Fatigue life prediction method of hull structure based on state-based peridynamics

By employing a state-based near-field dynamics approach, spontaneous bond breaking is used to handle bonds exceeding the critical elongation rate. This solves the difficulties of predicting crack propagation paths and fatigue life in three-dimensional ship hull structures using traditional methods, achieving efficient and accurate prediction results.

CN118656986BActive Publication Date: 2024-11-26HARBIN INST OF TECH AT WEIHAI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411125146.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-26
Estimated Expiration
2044-08-16

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately predict crack propagation paths and fatigue life in three-dimensional ship structures. Traditional methods, such as XFEM, are complex and yield inaccurate results, while the applicability of bond-type peri-field dynamics methods is limited by Poisson's ratio.

Method used

A state-based near-field dynamics method is adopted. By establishing a crack propagation model, spontaneous bond breaking is used to handle bonds exceeding the critical elongation. Damage value and cycle number are calculated, and crack propagation prediction and fatigue life prediction are output, avoiding the introduction of additional conditions and criteria.

Benefits of technology

It achieves efficient crack propagation path prediction and fatigue life forecasting for three-dimensional ship hull structures, improves computational efficiency, solves the computational difficulties of commercial software in multi-crack problems, and provides more accurate prediction results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118656986B_ABST
    Figure CN118656986B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of shipbuilding, and in particular to a method for predicting the fatigue life of a hull structure based on state-based peridynamics, which can accurately simulate and predict the crack propagation path of a hull structure, and further realize the prediction of the fatigue life of the hull structure. The method comprises: establishing a hull crack propagation model; judging whether a bond is broken: setting a bond breaking criterion of the hull crack propagation model, and spontaneously breaking the bond that exceeds a critical elongation, after which the bond is damaged and two material points connected by the bond lose interaction; calculating a damage value, and judging the crack state; selecting a fatigue bond breaking number according to the calculated number of cycles, and determining whether the bond is broken. n c The crack growth prediction diagram and fatigue life prediction a-N curve are output, which can realize the crack growth path prediction and fatigue life prediction of plate grid and stiffened plate structure, and provide a new and efficient method for fatigue life prediction of ship hull structure.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of shipbuilding technology, and in particular to a hull structure fatigue life prediction method based on state-based peri-field dynamics, which can accurately simulate and predict the crack propagation path of the hull structure, and further realize the fatigue life prediction of the hull structure. Background Art

[0002] There are structural components such as plate grids and stiffeners in the ship's hull structure. During the operation of the ship, these components will deteriorate due to prolonged shipping time or collisions, such as cracking. Such degradation may shorten the service life of the hull structure and require immediate replacement and repair. However, for minor initial damage, the overall replacement of related components will obviously cause a large-scale increase in the cost of ship maintenance. Therefore, predicting the crack propagation path of damage on ship structural components can provide strong support for the maintenance and repair of ship structural components.

[0003] In traditional continuous medium mechanics, there are situations where derivatives are not defined at discontinuities, which will cause difficulties in solving. At present, the main method for studying crack propagation is XFEM (Extended Finite Element Method), but this method requires adding functions related to displacement and introducing external fracture criteria to control discontinuous displacements. It is cumbersome to use and the calculation results are not accurate enough in complex situations. By establishing integral equations, peridynamics avoids this singularity and can simulate the spontaneous initiation and propagation of cracks. However, the current research on crack propagation by peridynamics is mainly focused on two-dimensional structures, and there is less research on three-dimensional structures. Most of the methods used are bond-type peridynamics, which are not very applicable due to the limitation of Poisson's ratio. Considering that most real structures are three-dimensional structures, and the bond-type peridynamics limits the Poisson's ratio to 0.25 and 0.33, it obviously cannot meet the needs of crack propagation path prediction and fatigue life prediction of structures. Summary of the invention

[0004] In order to overcome the deficiencies in the prior art, the present invention provides a method based on state-type peridynamics, which breaks through the limitation of Poisson's ratio of the original bond-type peridynamics method, and can spontaneously simulate crack propagation without the need to introduce additional conditions and criteria. The method is also applied to three-dimensional structures, which is a fatigue life prediction method for hull structures based on state-type peridynamics that is more in line with practical engineering applications.

[0005] The present invention is achieved by the following measures:

[0006] A method for predicting fatigue life of hull structure based on state-based peridynamics, characterized by comprising the following steps:

[0007] Step 1: Establish a hull crack growth model;

[0008] Step 2: Determine whether the bond is broken: Set the bond breaking criterion of the hull crack growth model, and spontaneously break the bond that exceeds the critical elongation. After the bond is broken, the bond is damaged and the two material points connected by the bond lose interaction;

[0009] Step 3: Calculate the damage value and determine the crack state;

[0010] Step 4: Select the number of fatigue bond breaks n according to the calculated number of cycles c The bonds within are broken;

[0011] Step 5: Output the crack growth prediction diagram and fatigue life prediction aN curve.

