Creep-fatigue Life Prediction Research Articles

A physics-informed neural network for creep-fatigue life prediction of components at elevated temperatures

Physics-informed neural network has strong generalization ability for small dataset, due to the inclusion of underlying physical knowledge. Two strategies are enforced to incorporate physics constraints to a deep neural network in this work. One is to obtain extended features through physics-informed feature engineering, and the other is to incorporate physics-informed loss function into deep neural network as constraints. Conventional machine learning models, deep neural network and physics-informed neural network are applied to predict creep-fatigue life of 316 stainless steel. Results show that physics-informed neural network presents better prediction accuracy than deep neural network and conventional machine learning models.

Accuracy of the predicting for creep-fatigue cyclic life based on parameters in a characteristic cycle

The purposes of the present work are to evaluate the prediction accuracy of characteristic cycle on creep-fatigue life prediction and to analyze the advantages and disadvantages of various models in tensile strain-dwell tests. The parameters of the first cycle, the 10% life, the half-life, and the characteristic cycle that we proposed (a turning point between the initial rapid softening and subsequent slight softening/hardening) are employed to linear-damage-summation (LDS) and energy-life prediction models based on the creep-fatigue data of directionally-solidified Nickel-based superalloy, DZ445, at 900 °C. It is found that the characteristic cycle parameters with clear physical significance have the highest life prediction accuracy. Moreover, the optimal critical value of fatigue and creep damage (the coordinate of the creep-fatigue envelope intersection) in the LDS is also determined. The prediction accuracy of creep damage based on the time-fraction, the simple ductility-exhaustion, and the strain-energy- density-exhaustion models is sequentially improved in the LDS rule. In the energy-life model, the life prediction accuracy based on damage mechanism and frequency correction is higher than the value without any correction. This investigation provides a new method of the parameter-selection for the creep-fatigue life prediction, which is accurate and convenient. It provides the theoretical-guidance for creep-fatigue life prediction in both laboratory experiment and actual components.

An Improved Creep-Fatigue Life Model Involving the Cyclic Softening/Hardening and Stress Relaxation Effect

Abstract In this work, improved creep-fatigue life prediction models based on time fraction (TF) and ductility exhaustion (DE) rule are established for creep-fatigue life prediction of components at elevated temperatures. Cyclic softening/hardening and stress relaxation models for cyclic softening/hardening material are incorporated in the model. Materials data conducted by authors and collected from references are employed for validations. Results indicated that the improved models proposed could present more realistic evolution behavior of materials than the half-life models. The differences of the predicted lives based on the improved models and half-life models are dependent on cyclic behaviors of materials. The predicted life by the improved model is lower than that gained by the half-life model for cyclic softening steels, while it is higher than that by the half-life model for cyclic hardening steels. In addition, the DE rule could overcome the influence of cyclic softening/hardening on creep-fatigue life prediction more significantly than the TF rule.

Grain boundary sliding model for assessing creep-fatigue life of Sn37Pb eutectic solder

This study discusses the creep-fatigue life prediction of eutectic Sn37Pb solder. Creep-fatigue lives were experimentally obtained at 313 K using fast-fast, slow-slow, slow-fast, fast-slow, and strain-hold waveforms. Conventional four creep-fatigue life prediction rules, time fraction rule, strain range partitioning method, frequency modified fatigue life, and ductility exhaustion model was used to predict the creep-fatigue lives. However, not all the rules were able to properly estimate them. Since the grain boundary sliding rates were closely related to creep-fatigue lives, a new life prediction method based on grain boundary sliding rates was proposed.

Creep-fatigue life prediction of perforated copper-alloy plate by using digital image correlation method

Creep-fatigue life prediction of perforated copper-alloy plate by using digital image correlation method

A crystal plasticity-based approach for creep-fatigue life prediction and damage evaluation in a nickel-based superalloy

A numerical process based on crystal plasticity finite element (CPFE) was implemented to predict creep-fatigue crack initiation life. CPFE-based model can describe the macroscopic cyclic deformation and reveal grain-level damage mechanism. A new life prediction approach was then constructed by introducing fatigue and creep indicator parameters. Furthermore, a series of strain-controlled creep-fatigue tests in GH4169 superalloy at 650℃ were used to validate predicted accuracy of this model, where most of the data points lied within ±1.5 error band. Finally, a potential methodology for conservative creep-fatigue life evaluation was provided by flexible creep-fatigue damage summation rule in engineering applications.

