Method for predicting fatigue life of structure under multistage stress loading

Abstract

The invention discloses a method for predicting fatigue residual life of a multi-stage stress loading structure. The method comprises the following steps: respectively determining values of weight factors of high and low loading and weight factors of low and high loading of two-stage stress based on an improved whale optimization algorithm; substituting the determined numerical value, the two-stage loading stress and the corresponding fatigue life into life ratio characteristic parameters of an improved M-H model for correction to obtain improved life ratio characteristics; and substituting the improved life ratio characteristic parameters into the M-H model to replace life ratio characteristic parameters in the classical M-H model to obtain a fatigue life prediction model under multi-stage stress loading. According to the method for predicting the fatigue life of the structure under the multi-stage stress loading based on the improved M-H model, the loading sequence effect of different materials and the interaction effect between loads can be fully represented based on the optimization algorithm instead of empirical correction, so that the prediction precision is stably improved, and the application range is wider.

Description

Technical Field

[0001] The scheme belongs to the technical field of structural fatigue damage assessment, and particularly relates to a method for predicting the fatigue life of a structure under multi-level stress loading. Background Art

[0002] The service environment of the structure is complex, and the loading state is usually multi-level stress level loading, so the structure is prone to fatigue damage accumulation, performance degradation, and finally fracture. However, due to the influence of factors such as material properties, loading conditions and structural geometry, the accumulation process of structural fatigue damage is nonlinear, which increases the uncertainty of the prediction of structural fatigue remaining life. Therefore, a reliable fatigue remaining life prediction method is needed to effectively evaluate the fatigue damage state of the structure, thereby enhancing the service safety of the material.

[0003] The most commonly used method for predicting the fatigue residual life of structures under multi-level stress is based on the Manson-Halford (MH) nonlinear fatigue cumulative damage model. However, the existing MH nonlinear fatigue cumulative damage models have the following problems: first, the calculation formula is complex, the amount of calculation increases, and it is limited in practical engineering applications; second, most of the improvements are empirical, and it is difficult to accurately reproduce the impact of the interaction between multi-level stress loads on fatigue damage accumulation, which makes the prediction error volatile and the prediction accuracy unstable. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide an improved method for predicting the fatigue life of a structure under multi-level stress loading.

[0005] To this end, some embodiments of the present application propose a method for predicting the fatigue life of a structure under multi-level stress loading, including step 1: conducting a uniaxial multi-level stress structure fatigue loading experiment, recording the loading stress (σ i ), number of cyclic loading (n i ) and the corresponding fatigue life (N i ), where i represents one level; Step 2: Select the experimental data of any two adjacent stress loadings of different material samples, and determine the values ​​of the weight factors (γ1) of the high and low loading and the weight factors (γ2) of the low and high loading of the two stresses based on the improved whale optimization algorithm; wherein the optimization objective function is set to minimize the average absolute error between the predicted value and the experimental value; Step 3: Based on the weight factors (γ1) of the high and low loading and the weight factors (γ2) of the low and high loading, the two-level loading stress and the corresponding fatigue life are brought into the life ratio characteristic parameters of the improved MH model for correction, and an improved life ratio characteristic is obtained, whose expression is: And step 4: Bring the improved life ratio characteristic parameters into the MH model to replace the life ratio characteristic parameters in the classic MH model to obtain the fatigue life prediction model under multi-level stress loading, the expression is Where N i is the fatigue life under the corresponding stress loading level.

[0006] In some embodiments, the improved whale optimization algorithm includes: step S21, based on the improved Circle chaos mapping theory Improved population initialization, where n is the population size, x n is the original position of the target, x n+1 is the new position after chaotic mapping; Step S22, the dynamically changing inertia weight ω is introduced into the whale optimization algorithm, and its calculation method is: In the formula, B(1,2) represents a random number that obeys the Beta distribution; * is represented as the current best candidate whale position vector, X is represented as the position vector, X rand It is represented as a reference whale position vector randomly selected by forced selection, based on which a new position update strategy can be obtained, the position update strategy includes update strategy 1: X(t+1)=ωX * (t)-AD, update strategy 2: X(t+1)=ωX rand -AgD rand , and update strategy 3: X(t+1)=D′e bl cos(2πl)+ωX * (t), b is the constant coefficient used to define the shape of the logarithmic spiral, l is a random number on [-1,1], and D is the movement of the individual whale to X * The step vector close to each other can be calculated according to D = |CX * (t)-X(t)|, where C is a random number on [0,2], and D′ is the difference between the individual whale and X * distance vector; step S23, using the pinhole imaging reverse learning strategy to help the algorithm jump out of the local optimum.

[0007] In some embodiments, the weight factor (γ) value has a predetermined precision.

[0008] In some embodiments, the weighting factor for high-low stress loading is 0.00901, and the weighting factor for low-high stress loading is 0.01011.

