CN115169241A - TiAl alloy fatigue life prediction method driven by data-model - Google Patents
TiAl alloy fatigue life prediction method driven by data-model Download PDFInfo
- Publication number
- CN115169241A CN115169241A CN202210886715.6A CN202210886715A CN115169241A CN 115169241 A CN115169241 A CN 115169241A CN 202210886715 A CN202210886715 A CN 202210886715A CN 115169241 A CN115169241 A CN 115169241A
- Authority
- CN
- China
- Prior art keywords
- fatigue life
- model
- tial alloy
- data
- verifying
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Medical Informatics (AREA)
- Development Economics (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention provides a method for predicting fatigue life of TiAl alloy driven by data-model interaction. The method combines the accurate prediction capability of model drive on the fatigue life of a single working condition and the generalization capability of data drive on the fatigue life of multiple working conditions, and establishes an accurate prediction model of the fatigue life of the TiAl alloy under the conditions of small samples and multiple working conditions. According to the method, theoretical modeling is carried out on fatigue life test data of small samples under various working conditions through model driving, a large amount of multi-working condition data are obtained according to theoretical model interpolation, and mining learning is carried out on the obtained data by using machine learning, so that an accurate TiAl alloy fatigue life prediction model with generalization capability is obtained. The method overcomes the defect that the fatigue life can only be predicted at a specific loading angle and temperature when the fatigue life of the TiAl alloy is predicted by the traditional method, and has the advantages of small sample, high precision and strong generalization capability.
Description
Technical Field
The invention relates to a method for predicting fatigue life of TiAl alloy driven by data-model interaction, belonging to the field of prediction of fatigue life of materials.
Background
The TiAl alloy has excellent creep resistance, oxidation resistance and higher specific strength, and is a light heat-resistant high-temperature structural material with application potential. The microstructure of the TiAl alloy sheet layer has large anisotropy of mechanical properties, wherein the resistance of the initiation and the propagation of fatigue cracks is random, so that the fatigue life of the TiAl alloy sheet layer shows obvious difference under different loading angles and temperatures.
Although the fatigue life of the TiAl alloy can be measured by an experimental method and life prediction is carried out by a theoretical model, the experimental cost is high, and only small sample data can be obtained, so that the traditional model prediction method can only carry out prediction under a single working condition.
In recent years, machine learning methods based on data science have been increasingly applied to the fields of material performance prediction and new material development, and have significant advantages in prediction efficiency and cost control. However, machine learning requires a large number of reliable data sets for mining learning, and fatigue data can only be obtained from small sample data, so that the accuracy of the fatigue life prediction method based on data driving is limited. Based on the reasons, an accurate fatigue life prediction method for small samples and multiple working conditions is urgently needed to be developed.
Disclosure of Invention
The invention aims to provide a data-model mutual driving TiAl alloy fatigue life prediction method so as to accurately predict the fatigue life of a TiAl alloy with a small sample, low cost and high precision.
In order to achieve the purpose, the invention provides a TiAl alloy fatigue life prediction method driven by a data-model, which specifically comprises the following steps:
the method comprises the following steps: acquiring a TiAl alloy working condition and fatigue life original data set;
step two: establishing a theoretical model, and performing data processing on the original data set in the step one according to the theoretical model to obtain a standard data set;
step three: constructing a random forest model, training the standard data set in the second step, and testing the optimal parameter combination to obtain a fatigue life prediction model;
step four: inputting the working condition of the TiAl alloy to be predicted into the fatigue life prediction model in the step three to obtain the fatigue life prediction value of the TiAl alloy.
Preferably, in the first step, the operating conditions of the TiAl alloy are obtained, and the operating conditions include loading angle, temperature, and stress.
Preferably, in the second step, a theoretical model is established, where lgN = a + bS, where a and b are constants, N is fatigue life, and S is stress.
Preferably, in the second step, the original data set is interpolated to obtain a standard data set.
Preferably, in the third step, the parameter combinations of the constructed random forest model include the number of decision trees, the maximum depth of the decision trees, and the maximum feature number of the decision trees.
