Maintenance policy for a system with a weighted linear combination of degradation processes

Shaomin Wu and Inma T. Castro

arXiv: 2401.12263v1 · PDF · 22 January 2024
Categories: eess.SY, cs.SY, math.PR
Licence: CC BY 4.0 The site shows the text.

Abstract

This paper develops maintenance policies for a system under condition monitoring. We assume that a number of defects may develop and the degradation process of each defect follows a gamma process, respectively. The system is inspected periodically and maintenance actions are performed on the defects present in the system. The effectiveness of the maintenance is assumed imperfect and it is modelled using a geometric process. By performing these maintenance actions, different costs are incurred depending on the type and the degradation levels of the defects. Furthermore, once a linear combination of the degradation processes exceeds a pre-specified threshold, the system needs a special maintenance and an extra cost is imposed. The system is renewed after several preventive maintenance activities have been performed. The main concern of this paper is to optimise the time between renewals and the number of renewals. Numerical examples are given to illustrate the results derived in the paper.

Items 4

Example

Example 1

BG1J

We consider a system subject to three degradation processes \left\{X_1(t), t \geq 0\right\}, \left\{X_2(t), t \geq 0\right\} and \left\{X_3(t), t \geq 0\right\}, respectively. These degradation processes start at random times according to a homogeneous Poisson process with parameter \lambda=1. The degradation processes develop according to non-homogeneous gamma process with parameters \alpha_1=1.1, \beta_1=1.1, \alpha_2=1.2, \beta_2=1.2 and \alpha_3=1.3, \beta_3=1.3. Figure ? shows these degradation processes and the process Y(t)=\sum_{j=1}^{3} b_j X_j(t) with b_1=1, b_2=0.8, and b_3=0.9.

The paper has a figure here, which this page does not show.

Lemma

Lemma 1

3PXT

The conditional probability f_{U(t)|Y(t)}(y,z) is given by

\begin{aligned} f_{U(t)|Y(t)}(y,u) &=\frac{1}{4\pi^2 g_{Y(t)}(y)} \int_{-\infty}^{\infty} \int_{-\infty}^{\infty} \left( \prod_{k=1}^n (1-i(b_k t_1 +c_k t_2) \beta_k)^{-\alpha_k(t)} \right) e^{-it_1 y-it_2 u} \d t_1\d t_2. \end{aligned}

Definition

Definition 1

F3WQ

[Lam] Given a sequence of non-negative random variables \{X_j,j=1,2, \dots\}, if they are independent and the cdf of X_j is given by F(a^{j-1}x) for j=1,2, \dots, where a is a positive constant, then \{X_j,j=1,2,\cdots\} is called a geometric process (GP).

Lemma

Lemma 2

3TZS

If \alpha_k(T) is convex in T for all k with \alpha_k(0)=0 and \alpha_k'(0)<\infty, then

\mathrm{CV}(N,T) is increasing in T for fixed N, and
\mathrm{CV}(N,T) is increasing in N for fixed T.

Source and licence

The text on this page is from arXiv:2401.12263v1 by Shaomin Wu and Inma T. Castro, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2401/2401.12263.jsonl.