Method for simulating buckling state of storage tank under combined action of typhoon and rainfall

Abstract

The invention discloses a method for simulating the buckling state of a storage tank under the combined action of typhoon and rainfall, and the method comprises the steps: obtaining the wind pressure pulsation and liquid level change characteristics of the typhoon, and generating a non-stationary load sequence; constructing a three-dimensional finite element model of the storage tank, coupling multiple physical fields, and determining initial stress distribution; the overall buckling instability process is simulated, and key node displacement is obtained; judging local stress concentration according to the displacement threshold value to obtain a stress concentration distribution diagram; calculating accumulated damage and resistance deterioration degree under multiple typhoons based on a damage accumulation model; simulating a damage extension track from integral buckling to local damage, analyzing a high-risk area formation mechanism, and predicting a failure mode; and outputting the residual service life index of the storage tank under the composite load in combination with the residual service life evaluation framework. According to the method, accurate pre-judgment and quantitative evaluation of the full-life-cycle failure risk of the storage tank under the typhoon precipitation composite load are realized.

Description

Technical Field

[0001] This invention belongs to the field of storage tank condition simulation technology, and particularly relates to a method for simulating the buckling state of storage tanks under the combined effects of typhoon and precipitation. Background Technology

[0002] Current research on the mechanical behavior of storage tanks under extreme weather conditions is mostly limited to analyses under single wind loads or hydrostatic pressure, with little consideration given to the real dynamic processes of rapid rises and falls in liquid level and intense instantaneous coupling of internal and external pressures caused by wind pressure and precipitation during typhoons. This simplification makes it difficult for existing methods to accurately capture the instantaneous evolution of stress distribution in storage tanks under actual typhoon and precipitation combined loads and the true formation mechanism of high-risk areas. In particular, it is difficult to effectively reveal how local areas rapidly enter a state of failure due to stress concentration and redistribution after the overall structure has buckled and become unstable. Typhoon and precipitation combined loads are highly non-stationary and have strong coupling characteristics with multiple physics fields. The instantaneous pulsation of wind pressure and the rapid changes in liquid level superimpose each other, causing the magnitude and direction of the load on the tank wall to reverse drastically in a short period of time, resulting in a significant difference between the overall buckling morphology and the results of traditional static analysis. Furthermore, when the overall structure enters the buckling state, extremely high local stress concentrations occur at local structures such as welds and reinforcing rings due to geometric abrupt changes and nonlinear material responses, which in turn triggers plastic deformation, damage initiation, and even crack propagation. Currently, there is a lack of effective cross-scale correlation description methods for this continuous evolution process from macroscopic overall instability to microscopic local damage.

[0003] Therefore, accurately simulating the entire process of a storage tank from overall buckling instability to the initiation and propagation of local damage under the combined load of typhoon precipitation, and further considering the damage accumulation and continuous deterioration of resistance caused by multiple typhoon events, has become a key issue in predicting the actual failure mode of the storage tank and assessing its remaining service life. Summary of the Invention

[0004] This invention proposes a method for simulating the buckling state of storage tanks under the combined effects of typhoon precipitation, in order to solve the problems existing in the prior art.

[0005] To achieve the above objectives, this invention provides a method for simulating the buckling state of a storage tank under the combined effects of typhoon and precipitation, comprising the following steps:

[0006] Collect historical records and real-time monitoring data of typhoon events to obtain composite load parameters containing wind pressure pulsation and liquid level change characteristics, and generate non-stationary load sequences;

[0007] Based on the aforementioned non-stationary load sequence, a three-dimensional geometric model of the storage tank is constructed using the finite element analysis method, and a multi-field coupling mechanism is incorporated to determine the initial stress distribution state of the structure under composite loads.

[0008] A dynamic load sequence is applied to the initial stress distribution state to simulate the overall buckling instability process and obtain the structural deformation morphology and key node displacement data.

[0009] If the displacement data of key nodes exceeds the preset threshold, it is determined that the local stress concentration stage has been entered, and the stress concentration distribution map of the weld and the reinforcing ring is obtained.

[0010] Based on the stress concentration distribution map, the cumulative damage value under multiple typhoon events is calculated using a damage accumulation model to determine the degree of resistance degradation.

[0011] Based on the aforementioned degree of resistance degradation, a continuous evolution path from overall buckling instability to the initiation of local damage is simulated to obtain the damage propagation trajectory;

[0012] If the damage propagation trajectory shows the initiation of localized damage, the formation mechanism of the high-risk area is analyzed to obtain the failure mode prediction results;

[0013] Based on the failure mode prediction results and combined with the remaining service life assessment framework, the remaining service life index of the storage tank under combined loads is obtained.

[0014] Optionally, generating the non-stationary load sequence includes:

[0015] Collect raw wind pressure and liquid level sequences;

[0016] The trend component and the pulsation component are separated by the time series decomposition method to obtain the wind pressure pulsation sequence and the liquid level fluctuation sequence;

[0017] The non-stationarity of the sequence is determined. If the variance of the sequence changes with time window exceeding a preset threshold, it is marked as a non-stationary sequence.

[0018] Composite load components are extracted from non-stationary labeled sequences to form composite load sequences;

[0019] A random forest model was used to rank the features of the composite load sequence to determine the dominant load parameter sequence.

[0020] The composite load sequence is weighted and synthesized based on the dominant parameter sequence to generate the final non-stationary load sequence.

[0021] Optionally, determining the initial stress distribution state of the structure under combined load includes:

[0022] Divide the time step according to the non-stationary load sequence to obtain the discrete load time point;

[0023] A three-dimensional geometric model of the storage tank is constructed based on discrete load time points, and a mesh is generated to obtain a finite element model;

[0024] A multi-field coupling mechanism between the temperature field and the flow field is incorporated into the finite element model to determine the coupling influence coefficient;

[0025] The composite load is applied based on the coupling influence coefficient, and the instantaneous stress field at each load moment is obtained.

[0026] The instantaneous stress field is superimposed over time to determine the cumulative initial stress distribution.