Preferably, in the third step, the optimal parameter combination is tested by adopting a cross-validation method, and the specific steps are as follows:
dividing the standard data set into ten parts in a random average way, and respectively naming the ten parts as D 1 ,D 2 ,…,D i I =1,2, \8230;, 10, D 1 Training the random forest model as a verification set and the rest nine training sets by using a decision coefficient R 1 2 And root mean square error RMSE 1 Evaluating the model accuracy; will D 2 Training the random forest model as a verification set and the rest nine training sets by using a decision coefficient R 2 2 And root mean square error RMSE 2 Evaluating the model accuracy; by analogy, respectively obtaining D according to the training and evaluating steps i The accuracy of the model is taken as a verification set, the total accuracy of the model is obtained after averaging,
wherein D is i Determining coefficient R as verification set i 2 (i =1,2, \8230;, 10) is:
wherein n is i Is D i Verifying the fatigue life data number, y, of the concentrated TiAl alloy ij Is D i Verifying the fatigue life experimental value of the concentrated TiAl alloy,
is D i Verifying the fatigue life prediction value of the concentrated TiAl alloy,
is D i Verifying the fatigue life average value of the concentrated TiAl alloy;
D i root mean square error RMSE as a validation set i (i =1,2, \8230;, 10) is:
wherein n is i Is D i Verifying the fatigue life data number, y, of the concentrated TiAl alloy ij Is D i Verifying the fatigue life experimental value of the concentrated TiAl alloy,
is D i Verifying the predicted value of the fatigue life of the concentrated TiAl alloy.
Compared with the prior art, the invention has the following remarkable advantages:
according to the invention, the fatigue life of the TiAl alloy under different working conditions can be predicted with smaller error by establishing a proper random forest model. Compared with the traditional method, the method saves the material and time cost and can quickly and accurately predict the fatigue life of the TiAl alloy.
Drawings
FIG. 1 is a comparison graph of predicted fatigue life and actual fatigue life of a TiAl alloy fatigue life prediction method driven by data-models.
FIG. 2 is a comparison graph of the fatigue stress-fatigue life prediction and the real fatigue stress-fatigue life of the TiAl alloy fatigue life prediction method driven by the data-model.
FIG. 3 is a schematic diagram showing the importance of working condition parameters to fatigue life of the TiAl alloy fatigue life prediction method driven by data-models in the invention.
Detailed Description
The present invention will be described in more detail with reference to the following examples and the accompanying drawings.
Examples
The invention relates to a method for predicting fatigue life of TiAl alloy driven by data-model interaction, which specifically comprises the following steps:
the method comprises the following steps: acquiring 30 groups of TiAl alloy working condition and fatigue life original data sets;
specifically, the obtained working conditions of the TiAl alloy comprise a loading angle, temperature and stress;
step two: establishing a theoretical model, and performing data processing on the original data set in the step one according to the theoretical model to obtain a standard data set;
specifically, the theoretical model is lgN = a + bS, wherein a and b are constants, N is fatigue life, and S is stress, values of a and b are determined, and then an original data set is subjected to interpolation processing to obtain 78 groups of data to form a standard data set;
step three: constructing a random forest model, training the standard data set in the step two, and testing the optimal parameter combination so as to obtain a final fatigue life prediction model;
specifically, the parameter combination of the constructed random forest model comprises the number of decision trees, the maximum depth of the decision trees and the maximum feature number of the decision trees;
the optimal parameters are as follows: the number of the decision trees is 90, the maximum depth of the decision trees is 10, and the maximum characteristic number of the decision trees is 2;
