Metal damper fatigue life prediction method and structure optimization method

Abstract

The invention discloses a method for predicting the fatigue life of a metal damper. The method comprises the following steps: acquiring data information of a target metal damper; carrying out material test and parameter calibration on the material of the target metal damper; carrying out hybrid hardening elastic-plastic finite element modeling on the target metal damper; and plastic finite element calculation and post-processing analysis are carried out to complete fatigue life prediction of the target metal damper. The invention further discloses a structure optimization method comprising the metal damper fatigue life prediction method. According to the method, fatigue life prediction and structure optimization of the target metal damper are achieved, and the scheme is higher in reliability and better in accuracy.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of finite element analysis, and particularly relates to a metal damper fatigue life prediction method and a structure optimization method. BACKGROUND

[0002] Under the action of an earthquake, a bridge steel structure and a damping device suffer from an alternating dynamic load caused by a reciprocating ground motion, a stress concentration area thereof experiences a large cyclic plastic strain and thus plastic fracture occurs, and a material thereof experiences a high-strain low-cycle fatigue process. Therefore, the fatigue life of a bridge damping device is a key index for safe operation and service of the bridge damping device. In a pier of a small and medium span bridge, the pier has a low reinforcement ratio and poor ductility, and thus a metal damper (such as a steel energy dissipation support or an E-shaped steel device) is needed to dissipate seismic energy and prevent the structure from instability or beam falling. Therefore, the fatigue life prediction of the metal damper is particularly important.

[0003] However, the existing fatigue life prediction scheme of the metal damper lacks a fusion scheme of material fatigue performance, structural geometric characteristics and a load spectrum, and thus the prediction reliability and accuracy of the existing scheme are relatively poor. SUMMARY

[0004] One of the objects of the present application is to provide a metal damper fatigue life prediction method with high reliability and good accuracy.

[0005] The second object of the present application is to provide a structure optimization method comprising the metal damper fatigue life prediction method.

[0006] The metal damper fatigue life prediction method provided by the present application comprises the following steps:

[0007] S1. obtaining data information of a target metal damper;

[0008] S2. performing material testing and parameter calibration on a material of the target metal damper according to the data information obtained in step S1;

[0009] S3. performing mixed hardening elastoplastic finite element modeling on the target metal damper according to the data information obtained in step S2;

[0010] S4. performing plastic finite element calculation and post-processing analysis according to the modeling result of step S3 to complete fatigue life prediction of the target metal damper.

[0011] The data information of the target metal damper in step S1 specifically comprises material data information of the target metal damper and structure data information of the target metal damper.

[0012] According to the data information obtained in step S1, material testing and parameter calibration are performed on the material of the target metal damper, and the specific steps include the following steps:

[0013] According to the set standard, engineering stress data and engineering strain data of the material of the target metal damper at room temperature are obtained;

[0014] The engineering stress data and the engineering strain data are converted into real stress and strain data by using the following formula:

[0015] In the formula, is the real stress; is the engineering stress; is the real strain; is the engineering strain.

[0016] According to the obtained real stress and strain data, a stress-strain curve is drawn, and the slope of the linear segment of the linear region of the stress-strain curve is taken as the elastic modulus E;

[0017] According to the stress-strain curve, the lower yield strength is determined according to GB / T 228.1, and the lower yield strength value is set as the yield strength , and the corresponding experimental data point is set as the yield point;

[0018] The data curve from the yield point to the maximum stress point is nonlinearly fitted by using the formula to obtain the hardening index n and the strength coefficient K:

[0019] According to the set standard, the material of the target metal damper is subjected to strain-controlled cyclic loading to obtain a stable cyclic hysteresis curve, and the Basquin-Coffin-Manson curve parameters are fitted:

[0020] In the formula, is the total strain amplitude; is the fatigue strength coefficient; is the cycle life, i.e., the actual cycle number reaching the fatigue failure criterion in the experiment; b is the fatigue strength index; is the fatigue ductility coefficient; and c is the fatigue ductility index.

[0021] According to the data information obtained in step S2, a mixed hardening elastoplastic finite element modeling is performed on the target metal damper, and the specific steps include the following steps:

[0022] The Chaboche hybrid hardening model is used to describe the isotropic hardening and kinematic hardening effects under cyclic loading.