[0027] Optionally, obtaining the structural deformation morphology and key node displacement data includes:

[0028] A finite element model is constructed based on the initial stress distribution state, and boundary conditions are set.

[0029] Apply a dynamic load sequence and perform loading simulation at a preset time step to obtain stress response data;

[0030] If the stress response exceeds the preset threshold, record the instability data at the corresponding time point;

[0031] Extract deformation morphology features from instability state data to determine deformation evolution patterns;

[0032] Key nodes are located based on the deformation evolution law, and their displacement data are calculated.

[0033] Optionally, obtaining the stress concentration distribution map at the weld and reinforcing ring includes:

[0034] If the displacement data of critical nodes exceeds the threshold, it is determined that the local stress concentration stage has been entered.

[0035] Obtain geometric information of the weld and reinforcing ring, and establish a preliminary stress distribution model;

[0036] Fine meshing was performed on the weld and reinforcing ring areas to obtain high-precision stress distribution data;

[0037] Stress distribution charts are constructed based on high-precision stress distribution data to visually represent areas of stress concentration.

[0038] Update the response data of key nodes based on the distribution chart, iteratively optimize the judgment results, and generate the final stress concentration distribution map.

[0039] Optionally, determining the degree of resistance degradation includes:

[0040] Based on the stress concentration distribution map, extract data on key stress concentration locations;

[0041] A damage accumulation model was used to calculate the preliminary value of cumulative damage by combining the frequency and intensity of multiple typhoon events.

[0042] By incorporating environmental factor data, the initial value of cumulative damage is corrected, and the trend of damage accumulation is determined.

[0043] The level of resistance degradation is determined based on the trend of cumulative damage.

[0044] By combining historical typhoon event data, key structural safety risks are identified;

[0045] The degree of resistance degradation is output after comprehensive evaluation.

[0046] Optionally, obtaining the damage propagation trajectory includes:

[0047] Based on the resistance degradation data, key points of overall buckling are extracted;

[0048] The buckling behavior was simulated using the finite element method to determine the stress concentration region under the unstable state;

[0049] Analyze the conditions for the initiation of local damage; if the stress value exceeds the preset threshold, record the point of failure initiation.

[0050] Simulate the process of damage initiation and evolution, and track the changing trajectory of stress concentration areas;

[0051] By extracting the damage propagation path during the evolution process and analyzing the stress transmission law, the damage propagation trajectory is obtained.

[0052] Optionally, obtaining the failure mode prediction results includes:

[0053] Determine the extent of damage distribution based on the damage propagation trajectory;

[0054] Based on the distribution range of damage, the local damaged areas are finely segmented to obtain specific location data;

[0055] Based on specific location data, determine whether the signs of damage exceed a preset threshold; if they do, mark it as a high-risk area.

[0056] Based on high-risk areas, feature data are extracted, and classification algorithms are used to determine the failure mode type.

[0057] Based on the failure mode type, track the damage propagation trend and obtain propagation trend prediction data;

[0058] Based on the extended trend forecast data, if the risk is determined to exceed the safe range, a regional adjustment plan is generated.

[0059] Based on the regional adjustment plan, the failure mode evolution path is simulated to obtain the failure mode prediction results.

[0060] Optionally, obtaining the remaining service life index of the storage tank under combined loads includes:

[0061] Based on the failure mode prediction results, key data are extracted and categorized.

[0062] Based on classification information and combined with preset life assessment standards, preliminary assessment results are obtained through data matching.

[0063] Based on the preliminary assessment results, if an abnormal service condition is detected under combined loads, detailed change data will be extracted.

[0064] Based on detailed change data, a classification algorithm is used to analyze the load impact and determine the trend of state change;

[0065] Based on the trend of state change, a correlation mapping between failure modes and life indicators is established, and preliminary life indicators are derived.

[0066] Based on the preliminary service life indicators, the long-term performance of the storage tank under combined loads is simulated to determine the remaining service life indicators.

[0067] Based on the remaining service life indicators, they are archived into the assessment database to form complete assessment information.

[0068] Compared with the prior art, the present invention has the following advantages and technical effects:

[0069] This invention integrates and analyzes historical typhoon data and real-time monitoring information to generate a non-stationary load sequence that includes wind pressure pulsations and liquid level changes. A three-dimensional finite element model incorporating multi-field coupling mechanisms is established to realistically reproduce the stress response and overall buckling evolution of the storage tank under dynamic composite loads. Based on displacement thresholds, local stress concentration is determined, and a damage accumulation model is used to assess the deterioration of resistance and damage propagation under multiple typhoon events. The system systematically reveals the continuous evolution path from overall instability to the initiation of local damage and the formation mechanism of high-risk areas. Finally, by integrating failure mode prediction and remaining life assessment frameworks, a quantitative evaluation and risk warning of the storage tank's full life-cycle service status under extreme composite environments is achieved, significantly improving the structural safety and operational reliability of the storage tank in typhoon-prone areas. Attached Figure Description

[0070] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0071] Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0072] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0073] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0074] Example 1

[0075] like Figure 1 As shown, this embodiment provides a method for simulating the buckling state of a storage tank under the combined effects of typhoon and precipitation, including the following steps:

[0076] Collect historical records and real-time monitoring data of typhoon events to obtain composite load parameters containing wind pressure pulsation and liquid level change characteristics, and generate non-stationary load sequences;

[0077] Based on the aforementioned non-stationary load sequence, a three-dimensional geometric model of the storage tank is constructed using the finite element analysis method, and a multi-field coupling mechanism is incorporated to determine the initial stress distribution state of the structure under composite loads.

[0078] A dynamic load sequence is applied to the initial stress distribution state to simulate the overall buckling instability process and obtain the structural deformation morphology and key node displacement data.

[0079] If the displacement data of key nodes exceeds the preset threshold, it is determined that the local stress concentration stage has been entered, and the stress concentration distribution map of the weld and the reinforcing ring is obtained.

[0080] Based on the stress concentration distribution map, the cumulative damage value under multiple typhoon events is calculated using a damage accumulation model to determine the degree of resistance degradation.