[0012] Step 1 of the present invention specifically includes the following contents:

[0013] Step 1-1: Determine and input material parameters, including elastic modulus, Poisson's ratio, density, determine the near-field neighborhood size, the number of material points in each direction, and determine the material point spacing;

[0014] Step 1-2: Determine the number and coordinates of material points: Combine the geometric dimensions of the structure, calculate and determine the number of material points and the coordinates of each material point, and establish a structural geometric model;

[0015] Step 1-3: Discretize and traverse each material point, and determine the number of material points within the near field: use a two-dimensional array to represent the jth near-field material point of the i-th central material point, traverse each material point in the modeling order, perform the first numbering process of all material points according to i, and store them in the array, search for material points in sequence according to the numbering order, set the i-th material point searched as the central material point, find out the number of near-field material points in the near-field neighborhood of the central material point, and perform the second numbering process according to j, store the searched near-field material points in another array, which now represents the jth near-field material point of the i-th central material point, wait until all near-field material points in the near-field area are searched, and start searching for the next central material point until all i-numbered material points are searched and second numbered;

[0016] Step 1-4: Define the initial crack. The 2D structure sets the crack by breaking the bonds between material points on the crack path, while the 3D structure sets the crack by removing the material points at the initial crack position.

[0017] Step 1-5: Perform volume correction and introduce volume correction factor v c (i) Perform volume correction on material points, see formula (1):

[0018]

[0019] Where: i, j is the jth near-field material point of the i-th central material point,

[0020] ξ is the initial bond length, representing the relative position of x and x′.

[0021] δ is the near-field neighborhood size,

[0022] Δ is the distance between material points,

[0023] ξ (i)(j) <δ-Δ means that the material point is completely within the near-field neighborhood,

[0024] δ-Δ≤ξ (i)(j) <δ+Δ means that part of the material point is located in the near field neighborhood.

[0025] ξ (i)(j) ≥δ+Δ means that the material point is completely outside the near field;

[0026] Step 1-6: Calculate the weighted volume and define the weighted volume scalar q (i) , calculate the weighted volume of each central material point, see formula (2),

[0027]

[0028] Where: x (i)(j) is the reference position scalar state field of the bond, representing the scalar bond length between the jth near-field material point of the i-th central material point, ω is the influence function, · is the dot product symbol;

[0029] Step 1-7: Time domain integration: Start the calculation according to the set total time step and time interval. Each step updates the corresponding boundary conditions and force loads, calculates the new expansion rate and force state, and then calculates the new acceleration, velocity, and displacement. Repeat until the set total time step is reached;

[0030] Step 1-8: Apply boundary conditions and force loads. Boundary conditions and force loads are set and input directly in the program, or read and input through a txt file. The force load is in the form of Newton force, pressure and peridynamic body force, and the boundary condition is in the form of displacement, velocity and acceleration.

[0031] Step 1-9: Calculate the expansion rate: Expansion rate θ (i) For small deformations, the expansion rate measures the volumetric strain of the object, representing the relationship between the deformation of the volume around the central material point and the central material point. x (i)(j) The influence of the force state on is defined as formula (3),

[0032]

[0033] In the formula, y (i)(j) is the deformed scalar state field of the bond,

[0034] e (i)(j) is the elongated scalar state;

[0035] Step 1-10: Calculate the force state. Use the force state and motion equation to calculate the displacement, velocity, and acceleration of the material point. The expression of the force state is formula (4):

[0036] In the formula, v is Poisson's ratio, E is elastic modulus, κ is bulk elastic modulus, under small deformation conditions,

[0037] α is the parameter of shear modulus μ, α=15μ / q,

[0038] To stretch the scalar state, all terms are equal.

[0039] e d = e - e i is the deflection of each item in the stretched scalar state, χ is the reference position scalar state field of the key, M is the unit vector direction of the force state, indicating the direction of the force, e represents the elongated scalar state;

[0040] The adaptive dynamic relaxation method in the implicit algorithm is used to solve the motion equation. Artificial damping is introduced and the damping term is determined in each time integral. The motion equation of the material point x with virtual damping is shown in formula (5).

[0041] Where c is the damping coefficient term, the superscript n represents the current time step, and ρ is the current material point density. The velocity vector of the material point χ at time t is,

[0042] T (i)(j) [x, t]<x′-x> represents the force state of material point χ' on the central material point χ at time t, is the acceleration vector of the material point x at time t, b(x, t) is the external volume force density of the material point x at time t,

[0043] H=H(x,δ): {x′∈R:||x′-x||≤δ} represents the neighborhood of the material point χ, and the displacement, velocity, and acceleration corresponding to the material point are calculated according to the equation of motion.