[0023] The parameters set include: initial yield stress Total saturation value of back stress Yield stress evolution rate parameter The first parameter for fitting back stress components Fitting the second parameter to the back stress component ;

[0024] Strain control cycle tests were conducted according to the set standards, and the stable hysteresis loop was recorded to obtain the stress extreme difference of the stable hysteresis loop; the stress extreme difference was recorded as the saturation stress amplitude. and cyclic yield strength ;

[0025] Fitting using the following formula and :

[0026] In the formula The average plastic strain is the cyclic strain.

[0027] For the low-cycle fatigue plastic process, the kinematic hardening modulus is taken as the second-order back stress; the back stress is separated from the unloading section of the hysteresis loop using the following formula:

[0028] In the formula This is the back stress value; The total stress that can be directly measured in the experiment; This represents the maximum stress in a single cycle. The minimum stress for a single cycle;

[0029] Using the Levenberg-Marquardt algorithm, minimize the following objective function:

[0030] In the formula , These are the parameters of the first set of back stress components; , These are the parameters for the second set of back stress components; This represents the plastic strain value;

[0031] By minimizing the objective function, the back stress fitting parameters are obtained ( , , , );

[0032] Substitute all the parameters obtained into the finite element model, set the loading mode and boundary conditions, and complete the mixed hardening elastoplastic finite element modeling of the target metal damper:

[0033] Geometric modeling method: according to the shape of the target metal damper, the three-dimensional structure is reconstructed, and the three-dimensional structure is discretized by using a structure grid;

[0034] The set material constitutive parameters include: elastic modulus E, yield strength , hardening index , and strength coefficient K.

[0035] The set mixed hardening parameters include: initial yield stress , total saturation value of back stress , yield stress evolution rate parameter , back stress component fitting first parameter , and back stress component fitting second parameter .

[0036] Step S4: According to the modeling results of step S3, perform plastic finite element calculation and post-processing analysis to complete the fatigue life prediction of the target metal damper, which specifically includes the following steps:

[0037] According to the modeling results of step S3, use an implicit solver to solve the displacement, stress and strain of the cycle process; when the hysteresis loop is stable or the maximum stress exceeds the set strength limit , stop the cycle process;

[0038] Post-processing analysis:

[0039] The strain tensor output by Gauss is decomposed into elastic and plastic parts, which is expressed as:

[0040] In the formula, is the elastic strain; is the plastic strain;

[0041] The strain increment norm of each time step is expressed as:

[0042] In the formula, is the norm of the strain increment; is the plastic normal strain in the 1 direction; is the plastic normal strain in the 2 direction; is the plastic normal strain in the 3 direction; is the plastic shear strain in the 1-2 direction; is the plastic shear strain in the 1-3 direction; plastic shear strain in 2-3 direction;

[0043] Setting history variable , and accumulated at each time step, expressed as:

[0044] In the formula is plastic strain accumulation value at the moment; is plastic strain accumulation value at the moment;

[0045] The history variable Visualize as a cloud chart, traverse all points in the cloud chart, and take the node corresponding to the maximum value as a dangerous point;

[0046] Stress-strain history extraction:

[0047] Extract the corresponding stress data and strain data at the dangerous point, and draw the corresponding stress-strain curve; adopt Basquin and Coffin-Manson formula to carry out life prediction;

[0048] Life prediction:

[0049] According to the stress data and strain data of the obtained dangerous point, the stress amplitude, strain amplitude and average stress corresponding to each cycle are calculated;

[0050] The fatigue life corresponding to each cycle is fitted by using the following formula

[0051] In the formula is the plastic strain amplitude of the Ith cycle; is the fatigue strength coefficient; is the cycle number experienced by the material under constant strain amplitude loading until fatigue failure occurs; is the fatigue ductility coefficient;

[0052] Set the total cycle number accumulated expressed as , is the serial number of the load block;

[0053] Select And The corresponding N is the fatigue life prediction result of the target metal damper.