[0081] Based on the aforementioned degree of resistance degradation, a continuous evolution path from overall buckling instability to the initiation of local damage is simulated to obtain the damage propagation trajectory;

[0082] If the damage propagation trajectory shows the initiation of localized damage, the formation mechanism of the high-risk area is analyzed to obtain the failure mode prediction results;

[0083] Based on the failure mode prediction results and combined with the remaining service life assessment framework, the remaining service life index of the storage tank under combined loads is obtained.

[0084] Specifically, the following steps are included:

[0085] Step S101: By collecting historical records of typhoon events and real-time monitoring data, composite load parameters, including wind pressure pulsation and liquid level change characteristics, are obtained to obtain a non-stationary load sequence for subsequent simulation input.

[0086] Specifically, raw wind pressure and liquid level sequences are obtained by collecting historical typhoon data and real-time monitoring data. Based on these sequences, time series decomposition is used to separate trend and pulsation components, resulting in wind pressure pulsation and liquid level fluctuation sequences. Non-stationarity is assessed in these sequences; if the sequence variance exceeds a preset threshold over a time window, non-stationarity is confirmed, resulting in a non-stationary labeled sequence. Composite load components for corresponding time periods are extracted from the wind pressure pulsation and liquid level fluctuation sequences based on the non-stationary labeled sequence, resulting in a composite load sequence. A random forest model is used to rank the composite load sequences by feature importance to determine the dominant load parameters, resulting in a dominant parameter sequence. Finally, the composite load sequences are weighted and synthesized based on the dominant parameter sequence to generate the final non-stationary load sequence.

[0087] In the monitoring of coastal engineering structures under the influence of typhoons, the original wind pressure sequence and liquid level sequence are obtained by collecting historical typhoon records and real-time monitoring data.

[0088] For example, during a typhoon, the wind pressure values ​​recorded every minute form the original wind pressure sequence, such as gradually decreasing from 1000 hPa to 960 hPa and then rising again, while the liquid level sequence reflects the tidal changes caused by storm surge.

[0089] Specifically, a time series decomposition method is used to separate the trend component and the pulsation component.

[0090] For example, using the moving average method or STL decomposition, the slow changing trend (such as an overall decrease in air pressure) in the original wind pressure sequence can be separated, and the remaining part is the wind pressure pulsation sequence, which mainly captures the rapid fluctuations caused by gusts. Similarly, decomposing the liquid level sequence yields the liquid level fluctuation sequence, reflecting the instantaneous fluctuations of waves and swells. This decomposition helps isolate long-term environmental changes from short-term random disturbances, facilitating subsequent targeted analysis. Non-stationarity is then assessed for both the wind pressure pulsation sequence and the liquid level fluctuation sequence.

[0091] In this embodiment, a sliding window is used to calculate the variance. If the variance rises sharply from 0.5 to 2.5 within a certain time window, exceeding a preset threshold of 1.0, the segment is marked as non-stationary, thus generating a non-stationary labeled sequence. This step accurately identifies the intense turbulent phase when the typhoon eyewall passes, avoiding the application of the stationary assumption to a strongly non-stationary process and improving the reliability of the analysis. Based on the non-stationary labeled sequence, the composite load components for the corresponding time periods are extracted from the wind pressure pulsation sequence and the liquid level fluctuation sequence.

[0092] For example, only data from periods marked as non-stationary, such as the overlapping portion of wind pressure pulsations reaching a peak of 15 m / s and liquid level fluctuations reaching 1.2 m, are extracted to form a composite load sequence. This extraction focuses on the most dangerous coupled load periods, significantly reducing the amount of data and highlighting key risk intervals. A random forest model is used to rank the composite load sequence by feature importance to determine the dominant load parameters.

[0093] In this embodiment, the model inputs features such as the root mean square of wind pressure fluctuations, peak factor, and energy spectrum of liquid level fluctuations. The output shows that the importance score of the root mean square of wind pressure fluctuations is 0.65, and that of liquid level fluctuation energy is 0.25, thus selecting the dominant parameter sequence. This ranking objectively reveals that wind loads often dominate the structural response during extreme typhoons, helping to prioritize the main threat factors. The composite load sequence is then weighted and synthesized based on the dominant parameter sequence to generate the final non-stationary load sequence.

[0094] For example, by using importance scores as weights, wind pressure pulsation components are weighted by 0.65 and liquid level fluctuations by 0.25, and then linearly combined to obtain a synthesized non-stationary load sequence. This sequence more realistically reflects the comprehensive extreme loads actually borne by the structure and can be used for subsequent wind and wave resistance design verification, significantly improving the accuracy of safety margin and durability assessment of engineering structures in typhoons.

[0095] Step S102: Based on the obtained non-stationary load sequence, a three-dimensional geometric model of the storage tank is constructed using the finite element analysis method, incorporating a multi-field coupling mechanism to determine the initial stress distribution state of the structure under composite load.

[0096] Specifically, the time step is divided according to the non-stationary load sequence to obtain multiple discrete load time points. A three-dimensional geometric model of the storage tank is constructed using these discrete load time points, resulting in a meshed finite element model. By incorporating a multi-field coupling mechanism into the finite element model, the coupling influence coefficients of the temperature field and flow field on the structure are determined. Based on these coupling influence coefficients, a composite load is applied to obtain the instantaneous stress field of the storage tank structure at each load time. The instantaneous stress field sequence is then time-sequentially superimposed to determine the cumulative initial stress distribution of the storage tank structure.

[0097] For example, in dividing the time step based on a non-stationary load sequence, the entire load duration can be divided into multiple small time windows, each representing a discrete load moment. Assuming the typhoon's impact on the storage tank lasts 24 hours, this can be divided into 24 discrete moment points, with one time step per hour, to facilitate subsequent analysis of the load at each moment. This division method captures the load's time-varying characteristics more precisely, providing accurate data support for subsequent modeling.

[0098] For example, when constructing a three-dimensional geometric model of a storage tank, the geometry can be generated using computer-aided design tools based on data from discrete load time points, combined with the tank's actual dimensions and material properties. Assuming the tank is 10 meters high, 5 meters in diameter, and 0.02 meters thick, the model can be divided into tens of thousands of small elements through mesh generation for finite element analysis. This refined meshing method helps improve the accuracy of the simulation, especially in the stress distribution analysis of key parts of the tank under load.