[0044] In step 2 of the present invention, the three-dimensional structural bond breaking criterion based on the critical elongation is used to determine whether the bond is broken. The elongation is first calculated and then compared with the set critical elongation to determine whether the bond is broken. After the break, there is no interaction between the two points and the near-field force becomes 0. If there is no break, the interaction exists until the break or the calculation is completed. The critical elongation and judgment are shown in formulas (6) and (7).

[0045]

[0046]

[0047] In the formula, m (i)(j) ′<ξ> represents the connection state of the bond,

[0048] S (i)(j) is the bond elongation, S0 is the critical value of elongation, G0 is the critical energy release rate of the material, λ (i)(j) ——represents the remaining life of the bond, κ and μ are the bulk modulus and shear modulus respectively, and δ is the near-field neighborhood size.

[0049] The step 3 of calculating the damage value of the present invention is specifically as follows: calculating the damage value of each material point pair at each time step. The damage value of the material point pair is defined as a quantity describing the damage degree of the material point x at time t. The expression is shown in formula (8),

[0050] It represents the damage value of material point x at time t. The damage value of material point x ranges from 0 to 1. When the damage index of a material point x is 0, it means that no bonds in the structure are broken at this point. When the damage index of a material point x is 1, it means that all bonds connected to the material point are broken and the area is completely destroyed. By calculating the damage value, the degree of damage at this point is obtained to describe the degree of damage to the structure.

[0051] Step 4 of the present invention comprises the following contents:

[0052] Step 4-1: Determine the state of the material point and calculate the number of cycles for each bond: Calculate the number of cycles for each bond according to the following formulas (9), (10), and (11). If it is the crack initiation stage, use the A1 and P1 parameters for calculation. If it is the crack propagation stage: use the A2 and P2 parameters for calculation. By calculating the damage value, if it is below 0.5, it is considered to be the crack initiation stage, otherwise it is the crack propagation stage:

[0053]

[0054]

[0055]

[0056] Where: N (i)(j) is the number of cycles, λ (i)(j) is called the remaining life, ε (i)(j) is the cyclic strain of the bond, and are the maximum and minimum elongation of the bond in one cycle respectively, R is the stress ratio, and A and P are unknown constants.

[0057] Step 4-2: Perform bond breaking and select the n with the minimum number of cycles required c keys, so that this c The key is damaged, this c The largest number of cycles corresponding to the keys is recorded as ΔN1, and the total number of cycles also increases by ΔN1;

[0058] Step 4-3: For the bonds that have not been damaged, the damage generated in the ΔN1 load cycles is accumulated in the bonds, ready for the next quasi-static analysis;

[0059] Step 4-4: Determine whether the time step has been reached. If so, proceed to step 5, otherwise return to step 1-7.

[0060] Step 5 of the present invention is specifically to visualize and analyze the crack extension data obtained from the calculated damage value through tecplot software to obtain the crack extension path, and to use origin to draw the fatigue life obtained from the fatigue life analysis to obtain the aN life curve.

[0061] In step 1-1 of the present invention, the near-field size is represented by δ. A circle (sphere) with a radius of δ is drawn with the central material point as the center (two-dimensional) or the center of the sphere (three-dimensional). The circle (sphere) is the near-field neighborhood of the central material point. The material point spacing is Δ, which represents the distance between the two material points. The appropriate material point spacing is selected through convergence verification, that is, it is necessary to ensure that the material point spacing is not too large to affect the calculation results, and to ensure that the material point spacing is not too small to affect the calculation speed. According to current research, the near-field neighborhood size is three times the material point spacing, that is, δ=3Δ, but in order to avoid floating-point errors when writing the program, it will be set to 3.015 or other approximate values ​​as needed; the number of material points is the number of material points after the actual size of the structure is discretized according to the material point spacing.

[0062] In step 1-2 of the present invention, the two-dimensional model is discretized into a square, and the three-dimensional model is discretized into a cube; the discrete near-field range in the three-dimensional model is a sphere, and the two-dimensional model is a plane circle.

[0063] In step 2 of the present invention, the bond breakage is judged and the loss value is calculated. If the bond is broken, the interaction disappears. This criterion is the basis of the near-field dynamics quantitative fracture analysis. After exceeding the set critical elongation, the bond automatically breaks. By calculating the damage value, the degree of damage at that location can be obtained, and the degree of structural damage can be described according to the damage value through the drawing software.

[0064] In step 4-2 of the present invention, the n with the minimum number of required cycles is selected. c keys, and make these n c The number of cycles is calculated by step 4-1. c The largest number of cycles corresponding to the keys is recorded as ΔN1, and ΔN1 is the number of fatigue life cycles this time. Until the end of the cycle, the obtained crack length and the total number of cycles are plotted to form an aN curve to achieve fatigue life prediction.