[0054] The application also provides a structure optimization method comprising the metal damper fatigue life prediction method, further comprising the following steps:

[0055] S5. According to the obtained results, a multi-objective genetic algorithm is adopted to perform multi-objective optimization on the steel structure geometric size of the target metal damper, with minimum mass and maximum fatigue life as the targets;

[0056] S6. According to the optimization results obtained in step S5, process improvement of fatigue performance is performed.

[0057] The multi-objective optimization on the steel structure geometric size of the target metal damper according to the obtained results in step S5, with minimum mass and maximum fatigue life as the targets, specifically includes the following steps:

[0058] The constructed optimization objective function is:

[0059] In the formula, is the first objective function value; is the fatigue life function; is the second objective function value; is the material density; is the stress function of the structure; is a vector composed of design variables of the target metal damper;

[0060] The constructed optimization constraint condition is:

[0061] In the formula, is the maximum stress of the structure; is the allowable stress of the material; is the characteristic length of the damper;

[0062] The Kriging surrogate model is trained:

[0063] The initial sampling is performed on , and Sobol sequence is adopted to generate several groups of initial vectors;

[0064] The stress-strain hysteresis curve of the dangerous point is extracted by adopting the finite element calculation scheme;

[0065] The feature vector under each cycle is calculated; the feature vector is composed of the cycle number I, the minimum stress of the structure and the cumulative plastic strain ;

[0066] For the obtained new , it is expressed as ; a set of eigenvectors is calculated by using the Kriging surrogate model, and a first objective function value is calculated, and a second objective function value is calculated from the geometric dimensions;

[0067] Optimization process: two individuals are selected from the initial population, and the winner is selected by the size relationship between the respective first objective function value and the second objective function value, and the winner is crossed to generate offspring; a set of offspring is generated by disturbance with a set probability; the life of the offspring is calculated by the Kriging surrogate model; all parent individuals and offspring individuals are optimized to obtain the first front solution; finally, a stable solution is obtained by convergence after several generations of iteration;

[0068] Finally, the geometric size optimization of the target metal damper steel structure is completed.

[0069] The metal damper fatigue life prediction method and the structure optimization method provided by the application realize fatigue life prediction and structure optimization of the target metal damper through experimental analysis, finite element modeling, calculation and post-processing of the target metal damper, and the scheme has higher reliability and better accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0070] Figure 1 It is a method flowchart of the fatigue life prediction method of the application.

[0071] Figure 2 It is a method flowchart of the structure optimization method of the application.

[0072] Figure 3 It is a runway type (rectangular variable cross section) steel damper shape design scheme diagram of the embodiment of the application.

[0073] Figure 4 It is a boundary condition setting diagram of the ring damper of the embodiment of the application.

[0074] Figure 5 It is an equal cross section runway type transverse loading mode and a typical hysteresis loop diagram of the embodiment of the application.

[0075] Figure 6 It is an equal cross section runway type longitudinal loading mode and a typical hysteresis loop diagram of the embodiment of the application.

[0076] Figure 7 It is a dangerous point position checking diagram of the embodiment of the application.

[0077] Figure 8 It is a stress-strain cycle diagram at a typical dangerous point of the embodiment of the application.

[0078] Figure 9A set of typical low-cycle fatigue fracture schematic diagrams of the embodiment of the present application under lower strain amplitude (0.5%) loading.

[0079] Figure 10 A set of typical low-cycle fatigue fracture schematic diagrams of the embodiment of the present application under higher strain amplitude (2%) loading. DETAILED DESCRIPTION

[0080] As Figure 1 shown is a method flow schematic diagram of the fatigue life prediction method of the present application: the metal damper fatigue life prediction method disclosed by the present application comprises the following steps:

[0081] S1. Obtain data information of a target metal damper; specifically including material data information of the target metal damper and structure data information of the target metal damper;

[0082] S2. According to the data information obtained in step S1, perform material testing and parameter calibration on the material of the target metal damper; specifically comprising the following steps:

[0083] According to the set standard (GB / T 228-2010), obtain engineering stress data and engineering strain data of the material of the target metal damper at room temperature;

[0084] Convert the engineering stress data and the engineering strain data into true stress-strain data by using the following formula:

[0085] In the formula, is the true stress; is the engineering stress; is the true strain; is the engineering strain.