the optimal parameter combination is tested by adopting a cross-validation method to obtain a final fatigue life prediction model, and the method comprises the following specific steps of:
dividing the standard data set into ten parts in a random average way, and respectively naming the ten parts as D 1 ,D 2 ,…,D i I =1,2, \8230;, 10, D 1 As a verification set, the rest nine parts are used as training sets to train the random forest model, and a decision coefficient R is adopted 1 2 And root mean square error RMSE 1 Evaluating the model accuracy; will D 2 As a verification set, the rest nine parts are used as a training set to train the random forest model, and a decision coefficient R is adopted 2 2 And root mean square error RMSE 2 Evaluating the model accuracy; by analogy, the training and evaluation steps are carried out to obtain D in sequence i The accuracy of the model is obtained after ten times of averaging when the model is taken as a verification set,
wherein D is i Determining coefficient R as verification set i 2 (i =1,2, \8230;, 10) is:
wherein n is i Is D i Verifying the fatigue life data number, y, of the centralized TiAl alloy ij Is D i Verifying the fatigue life experimental value of the concentrated TiAl alloy,
is D i Verifying the fatigue life prediction value of the concentrated TiAl alloy,
is D i Verifying the fatigue life average value of the concentrated TiAl alloy;
D i root Mean Square Error (RMSE) as a validation set i (i =1,2, \8230;, 10) is:
wherein n is i Is D i Verifying the fatigue life data number, y, of the centralized TiAl alloy ij Is D i Verifying the fatigue life experimental value of the concentrated TiAl alloy,
is D i Verifying the fatigue life prediction value of the concentrated TiAl alloy;
FIG. 1 is a comparison graph of fatigue life prediction model for predicting fatigue life of TiAl alloy and actual fatigue life, and with the aid of the comparison graph in FIG. 1, the total accuracy of the model obtained through cross validation is R 2 0.801 for RMSE 0.729;
step four: inputting the working condition of the TiAl alloy to be predicted into the fatigue life prediction model in the third step to obtain a predicted value of the fatigue life of the TiAl alloy, wherein the predicted value is shown in figure 2;
FIG. 3 is a schematic diagram showing the importance of working condition parameters of the TiAl alloy fatigue life prediction method based on data-model mutual driving on fatigue life, and it can be seen that the influence factors of material stress on fatigue life are more than 80%.
Claims (6)
1. A TiAl alloy fatigue life prediction method driven by data-models is characterized by comprising the following steps:
the method comprises the following steps: acquiring a TiAl alloy working condition and fatigue life original data set;
step two: establishing a theoretical model, and performing data processing on the original data set in the step one according to the theoretical model to obtain a standard data set;
step three: constructing a random forest model, training the standard data set in the second step, and testing the optimal parameter combination to obtain a fatigue life prediction model;
step four: inputting the working condition of the TiAl alloy to be predicted into the fatigue life prediction model in the step three to obtain the fatigue life prediction value of the TiAl alloy.
2. The method of claim 1, wherein in the first step, the operating conditions of the TiAl alloy are obtained, and the operating conditions comprise loading angle, temperature and stress.
3. The method of claim 1, wherein in the second step, a theoretical model is established, the theoretical model is lgN = a + bS, wherein a and b are constants, N is fatigue life, and S is stress.
4. The method of claim 1, wherein in step two, the raw data set is interpolated to obtain a standard data set.
5. A method as claimed in claim 1, wherein in step three, the parameter combinations of the constructed random forest model comprise the number of decision trees, the maximum depth of the decision trees and the maximum number of features of the decision trees.