[0065] The present invention can realize the prediction of crack propagation path and fatigue life of plate grid and stiffened plate structure, solve the difficulty of singularity in continuous medium mechanics when dealing with discontinuous problems, and solve the problem that commercial software needs to introduce additional criteria when dealing with cracks. Especially in the problem of multiple cracks, the characteristic of natural crack initiation in peri-field dynamics greatly improves the calculation efficiency, and can provide a new and efficient method for fatigue life prediction of ship hull structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a flow chart of the present invention.

[0067] Figure 2 is a schematic diagram of a stiffened plate model in an embodiment of the present invention, Figure 2 In the figure a is the overall structure diagram. Figure 2 Figure b is a schematic diagram of crack details.

[0068] Figure 3 is a schematic diagram of a plate grid model in an embodiment of the present invention, Figure 3 In the figure a is the overall structure diagram. Figure 3 Figure b is a crack detail image.

[0069] Figure 4 is a crack growth diagram of a peridynamic stiffened plate according to an embodiment of the present invention, Figure 4 Figure a shows the crack extension diagram at step 125. Figure 4 Figure b shows the crack extension diagram at step 225. Figure 4 Figure c shows the crack extension diagram at step 325. Figure 4 Figure d shows a schematic diagram of crack propagation at step 425.

[0070] Figure 5 1 is a graph of the plate grid aN life in an embodiment of the present invention.

[0071] Figure 6 The plate grid damage in the embodiment of the present invention Figure 1 DETAILED DESCRIPTION

[0072] The present invention is further described with reference to the accompanying drawings and embodiments.

[0073] Example

[0074] This example provides a method for predicting the structural crack propagation path and fatigue life based on state-based peridynamics, performing steps (1) to (17):

[0075] Step (1): determine and input material parameters, structural material parameters include elastic modulus, Poisson's ratio, density, determine the near-field neighborhood size, the number of material points in each direction, and determine the material point spacing;

[0076] Step (2): Determine the number and coordinates of material points: Combine the geometric dimensions of the structure, calculate and determine the number of material points and the coordinates of each material point, and establish a structural geometric model;

[0077] Step (3): Discretize and traverse each material point, and determine the number of material points within the near field range: use a two-dimensional array to represent the jth near-field material point of the i-th central material point, traverse each material point in the modeling order, perform a first numbering process on all material points according to i, and store them in the array; search for material points in sequence according to the numbering order, set the i-th material point found as the central material point, find out the number of near-field material points of the central material point in its near-field neighborhood, and perform a second numbering process on them according to j, and store the searched near-field material points in another array, which now represents the jth near-field material point of the i-th central material point. After all near-field material points in the near-field area have been searched, start processing the next central material point until all i-numbered material points have been searched and the second numbering process to determine the near-field neighborhood material points is completed;

[0078] Step (4): Define the initial crack: When the initial crack needs to be preset, the size and position of the initial crack need to be input during the modeling stage. The two-dimensional structure sets the crack by breaking the bonds between the material points on the crack path, while the three-dimensional structure sets the crack by digging out the material points at the initial crack position.

[0079] Step (5): Perform volume correction: Because the determined near-field neighborhood is circular (spherical), and the discretized material points are square (cube), they cannot completely overlap at the boundary, so the volume correction factor v is introduced c (i) Perform volume correction on the material point to reduce this part of the error, see formula (1);

[0080]

[0081] Where: i, j represents the jth near-field material point of the i-th central material point, ξ is the initial bond length, represents the relative position of x and x', δ is the near-field neighborhood size, and Δ is the distance between material points;

[0082] ξ (i)(j) <δ-Δ means that the material point is completely within the near-field neighborhood,

[0083] δ-Δ≤ξ (i)(j) <δ+Δ means that part of the material point is located in the near field neighborhood.

[0084] ξ (i)(j) ≥δ+Δ means that the material point is completely outside the near field;

[0085] Step (6): Calculate the weighted volume: Since the bond lengths of each material point in the near field and the central material point are different, the interactions are also different; a weighted volume scalar q is defined here: (i) , calculate the weighted volume of each central material point, see formula (2),

[0086]

[0087] x (i)(j) represents the scalar state field of the reference position of the bond, that is, the scalar bond length between the jth near-field material point of the ith central material point,

[0088] ω is the influence function, which is set to be a function related to the near field range and bond length.

[0089] · is the dot product symbol, that is, · the integration of the previous and next terms;

[0090] Step (7): Time domain integration: Start the calculation according to the set total time step and time interval. Each step updates the corresponding boundary conditions and force loads, calculates the new expansion rate and force state, and then calculates the new acceleration, velocity, and displacement. Repeat until the set total time step is reached.

[0091] Step (8): Apply boundary conditions and force loads: boundary conditions and force loads can be set and input directly in the program, or read and input through a txt file. They are determined by the load and boundary conditions. The force load can be in the form of Newtonian force, pressure and peridynamic body force, and the boundary condition can be in the form of displacement, velocity and acceleration. The position of the boundary conditions and force loads is set by the program.