[0086] According to the obtained true stress-strain data, draw a stress-strain curve; extract the linear region of the stress-strain curve, and take the slope of the linear segment of the linear region as the elastic modulus E;

[0087] According to the stress-strain curve, determine the lower yield strength according to GB / T 228.1, and set the lower yield strength value as the yield strength , and the corresponding experimental data point is set as the yield point;

[0088] Perform nonlinear fitting on the data curve from the yield point to the maximum stress point by using the formula to obtain the hardening index n and the strength coefficient K:

[0089] According to the set standard (GB / T 15248-2008), the material of the target metal damper is subjected to strain control cyclic loading to obtain a stable cyclic hysteresis curve, and Basquin-Coffin-Manson curve parameters are fitted:

[0090] In the formula is the total strain amplitude; is the fatigue strength coefficient; is the cyclic life, i.e. the actual number of cycles in the experiment to reach the fatigue failure criterion; b is the fatigue strength index; is the fatigue ductility coefficient; c is the fatigue ductility index.

[0091] S3. According to the data information obtained in step S2, a mixed hardening elastoplastic finite element model of the target metal damper is established; specifically including the following steps:

[0092] The Chaboche mixed hardening model is used to describe the isotropic hardening and dynamic hardening effect under cyclic loading;

[0093] The set parameters include: initial yield stress , total saturation value of back stress , yield stress evolution rate parameter , fitting first parameter of back stress component , and fitting second parameter of back stress component ;

[0094] According to the set standard (GB / T 15248-2008), strain control cyclic test is carried out, and stable hysteresis loop is recorded to obtain the stress extreme difference of stable hysteresis loop; the stress extreme difference is recorded as saturation stress amplitude and cyclic yield strength ;

[0095] The following formulas are used to fit and :

[0096] In the formula is the cyclic average plastic strain;

[0097] For low-cycle fatigue plastic process, the dynamic hardening modulus is taken as two-level back stress; from the unloading section of the hysteresis loop, the following formula is used to separate the back stress:

[0098] In the formula is the back stress value; is the total stress that can be directly measured in the experiment; is the maximum stress for a single cycle; is the minimum stress for a single cycle;

[0099] The Levenberg-Marquardt algorithm is used to minimize the following objective function:

[0100] wherein , is the first set of back stress component parameters; , is the second set of back stress component parameters; is the plastic strain value;

[0101] The back stress fitting parameters (a1, a2, a3, a4) are obtained by minimizing the objective function; , , ,

[0102] All the obtained parameters are substituted into the finite element model, and the loading mode (including transverse displacement loading, longitudinal displacement loading) and boundary conditions (different boundary conditions can be set) are set to complete the mixed hardening elastoplastic finite element modeling of the target metal damper:

[0103] Geometric modeling method: according to the shape of the target metal damper, the three-dimensional structure is reconstructed, and the three-dimensional structure is discretized by using structural grid;

[0104] The set material constitutive parameters include: elastic modulus E, yield strength , hardening index , strength coefficient K;

[0105] The set mixed hardening parameters include: initial yield stress , total saturation value of back stress , yield stress evolution rate parameter , back stress component fitting first parameter and back stress component fitting second parameter ;

[0106] S4. According to the modeling results of step S3, perform plastic finite element calculation and post-processing analysis to complete the fatigue life prediction of the target metal damper; specifically including the following steps:

[0107] According to the modeling results of step S3, use the implicit solver to solve the displacement, stress and strain of the cycle process; when the hysteresis loop is stable or the maximum stress exceeds the set strength limit , stop the cycle process;

[0108] ​Post-processing analysis:

[0109] The strain tensor pointed out by Gauss is decomposed into elastic part and plastic part, expressed as:

[0110] In the formula is the elastic strain; is the plastic strain;

[0111] The strain increment norm of each time step is expressed as:

[0112] In the formula is the norm of strain increment; is the plastic normal strain in the 1 direction; is the plastic normal strain in the 2 direction; is the plastic normal strain in the 3 direction; is the plastic shear strain in the 1-2 direction; is the plastic shear strain in the 1-3 direction; is the plastic shear strain in the 2-3 direction;

[0113] Set the history variable , and accumulate at each time step, expressed as:

[0114] In the formula is the plastic strain accumulation value at the moment is the plastic strain accumulation value at the moment