[0099] For example, when incorporating multi-field coupling mechanisms into the finite element model, the influence of temperature and flow fields on the tank structure can be considered. Assuming a typhoon environment where the external temperature of the tank drops sharply from 25 degrees Celsius to 15 degrees Celsius, accompanied by changes in the flow field caused by strong winds, the coupling influence coefficient can be determined by analyzing the temperature gradient and wind speed distribution. This coefficient reflects the comprehensive impact of the multi-field environment on the tank structure, providing a basis for applying subsequent composite loads and helping to more realistically simulate the tank's response in complex environments.

[0100] For example, when applying composite loads and obtaining the instantaneous stress field, loads such as wind pressure and liquid level changes can be applied to the tank model based on the previously determined coupling effect coefficients. Assuming that at a certain point in time, the peak wind pressure is 2000 Pa and the liquid level fluctuation is 0.5 meters, the instantaneous stress distribution on the tank wall and bottom can be obtained through finite element simulation. This analysis can help identify weak areas of the tank at specific moments, providing an important reference for structural safety assessment.

[0101] For example, when performing time-series superposition of instantaneous stress field sequences, stress distribution data at 24 discrete time points can be cumulatively analyzed in chronological order. Assuming that the stress value in a certain region at the bottom of the tank continuously increases over several hours, reaching near the critical value of the material's yield strength, time-series superposition can clearly show the trend of this cumulative effect. This method helps assess the fatigue damage risk of the tank under long-term loading.

[0102] For example, to determine the cumulative initial stress distribution of a storage tank structure, the stress distribution diagram of each part of the tank can be drawn using the aforementioned time-series overlay results. The analysis suggests that the stress concentration is most significant at the top connection point of the tank, while the stress distribution is more uniform in the bottom region. This provides data support for subsequent structural optimization design and helps to specifically strengthen the wind resistance of key components. This analytical method can effectively improve the stability of storage tanks in extreme environments.

[0103] Step S103: Apply a dynamic load sequence to the initial stress distribution state using the finite element analysis method to simulate the overall buckling instability process and obtain the structural deformation morphology and key node displacement data.

[0104] Specifically, a structural model is constructed using the finite element method (FEM). Mesh generation and boundary condition settings are applied to the initial stress distribution to obtain a digital representation of it. Based on this digital representation, a dynamic load sequence is applied, and loading simulation is performed using a preset time step to obtain stress response data at different time points. The buckling instability behavior of the overall structure is analyzed based on this stress response data. If the stress response exceeds a preset threshold, the instability state data at the corresponding time point is recorded. Deformation morphology features are extracted from the instability state data to obtain the geometric deformation distribution of the structure during buckling instability and determine the evolution law of the deformation morphology. Based on the evolution law of the deformation morphology, the location information of key nodes is identified, and the dynamic change record of the displacement data of key nodes is obtained by calculating the coordinate changes of these key nodes at different time points.

[0105] For example, based on the initial stress distribution of the tank structure, when meshing, tetrahedral or hexahedral elements can be used to refine the tank wall, bottom, and top regions, ensuring higher mesh density at critical connection points such as welds, thereby obtaining a more accurate digital representation of the initial stress. This representation is usually stored in the form of nodal stress vectors, which facilitates subsequent dynamic loading simulations.

[0106] Specifically, when applying a dynamic load sequence based on the digital representation of the initial stress distribution, seismic waves or wind loads can be selected as non-stationary sequences, with a time step of 0.02 seconds to ensure the capture of rapidly changing load peaks.

[0107] In this embodiment, the initial stress is first imported into the finite element model as a prestress field, and then dynamic loads are superimposed step by step over time to calculate the stress response data at each moment. These data reflect the actual stress state of the storage tank under complex conditions, which helps to improve the simulation accuracy.

[0108] It should be noted that when analyzing the buckling instability behavior of the overall structure based on stress response data, this can be determined by monitoring whether the principal stress or equivalent stress exceeds the material's yield strength.

[0109] For example, when the equivalent stress in the middle of the tank wall reaches 350 MPa while the material's yield strength is 345 MPa, an exceedance of the threshold is detected, and the instability data at that point in time is recorded. This timely detection helps to identify potential risks early and avoid sudden structural failure.

[0110] In this embodiment, when extracting deformation morphology features from instability state data, attention can be paid to the distribution patterns of radial and axial displacements to obtain the non-uniform characteristics of geometric deformation in the height direction of the tank, thereby determining the law of deformation morphology evolving from local bulges to overall ellipticization. This evolution law reveals the gradual failure process of the storage tank under continuous load.

[0111] Preferably, when locating key nodes based on the evolution of deformation patterns, the circumferential weld node with the highest stress concentration can be selected as the monitoring object. By tracking its coordinate changes at different time points, a dynamic record of the displacement amplitude gradually increasing from an initial 2mm to 15mm can be obtained. This record provides a direct basis for subsequent reinforcement design.

[0112] For example, from another perspective, when key node displacement data indicates a sudden acceleration in the rate of lateral displacement, the critical moment of buckling instability can be further verified, corroborating the aforementioned stress threshold judgment and forming a complete instability assessment chain. These analyses collectively enhance the understanding of tank structural stability, effectively support the optimization of seismic or wind resistance capabilities, and significantly reduce safety hazards in actual operation.

[0113] Step S104: If the displacement data of the key node exceeds the preset threshold, it is determined that the local stress concentration stage has been entered, and the stress concentration distribution map at the weld and the reinforcing ring is obtained.