[0092] Step (9): Calculate the expansion rate: define a parameter expansion rate θ (i) For small deformations, this parameter measures the volume strain of the object, representing the bond between the deformation of the surrounding volume and the central material point. x (i)(j) The influence of the force state on is defined as formula (3):

[0093]

[0094] Where: x (i)(j) is the reference position scalar state field of the key, y (i)(j) is the deformed scalar state field of the bond, e (i)(j ) is the stretched scalar state, ω is The influence function is set as a function related to the near-field range and the bond length, · is the dot product symbol, that is, · is integrated before and after;

[0095] Step (10): Calculate the force state. Calculate the displacement, velocity, and acceleration of the material point through the force state and the equation of motion. The expression of the force state is formula (4).

[0096]

[0097] In the formula, v is Poisson's ratio, E is elastic modulus, and K is bulk elastic modulus. Under small deformation conditions,

[0098] α is the parameter of shear modulus μ, α=15μ / q,

[0099] ω is the influence function, which is set to be a function related to the near field range and bond length.

[0100] To stretch the scalar canine state,

[0101] e d = e - e i is the deflection of each item in the stretched scalar state,

[0102] x is the reference position scalar state field of the key,

[0103] M is the unit vector direction of the force state, indicating the direction of the force,

[0104] e represents the elongated scalar state;

[0105] The adaptive dynamic relaxation method in the implicit algorithm is used to solve the motion equation. Artificial damping is introduced and the damping term is determined in each time integral. The motion equation of the material point x with virtual damping is shown in formula (5):

[0106]

[0107] Where c is the damping coefficient term, the superscript n represents the current time step, and ρ is the current material point density.

[0108] The velocity vector of the material point x at time t is,

[0109] T (i)(j ) [x, t]<x′-x> represents the force state of material point x′ on the central material point x at time t,

[0110] is the acceleration vector of the material point x at time t,

[0111] b(x, t) is the external volume force density of material point x at time t,

[0112] H=H(x,δ): {x′∈R:||x′-x||≤δ} is the neighborhood range of the material point x,

[0113] Calculate the displacement, velocity, and acceleration of a material point according to the equation of motion

[0114] Step (11): To determine whether a bond is broken, the bond breaking criterion is used. The bond breaking criterion defines and determines the connection state of the bonds between material points and is the basis for the quantitative fracture analysis of peri-field dynamics. This method uses a three-dimensional structural bond breaking criterion based on the critical elongation. The elongation is first calculated and then compared with the set critical elongation to determine whether a break occurs. After the break, there is no interaction between the two points and the near-field force becomes 0; if not, the interaction exists until the break or the calculation is completed. The critical elongation and judgment are shown in formulas (6) and (7):

[0115]

[0116]

[0117] In the formula, m (i)(j ) ′<ξ> represents the connection state of the bond,

[0118] x (i)(j) is the reference position scalar state field of the key,

[0119] y (i)(j ) is the deformed scalar state field of the bond,

[0120] S (i)(j) is the bond elongation,

[0121] S0 is the critical value of elongation,

[0122] G0 is the critical energy release rate of the material,

[0123] λ (i)(j)Indicates the remaining life of the key.

[0124] K and μ are the bulk modulus and shear modulus, respectively.

[0125] δ is the near-field neighborhood size;

[0126] Step (12): Calculate the damage value: Calculate the damage value of each material point pair at each time step. The damage value of the material point pair is defined as a quantity that describes the damage degree of the material point x at time t. Its expression is shown in formula (8);

[0127]

[0128] It represents the damage value of material point x at time t. The damage value of material point x is related to the bonds connected to the material point, and its value range is 0 to 1. When the damage index of a material point x is 0, it means that no bonds in the structure are broken at this point. When the damage index of a material point x is 1, it means that all bonds connected to the material point are broken and the area is completely destroyed. Usually, the damage index rarely reaches 1, but in the damaged area near the crack path, the material point index in the area can basically reach more than 0.5; the damage value is the ratio of the number of broken bonds between the central material point and the surrounding material points to all the bonds. By calculating the damage value, the degree of damage at this point can be obtained, and the degree of damage of the structure can be described according to the damage value through the drawing software;

[0129] Step (13): Determine the state of the material point and calculate the number of cycles of each bond: Calculate the number of cycles of each bond according to the following formulas (9), (10), and (11). If it is the crack initiation stage, use A1 and P1 parameters for calculation; if it is the crack propagation stage, use A2 and P2 parameters for calculation: By calculating the damage value, if it is below 0.5, it is considered to be the crack initiation stage, otherwise it is the crack propagation stage:

[0130]

[0131]

[0132]

[0133] Where: N (i)(j) is the number of cycles;

[0134] λ (i)(j) It is called the remaining life and is related to the number of cyclic loading N (i)(j) related,

[0135] ε (i)(j) is the cyclic strain of the bond;