[0115] Visualize the history variable as a cloud chart, traverse all points in the cloud chart, and take the node corresponding to the maximum value as the dangerous point;

[0116] Stress-strain history extraction:

[0117] Extract the corresponding stress data and strain data at the dangerous point, and draw the corresponding stress-strain curve; adopt Basquin and Coffin-Manson formula to carry out life prediction;

[0118] Life prediction:

[0119] According to the stress data and strain data of the dangerous point obtained, the stress amplitude, strain amplitude and average stress corresponding to each cycle are calculated;

[0120] The fatigue life corresponding to each cycle is fitted by using the following formula ​​

[0121] wherein is the plastic strain amplitude of the first cycle; is the fatigue strength coefficient; is the material density; is the number of cycles experienced by the material under constant strain amplitude is the fatigue ductility coefficient;

[0122] the total number of cycles accumulated is set as is represented as , is the serial number of the load block;

[0123] is selected and the corresponding N is the fatigue life prediction result of the target metal damper.

[0124] As shown in Figure 2 the method flowchart of the structure optimization method of the present application: the structure optimization method disclosed by the present application including the metal damper fatigue life prediction method, comprising the following steps:

[0125] S1. obtaining the data information of the target metal damper;

[0126] S2. according to the data information obtained in step S1, material testing and parameter calibration are performed on the material of the target metal damper;

[0127] S3. according to the data information obtained in step S2, mixed hardening elastoplastic finite element modeling is performed on the target metal damper;

[0128] S4. according to the modeling result of step S3, plastic finite element calculation and post-processing analysis are performed to complete the fatigue life prediction of the target metal damper;

[0129] S5. according to the obtained result, a multi-objective genetic algorithm is adopted to take the minimum mass and the maximum fatigue life as the target to perform multi-objective optimization on the geometric size of the steel structure of the target metal damper; specifically comprising the following steps:

[0130] The optimization objective function constructed is:

[0131] wherein is the first objective function value; is the fatigue life function; is the second objective function value; is the material density; is the objective function of the structure; is a vector composed of design variables of the target metal damper;

[0132] The optimization constraint condition constructed is:

[0133] In the formula is the maximum stress of the structure; is the allowable stress of the material; is the characteristic length of the damper;

[0134] The Kriging surrogate model is trained:

[0135] The initial sampling is performed on , and a plurality of groups of initial vectors are generated by using a Sobol sequence;

[0136] The stress-strain hysteresis curve of the dangerous point is extracted by using a finite element calculation scheme;

[0137] The characteristic vector under each cycle is calculated; the characteristic vector is composed of the cycle number I, , the minimum stress of the structure , and the cumulative plastic strain ;

[0138] For the new obtained, it is expressed as ; a plurality of characteristic vectors are calculated by using the Kriging surrogate model, and a first objective function value is calculated, and a second objective function value is calculated from the geometric size;

[0139] The optimization process: two individuals are selected from the initial population, and the winners are selected by the size relationship between the respective first objective function values and the second objective function values, and the winners are crossed to generate offspring; a plurality of offspring are generated by perturbation with a set probability; the life of the offspring is calculated by the Kriging surrogate model; the first front overall solution is selected from all the parent individuals and offspring individuals; finally, a stable solution is obtained after a plurality of generations of iteration;

[0140] The optimization of the geometric size of the steel structure of the target metal damper is finally completed;

[0141] S6. According to the optimization result obtained in step S5, the process improvement of the fatigue performance is performed.

[0142] The present application is described below in combination with an embodiment:

[0143] Figure 3 is a schematic diagram of a certain type of steel damper structure in this embodiment, wherein the long arm and the short arm are both 600 mm long.

[0144] After the corresponding finite element model is constructed, the corresponding boundary conditions are set as shown in Figure 4 ; the transverse loading and typical hysteresis loop are shown in Figure 5 ; the transverse loading displacement is slightly larger than the working condition according to the fatigue test working condition. Figure 5 The entire device is 980*1280, the design displacement is 200mm, and the loading displacement value is 250mm. The longitudinal loading and typical hysteresis loop are shown in Figure 6 ; the longitudinal loading displacement is slightly larger than the working condition according to the fatigue test working condition. Figure 5 The entire device is 980*1280, the design displacement is 200mm, and the loading displacement value is 250mm.

