, or a more complex multiaxial damage model) to... , , The calculation, along with its corresponding weights, ultimately outputs a single, dimensionless assessment value for the remaining service strength of a given region. This value is typically expressed relative to the intact state ( The remaining strength proportion of the midship cargo hold. For example, referring to Table 1, the weight allocation and fusion calculation yield the weight of the midship cargo hold partition. .

[0065] Table 1: Weighting of Damage Factors and Residual Strength Assessment for Midship Cargo Hold Zones

[0066]

[0067] Note: This table shows the specific calculation process for an example partition. The weight coefficients sum to 1.0. The example fusion method is to subtract 1 from the weighted sum (i.e., ...). The actual algorithm may be more complex.

[0068] When performing health status assessments for each partition sequentially using a sequential diagnostic approach, setting the warning trigger offset margin is a crucial step. This margin aims to establish a buffer or warning zone near the partition boundary to detect potential risks earlier. The system sets the warning trigger offset margin on the dynamic partition boundary trajectory line. This margin is not a fixed value but is based on the partition's location, structural importance, historical damage records, and the currently calculated remaining service strength assessment value of the partition. Dynamically adjusted. For example, for zones with low residual strength assessment values ​​(such as...). For partition boundaries located in high-stress areas (such as cargo hold corners or large opening edges), the system will set a smaller offset margin (i.e., closer to the boundary line) to trigger warnings more sensitively; for partition boundaries in good condition or non-critical areas, a larger offset margin can be set. The offset margin can be expressed as a fixed distance value (e.g., 50 mm) relative to the boundary trajectory line in the normal direction, or a proportional value based on structural feature dimensions near the boundary (e.g., plate thickness, stiffener spacing) (e.g., 0.5 times the plate thickness), or a value related to the remaining strength of the partition. Related function values.

[0069] The activation of a local structural damage diagnosis command is based on the spatial relationship between real-time strain data and the warning trigger offset margin. The system continuously monitors real-time strain data within a zone, especially at monitoring points near the zone boundary. This data is mapped to the spatial coordinate system of the digital twin model. The system calculates the difference or positional relationship between the real-time strain value (or equivalent stress or deformation value) of each monitoring point and the warning trigger offset margin threshold corresponding to its location (this threshold defines the upper limit of the offset margin region). When the real-time strain value of a monitoring point reaches or exceeds the warning trigger offset margin threshold set for its location, it indicates that the structural response at that location has entered the preset warning region. At this time, the system activates a local structural damage diagnosis command for that zone, or more precisely, for the local area where the monitoring point is located. This command triggers a more refined analysis process, which may include: calling a higher-resolution digital twin sub-model for local stress concentration analysis; performing denser virtual sensor data interpolation calculations for the area surrounding the point; initiating a special assessment based on fracture mechanics or fatigue accumulation theory; or issuing a maintenance inspection recommendation notification. For example, at a monitoring point in the corner of a cargo hold hatch, the set warning trigger offset margin threshold is based on 80% of the allowable design stress at that point. When the stress value calculated from the real-time monitored strain reaches 105% of this threshold, the system immediately activates a local diagnostic command for that corner area, performing a detailed assessment of stress distribution and potential crack propagation, without waiting for the sequential assessment of the entire zone to complete. This offset margin-based warning mechanism enables the health monitoring system to proactively and promptly focus on high-risk areas, improving the efficiency and targeting of monitoring.