[0136] and The maximum and minimum elongation of the key in one cycle respectively;

[0137] R is the stress ratio;

[0138] A and P are unknown constants, corresponding to different A and P in the crack initiation stage and crack extension stage respectively;

[0139] Step (14): Perform bond breaking to select the n with the minimum number of cycles required c keys, and make these n c The key is damaged, this c The largest number of cycles corresponding to the keys is recorded as ΔN1, and the total number of cycles also increases by ΔN1;

[0140] Step (15): Determine whether it is greater than the maximum number of cycles: For other bonds that have not been damaged, the damage generated in these ΔN1 load cycles is accumulated in the bond (reflected in the decrease of λ), and then prepare to enter the next quasi-static analysis;

[0141] Step (16): Determine whether the specified time step is reached. If so, proceed to step (16); otherwise, return to step (7); Step (17): Post-processing of results: visualize the numerical results, visualize the crack extension data obtained from the calculated damage value through tecplot software to obtain the crack extension path; use origin to draw a graph of the fatigue life obtained from the fatigue life analysis to obtain the aN life curve.

[0142] In step (1) of this example, the size of the near-field domain is represented by δ. A circle (sphere) with a radius of δ is drawn with the central material point as the center (two-dimensional) or the center of the sphere (three-dimensional). The circle (sphere) is the near-field neighborhood of the central material point. The material point spacing is Δ, which represents the distance between two material points. The appropriate material point spacing is selected through convergence verification, that is, it is necessary to ensure that the material point spacing is not too large to affect the calculation results, and to ensure that the material point spacing is not too small to affect the calculation speed. According to current research, the near-field neighborhood size is three times the material point spacing, that is, δ = 3Δ, but in order to avoid floating-point errors when writing the program, it will be set to 3.015 or other approximate values ​​as needed; the number of material points is the number of material points after the actual size of the structure is discretized according to the material point spacing.

[0143] In step (2) of this example, the two-dimensional model is discretized into a square, and the three-dimensional model is discretized into a cube; the discretized near-field range in the three-dimensional model is a sphere, and the two-dimensional model is a plane circle.

[0144] In steps (11) and (12) of this example, the bond fracture is judged and the loss value is calculated. If the bond fractures, the interaction disappears. This criterion is the basis of the near-field dynamic quantitative fracture analysis. After exceeding the set critical elongation, the bond automatically breaks. The damage value is calculated to obtain the damage degree at that location. The damage degree of the structure is described by the drawing software according to the damage value.

[0145] In step (14) of this example, the n with the minimum number of cycles is selected. c keys, and make these n c The number of cycles is calculated by step (13). c The largest number of cycles corresponding to the keys is recorded as ΔN1, and ΔN1 is the number of fatigue life cycles of this time. Until the end of the cycle, the crack length and the total number of cycles are plotted to form an aN curve to achieve fatigue life prediction.

[0146] The following is an analysis of the performance of this example:

[0147] First, modeling and crack presetting are implemented: a crack propagation model of a hull plate with a central through-crack is calculated based on Fortran programming, and modeling, crack and presetting links are added. After defining the x, y, and z directions of the space, by inputting the length, width, thickness, and material parameters of the plate and stiffened plate, modeling can be automatically performed and discretized, numbered, and searched according to the set near-field range; by inputting the size and position of the crack, the material points at the crack can be removed to complete the crack presetting. For a two-dimensional model, only the bonds at the crack need to be broken, and the near-field range in the three-dimensional model is a sphere.

[0148] First determine the one-dimensional distance of three material points to narrow the range, and then perform three-dimensional distance calculation in this square to determine the points around the central material point, thereby reducing the amount of calculation. The larger the model, the higher the efficiency. Set the bond breaking criterion, and spontaneously break the bonds that exceed the critical elongation. After the bond is broken, the bond is damaged and the two material points connected by the bond lose interaction. According to the calculated number of cycles, select the number of fatigue bond breaks n c The bonds within are broken; the damage value is calculated to determine the crack state; the crack extension prediction diagram and fatigue life prediction aN curve are output through drawing software.

[0149] Execute steps (1)-(4) of this example to model the structure and preset cracks, and complete the discretization, numbering, and search of peridynamics, such as Figure 2 , Figure 3 Modeling of the stiffened plate and panel model and crack preset schematic diagram, where Figure 2Where 1 is the length of the bottom plate of the stiffened plate; w is the width of the bottom plate of the stiffened plate; t is the thickness of the bottom plate of the stiffened plate; lw is the length of the web of the stiffened plate; ww is the width of the web of the stiffened plate; hw is the height of the web of the stiffened plate; lf is the length of the face plate of the stiffened plate; wf is the width of the face plate of the stiffened plate; tf is the thickness of the face plate of the stiffened plate; cl is the length of the crack in the bottom plate; cx is the starting x-coordinate of the crack in the bottom plate; cy is the starting y-coordinate of the crack in the bottom plate; clw is the length of the crack in the web; cxw is the starting x-coordinate of the crack in the web; cyw is the starting y-coordinate of the crack in the web; czw is the starting z-coordinate of the crack in the web, Figure 3 Where 1 is the plate grid length; w is the plate grid width; t is the plate grid thickness; cl is the crack length; cx is the x-coordinate of the crack initiation; cy is the y-coordinate of the crack initiation.