[0009] The beneficial effects of the embodiments of the present invention include that the method for predicting the fatigue life of a structure under multi-level stress loading based on the improved MH model proposed in some embodiments is not an empirical correction, but is based on an optimization algorithm, which can fully characterize the loading sequence effect of different materials and the interaction effect between loads, so that the prediction accuracy is steadily improved and the application range is wider. In some embodiments, the optimization algorithm is improved to have a higher optimization speed and optimization accuracy, which helps to accurately determine the value of the weight factor, thereby improving the prediction accuracy of the method for predicting the fatigue life of a structure under multi-level stress loading. In some embodiments, the weight factor determined based on the secondary stress level can be extended to stress levels above the secondary level, and fewer parameters are introduced, which simplifies the calculation process of the MH model, and can obtain a higher-precision prediction result in a simpler way, which has a significant impact on predicting the fatigue life under multi-level stress level loading. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A flow chart of a method for predicting structural fatigue life under multi-level stress loading according to an embodiment of the present invention.

[0011] Figure 2 Flowchart of an improved whale optimization algorithm according to an embodiment of the present invention.

[0012] Figure 3 Optimization effect diagram of the improved whale optimization algorithm according to an embodiment of the present invention.

[0013] Figure 4 Comparison diagram of predicted values ​​and experimental values ​​according to an embodiment of the present invention. DETAILED DESCRIPTION

[0014] This scheme provides a method for predicting the fatigue residual life of multi-level stress loading structures. The process is as follows: Figure 1 As shown, it mainly includes the following steps:

[0015] Step 1: Carry out uniaxial multi-level stress structure fatigue loading experiment and record the loading stress σ at each level i , number of cyclic loading times n i And the corresponding fatigue life N i .

[0016] Step 2: Select the experimental data of two-level stress loading of different material samples, and set the optimization objective function as the minimum average absolute error between the predicted value and the experimental value based on the improved whale optimization algorithm, and determine the weight factor γ values ​​of the two-level stress high and low and low and high loading, respectively, accurate to 5 decimal places. The flowchart of the improved whale optimization algorithm is as follows: Figure 2 As shown, it includes: step S21, based on the improved Circle chaos mapping theory Improved population initialization, where n is the population size, x n is the original position of the target, x n+1 is the new position after chaotic mapping; Step S22, the dynamically changing inertia weight ω is introduced into the whale optimization algorithm, and its calculation method is: In the formula, B(1,2) represents a random number that follows a Beta distribution; * is represented as the current best candidate whale position vector, X is represented as the position vector, X rand It is represented as a reference whale position vector randomly selected by forced selection, based on which a new position update strategy can be obtained, the position update strategy includes update strategy 1: X(t+1)=ωX * (t)-AD, update strategy 2: X(t+1)=ωX rand -AgD rand , and update strategy 3: X(t+1)=D′e bl cos(2πl)+ωX * (t), b is the constant coefficient used to define the shape of the logarithmic spiral, l is a random number on [-1,1], and D is the movement of the individual whale to X * The step vector close to each other can be calculated according to D = |CX * (t)-X(t)|, where C is a random number on [0,2], and D′ is the difference between the individual whale and X * distance vector; step S23, using the pinhole imaging reverse learning strategy to help the algorithm jump out of the local optimum.

[0017] like Figure 3 As shown in the figure, compared with the classic whale optimization algorithm, the improved whale optimization algorithm has been confirmed by the test function to have faster optimization speed and higher optimization accuracy.

[0018] In this embodiment, the Circle chaotic map is used to generate the initial group. The Circle chaotic map is used to generate the initial population. Compared with the randomly distributed population, the improved population initial position distribution ratio is more uniform, which expands the search range of the whale group in space and increases the diversity of the overall position.

[0019] The pinhole imaging reverse learning strategy proposed in this embodiment combines pinhole imaging reverse learning with a dimension-by-dimension reverse learning strategy to obtain a reverse solution dimension by dimension for a feasible solution, so that the algorithm can obtain the ability to escape from local extreme values.

[0020] Step 3: Based on the weight factor, loading stress and corresponding fatigue life, the life ratio characteristic parameters of the improved MH model are brought in for correction. The expression is:

[0021] Step 4: Substitute the improved life ratio characteristic parameters into the classic MH model to replace the life ratio characteristic parameters, and then improve the MH model to obtain the fatigue life prediction model under multi-level stress loading, which is expressed as The N on the left side of the equation i is the value to be evaluated, that is, the fatigue life under the corresponding grade stress loading level.

