Time-dependent Failure Research Articles

A new learning function for estimating time-dependent failure possibility under fuzzy uncertainty

A new learning function for estimating time-dependent failure possibility under fuzzy uncertainty

Investigation of time-dependent gate dielectric breakdown in recessed E-mode GaN MIS-HEMTs using ferroelectric charge trap gate stack (FEG-HEMT)

The aim of achieving E-mode operations has led to the proposal of novel hybrid ferroelectric charge trap gate stack schemes. These schemes involve utilizing a combination of the charge trapping layer (CTL) and the ferroelectric laminated layer to positively shift the threshold voltage to a normally-off margin after initializing the gate pulse, thus achieving the Enhancement-mode (E-mode) characteristic. This study focuses on the forward gate bias time-dependent dielectric breakdown (TDDB) gate reliability in the E-Mode GaN MIS-HEMTs using ferroelectric charge trap gate stack (FEG-HEMT) with and without recess and investigates the impact of temperature on TDDB. The gate voltages required for operation with a 1 % failure rate over a 10-year period are 12.95 V and 10.22 V at 25 °C and 150 °C for both samples, respectively. Also, the examination of the stability of the recessed process by calculating the activation energies (0.7 eV) of both samples are demonstrated. The article also investigates the relationship between long-term reliability and short-term initialization by examining the transition point of the time-dependent failure curve.

Life-Cycle Seismic Reliability Analysis of a Railway Cable-Stayed Bridge Considering Material Corrosion and Degradation

To study the life-cycle seismic reliability analysis (SRA) of cable-stayed bridges (CSBs) taking into account chloride-induced corrosion and degradation of components, an actual railway CSB with uncertainties in structural geometry and material corrosion coefficients was employed in this investigation, and time-dependent models of CSB components at different service times were studied. Based on the OpenSees batch program, we adapted a mass numerical computation to obtain time-dependent non-linear seismic response and probability density function (PDF) of response via the multiplier dimensional-reduction method (MDRM) and the maximum entropy method with fractional moments (FM-MEM). Next, the time-dependent failure possibility of every component and the association coefficient between the failure modes of different parts were acquired. In the end, the product of the conditional marginal (PCM) approach was employed to obtain the life-cycle failure possibility of the CSB system. The results showed that the system failure possibility of the CSB in a corrosive environment increases significantly with increasing servicing time.

Evolution law and mechanism of time-delayed spalling in a deep TBM tunnel: A case study

Time-delayed spalling is proposed to describe the time-dependent splitting failure near the tunnel periphery, which occurs in several days or several months after excavation. This kind of failure has been frequently encountered in a deep TBM tunnel, endangering workers' security and delaying the construction schedule. In this paper, two typical time-delayed spalling cases were systematically introduced, including geological and supporting conditions. Based on field observation, microseismic monitoring and laboratory creep tests, the characteristics and mechanism of time-delayed spalling were investigated. Results showed that time-delayed spalling is a stress-induced time-dependent failure in the vicinity of tunnel surface and appears as flake or plate-like splitting fracture subparallel to the free surface. Although the failure region where time-delayed spalling happened is intact, discontinuities are scattered in the peripheral region. Time-delayed spalling is sensitive to boundary conditions and timely support is an essential method for preventing time-dependent spalling. This research is helpful to the tunnel excavation under similar geological conditions and provides foundations for the prevention of time-delayed spalling.

Redundancy Allocation of Components with Time-Dependent Failure Rates

The Redundancy Allocation Problem (RAP) is well-known in the field of reliability optimization. In this paper, RAP is investigated assuming that the distribution of the time to failure of the components has the form of an Erlang distribution with a time-dependent rate parameter and considering that the choice of redundancy for each subsystem can be none, active, standby or mixed. A genetic algorithm is used to solve the problem of optimal allocation. To analyze the effect of the time dependence, some numerical examples are worked out. Then, a case study of RAP from the literature is analyzed. The obtained results show that time dependence of the failure time distribution parameters can lead to significant differences in the optimal redundancy allocation.

Uncertainty quantification of a deep learning model for failure rate prediction of water distribution networks

Predicting the time-dependent pipe failure rate of the water distribution networks (WDNs) is important for planning its renewal budget but also challenging due to the complex factors involved. The recent development of machine learning techniques provides a novel approach for accurate failure rate prediction based on historical data. However, the inherited randomness and uncertainty of water pipe failures and machine learning algorithms are often ignored in the training and prediction process. This article develops an uncertainty quantification framework for deep learning-based WDN system-wide failure rate prediction. The framework integrates a probabilistic long short-term memory (LSTM) model with a Monte Carlo method. The historical climate data and WDN pipe maintenance data over the past 35 years for Cuyahoga County, USA, are used to illustrate this deep-learning model-based uncertainty quantification framework in this study. A statistical time series regression model, ARIMAX, is used as a comparison benchmark. The results show that the LSTM model outperforms the ARIMAX model in prediction accuracy in most years by considering the uncertainties. Besides, the uncertainty range of the LSTM model prediction is 50% of that of the ARIMAX model. The results also identified the major contributing factors to the uncertainties of LSTM machine learning model prediction. The proposed uncertainty framework features excellent extensibility and can be adapted to quantify uncertainties with other types of machine learning models.