[0150] Further executing steps (5)-(11) of this example, the crack propagation path prediction of the structure is completed, such as Figure 4 To simulate the crack propagation diagram of the stiffened plate according to the process, continue to perform steps (12)-(17) to complete the fatigue life prediction of the structure, such as Figure 5 is the fatigue life aN diagram of the plate grid model, Figure 6 It is the damage diagram under fatigue life.

[0151] The present invention has the following advantages: the present invention can realize the crack propagation path prediction and fatigue life prediction of plate grid and stiffened plate structures, solves the difficulty of singularity in continuous medium mechanics when dealing with discontinuous problems, and solves the problem that commercial software needs to introduce additional criteria when dealing with cracks. Especially in the problem of multiple cracks, the feature of natural crack initiation in near-field dynamics greatly improves the calculation efficiency, and can provide an efficient new method for fatigue life prediction of ship hull structure.

Claims

1. A method for predicting fatigue life of hull structure based on state-based peridynamics, characterized in that: The following steps are involved: Step 1: Establish a hull crack growth model; Step 2: Determine whether the bond is broken: Set the bond breaking criterion of the hull crack growth model, and spontaneously break the bond that exceeds the critical elongation. After the bond is broken, the bond is damaged and the two material points connected by the bond lose interaction; Step 3: Calculate the damage value and determine the crack state; Step 4: Select the number of fatigue bond breaks n according to the calculated number of cycles c The bonds within are broken; Step 5: Output crack growth prediction diagram and fatigue life prediction aN curve; Step 1 specifically includes the following: Step 1-1: Determine and input material parameters: Material parameters include elastic modulus, Poisson's ratio, density; determine the near-field neighborhood size, the number of material points in each direction, and the distance between material points; Step 1-2: Determine the number and coordinates of material points: Combine the geometric dimensions of the structure, calculate and determine the number of material points and the coordinates of each material point, and establish a structural geometric model; Step 1-3: Discretize and traverse each material point, and determine the number of material points within the near field: use a two-dimensional array to represent the jth near-field material point of the i-th central material point, traverse each material point in the modeling order, perform the first numbering process of all material points according to i, and store them in the array, search for material points in sequence according to the numbering order, set the i-th material point searched as the central material point, find out the number of near-field material points in the near-field neighborhood of the central material point, and perform the second numbering process of the near-field material points according to j, store the searched near-field material points in another array, representing the j-th near-field material point of the i-th central material point, and after all near-field material points in the near-field area have been searched, start searching for the next central material point until all i-numbered material points have completed the search and the second j-numbering process; Step 1-4: Define the initial crack. The 2D structure sets the crack by breaking the bonds between material points on the crack path, while the 3D structure sets the crack by removing the material points at the initial crack position. Step 1-5: Perform volume correction and introduce volume correction factor ν c(i) , volume correction is performed on material points, Where: i,j represents the jth near-field material point of the i-th central material point, ξ (i)(j) is the initial bond length, representing the relative position of x and x', δ is the near-field neighborhood size, Δ is the material point spacing, ξ (i)(j) <δ-Δ means that the material point is completely located in the near-field neighborhood, δ-Δ≤ξ (i)(j) <δ+Δ means that part of the material point is located in the near-field neighborhood, ξ (i)(j) ≥δ+Δ means that the material point is completely outside the near field; Step 1-6: Calculate the weighted volume and define the weighted volume scalar q (i) , calculate the weighted volume of each central material point, see formula (2), q (i) = (ωx (i)(j) )·x (i)(j) (2), Where: x (i)(j) is the scalar state field of the reference position of the bond, represents the scalar bond length between the jth near-field material point of the ith central material point, ω is the influence function, and · is the dot product symbol; Steps 1-7: Time domain integration, start calculation according to the set total time step and time interval, update the corresponding boundary conditions and force loads at each step, calculate the new expansion rate and force state, and then calculate the new acceleration, velocity, and displacement, and repeat until the set total time step is reached; Step 1-8: Apply boundary conditions and force loads. Boundary conditions and force loads are set and input directly in the program, or read and input through a txt file. The force load is in the form of Newton force, pressure and peridynamic body force, and the boundary condition is in the form of displacement, velocity and acceleration. Step 1-9: Calculate the expansion rate and define the expansion rate θ (i) For small deformations, the expansion rate measures the volume strain of the object, representing the deformation of the volume around the central material point with respect to the reference position scalar state field x of the bond to the central material point. (i)(j) The force effect of is defined as formula (3), Where: y (i)(j) is the deformed scalar state field of the bond, e (i)(j) is the elongated scalar state; Step 1-10: Calculate the force state. Use the force state and motion equation to calculate the displacement, velocity, and acceleration of the material point. The expression of the force state is shown in formula (4): Where: ν is Poisson's ratio, E is elastic modulus, κ is bulk elastic modulus. Under small deformation conditions, α is the parameter of shear modulus μ, α=15μ / q (i) , The stretched scalar state is the same as the other states. e d (i)(j) = e (i)(j) - e i (i)(j) The various deflections of the stretched scalar state, x (i)(j) is the reference position scalar state field of the key, M (i)(j) is the unit vector direction of the force state, indicating the direction of the force, e (i)(j) represents the elongated scalar state; The adaptive dynamic relaxation method in the implicit algorithm is used to solve the motion equation. Artificial damping is introduced and the damping term is determined in each time integral. The motion equation of the material point x with virtual damping is shown in formula (5). Where: c n (i) (x, t) is the damping coefficient term, and the superscript n represents the current time step. ρ (i) is the current material point density, The velocity vector of the material point x at time t, the superscript n represents the current time step, T (i)(j) [x,t]<x'-x> represents the force state of material point x' on the central material point at time t, is the acceleration vector of material point x at time t, b (i) (x, t) is the external volume force density of material point x at time t, H=H(x,δ):{x'∈R:||x'-x||≤δ}, which represents the neighborhood range of the material point x, and calculates the displacement, velocity, and acceleration corresponding to the material point according to the equation of motion; In step 2, the three-dimensional structural bond breaking criterion based on the critical elongation is used to determine whether the bond is broken. The elongation is calculated first, and then compared with the set critical elongation to determine whether the bond is broken. After the bond is broken, there is no interaction between the two points, and the near-field force becomes 0. If the bond is not broken, the interaction exists until the bond is broken or the calculation is completed. The critical elongation and judgment are shown in formulas (6) and (7). Where: μ (i)(j) '<ξ> represents the connection state of the bond; x (i)(j) is the reference position scalar state field of the key, y (i)(j) is the deformed scalar state field of the bond, s (i)(j) is the bond elongation, s0 is the critical value of elongation, G0 is the critical energy release rate of the material, λ (i)(j) Indicates the remaining life of the key. κ and μ represent bulk elastic modulus and shear modulus; In step 1-1, the size of the near field is represented by δ. A circle or a sphere with a radius of δ is drawn with the central material point as the center. The circle or sphere is the near field neighborhood of the central material point. The distance between the material points is Δ, δ=3Δ. In step 1-2, the two-dimensional model is discretized into a square, the three-dimensional model is discretized into a cube, the discrete near-field range in the three-dimensional model is a sphere, and the two-dimensional model is a plane circle.