[0114] Specifically, regarding the dynamic changes in displacement data of key nodes, if the data exceeds a preset threshold, the structural judgment process is triggered to determine the state of entering the local stress concentration stage. Based on the state of the local stress concentration stage, the geometric information of the weld location and the reinforcing ring is obtained, and a preliminary stress distribution model of the corresponding area is generated. For the preliminary stress distribution model, the weld location and the reinforcing ring are meshed using the finite element analysis method to obtain high-precision stress distribution data. Using the high-precision stress distribution data, a stress distribution chart is constructed to visualize the characteristics of the local stress concentration stage and determine the specific shape of the distribution chart. Based on the specific shape of the distribution chart, the analysis results of key nodes are updated, and the response data of the nodes in the local stress concentration stage is recorded to obtain detailed basis for stage division. Based on the detailed basis for stage division, if the response data continues to deviate from the preset threshold, the structural judgment results are iteratively processed to generate the final stress concentration distribution map.

[0115] For example, after the overall buckling instability simulation of the structure is completed by finite element analysis, monitoring the dynamic changes of displacement data at key nodes is a common follow-up processing method.

[0116] Specifically, when the displacement curve suddenly accelerates and exceeds a preset threshold, such as five times the initial displacement, the structural assessment process is triggered to confirm that the system has entered the local stress concentration stage. This threshold is selected based on historical simulation experience and can effectively capture the transition from overall deformation to local failure, helping to identify potential risk points in advance.

[0117] In this embodiment, after entering the local stress concentration stage, the geometric information of the weld location and the reinforcing ring is first obtained. For example, the weld is distributed along the longitudinal direction of the cylinder in multiple circumferential sections, while the reinforcing ring is located in the critical support area. A preliminary stress distribution model is generated using these geometric parameters. This model can simplify the overall structure to the area of ​​interest, significantly reducing the amount of subsequent calculations while maintaining the accuracy of local features.

[0118] For example, when meshing welds and reinforcing rings, a finer mesh is preferred, such as keeping the unit size within 5mm near the weld and gradually increasing it to 20mm further away. This high-precision mesh can capture dramatic changes in stress gradients, ensuring reliable and accurate stress distribution data. The resulting data reveals that the maximum principal stress at the weld is often more than 1.5 times higher than in the surrounding area, thus providing a solid foundation for subsequent visualization.

[0119] Specifically, stress distribution charts are constructed using high-precision stress distribution data. For example, cloud maps can be used to present the characteristics of localized stress concentration stages, with a color gradient from blue to red indicating that the stress increases from low to high. This visualization can intuitively show the specific shape of stress concentration areas, such as elliptical or strip-shaped distributions, helping engineers quickly locate high-risk areas and improve analysis efficiency.

[0120] In this embodiment, the analysis results of key nodes are updated based on the shape of the distribution chart. For example, nodes that originally had large displacements during the overall buckling stage may experience stress peaks shifting to adjacent weld nodes during the local stage. The response data of these nodes during the local stress concentration stage are recorded. If the stress increase reaches 30%, it can serve as a detailed basis for stage division, supporting more accurate failure mode determination.

[0121] For example, if the response data continuously deviates from the preset threshold, such as exceeding the safety factor of 1.2 times for 10 consecutive time steps, the structural assessment results are iteratively processed. Through multiple iterations and convergence, a final stress concentration distribution map can be generated. This map not only integrates the entire process data from the overall to the local level but also highlights multiple potential crack initiation points. This iterative mechanism helps improve the robustness of the simulation, avoids misjudgments caused by a single threshold judgment, and ultimately provides more comprehensive technical support for structural safety assessment.

[0122] Step S105: Based on the stress concentration distribution diagram, use the damage accumulation model to calculate the cumulative damage value under multiple typhoon events and determine the degree of resistance degradation.

[0123] Specifically, stress concentration distribution data is acquired graphically, and data is extracted from the affected areas of typhoon events. Key locations of stress concentration are identified through comparison with a pre-established database. Based on these key locations, a damage accumulation model is used for calculation. Data on the frequency and intensity of multiple typhoon events is integrated to obtain preliminary values ​​for cumulative damage. These preliminary values ​​are then combined with environmental factor impact data, and a linear regression model is used to correct the damage values, determining the trend of damage accumulation. Based on this trend, relevant parameters for structural resistance degradation are obtained. If these parameters exceed preset thresholds, structural safety is assessed in a tiered manner to determine the specific level of structural resistance degradation. For each level of degradation, impact data from multiple typhoon events is acquired, and key risk points for structural safety are identified through comparative analysis of historical records. Based on these key risk points and the distribution data of event impacts, automated analysis tools are used to comprehensively process the assessment results, yielding the final degree of structural resistance degradation.

[0124] For example, when acquiring stress concentration distribution data, preliminary processing of stress data collected by monitoring equipment in typhoon-affected areas can be performed. Combined with a geographic information system (GIS), the data can be mapped to specific structural locations. Suppose a steel bridge in a coastal area experiences significant wind load during a typhoon, and monitoring points show a stress value of 120 MPa at a critical node, exceeding the normal range. Database comparison confirms this node as a critical location of stress concentration. This method can quickly pinpoint risk areas, providing accurate data for subsequent analysis.

[0125] For example, when using a damage accumulation model to calculate stress concentration at critical locations, fatigue damage to the structure can be estimated based on the frequency and intensity data of historical typhoon events. Assuming the bridge experienced 10 typhoons in the past 5 years, with an average wind speed of 30 meters per second per typhoon, and combining this with the material's fatigue life curve, the preliminary calculated cumulative damage value is approximately 0.3. This analysis helps to understand the damage accumulation of the structure under multiple external forces, providing data support for further corrections.

[0126] For example, when correcting damage values ​​for environmental factors, linear regression models can be used to analyze the impact of factors such as humidity and salt spray corrosion on damage trends. Assuming environmental monitoring shows that the average annual humidity in the area is 85% and the salt spray concentration is high, regression analysis reveals that humidity has a damage aggravation coefficient of 1.2. Therefore, the corrected damage value may increase to 0.36. This correction method more closely reflects the actual environmental impact on structures, enhancing the reliability of the assessment.

[0127] For example, when obtaining resistance degradation parameters based on changes in the cumulative damage trend, if a parameter such as the material's yield strength drops to 80% of the original design value, exceeding a preset threshold, a graded assessment needs to be initiated. Assuming the assessment results show a resistance degradation level of two, it indicates that the structure's load-bearing capacity has significantly decreased. This grading method provides a clear reference standard for subsequent risk management.