Safety Lifetime Analysis Using Two-Phase Subset Simulation Combined with Kriging Model

The safety lifetime analysis under the required time-dependent failure probability (R-FP) constraint is essential to ensure structure service safety, but existing methods are unaffordable for solving the safety lifetime under a small R-FP constraint. For this issue, a novel method is proposed by combining a two-phase subset simulation with an adaptive kriging model (TP-SS-AK). The innovations of the TP-SS-AK include three aspects. First, it creatively transforms the safety lifetime under a small R-FP constraint into that under a large conditional R-FP constraint by subset simulation, which can greatly improve efficiency because the large conditional R-FP constraint only needs a small candidate sample pool for searching the safety lifetime. Second, by deducing the misclassification probability of the time-dependent structure state in the first phase, the first new learning strategy is established to train the kriging model of the time-dependent performance function and further determine a reasonable lifetime search interval efficiently. Third, by deriving the misclassification probability of the first failure instant, the second new learning strategy is constructed to efficiently update the kriging model in the second phase and further search the safety lifetime. Because the proposed TP-SS-AK combines the advantages of the two-phase subset simulation and two new learning strategies, it is superior to the existing methods in efficiency, and its superiority is verified by the examples.

Current status data with two competing risks and time-dependent missing failure types

In competing risks data, in practice, there may be lack of information or uncertainty about the true failure type, termed as ‘missing failure type’, for some subjects. We consider a general pattern of missing failure type in which we observe, if not the true failure type, a set of possible failure types containing the true one. In this work, we focus on both parametric and non-parametric estimation based on current status data with two competing risks and the above-mentioned missing failure type. Here, the missing probabilities are assumed to be time-dependent, that is, dependent on both failure and monitoring time points, in addition to being dependent on the true failure type. This makes the missing mechanism non-ignorable. We carry out maximum likelihood estimation and obtain the asymptotic properties of the estimators. Simulation studies are conducted to investigate the finite sample properties of the estimators. Finally, the methods are illustrated through a data set on hearing loss.

Fracture-pattern growth in the deep, chemically reactive subsurface

Arrays of natural opening-mode fractures show systematic patterns in size and spatial arrangement. The controls on these factors are enigmatic, but in many cases the depth of formation appears to be critical. Physical, potentially depth-dependent factors that could account for these variations include confining stress, fluid pressure, and strain rate; these factors are common inputs to existing fracture models. However, temperature-dependent chemical processes likely exert an equally important control on patterns, and such processes have not yet been rigorously incorporated into models of fracture formation. Here we present a spring-lattice model that simulates fracturing in extending sedimentary rock beds, while explicitly accounting for cementation during opening of fractures, and for rock failure via both elastic and time-dependent failure criteria. Results illustrate three distinct fracturing behaviors having documented natural analogs, which we here term fracture facies. “Exclusionary macrofracturing” occurs at shallow levels and produces large, widely spaced, uncemented fractures; “multi-scale fracturing” occurs at moderate depth and produces partially cemented fractures having a wide range of sizes and spacings; and “penetrative microfracturing” occurs at great depth and produces myriad narrow, sealed fractures that are closely and regularly spaced. The effect of depth is primarily to accelerate both dissolution and precipitation reactions via increased temperature and porewater salinity; the specific depth range of each fracture facies will vary by host-rock lithology, grain size, strain rate, and thermal history.

Time-dependent reliability analysis under random and interval uncertainties based on Kriging modeling and saddlepoint approximation

Time-dependent reliability analysis is capable of evaluating the reliability of the system over its full life cycle, which is of highly concerned for designers. In practice, some uncertainties cannot be simply described by a deterministic distribution due to lack of knowledge or information, and generally, their upper and lower bounds can be obtained. Thus, this paper develops a novel computational method for time-dependent reliability under random and interval uncertainties. The first step of the proposed method is to construct the Kriging surrogating model of the most probable point trajectory, with which the bound of the limit-state function can be efficiently recast as an equivalent Gaussian process. In this case, the computation of the bound of time-dependent failure probability (TDFP) is converted to a high-dimensional Gaussian integral problem. To solve this problem, the expansion optimal linear estimation, sparse-grid Gaussian quadrature technique and saddlepoint approximation method are integrated so that the above integral can be efficiently computed. Finally, four engineering examples are studied to demonstrate the effectiveness of the proposed method by comparison with previous methods. The results show this novel method can provide accurate and efficient time-dependent reliability analysis in terms of random and interval uncertainties.