[0145] Under the transverse loading condition, the dangerous point is located inside the broken arm bending, as indicated by the red dot in Figure 7 (a). The estimated dangerous point position is also consistent with the actual macroscopic crack position, as shown in Figure 7 (b). The maximum value of the cumulative plasticity appears at the bending position close to the external link. The destruction of this point determines the fatigue life of the entire structure according to the present application scheme.

[0146] Draw the corresponding stress-strain curve, a typical example is shown in Figure 8 ;

[0147] Process improvement process of the present embodiment:

[0148] Fracture analysis is performed on the fatigue fracture under different strain amplitudes. According to the working condition and the position of fracture, it is divided into low strain amplitude internal cracking, low strain amplitude surface cracking, high strain amplitude internal cracking, and high strain amplitude external cracking. The actual working condition may be a combination of the above. The specific process improvement scheme should be determined jointly according to the fracture position and microstructure morphology. Since the typical fracture process of low strain amplitude working condition and high strain amplitude working condition is different, they are discussed respectively:

[0149] Figure 9 For the fracture morphology under lower strain amplitude, Figure 10 For the fracture morphology under higher strain amplitude; according to the design load condition (such as low strain amplitude or high strain amplitude), the scanning electron microscope local magnification pictures of the fatigue source area, fatigue propagation area and instantaneous fracture area are taken, the crystal phase microstructure morphology is analyzed, and the corresponding process improvement measures are proposed according to the microstructure morphology and the fracture mechanism.

[0150] 1) Fracture analysis of low strain amplitude working condition (0.5%)

[0151] Fatigue source area (corresponding to Figure 9(b)): cracks initiated from subsurface region (about 150 μm from surface), source sites existed with complex inclusions of about 15 μm in size. This indicates that the stress concentration effect of internal defects is easy to become the dominant failure factor under low strain amplitude cyclic loading.

[0152] Fatigue propagation zone (corresponding to Figure 9 (c)): narrow pitch fatigue striations (average pitch about 0.2 μm) were observed, and a large number of intergranular secondary cracks (length about 30 μm) were distributed between the striations, which penetrated through multiple grains. This indicates that grain boundary weakening accelerates crack propagation.

[0153] Instantaneous fracture zone (corresponding to Figure 9 (d)): mainly with large size equiaxed dimples (diameter 5-8 μm), and there were non-spheroidized long strip-shaped cementite (length about 3 μm) at the bottom of the dimples, and no cleavage fracture characteristics appeared.

[0154] Process improvement scheme:

[0155] For the low strain amplitude failure mechanism, the following measures are taken:

[0156] Melting process upgrade: adopt vacuum deoxidization to control oxygen content ≤15 ppm, and add rare earth elements to modify the morphology of inclusions;

[0157] Grain boundary strengthening: desulphurization pretreatment of molten steel, and controlled cooling process to accelerate cooling and refine grains;

[0158] Carbide spheroidization: 780 °C x 4 h annealing treatment makes the diameter of carbide ≤0.8 μm;

[0159] Multi-directional forging: forging ratio ≥6 to break the inclusion chain.

[0160] (2) High strain amplitude working condition (2%) fracture analysis

[0161] Fatigue source zone (corresponding to Figure 10 (b)): cracks originated from the root of surface machining tool marks (tool mark depth about 12 μm), and obvious radial ridges were observed in the source zone. High strain amplitude load amplifies the surface stress concentration effect, resulting in cumulative damage of slip bands at the root of tool marks.

[0162] Fatigue propagation zone (corresponding to Figure 10 (c)): fatigue striation pitch increased to 1-3 μm and was discontinuously distributed, and shallow dimple bands (depth-diameter ratio ≤0.3) appeared between the striations. This feature is directly related to the plastic deformation concentration caused by banded segregation structure.

[0163] Instantaneous fracture zone (corresponding to Figure 10 (d)): showed a mixed fracture morphology of dimples and cleavage, and the proportion of cleavage surface was about 40%, and large-angle river patterns were locally observed.