[0128] For example, when analyzing the impact data of multiple typhoon events, historical data comparison can reveal that a critical connection point has exhibited minor deformations in several past typhoons, with a cumulative deformation of 2 millimeters, identifying it as a critical risk point for structural safety. This comparative analysis can help identify potential weak points and provide a basis for targeted maintenance.

[0129] For example, for key structural safety risks, automated analysis tools can integrate multi-dimensional data such as wind load, stress, and deformation, combined with event impact distribution data, to generate a comprehensive assessment report. Suppose the report indicates that the resistance degradation level of a certain area is moderate, it recommends increasing the monitoring frequency. This comprehensive approach improves assessment efficiency and ensures full coverage of risk points. Through the above analysis and examples, it can be seen that the entire process, from obtaining stress distribution data to the final assessment of resistance degradation, is interconnected and logically rigorous, providing a scientific basis for structural safety management while effectively reducing potential risks from natural disasters such as typhoons.

[0130] Step S106: Using the degree of resistance degradation output by the damage accumulation model, simulate the continuous evolution path from overall buckling instability to the initiation of local damage to obtain the damage propagation trajectory.

[0131] Specifically, damage accumulation models are used to obtain structural resistance degradation data. Stress distribution under the initial state is calculated to obtain preliminary results of resistance changes. Key points of overall buckling are extracted from these preliminary results, and finite element analysis is used to simulate buckling behavior, identifying stress concentration areas under instability. For these stress concentration areas under instability, the conditions for local failure initiation are analyzed. If the stress value exceeds a preset threshold, the location of failure initiation is recorded, obtaining data on the failure initiation point. Based on the failure initiation point data, the evolution of failure initiation is simulated, tracking the trajectory of stress concentration areas and determining the continuity characteristics of the evolution process. Damage propagation paths are extracted from these continuity characteristics, and the stress transfer patterns along these paths are analyzed to obtain the distribution pattern of the propagation trajectory.

[0132] For example, when obtaining structural resistance degradation data using damage accumulation models, one can first understand the basic principle of damage accumulation models from a theoretical perspective: the gradual accumulation of microscopic damage under long-term stress reflects the overall performance decline of the structure. Assuming a steel bridge in a coastal area initially has stress distribution data collected by sensors at 50 MPa per square centimeter, combined with historical typhoon data, the model will simulate the stress accumulation effect after multiple typhoons, and preliminarily determine the extent of resistance degradation. This method helps to identify potential performance problems early.

[0133] For example, to extract the key points of overall buckling from the preliminary results of resistance changes, structural mechanics analysis can be used to identify the connection nodes in the middle of the main beam of the bridge as key points, because these points often bear the maximum bending moment.

[0134] In this embodiment, finite element analysis software was used to model the location and simulate its deformation behavior under typhoon load. This revealed the possibility of local buckling in the middle of the main beam, thus identifying the stress concentration area under instability conditions. This simulation provides precise location information for subsequent analysis.

[0135] For example, when analyzing the initiation conditions of localized failure in stress concentration areas under unstable conditions, a stress threshold of 80 MPa per square centimeter can be set. If the sensor detects a stress of 85 MPa at a certain node, this location is recorded as the failure initiation point. The likelihood of failure initiation is then analyzed in conjunction with environmental factors such as salt spray corrosion. This method helps identify the most vulnerable parts of the structure, providing a targeted basis for maintenance.

[0136] For example, simulations of the evolution of the damage initiation point can use time-step analysis to track the changing trajectory of stress concentration areas under successive typhoons. Assuming that after the third typhoon, the damaged area expands from an initial 2 square centimeters to 5 square centimeters, exhibiting a continuous expansion characteristic, this tracking helps determine whether the damage will further deteriorate, providing data support for preventative measures.

[0137] For example, when extracting the damage propagation path and analyzing the stress transfer pattern, it can be observed that the damage extends from the middle of the main beam towards the supports on both sides, forming a linear path. By analyzing the stress values ​​at each point along the path, it is found that the stress transfer shows a gradually weakening trend. This analysis can reveal the potential direction of damage propagation and provide a reference for reinforcement design.

[0138] For example, based on the above analysis, we can start from the core damage accumulation model and gradually expand to the analysis of buckling simulation, failure initiation and propagation paths, forming a complete assessment system.

[0139] In this embodiment, to address the issue of structural weakness in coastal steel bridges, a phased monitoring plan can be developed by combining typhoon frequency and intensity data to ensure structural safety. The benefit of this multi-level analysis is that it allows for a comprehensive understanding of structural performance changes, enabling timely intervention and extending service life.

[0140] Step S107: If the damage propagation trajectory shows the initiation of local damage, determine the formation mechanism of the high-risk area and obtain the failure mode prediction result.

[0141] Specifically, trajectory monitoring technology is used to obtain dynamic information on damage propagation from structural data, determining the initial damage distribution range. Based on the damage distribution range, image processing methods are used to finely segment local damaged areas, obtaining specific location data for local damage. For the specific location data of local damage, if the detected signs of nascent damage exceed a preset threshold, a high-risk area marking process is triggered to determine the potential high-risk area range. From the high-risk area range, relevant feature data of the formation mechanism are obtained, and the feature data is classified using a support vector machine algorithm to determine the main types of failure modes. For the main types of failure modes, trajectory analysis technology is used to dynamically track the trend of damage propagation, obtaining predicted data of the propagation trend. Based on the predicted data of the propagation trend, if the predicted data indicates that the risk assessment value exceeds the safe range, an adjustment plan for area division is generated, determining new key protection areas. Using data from the key protection areas, combined with mode derivation technology, the evolution path of failure modes is simulated to obtain the final failure mode prediction results.

[0142] For example, when acquiring dynamic information about structural data through trajectory monitoring technology, consider a scenario for long-term monitoring of a large steel bridge. Multiple sensors are deployed to collect stress change data in real time. If the stress value at a critical node continuously rises over a period of time, reaching a preset warning value of 50 MPa, the system will automatically mark that area as the initial damage distribution range. This approach can promptly identify potential risk points, laying the foundation for subsequent analysis.