2. A method for predicting fatigue life of hull structure based on state-based peridynamics according to claim 1, characterized in that: The damage value calculation in step 3 is as follows: the damage value of each material point pair is calculated at each time step. The damage value of the material point pair is defined as a quantity that describes the damage degree of the material point x at time t. The expression is shown in formula (8): It represents the damage value of material point x at time t. The damage value of material point x ranges from 0 to 1. When the damage value of a material point x is 0, it means that no bonds in the structure are broken at that location. When the damage value of a material point x is 1, it means that all bonds connected to the material point are broken and the area is completely destroyed. The degree of structural damage can be described by calculating the damage value.

3. The method for predicting fatigue life of hull structure based on state-based peridynamics according to claim 2 is characterized in that: Step 4 includes the following: Step 4-1: Determine the state of the material point and calculate the number of cycles of each bond. Calculate the number of cycles of each bond according to the following formulas (9), (10), and (11). If it is the crack initiation stage, use the A1 and P1 parameters for calculation. If it is the crack propagation stage, use the A2 and P2 parameters for calculation. By calculating the damage value, if it is below 0.5, it is considered to be the crack initiation stage, otherwise it is the crack propagation stage: Where: N (i)(j) is the number of cycles; ε (i)(j) is the cyclic strain of the bond, and are the maximum and minimum elongation of the key in one cycle, Z is the stress ratio, A and P are unknown constants, corresponding to different A and P in the crack initiation stage and crack extension stage respectively; Step 4-2: Perform bond breaking: select the n with the minimum number of cycles required c keys, so that this c The key is damaged, this c The largest number of cycles corresponding to the keys is recorded as ΔN1, and the total number of cycles also increases by ΔN1; Step 4-3: For the bonds that have not been damaged, the damage generated in the ΔN1 load cycles is accumulated in the bonds, ready for the next quasi-static analysis; Step 4-4: Determine whether the time step has been reached. If so, proceed to step 5; otherwise, return to step 1-7.