[0164] Fatigue fracture analysis - process improvement mapping relationship as shown in Table 1:

[0165] (3) Common process optimization

[0166] Surface strengthening: shot blasting treatment is implemented on all damper working sections, and the residual compressive stress is greater than 350 MPa;

[0167] Microstructure control: controlled rolling and controlled cooling process makes the banded structure ≤ 2 levels;

[0168] Verification effect: after the above improvement, the improved Q355B is used for damper, and the fatigue life of the damper is improved to times under the strain amplitude of times, The life under the strain amplitude of

[0169] The final optimization result is: the length of the short arm is increased from 600 mm to 650 mm, the bending radius is expanded from 50 mm to 70 mm, the fatigue life is improved from 22 times to 35 times, and the total weight can be reduced by 18.5% after structure optimization.

Claims

1. A method of predicting fatigue life of a metal damper, characterized by The method comprises the following steps: S1. obtaining data information of the target metal damper; S2. performing material testing and parameter calibration on the material of the target metal damper according to the data information obtained in step S1; S3. performing mixed hardening elastoplastic finite element modeling on the target metal damper according to the data information obtained in step S2; S4. performing plastic finite element calculation and post-processing analysis according to the modeling result of step S3 to complete fatigue life prediction of the target metal damper.

2. The metal damper fatigue life prediction method according to claim 1, characterized by The data information of the target metal damper in step S1 specifically comprises material data information of the target metal damper and structure data information of the target metal damper.

3. The metal damper fatigue life prediction method according to claim 2, characterized by The material testing and parameter calibration on the material of the target metal damper according to the data information obtained in step S1 in step S2 specifically comprises the following steps: According to the set standard, the engineering stress data and the engineering strain data of the material of the target metal damper at room temperature are obtained; The engineering stress data and the engineering strain data are converted into true stress-strain data by using the following formula: is the true stress; is the engineering stress; is the true strain; is the engineering strain. According to the obtained true stress-strain data, a stress-strain curve is drawn; The linear region of the stress-strain curve is extracted, and the slope of the linear segment of the linear region is taken as the elastic modulus E; According to the stress-strain curve, the lower yield strength is determined according to GB / T 228.1, and the lower yield strength value is set as the yield strength The corresponding experimental data point is defined as the yield point; The data curve from the yield point to the maximum stress point is fitted non-linearly to obtain the hardening exponent n and the strength coefficient K: n and K. According to the set standard, the stable cyclic hysteresis curve of the material of the target metal damper is obtained through strain control cyclic loading, and the Basquin-Coffin-Manson curve parameters are fitted: wherein is the total strain amplitude; is the fatigue strength coefficient; is the cycle life;b is the fatigue strength exponent; is the fatigue ductility coefficient; c is the fatigue ductility exponent.

4. The metal damper fatigue life prediction method according to claim 3, characterized by The mixed hardening elastoplastic finite element modeling on the target metal damper according to the data information obtained in step S2 in step S3 specifically comprises the following steps: The Chaboche mixed hardening model is used to describe the isotropic hardening and follow-up hardening effect under cyclic loading; Parameters that are set include: initial yield stress , total saturation value of back stress , yield stress evolution rate parameter , back stress component fitting first parameter , and back stress component fitting second parameter ; According to the set standard, strain control cyclic test is carried out, stable hysteresis loop is recorded, and stress extreme difference of the stable hysteresis loop is obtained; the stress extreme difference is recorded as a saturation stress amplitude and cyclic yield strength ; The following equation was used to fit and : In the formula is the cyclic average plastic strain; For the low-cycle fatigue plastic process, the follow-up hardening modulus is taken as the second-order back stress; from the unloading segment of the hysteresis loop, the back stress is separated by using the following formula: wherein is the back stress value; is the total stress that can be measured directly in the experiment; is the maximum stress of a single cycle; is the minimum stress of a single cycle; The Levenberg-Marquardt algorithm is used to minimize the following objective function: In the formula , These are the parameters of the first set of back stress components; , These are the parameters for the second set of back stress components; This represents the plastic strain value; By minimizing the objective function, the back stress fitting parameters (a, b, c, d) are obtained , , , ); All the obtained parameters are substituted into the finite element model, the loading mode and the boundary conditions are set, and the mixed hardening elastoplastic finite element modeling on the target metal damper is completed: The geometric modeling mode is to reconstruct the three-dimensional structure according to the shape of the target metal damper, and the three-dimensional structure is discretized by using a structure grid; The set material constitutive parameters include: elastic modulus E, yield strength , hardening index , strength coefficient K; The set of mixed hardening parameters comprises: an initial yield stress a total saturation value of back stress a yield stress evolution rate parameter a back stress component fitting first parameter and a back stress component fitting second parameter .