[0143] For example, image processing methods targeting the extent of damage can use high-resolution cameras to capture surface images of localized areas of the bridge. Combined with edge detection algorithms, the crack width in the damaged area can be precisely determined to the 0.1 mm level, segmenting the specific damage location. This refined segmentation helps identify areas requiring focused attention, improving the targeting of subsequent processing.

[0144] For example, when detecting localized damage locations and identifying signs of budding, if the crack length in a certain area exceeds a preset threshold of 5 millimeters, the system will trigger a high-risk area marking process, automatically designating a 2-meter radius around that area as a potential high-risk zone. This method can quickly pinpoint areas that may deteriorate further, buying time for protective measures.

[0145] For example, when classifying characteristic data for high-risk areas using the support vector machine algorithm, features such as crack propagation rate and stress concentration can be extracted. Assuming a crack propagation rate of 0.2 mm per day in a certain area, and combining this with historical data, it can be determined that the failure mode is fatigue. This classification method helps to clarify the failure mechanism and provides a basis for subsequent predictions.

[0146] For example, when dynamically tracking damage propagation trends using trajectory analysis technology, the direction of crack propagation can be predicted based on sensor data. If the prediction indicates that the crack will propagate to a critical load-bearing node within the next 30 days, and the risk assessment score reaches 80 points, exceeding the safety threshold of 60 points, the system will generate an adjustment plan. This prediction provides data support for adjusting protective measures.

[0147] For example, when generating a zone division adjustment plan, if the predicted data indicates that the risk exceeds the standard, the key protection area can be expanded from the original 10 square meters to 20 square meters, and reinforcement materials can be deployed first. This adjustment can effectively cover potentially risky areas and improve the safety of the structure.

[0148] For example, simulations of failure mode evolution paths in key protected areas can use mode derivation techniques, combined with historical stress data and environmental factors, to predict the overall instability modes that crack propagation may lead to. This could be achieved by assuming that the simulation results show the crack will induce local buckling within 60 days. Such simulations help to develop proactive countermeasures and reduce the risk of failure.

[0149] Step S108: Based on the failure mode prediction results, integrate them into the remaining service life assessment framework to obtain the remaining service life index of the storage tank under combined loads.

[0150] Specifically, for failure mode analysis, key data is extracted from the prediction results. Information extraction techniques are used to classify and organize the data to obtain mode classification information. Based on the mode classification information, combined with the remaining service life assessment framework, data matching techniques are used to compare the classification information with preset service life assessment standards to determine the preliminary assessment results of the tank status. If an abnormal service status is detected under combined load conditions based on the preliminary assessment results of the tank status, a deep data mining process is triggered to obtain detailed change data of the service status. Based on the detailed change data of the service status, a support vector machine algorithm is used to classify the load impact and determine the status change trend under the load influence. Based on the status change trend under the load influence, mode analysis techniques are used to map the correlation between failure modes and service life indicators to obtain preliminary derivation results of the service life indicators. Based on the preliminary derivation results of the service life indicators, combined with status derivation techniques, the long-term performance of the tank under combined loads is simulated to determine the final remaining service life indicators. For the final remaining service life indicators, data storage techniques are used to archive them into the assessment framework's database to obtain the archived complete assessment information.

[0151] For example, in failure mode analysis, information extraction techniques can be used to extract key data from prediction results. Information extraction techniques primarily involve structuring large amounts of data to transform unstructured prediction information into categorizable pattern data.

[0152] Specifically, assuming the tank failure prediction results contain multiple failure signals, such as crack propagation rate and stress concentration point location data, information extraction technology will categorize and organize these signals into hierarchical patterns, such as classifying crack propagation rate as "structural damage" and stress concentration points as "local failure." This helps to quickly locate the core failure issue.

[0153] For example, when combining pattern classification information with a remaining useful life assessment framework, data matching technology can compare the classification information with preset standards. Suppose the preset standard defines a crack propagation rate exceeding 0.5 mm / year as a high-risk condition; data matching technology can quickly determine whether the current tank's crack propagation rate meets this standard. If a tank's rate is 0.7 mm / year, the preliminary assessment will mark it as an area requiring close monitoring. This method improves the accuracy of the assessment.

[0154] For example, when detecting abnormal service conditions under combined load conditions, the data deep mining process further analyzes the detailed change data of the storage tank. Suppose the storage tank exhibits abnormal vibration signals under high temperature and high pressure combined loads, data deep mining will extract detailed information such as vibration frequency and amplitude to determine whether the anomaly is caused by the superposition of loads. This refined analysis helps to uncover hidden risks.

[0155] For example, when using the support vector machine algorithm to classify load effects, different load types, such as temperature stress and mechanical stress, can be distinguished. Assuming that temperature stress causes localized deformation of the tank, while mechanical stress induces overall fatigue, the algorithm can classify which load has a greater impact on the trend of state changes, thus providing a targeted basis for subsequent protection.

[0156] For example, by mapping the correlation between failure modes and life indicators using pattern analysis, the specific impact of different failure modes on life can be deduced. Assuming a positive correlation between crack propagation mode and life reduction, the analysis technique can simulate, based on historical data, a preliminary deduction that every 1 mm increase in crack size may lead to a 100-hour reduction in life. This mapping relationship provides data support for life prediction.

[0157] For example, when simulating the long-term performance of storage tanks under combined loads using state derivation techniques, future performance can be predicted based on current state data. Assuming the tank operates at sustained high temperatures, the derivation technique simulates accelerated material fatigue, ultimately determining the remaining service life to be two years. This simulation provides a reference for maintenance planning.

[0158] For example, when archiving the final remaining service life indicators, data storage technology integrates them into the assessment framework database. Assuming a storage tank has a service life indicator of 2 years, after archiving, it can be linked with other historical assessment information to form a complete assessment information database. This archiving method facilitates subsequent querying and comparative analysis, improving management efficiency.