Thermal performance and fatigue life prediction of POP stacked chip assembly under thermal cycling load

PurposeThis paper aims to perform experimental test on fatigue characteristics of package on package (POP) stacked chip assembly under thermal cycling load. Some suggestions for design to prolong fatigue life of POP stacked chip assembly are provided.Design/methodology/approachThe POP stacked chip assembly which contains different package structure mode and chip position was manufactured. The fatigue characteristics of POP stacked chip assembly under thermal cycling load were tested. The fatigue load spectrum of POP stacked chip assembly under thermal cycling load was given. The fatigue life of chips can be estimated by using the creep–fatigue life prediction model based on different stress conditions.FindingsThe solder joint stress of top package is significantly less than that of bottom solder joints, and the maximum value occurs in the middle part of the solder joints inner ring.Originality/valueThis paper fulfils useful information about the thermal reliability of POP stacked chip assembly with different structure characteristics and materials parameters.

Experimental study on creep-fatigue behaviors of chinese P92 steel with consideration of several important factors

In this paper, the creep fatigue test at 550–650 °C of P92 steel is carried out, which not only reveals the damage evolution mechanism, but also analyzed the creep fatigue behaviors. The influence of temperature, load holding time and pre-fatigue damage on the creep fatigue behavior of P92 steel was studied in detail. The effects of the temperature and holding time on the hysteresis loops of stress–strain are evaluated quantitatively. Actually, the load holding time mainly affects the stress relaxation behavior. Secondly, pre-fatigue damage will replace the initial stage of cycle stress response and significantly reduce the duration of the stable stage. The linear cumulative damage model (LCD), strain range division model (SRD), strain energy density exhaustion model (SEDE) were used to predict the life of P92 steel. Actually, the SEDE model is found to be most suitable for the creep fatigue life prediction. Furthermore, the microstructure of the specimen and fracture at pre- and post- test was observed to explain the damage evolution mechanism of the creep fatigue interaction. The microstructure analysis shows that the main body of P92 steel is tempered martensite in lath shape, and a large number of M23C6 carbides are dispersed at the grain boundary. The M23C6 carbides have poor thermal stability, which will coarsen and connect after creep, reducing the material properties. The analysis of fracture morphology shows that the creep fatigue specimen with short-term load holding has notable characteristics of fatigue fracture, while the creep fatigue specimen with long-term load holding has notable characteristics of creep fracture.

Creep-fatigue life prediction in nickel-based superalloy GH4169 based on microstructural damage quantification with the help of electron backscatter diffraction

A series of strain-controlled creep-fatigue tests under different loading waveforms are carried out on GH4169 superalloy at 650 °C. In the macroscopic view, a unified viscoplastic constitutive framework is used to describe cyclic deformation process. Particularly, a modified kinematic hardening rule considering loading-dependent effect is developed to simulate the stress relaxation behavior during hold periods. Then, creep-fatigue assessment based on interaction diagram is quantitatively determined by the strain energy density exhaustion (SEDE) approach. In order to elucidate damage mechanisms under various loading conditions, the characterization of damage mechanism is observed from the post-test examination. Main-crack-failure modes from fracture appearance observations and cracking modes from longitudinal sections are studied via scanning electron microscopy (SEM) and electron backscatter diffraction (EBSD), respectively. Furthermore, image-based creep-fatigue diagram is proposed based on metallographic interpretation of mechanisms.