[0070] Example 5: After generating the remaining service strength assessment value for a partition, the system performs the operation of calculating the service safety margin index for the dynamic partition. This index is a quantitative measure of the current load-bearing capacity of the partition structure relative to its original design requirements. The system calls the design strength threshold stored in the ship's digital twin model. This threshold is the lower limit of the structural load-bearing capacity that the partition should meet when it is intact and in an ideal initial design state, usually corresponding to a specific design load condition. The system reads the latest calculated remaining service strength assessment value for a specific partition, which represents the current estimated retained load-bearing capacity level of that partition. The calculation process of the safety margin index is to directly compare these two values. A typical method is to calculate the ratio of the two, for example: the remaining service strength assessment value of the partition divided by the design strength threshold. A result of 1.0 indicates that the current load-bearing capacity of the partition just meets the minimum design requirements; a result greater than 1.0 indicates that the current capacity is higher than the design requirements, with a margin; a result less than 1.0 indicates that the current load-bearing capacity is lower than the minimum design requirements, facing a safety risk. Another possible method is to calculate the difference between the two, subtracting the design strength threshold from the remaining strength assessment value. A result of zero indicates that the requirements are met, a positive value indicates a surplus, and a negative value indicates a deficiency. Regardless of the algorithm used, the output dynamic partition's service safety margin index is a numerical value that intuitively reflects the safe carrying capacity of the partition structure in its current state. The system associates and stores this index value with the spatial location information of the partition. On the visualization interface of the ship's digital twin model, this index is usually displayed as a color-coded overlay on the partition model, allowing operators to easily identify risk partitions with low or even negative safety margins. For example, a cargo hold area amidships is displayed in warning red, indicating that its carrying capacity is close to or below the design threshold.

[0071] Before establishing a digital twin model of the hull, the configuration of multiphysics coupling rules must be completed. The hull structure operates in a real marine environment, and its mechanical response is the result of the interaction of multiple physical fields. The constitutive characteristics of the hull materials describe their fundamental physical laws and response characteristics under various loads and environmental conditions. This includes, but is not limited to: the elastic, plastic, and viscoelastic behavior of materials under static and dynamic loads; the fatigue characteristics of materials under cyclic loads; creep or relaxation phenomena that may occur during long-term service; the thermal expansion and thermal conductivity characteristics of materials under different temperature conditions and their impact on mechanical properties; and the electrochemical corrosion behavior of materials in seawater or specific media environments and its impact on strength. Based on the detailed constitutive model and data of the specific material system used in the hull structure, the system defines and configures the rules for the interaction between various physical fields. These rules explicitly describe the causal relationships and energy transfer mechanisms between the structural mechanical field and other physical fields. The core coupling relationships typically include: coupling between the structural mechanical field and the temperature field through the thermal stress effect caused by thermal expansion; coupling between the structural mechanical field and the fluid pressure field through the hydrodynamic pressure on the hull surface, which dynamically changes with hull motion and wave conditions; and coupling between the structural mechanical field and the corrosion field through cross-sectional loss or surface morphology changes caused by corrosion, leading to stress redistribution and material strength degradation. The configuration of the rules involves defining the input-output relationship of each coupling action, the action intensity coefficient, and the coupling solution strategy during the solution process.

[0072] Based on the configured multiphysics coupling rules, the system initializes the load transfer topology of the ship's digital twin model. This topology describes the physical path mechanism by which various external loads (such as hydrostatic pressure, wave impact force, cargo weight, propulsion machinery vibration load, temperature gradient, etc.) are transferred and distributed through the internal load-bearing component network of the ship structure as a whole system. Constructing the load transfer topology first requires identifying all possible load application points and surfaces. Then, based on the ship's geometric configuration and the connection relationships between components, the basic load transfer paths are defined. For example, the cargo weight on the deck is transferred to the longitudinal girder, then to the transverse bulkheads and side structures; the water pressure borne by the outer plating is transferred inward to the rib frame, then to the side longitudinals and longitudinal bulkheads. Each transfer path involves specific structural components and their connection nodes. Using the initialized topology, the system defines the load transfer ratio between different components, boundary constraints, and load balance relationships at the nodes. This topology establishes a mathematical model framework between external load inputs and the stress response states at key locations within the ship structure. For example, a typical longitudinal load transfer path might be defined as: bow wave impact force -> bowpost -> bottom flat keel -> inner bottom longitudinal ribs -> transverse strong frame -> longitudinal bulkhead -> to the stern structure. During the transfer process, the system calculates the additional thermal stress that may arise in adjacent areas due to uneven temperature distribution, or considers the impact of local stiffness reduction caused by material loss in specific corrosion areas on load distribution, based on coupling rules. This initial load transfer topology constitutes the core framework for dynamic simulation prediction of the hull digital twin model, enabling the model to simulate the real structural response process under load input, considering the effects of multi-field coupling.