Early-age strength and failure characteristics of 3D printable polymer concrete

The time-dependent rheological and early-age strength parameters of polymer concrete (PC) were investigated. PC flow decreased, and the static yield stress and thixotropy increased as the specimen aged, while the dynamic yield stress remained unchanged. The study found that the Herschel-Bulkley model can accurately describe the PC’s rheological behavior over time. The change in cohesion strength and internal friction angle over time was observed using uniaxial unconfined compression and direct shear tests of fresh PC. A time-dependent Mohr-Coulomb failure criterion was established. Temperature analysis was used to determine the gel time and explain the evolution of the time-dependent early-age strength of PC.

Damage Time-Dependent Characteristics and Failure Mechanism of Limestone with Different Mineral Compositions Under Acid Hydro-Chemical Solution Erosion

Taking two types of limestone with different mineral compositions as research objects, we conducted experiments on the mass, longitudinal wave velocity, ion concentration of solution, fracture morphology, and micromorphology of two types of limestone specimens eroded by acidic hydro-chemical solution. We investigated specimens’ damage time-dependent physical properties, dissolution, structure, discussing mineral composition effects and destruction. Results show that the mass loss rate of two types of limestone specimens eroded by acidic hydro-chemical solution increases with the increase of immersion time in a power function, and the longitudinal wave velocity decreases and then fluctuates in a small range. The specimen fracture degree increases with the increase of the immersion time, and the rock micro-structure changes from dense and uniform into loose and porous. The mass, the longitudinal wave velocity, the fracturing morphology, and damage to microscopic morphology of specimens increase with the increase of solution ion concentration. The acid rock reaction rate of limestone A is lower than that of limestone B, while the hydrolysis reaction rate is higher. The mass loss rate, the longitudinal wave velocity loss rate, the ion concentration, and the damage degree of the macro-and micro-structure of limestone A are higher than that of limestone B.

A simplified log creep model for describing triaxial creep of mélange rocks

This study aims to develop a simplified log creep model (LgCM) for predicting the triaxial three-stage creep behaviors of mélange rocks. The model was deduced from the creep deformation mechanism by considering the competition of strain rate hardening and damage during the steady and accelerating creep stages and was described by two simplified fractal functions. The model was then compared with the previous creep models on the uniaxial three-stage creep data of mortar, rock salt, and sandy shale, as well as the triaxial low-stress creep data of claystone. Afterward, the triaxial creep experimental results of the mélange rock samples were introduced to illustrate the process of calibrating the model in predicting the triaxial three-stage creep behaviors of mélange rocks. It was found the developed LgCM model showed good performance in predicting both the uniaxial and triaxial three-stage creep behaviors of rocks. The investigation reveals that the trend of the parameter β can indicate three thresholds of the hardening and damaging effects, and provide the equation to reproduce the creep behavior of the mélange rock. The work contributes to understanding the time-dependent failure of underground rock mass in mélange rock formations.

Time-dependent virtual crack model of rock with application to slope stability

Time-dependent deformation of rock involves the continuous to discontinuous deformation of joints, cracks, holes, and other defects under long-term stress conditions. Thus, it is of great significance to investigate the correlation between internal crack propagation in rock and external time-dependent damage and failure of rock to better understand the long-term stability of engineering rock mass. Here, we put forward a time-dependent virtual crack model to achieve the entire continuous-discontinuous time-dependent deformation progress of rock from the microscopic to the macroscopic scale. Simulated stress-strain curves for granite, including the mode of failure (axial splitting and Brazilian splitting), agree well with experimental data. We also used the virtual crack model to study the stability of one of the slopes of the Xiaowan Hydropower Station (China). The simulation results indicate that the weakly disturbed zone is the potential sliding surface, and that the deformation of the slope is similar to in-situ measurements. In the model, the vertical displacement decreased by 500 mm in 180 days and the axial strain rate decreased from 25 mm/d to zero, suggested that the slope is currently stable.

Resilience of wave energy farms using metocean dependent failure rates and repair operations

Emerging offshore renewable energy technologies are expected to become an important part of the future energy system, and reliability for these new technologies in different metocean scenarios must be guaranteed. This poses a challenge in extreme weather scenarios like storms, in particular for less mature technologies such as wave energy. Not only the offshore survivability must be controlled; the restoration after disruptive events and failures should be addressed and optimized. Offshore operations are costly and cannot be carried out if the weather is too harsh, and the resulting downtime after failures may be financially devastating for projects. In this paper, the resilience of large wave energy systems is studied with respect to wave conditions, metocean dependent failure rates, and weather windows available for offshore repair operations. A metocean- and time-dependent failure rate is derived based on a Weibull distribution, which is a novelty of the paper. The performance of the farm is assessed using the varying failure rates and metocean data at different offshore sites. Critical metocean thresholds for different offshore vessels are considered, and the resilience is quantified using relevant measures such as unavailability and expected energy not supplied. The resilience analysis is coupled to an economic assessment of the wave farm and different repair strategies. Our results show that the commonly used assumption of constant failure rates is seen to overestimate the annual energy production than when a more realistic varying failure rate is used. Two offshore sites are compared, and the availability is found to be higher at the calmer site. Most of the evaluated repair strategies cannot be considered to be economically justified, when compared to the cost of the energy not supplied.