5. The metal damper fatigue life prediction method according to claim 4, characterized by The post-processing analysis is performed: According to the modeling results of step S3, the displacement, stress and strain of the cyclic process are solved by using an implicit solver; when the hysteresis loop is stable or the maximum stress exceeds the set strength limit , the cyclic process is stopped; The strain tensor output by the Gauss point is decomposed into an elastic part and a plastic part, and is expressed as: The strain increment norm of each time step is expressed as: In the formula is the elastic strain; is the plastic strain; Stress-strain history extraction: wherein is the norm of the strain increment; is the plastic normal strain in direction 1 ; is the plastic normal strain in direction 2; is the plastic normal strain in direction 3; is the plastic shear strain in direction 1-2; is the plastic shear strain in direction 1-3; is the plastic shear strain in direction 2-3; Setting the history variable and accumulated at each time step, denoted as: In the formula is the plastic strain accumulation value at the time point; is the plastic strain accumulation value at the time point; history variable Visualize as a cloud chart, traverse all points in the cloud chart, take the node corresponding to the maximum value as the dangerous point; The corresponding stress data and strain data at the dangerous point are extracted, and the corresponding stress-strain curve is drawn; the Basquin and Coffin-Manson formulas are used for life prediction; Life prediction: According to the obtained stress data and strain data of the dangerous point, the stress amplitude, the strain amplitude and the average stress corresponding to each cycle are calculated; The fatigue life corresponding to each cycle is fitted by using the following formula ​ wherein is the plastic strain amplitude for the first cycle; is the fatigue strength coefficient; is the material's experienced number of cycles at constant strain amplitude under loading until fatigue failure occurs; is the fatigue ductility coefficient; Setting the total number of cycles accumulated is represented as , is the serial number of the load block; Selecting And The corresponding N is the fatigue life prediction result of the target metal damper.

6. A structural optimization method including the metal damper fatigue life prediction method according to any one of claims 1 to 5, characterized by Further comprising the following steps: S5. According to the obtained results, a multi-objective genetic algorithm is used to perform multi-objective optimization on the steel structure geometric size of the target metal damper, with the minimum mass and the maximum fatigue life as the targets; S6. According to the optimization results obtained in step S5, process improvement of the fatigue performance is performed.

7. The method of structural optimization according to claim 6, characterized in that The multi-objective optimization on the steel structure geometric size of the target metal damper according to the obtained results, with the minimum mass and the maximum fatigue life as the targets, as described in step S5, specifically comprises the following steps: The constructed optimization objective function is: wherein is a first objective function value; is a fatigue life function; is a second objective function value; is a material density; is a mention function of the structure; is a vector consisting of design variables of the target metal damper; The constructed optimization constraint condition is: wherein is the maximum stress of the structure; is the allowable stress of the material; is the characteristic length of the damper; The Kriging surrogate model is trained: To perform initial sampling and generate several groups of initial vectors using Sobol sequence; The stress-strain hysteresis curve of the dangerous point is extracted by using the finite element calculation scheme; calculating a feature vector for each cycle; the feature vector being composed of cycle number I, , minimum stress of the structure and accumulated plastic strain ; For the new , expressed as ; a set of feature vectors is calculated using the Kriging surrogate model, and a first objective function value is calculated, while a second objective function value is calculated from the geometric dimensions; Optimization process: two individuals are selected from the initial population, and the winner is selected for crossover and generation of offspring by the size relationship of the respective first objective function value and second objective function value; a certain probability is disturbed to generate several offspring; the life of the offspring is calculated by the Kriging surrogate model; for all parent individuals and offspring individuals, the first front all solutions are optimized; finally, after several generations of iteration, stable solutions are obtained by convergence; The optimization of the steel structure geometric size of the target metal damper is finally completed.

Images