A new fatigue life prediction model considering the creep-fatigue interaction effect based on the Walker total strain equation

The performance of high-temperature components of aero-engines under the Creep-Fatigue Interaction (CFI) behavior gets more attention recently. In this research, the creep-fatigue tests of two superalloys of Powder Metallurgy (PM) FGH96 and direct aging GH4169 were performed at 650 °C with different types of dwell, and the fracture morphology of FGH96 specimens was observed by Scanning Electron Microscopy (SEM) to analyze the creep-fatigue fracture feature and crack initiation. Additionally, according to phenomenology, the effect of dwell was introduced to develop a new uniaxial fatigue life prediction model based on the total strain equation, which has capability to take dwell time and load ratio into account together. The equations were utilized to model the test data of PM FGH96 and GH4169, together with data of another superalloy PM FGH95 conducted previously. A prominent prediction ability of the model in creep-fatigue life prediction of different superalloys has been manifested. Most data points of test data and estimated data are located within two times scatter band, which is ideal in engineering.

Creep-fatigue strength design of plate-fin heat exchanger by a homogeneous method

Creep-fatigue life prediction is essential to plate-fin heat exchangers (PFHE) work at high temperatures and pressures in addition to thermal cycles. As the core of PFHE, the plate-fin structure is a complex pore structure and it is pretty difficult to carry out the stress analysis by conventional finite element method. Thus in this paper, a homogeneous method is proposed to do the strength design for plate-fin structure. This method treats the pore structure as an equivalent solid structure, and the key is to calculate the equivalent material properties and the equivalent stress. The equivalent mechanical properties have been obtained in our previous work [1], and in this paper the equivalent thermophysical properties including thermal conductivity, coefficient of thermal expansion, specific heat and density have been derived. This equivalent method is also verified by the three-dimensional finite element method. Besides, creep and fatigue tests of plate-fin structure are applied to calculate the stress and strain magnification factors. Basing on the equivalent material properties, a full scale stress analysis of PFHE by the equivalent finite element model has been carried out successfully to calculate the equivalent stress. After that, we multiply the equivalent stress-strain by the stress and strain magnification factors to calculate the local stress and strain. Then the creep and fatigue damages are predicted according to design curves of base metal. The result of equivalent model agrees well with that of the classical model, which concludes that this homogeneous method is effective to predict the macroscopic performance of plate-fin structure.

Creep-Fatigue Life Prediction of 316H Stainless Steel by Utilizing Non-Unified Constitutive Model

True stress and strain are necessary to estimate the rupture life under creep-fatigue conditions. Finite element analysis (FEA) is one of the most reliable method to calculate true stress and strain, but the accuracy of the obtained result is often greatly dependent on the constitutive model used. Non-unified constitutive model has been proposed, where inelastic strain is decomposed into creep strain and visco-plastic strain. In this paper, a cyclic hardening effect and a plastic strain range dependency were introduced into the non-unified constitutive model to predict the creep-fatigue damage of 316H stainless steel. This model was implemented in a finite element program and FEA were conducted to develop a life assessment method and calculate the creep-fatigue damage by modified ductile exhaustion method. As a consequence, it was revealed that the predicted creep-fatigue lives showed high correspondence with the experimental results by the modified ductile exhaustion method utilizing the non-unified constitutive model.

Creep-fatigue life prediction of perforated copper-alloy plate by using Digital Image Correlation Method

Creep-fatigue life prediction of perforated copper-alloy plate by using Digital Image Correlation Method

Creep-fatigue life prediction and interaction diagram in nickel-based GH4169 superalloy at 650 °C based on cycle-by-cycle concept

The purposes of the present work are to extend the previous energy-based model on the basis of strain energy density exhaustion and to estimate creep-fatigue endurance and accumulated damage using cycle-by-cycle concept in tension-hold-only tests. The nickel-based GH4169 superalloy at 650°C is employed to fit material constants and to verify the prediction capacity of the present model under various loading conditions. The present model exhibits a higher accuracy than some existing models, especially for certain loading waveforms, where the half-life cycle cannot be considered as a representative one. In addition, the trajectory for time-dependent accumulated damage during the operating period can be monitored in the creep-fatigue interaction diagram.