[0159] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for simulating the buckling state of a storage tank under the combined effects of typhoon precipitation, characterized in that, The method comprises the following steps: Collecting typhoon event history records and real-time monitoring data, obtaining composite load parameters containing wind pressure fluctuation and liquid level change characteristics, and generating a non-stationary load sequence; Based on the non-stationary load sequence, a three-dimensional geometric model of the storage tank is constructed by using a finite element analysis method, and a multi-field coupling mechanism is integrated to determine the initial stress distribution state of the structure under the composite load; A dynamic load sequence is applied to the initial stress distribution state to simulate the overall buckling instability process and obtain the structural deformation morphology and key node displacement data; If the key node displacement data exceeds a preset threshold, it is determined that the local stress concentration stage is entered, and a stress concentration distribution map at the weld and the reinforcing ring is obtained; Based on the stress concentration distribution map, a damage accumulation model is used to calculate the cumulative damage value under multiple typhoon events to determine the resistance degradation degree; Based on the resistance degradation degree, the continuous evolution path from overall buckling instability to local damage initiation is simulated to obtain the damage propagation trajectory; If the damage propagation trajectory shows local damage initiation, the formation mechanism of the high-risk area is analyzed to obtain the failure mode prediction result; Based on the failure mode prediction result, combined with a residual life assessment framework, the residual service life index of the storage tank under the composite load is obtained.

2. The method of claim 1, wherein, The generation of the non-stationary load sequence comprises: Collecting original wind pressure sequence and liquid level sequence; Using a time series decomposition method to separate trend components and fluctuation components to obtain wind pressure fluctuation sequence and liquid level fluctuation sequence; Judging the non-stationarity of the wind pressure fluctuation sequence and the liquid level fluctuation sequence, if the sequence variance changes with the time window exceeding a preset threshold, it is marked as a non-stationary sequence; Based on the non-stationary marked sequence, composite load components are extracted to form a composite load sequence; Using a random forest model to sort the feature importance of the composite load sequence to determine the dominant load parameter sequence; Based on the dominant parameter sequence, the composite load sequence is weighted and synthesized to generate the final non-stationary load sequence.

3. The method of claim 1, wherein, The determination of the initial stress distribution state of the structure under the composite load comprises: According to the non-stationary load sequence, time steps are divided to obtain discrete load time points; Based on the discrete load time points, a three-dimensional geometric model of the storage tank is constructed and meshed to obtain a finite element model; The multi-field coupling mechanism of temperature field and flow field is integrated into the finite element model to determine the coupling influence coefficient; According to the coupling influence coefficient, the composite load is applied to obtain the instantaneous stress field at each load time; The instantaneous stress field is time-series superimposed to determine the cumulative initial stress distribution.

4. The method of claim 1, wherein, The obtaining of the structural deformation morphology and key node displacement data comprises: Based on the initial stress distribution state, a finite element model is constructed and boundary conditions are set; The dynamic load sequence is applied, and the loading simulation is performed at a preset time step to obtain stress response data; If the stress response exceeds a preset threshold, the instability state data at the corresponding time point is recorded; From the instability state data, the deformation morphology characteristics are extracted to determine the deformation evolution law; Based on the deformation evolution law, the key nodes are located, and the displacement data thereof is calculated.

5. The method of claim 1, wherein, The stress concentration distribution map at the weld and the reinforcing ring comprises: If the key node displacement data exceeds a threshold, it is determined that the local stress concentration stage is entered; Obtain the geometric information of the weld and the reinforcing ring, and establish a preliminary stress distribution model; Perform fine meshing on the weld and reinforcing ring area to obtain high-precision stress distribution data; Based on the high-precision stress distribution data, construct a stress distribution chart to visually present the stress concentration area; Update the key node response data based on the distribution chart, iteratively optimize the determination result, and generate the final stress concentration distribution chart.

6. The method of claim 1, wherein, The determination of the resistance degradation degree includes: Based on the stress concentration distribution chart, extract the key stress concentration position data; Use the damage accumulation model to calculate the preliminary cumulative damage value by combining the occurrence frequency and intensity of multiple typhoon events; Introduce environmental factor data to correct the preliminary cumulative damage value and determine the damage accumulation trend; Determine the resistance degradation level based on the damage accumulation trend; Identify the structure safety critical risk points by combining historical typhoon event data; Output the resistance degradation degree after comprehensive evaluation.

7. The method of claim 1, wherein, The damage propagation trajectory includes: Based on the resistance degradation data, extract the overall buckling key points; Use the finite element analysis method to simulate the buckling behavior and determine the stress concentration area in the instability state; Analyze the local damage initiation conditions. If the stress value exceeds the preset threshold, record the damage initiation point; Simulate the damage initiation and evolution process, and track the change trajectory of the stress concentration area; Extract the damage propagation path from the evolution process, analyze the stress transmission rule, and obtain the damage propagation trajectory.

8. The method of claim 1, wherein, The failure mode prediction result includes: Determine the damage distribution range based on the damage propagation trajectory; Based on the damage distribution range, perform fine segmentation on the local damage area to obtain specific location data; Determine whether the damage initiation signs exceed the preset threshold based on the specific location data. If they do, mark them as high-risk areas; Extract the feature data of the high-risk areas and determine the failure mode type using a classification algorithm; Track the damage propagation trend based on the failure mode type to obtain the propagation trend prediction data; If the risk exceeds the safety range based on the propagation trend prediction data, generate a regional adjustment scheme; Simulate the failure mode evolution path based on the regional adjustment scheme to obtain the failure mode prediction result.

9. The method of claim 1, wherein, The remaining service life index of the storage tank under combined load includes: Extract key data and classify them based on the failure mode prediction result; Based on the classification information, combine the preset life evaluation standard, and obtain the preliminary evaluation result through data matching; If the service state is abnormal under combined load based on the preliminary evaluation result, extract detailed change data; Use a classification algorithm to analyze the load influence based on the detailed change data and determine the state change trend; Based on the state change trend, establish the association mapping between the failure mode and the life index, and derive the preliminary life index; Simulate the long-term performance of the storage tank under combined load based on the preliminary life index to determine the remaining service life index; Archive the remaining service life index to the evaluation database to form complete evaluation information.

Images