[0022] A verification case is provided below to describe the application effect of this scheme. Obviously, the verification case provided is only one of the many verification cases used in the proposal and verification process of this scheme. The verification case selected LY12CZ as the research object to verify the feasibility of the proposed structural fatigue prediction method under multi-level stress loading. LY12CZ is a high-strength hard aluminum used to make various high-load parts and components, such as skeleton parts, skins, bulkheads, wing ribs, wing spars, rivets, etc. on aircraft. The uniaxial symmetric cyclic fatigue test scheme and experimental results under two-level stress level loading of LY12CZ are shown in Table 1. The uniaxial symmetric cyclic fatigue test scheme and experimental results under three-level stress level loading of LY12CZ are shown in Table 2.

[0023] Table 1

[0024]

[0025] Table 2

[0026]

[0027] The weight factors were determined based on the improved whale optimization algorithm: the weight factor for high and low stress loading was 0.00901, and the weight factor for low and high stress loading was 0.01011.

[0028] Apply this scheme to predict the data in Table 1 and Table 2, such as Figure 4 It can be seen that the predicted values ​​are all within the 50% error range, and most of them can even meet the 25% error range. And compared with the classic MH model, its prediction error is greatly reduced, see Table 3.

[0029] Table 3

[0030]

[0031] Compared with existing technologies, the advantages of this solution include:

[0032] First, the prediction method of structural fatigue life under multi-level stress loading based on the improved MH model proposed in this scheme is not an empirical correction, but is based on an optimization algorithm. It can fully characterize the loading sequence effect and the interaction effect between loads of different materials, so that the prediction accuracy is steadily improved and the application range is wider.

[0033] Second, the optimization algorithm has been improved to have higher optimization speed and accuracy, which helps to accurately determine the value of the weight factor, thereby improving the prediction accuracy of the method for predicting the fatigue life of the structure under multi-level stress loading.

[0034] Third, the weight factor determined based on the secondary stress level can be extended to stress levels above the secondary level, and fewer parameters are introduced, which simplifies the calculation process of the MH model and can obtain higher-precision prediction results in a simpler way, which has a significant impact on predicting fatigue life under loading at multiple stress levels.

[0035] The above description is only a preferred specific implementation manner of the present invention, and the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for predicting the fatigue residual life of a multi-level stress loading structure, characterized by: The following steps are involved: Step S1: Carry out uniaxial multi-level stress structure fatigue loading experiment and record the loading stress (σ i ), number of cyclic loading (n i ) and the corresponding fatigue life (N i ), where i represents the first level; Step S2: Select any two adjacent stress loading experimental data of different material samples, and determine the values ​​of the weight factors (γ1) of the high and low stress loading and the weight factors (γ2) of the low and high stress loading respectively based on the improved whale optimization algorithm; wherein the optimization objective function is set to minimize the average absolute error between the predicted value and the experimental value; Step S3: Based on the weight factor (γ1) of the high and low loading and the weight factor (γ2) of the low and high loading, the two-level loading stress and the corresponding fatigue life are brought into the life ratio characteristic parameter of the improved MH model for correction to obtain an improved life ratio characteristic, which is expressed as follows: Step S4: Bring the improved life ratio characteristic parameters into the MH model to replace the life ratio characteristic parameters in the classic MH model to obtain a prediction model for fatigue life under multi-level stress loading, expressed as Where N i is the fatigue life under the corresponding stress loading level.

2. The method for predicting the fatigue residual life of a multi-level stress loading structure according to claim 1 is characterized in that: The improved whale optimization algorithm comprises: step S21, based on the improved Circle chaos mapping theory Improved population initialization, where n is the population size, x n is the original position of the target, x n+1 is the new position after chaotic mapping; Step S22, the dynamically changing inertia weight ω is introduced into the whale optimization algorithm, and its calculation method is: In the formula, B(1,2) represents a random number that follows the Beta distribution; X * is represented as the current best candidate whale position vector, X is represented as the position vector, X rand It is represented as a reference whale position vector randomly selected by forced selection, based on which a new position update strategy can be obtained, the position update strategy includes update strategy 1: X(t+1)=ωX * (t)-AD, update strategy 2: X(t+1)=ωX rand -AgD rand , and update strategy 3: X(t+1)=D′e bl cos(2πl)+ωX * (t), b is the constant coefficient used to define the shape of the logarithmic spiral, l is a random number on [-1,1], and D is the movement of the individual whale to X * The step vector close to each other can be calculated according to D = |CX * (t)-X(t)|, where C is a random number on [0,2], and D′ is the difference between the individual whale and X * distance vector; step S23, using the pinhole imaging reverse learning strategy to help the algorithm jump out of the local optimum.

3. The method for predicting the fatigue residual life of a multi-level stress loading structure according to claim 1 is characterized in that: The weight factor (γ) value has a predetermined accuracy.

4. The method for predicting the fatigue residual life of a multi-level stress loading structure according to claim 1 is characterized in that: The weight factor for high-low stress loading is 0.00901, and the weight factor for low-high stress loading is 0.01011.

Images