skip to main content

Attention:

The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Thursday, February 13 until 2:00 AM ET on Friday, February 14 due to maintenance. We apologize for the inconvenience.


Title: Multi-dimensional Lévy Processes for Multiple Dependent Degradation Processes in Lifetime Analysis
The analysis of multiple dependent degradation processes is a challenging research work in the reliability field, especially for complex degradation with random jumps. To integrally handle the jump uncertainties in degradation and the dependence among degradation processes, we construct multi-dimensional Lévy processes to describe multiple dependent degradation processes in engineering systems. The evolution of each degradation process can be modeled by a one-dimensional Lévy subordinator with a marginal Lévy measure, and the dependence among all dimensions can be described by Lévy copulas and the associated multiple-dimensional Lévy measure. This Lévy measure is obtained from all its one-dimensional marginal Lévy measures and the Lévy copula. We develop the Fokker-Planck equations to describe the probability density in stochastic systems. The Laplace transforms of both reliability function and lifetime moments are derived. Numerical examples are used to demonstrate our models in lifetime analysis.  more » « less
Award ID(s):
1728321
PAR ID:
10064652
Author(s) / Creator(s):
; ;
Date Published:
Journal Name:
Proceedings of Industrial and Systems Engineering Research Conference
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract

    Many engineering systems experience complex degradation processes along with random shocks that often pose a risk to the safe and reliable operations of the system. The analysis of dependence among multiple degradation processes has been a challenging task in the field of reliability engineering. To study dependence behaviors and jump uncertainties among degradation processes, one of the effective approaches is to model these degradation processes using multidimensional Lévy subordinators. However, a critical obstacle arises in determining the convergent mathematical form of the reliability function under Lévy measures and distributions. To obtain a closed form of reliability function, this study investigates the Lévy measure and the characteristic function of multiple dependent degradation processes. Each degradation process is modeled by a one‐dimensional Lévy subordinator utilizing the marginal Lévy measure and characteristic function. The dependence among all dimensions is described by the Lévy copula and the associated multidimensional Lévy measure. For a two‐dimensional case, we derive the reliability function and probability density function (PDF) from the characteristic function and the inverse Fourier transform. Numerical examples are provided to demonstrate the proposed models for reliability and lifetime analysis of multidimensional degradation processes in engineering systems.

     
    more » « less
  2. Lévy processes are useful tools for analysis and modeling of jump‐diffusion processes. Such processes are commonly used in the financial and physical sciences. One approach to building new Lévy processes is through subordination, or a random time change. In this work, we discuss and examine a type of multiply subordinated Lévy process model that we term a deep variance gamma (DVG) process, including estimation and inspection methods for selecting the appropriate level of subordination given data. We perform an extensive simulation study to identify situations in which different subordination depths are identifiable and provide a rigorous theoretical result detailing the behavior of a DVG process as the levels of subordination tend to infinity. We test the model and estimation approach on a data set of intraday 1‐min cryptocurrency returns and show that our approach outperforms other state‐of‐the‐art subordinated Lévy process models.

     
    more » « less
  3. This paper investigates degradation modeling under dynamic conditions and its applications. Both univariate and multiple competing degradation processes are considered with individual degradation paths being described by Wiener processes. Parametric and non-parametric approaches are used to capture the effect of environmental conditions on process parameters. For competing degradation processes, we obtain the probability that a particular process reaches a pre-defined threshold, before other processes, over future time intervals. In particular, we consider the potential statistical dependence among the latent remaining lifetimes of multiple degradation processes due to unobserved future environmental factors. Two case studies, aircraft piston pump wear and US highway performance deterioration, are presented. Comprehensive comparison studies are also performed to generate some critical insights on the proposed approach. Data have been made available on GitHub. 
    more » « less
  4. Normalizing flows—a popular class of deep generative models—often fail to represent extreme phenomena observed in real-world processes. In particular, existing normalizing flow architectures struggle to model multivariate extremes, characterized by heavy-tailed marginal distributions and asymmetric tail dependence among variables. In light of this shortcoming, we propose COMET (COpula Multivariate ExTreme) Flows, which decompose the process of modeling a joint distribution into two parts: (i) modeling its marginal distributions, and (ii) modeling its copula distribution. COMET Flows capture heavy-tailed marginal distributions by combining a parametric tail belief at extreme quantiles of the marginals with an empirical kernel density function at mid-quantiles. In addition, COMET Flows capture asymmetric tail dependence among multivariate extremes by viewing such dependence as inducing a low-dimensional manifold structure in feature space. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of COMET flows in capturing both heavy-tailed marginals and asymmetric tail dependence compared to other state-of-the-art baseline architectures. All code is available at https://github.com/andrewmcdonald27/COMETFlows.

     
    more » « less
  5. Consider non-homogeneous Markov-dependent components in an m-consecutive- k-out-of- n:F (G) system with sparse [Formula: see text], which consists of [Formula: see text] linearly ordered components. Two failed components are consecutive with sparse [Formula: see text] if and if there are at most [Formula: see text] working components between the two failed components, and the m-consecutive- k-out-of- n:F system with sparse [Formula: see text] fails if and if there exist at least [Formula: see text] non-overlapping runs of [Formula: see text] consecutive failed components with sparse [Formula: see text] for [Formula: see text]. We use conditional probability generating function method to derive uniform closed-form formulas for system reliability, marginal reliability importance measure, and joint reliability importance measure for such the F system and the corresponding G system. We present numerical examples to demonstrate the use of the formulas. Along with the work in this article, we summarize the work on consecutive- k systems of Markov-dependent components in terms of system reliability, marginal reliability importance, and joint reliability importance.

     
    more » « less