Effect of cyclic loading on subsequent creep behaviour and its implications in creep–fatigue life assessment

Evaluation of creep–fatigue failure is essential in design and fitness evaluation of high-temperature components in power generation plants. Cyclic deformation may alter the creep properties of the material and taking cyclic effects into account may improve the accuracy of creep–fatigue failure life prediction. To evaluate such a possibility, creep tests were conducted on 316FR and modified 9Cr–1Mo steel specimens subjected to prior cyclic loading; their creep deformation and rupture behaviours were compared with those of as-received materials. It was found that creep rupture life and elongation generally decreased following cyclic loading in both materials. In particular, the rupture elongation of 316FR in long-term creep conditions drastically decreases as a result of being cyclically deformed at a large strain range. Use of creep rupture properties after cyclic deformation, instead of those of as-received material, in strain-based and energy-based life estimation approaches brought about a clear improvement of creep–fatigue life prediction.

Effect of dwell time on creep-fatigue life of a high-Nb TiAl alloy at 750 °C

The creep-fatigue life of a cast Ti-45Al-8.0Nb-0.2W-0.2B-0.1Y (at.%) alloy was investigated at 750°C by the consideration of effective dwell time. The life to failure decreases linearly with the increase of effective dwell time, and decreases exponentially with the increase of minimum creep rate. A linear formula and a power-law formula for the life prediction of creep-fatigue were proposed. Both creep-fatigue life prediction formulae have a good agreement with the experimental data.

Creep-Fatigue Life Prediction and Reliability Analysis of P91 Steel Based on Applied Mechanical Work Density

A creep-fatigue (CF) life prediction model and its simplified expression were developed based on the applied mechanical work density (AMWD). The foundation of this model was an integration of N-S curve. Comparisons of the model predicted fatigue lifetimes with the experimental data of load-controlled CF tests on P91 base metal and welded metal at 848 K from the reference were made and apparently illustrated that the model predictions were in a good agreement with the experimental fatigue lifetimes. In addition, the curve of the numbers of cycles to failure versus AMWD at the associated probability was deduced. A reliability model was constructed by combining the curve and the simplified life prediction model.

Creep–fatigue endurance of 304 stainless steels

In this paper, the effects of different factors on the combined creep–fatigue endurance of 304 stainless steel as well as the strength and limitation of the different approaches for creep–fatigue life prediction are reviewed. The factors, which influence on the creep–fatigue endurance, include hold time, strain range, temperature, and stress relaxation behavior. Some relationships between time or number of cycles to failure and the individual factors are summarized from the available data. A phenomenological approach is proposed to describe the synergistic effects of different factors. The approaches used for creep–fatigue life prediction range from purely empirical to mechanistic models based on microstructural observations. Each approach does have some degree of success in dealing with a specific set of creep–fatigue data. The comparison of the prediction capacities of linear damage summation method, strain-range partitioning method, and damage rate approach is made for the same set of creep–fatigue data through Bayesian information criterion analysis. The damage rate model exhibits higher accuracy than other two models.

Research on Creep-Fatigue Life Prediction for P92 Steel under Stress-Controlled State

Taking the P92 steel as the object,Creep-Fatigue (CF) tests of P92 steel at 873K under stress-controlled were carried out with GWT2504 equipment to investigate the CF life prediction. The life prediction model based on Applied Mechanical Work Density (AMWD) was developed in this study,and introduce the effective coefficient ƞ to modify the former. To verify the prediction capability of the AMWD-based and the modified model, comparisons of the models predicted lives with the experimental data of CF tests on P92 steel at 873K were made, it is found out that the AMWD-based model predictions for CF are in agreement with the experimental lives with the factors of 0.9013 and 1.0600, which verifies the model has a good predictability, and the Modified model with the factors of 0.9558 and 1.0469.

Creep Fatigue Life Simulation for Cylinder Head of Diesel Engine Based on the Concept of Continuum Damage Mechanics

The elasto-viscoplastic-creep constitutive equation based on the concept of continuum damage mechanics is formulated for the gray cast iron as used in diesel engine parts. It employs the viscoplastic strain given by Chaboche-Rousselier and the creep strain given by Perrin and Hayhurst. A part of the material parameters contained in the creep constitutive equation which was previously identified by Oka and Toi are assumed to be temperature-dependent in the present study. The identified model is applied to the creep fatigue life prediction of the cylinder head used in diesel engine parts which is made of gray cast iron. Although there are some points to be improved, the validity of the proposed model is illustrated by comparing the calculated results with the experimental results for the crack initiation locations and the crack propagation directions.