Experimental investigation of time dependent dielectric breakdown and failure mechanism for Hf0.5Zr0.5O2 ferroelectric field effect transistors

The experimental investigation of the time dependent dielectric breakdown (TDDB) was performed to study the dielectric breakdown mechanisms for Hf0.5Zr0.5O2 (HZO) ferroelectric field-effect transistors (FeFETs) with the SiO2 interfacial layer (IL). Constant voltage stress method combined with the TCAD device simulation was used to study the breakdown mechanism of the TaN/HZO/SiO2/Si gate stack. Due to the much higher electric field in the IL, the breakdown of the SiO2 IL is the major reason for the breakdown of the gate stack. In addition, it is found that the IL breakdown will speed up the disappearance of the memory window (MW), leading to the malfunction of the FeFET as a storage device. Finally, the Weibull distributions of time-to-breakdown (TBD) and charge-to-breakdown (QBD) were characterized, revealing decent TDDB performance of the FeFETs. Careful design of IL thickness is necessary to balance the tradeoff between the device performance and the reliability lifetime of HZO FeFETs.

Deep learning methods for time-dependent reliability analysis of reservoir slopes in spatially variable soils

The Three Gorges Reservoir Area (TGRA) is one of the most important landslide-prone regions in China, and rational stability evaluation of reservoir slopes in it is of great significance to design mitigation measures and prevent landslide disasters. It is well recognized that seasonal rainfall and periodic reservoir water level fluctuation are the two major factors influencing the stability of reservoir slopes, and thus the reservoir slope stability may be varying with the external triggering factors. Although geotechnical reliability analysis offers a novel means to quantify slope stability in a probabilistic manner, the previous research focuses more on the time-independent slope reliability analysis, ignoring the effects of time-varying factors. How to evaluate the time-dependent reliability of reservoir slopes accurately and efficiently remains an open question. This study proposes deep learning (DL)-based time-dependent reliability analysis approach, and a practical case adapted from the Bazimen landslide in the TGRA is used for illustration. The predictive performances of the three DL algorithms, namely Convolutional Neural Network (CNN), Long short-term memory (LSTM), and Light gradient boosting machine (LightGBM) are systematically investigated. Results show that the proposed DL-based approach can reasonably portray the variation tendency of the Bazimen landslide time-dependent failure probability, which provides a promising way to rationally evaluate the time-dependent failure probability of reservoir slopes considering the spatial variability of soil properties. Among the three DL algorithms, the CNN performs the best in the Bazimen landslide example.

A novel single-loop meta-model importance sampling with adaptive Kriging for time-dependent failure probability function

A novel single-loop meta-model importance sampling with adaptive Kriging for time-dependent failure probability function

Investigating time-dependent behavior of rocks using kinematic-constraint-inspired non-ordinary state-based peridynamics

In this paper, the kinematic-constraint-inspired non-ordinary state-based peridynamic formulation (KC-NOSBPD) is proposed to investigate the time-dependent behaviors of rocks, which play a critical role in evaluating the long-term stability of the surrounding rocks around underground opening quantitatively. The peridynamic differential operator (PDDO) is introduced to KC-NOSBPD to improve the precision of bond-associated deformation gradient and to suppress the skin effect. On this basis, a unified creep-damage-rupture KC-NOSBPD model is proposed, in which the Burgers-creep viscoplastic constitutive law, a visco-elastoplastic constitutive model, is used to describe the rheological deformation of rocks. A unified damage-rupture framework based on a strain-based isotropic scalar damage variable is established to simulate the creep damage and subsequent creep rupture. The fidelity of this proposed model without damage is verified by comparing it with the benchmark solution for the creep response of a 2D rock specimen subjected to various load histories. Further, the proposed model is employed to simulate uniaxial and triaxial compression creep tests. These present numerical results agree well with the experimental observation, implying that the proposed model successfully captures the accelerating creep behavior of rocks. Moreover, the proposed model performs the time-dependent deformation and failure analysis for the surrounding rocks around the tunnel.

A new acquisition function combined with subset simulation for active learning of small and time-dependent failure probability

A new acquisition function combined with subset simulation for active learning of small and time-dependent failure probability