[0073] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0074] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art 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 appended claims and their equivalents.

Claims

1. A method for monitoring the health of ship hull structures based on digital twins, characterized in that, include: Establish a digital twin model of the ship's hull and deploy monitoring points at key structural locations of the hull; Dynamic partitioning is performed based on the structural mesh division results of the ship's digital twin model and the real-time data collected from monitoring points; According to the order of dynamic partitioning, a sequential diagnostic approach is used to perform health status assessments for each partition until the entire ship's partition assessment is completed. The dynamic partitioning based on the structural mesh division results of the ship's digital twin model and the real-time data collected from monitoring points includes: matching the structural mesh volume threshold with the spatial distribution of real-time strain data collected from monitoring points based on the structural mesh volume threshold of the ship's digital twin model; the matching process must satisfy a positive containment relationship between the structural mesh volume threshold and the spatial distribution of real-time strain data. The dynamic partitioning based on the structural mesh division results of the ship's digital twin model and the real-time data collected from monitoring points also includes: identifying the arch foot positions of dynamic stress arches in the ship's digital twin model; delineating dynamic partition boundaries based on the geometric feature correlation between the arch foot positions of the dynamic stress arches and the structural mesh division results; wherein, the dynamic stress arches are formed by the ship's navigation load distribution characteristics. Identifying the arch foot positions of dynamic stress arches in the ship's digital twin model includes: extracting strain gradient data from continuous time series of monitoring points; determining the coordinate positions of the arch feet of the dynamic stress arches based on a preset strain gradient abrupt change threshold.

2. The method for monitoring the health of ship hull structures based on digital twins according to claim 1, characterized in that, Before performing the health status assessment of each partition sequentially according to the order of dynamic partitions, the process includes: obtaining the three-dimensional material property parameters of the ship's digital twin model; calculating the apparent structural deformation parameters in the transverse, longitudinal, and vertical directions respectively; wherein the three-dimensional material property parameters include the elastic modulus distribution and the yield strength threshold.

3. The method for monitoring the health of ship hull structures based on digital twins according to claim 2, characterized in that, After calculating the apparent structural deformation parameters in the transverse, longitudinal, and vertical directions respectively, the process includes: calculating the transverse structural damage factor, longitudinal structural damage factor, and vertical structural damage factor based on the apparent structural deformation parameters in the transverse, longitudinal, and vertical directions respectively.

4. The method for monitoring the health of ship hull structures based on digital twins according to claim 3, characterized in that, The process involves calculating the transverse structural damage factor, longitudinal structural damage factor, and vertical structural damage factor based on the apparent deformation parameters of the transverse structure, longitudinal structural damage factor, and vertical structural damage factor, respectively. Then, the process includes fusing the transverse structural damage factor, longitudinal structural damage factor, and vertical structural damage factor to generate the remaining service strength assessment value for the partition.

5. The method for monitoring the health of ship hull structures based on digital twins according to claim 1, characterized in that, The sequential diagnostic approach is used to perform health status assessments for each partition in turn, including: setting a warning trigger offset margin on the partition boundary trajectory line; and activating a local structural damage diagnosis command based on the spatial relationship between real-time strain data and the warning trigger offset margin.

6. The method for monitoring the health of ship hull structures based on digital twins according to claim 4, characterized in that, After generating the remaining service strength assessment value of the partition, the following steps are taken: based on the remaining service strength assessment value of the partition and the design strength threshold of the hull digital twin model, the service safety margin index of the dynamic partition is calculated.

7. The method for monitoring the health of ship hull structures based on digital twins according to claim 1, characterized in that, Before establishing the digital twin model of the hull, the process includes: configuring multiphysics coupling rules based on the constitutive characteristics of the hull material; and initializing the load transfer topology of the digital twin model of the hull based on the multiphysics coupling rules.

Images