6. The method of claim 1, wherein in step three, the optimal parameter combination is tested by using a cross-validation method, which comprises the following specific steps:
dividing the standard data set into ten parts in random average, and respectively naming the ten parts as D 1 ,D 2 ,…,D i I =1,2, \ 8230;, 10, and D 1 As a verification set, the rest nine parts are used as a training set to train the random forest model, and a decision coefficient R is adopted 1 2 And root mean square error RMSE 1 Evaluating the model accuracy; will D 2 Training the random forest model as a verification set and the rest nine training sets by using a decision coefficient R 2 2 And root mean square error RMSE 2 Evaluating the model accuracy; by analogy, according to the training and evaluating steps, respectively obtaining D i The accuracy of the model is taken as a verification set, the total accuracy of the model is obtained after averaging,
wherein D is i Determining coefficient R as verification set i 2 (i =1,2, \8230;, 10) is:
wherein n is i Is D i Verifying the fatigue life data number, y, of the centralized TiAl alloy ij Is D i Verifies the concentrated fatigue life experimental value of the TiAl alloy,
is D i Verifying the fatigue life prediction value of the concentrated TiAl alloy,
is D i Verifying the fatigue life average value of the concentrated TiAl alloy;
D i root Mean Square Error (RMSE) as a validation set i (i =1,2, \8230;, 10) is:
wherein n is i Is D i Verifying the fatigue life data number, y, of the centralized TiAl alloy ij Is D i Verifying the fatigue life experimental value of the concentrated TiAl alloy,
is D i And verifying the predicted value of the fatigue life of the concentrated TiAl alloy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210886715.6A CN115169241A (en) | 2022-07-26 | 2022-07-26 | TiAl alloy fatigue life prediction method driven by data-model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210886715.6A CN115169241A (en) | 2022-07-26 | 2022-07-26 | TiAl alloy fatigue life prediction method driven by data-model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115169241A true CN115169241A (en) | 2022-10-11 |
Family
ID=83497638
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210886715.6A Pending CN115169241A (en) | 2022-07-26 | 2022-07-26 | TiAl alloy fatigue life prediction method driven by data-model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115169241A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117150918A (en) * | 2023-09-08 | 2023-12-01 | 南京理工大学 | High-precision fatigue life prediction method based on small sample machine learning |
-
2022
- 2022-07-26 CN CN202210886715.6A patent/CN115169241A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117150918A (en) * | 2023-09-08 | 2023-12-01 | 南京理工大学 | High-precision fatigue life prediction method based on small sample machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104820781B (en) | Consider that temperature follows the BGA thermal fatigue life of solder joint Forecasting Methodologies of load sequence loading effect | |
| CN106354898B (en) | A kind of weld fatigue Life Calculating Methods based on total strain energy metric density | |
| CN103970965A (en) | Test run method for accelerated life test of gas turbine engine | |
| CN111310348A (en) | Material constitutive model prediction method based on PSO-LSSVM | |
| CN111339594B (en) | DIC technology-based near-field dynamics parameter experiment inversion system and use method | |
| WO2013026500A2 (en) | System and method for generating a combined model for isothermal and anisothermal fatigue life | |
| CN112307664B (en) | CT sample stress field introducing residual stress and crack propagation analysis method | |
| CN113033105B (en) | Method for predicting fatigue performance of steel material based on mechanical theory guiding transfer learning | |
| CN115169241A (en) | TiAl alloy fatigue life prediction method driven by data-model | |
| CN112651562B (en) | Laser cladding process optimization method based on signal-to-noise ratio and improved gray correlation degree | |
| CN115950609B (en) | Bridge deflection anomaly detection method combining correlation analysis and neural network | |
| CN110987676A (en) | Full-life prediction method considering crack closure effect under random multi-axis load | |
| CN114323940A (en) | Method for testing fatigue crack propagation rate of compact tensile sample under trapezoidal wave loading condition | |
| CN114295491B (en) | Prediction method for creep damage and deformation evolution behavior along with time | |
| CN114323994A (en) | Method for testing fatigue crack propagation rate of three-point bending sample in T-shaped wave load-holding test | |
| CN111143934A (en) | Structural deformation prediction method based on time convolution network | |
| CN115063051A (en) | Steel pipe performance evaluation method and system based on metal element analysis | |
| CN106991074A (en) | Accelerated degradation test Optimization Design based on accelerated factor principle of invariance | |
| CN105158147B (en) | A kind of device and method for being used to test sealing ring material aging | |
| CN117594164A (en) | Metal structure residual fatigue life calculation and evaluation method and system based on digital twin | |
| CN109918699B (en) | Method for identifying key parameters of performance of throat insert made of C/C composite material | |
| CN110457746B (en) | BP neural network-based structural plane peak shear strength prediction model construction method | |
| CN107704691A (en) | A kind of accelerated stress reliability compliance test preferred scheme design method | |
| CN108106873A (en) | The method and system of servo-drive system reliability test evaluation | |
| CN116842834A (en) | Interpretable creep rupture life prediction method based on machine learning and SHAP value |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |