Reliability Modeling and Analysis of Load-Sharing Systems With Continuously Degrading Components
read more
Citations
Reliability of response region: A novel mechanism in visual tracking by edge computing for IIoT environments
Reliability and availability analysis of stochastic degradation systems based on bivariate Wiener processes
Remaining useful life prediction and predictive maintenance strategies for multi-state manufacturing systems considering functional dependence
Rebooting data-driven soft-sensors in process industries: A review of kernel methods
Reliability modeling for degradation-shock dependence systems with multiple species of shocks
References
Statistical Methods for Reliability Data
Statistical Methods for Reliability Data
Optimal Reliability Modeling: Principles and Applications
The Inverse Gaussian Process as a Degradation Model
Related Papers (5)
Reliability modeling and preventive maintenance of load-sharing systems with degrading components
Stochastic modelling and analysis of degradation for highly reliable products
A load sharing system reliability model with managed component degradation
Frequently Asked Questions (20)
Q2. What are the future works in this paper?
Random and cumulative loads as well as various load-sharing rules are of interest to study. The data modeling framework can be extended to deal with test data involving censoring. Additionally, as mentioned before, various management problems related to system reliability can be formulated for such systems, and to solve the optimization problem robustly by considering parameter uncertainty should be very useful for decision makers in various industries.
Q3. What methods are used to evaluate the variability of parameter estimates?
Regarding the estimation uncertainty, the authors use both bootstrapping and large-sample approximation methods to evaluate the variability of parameter estimates.
Q4. What is the common technique used to remove organic matter and nutrients in the wastewater plant?
Activated sludge process (ASP) is the most commonly used technique to remove organic matter and nutrients (mainlyTR-2017-587R23nitrogen and phosphorus) in the wastewater plant [24].
Q5. How can the authors evaluate the likelihood contribution of XXiiii?
By utilizing the independence of non-overlapping increments of Wiener process, the authors can evaluate the likelihood contribution of Δ𝑿𝑿𝑖𝑖𝑖𝑖 conveniently by computing the product of likelihoods contributed by Δ𝑋𝑋𝑖𝑖𝑖𝑖𝑖𝑖 for all 𝑘𝑘.
Q6. What is the common use of the log-linear link function in degradation modeling?
In reliability analysis, the log-linear link functions are commonly used in degradation modeling and accelerated tests [25], [31].
Q7. What method is used to describe the MLEs as random variables?
To investigate the variability of parameter estimates, the authors use both the bootstrapping (BS) method and large-sample approximation (LS) to describe the MLEs as random variables.
Q8. How do the authors obtain the MLEs of unknown parameters?
To obtain the MLEs of the unknown parameters, the authors need to rewrite the likelihood function into the log-likelihood function as shown in (11).
Q9. What is the probability contribution of a Wiener process?
Since the FPT of a Wiener process follows the inverse Gaussian distribution, by conditioning on observing the last degradation measure 𝑋𝑋𝑖𝑖𝑖𝑖𝐿𝐿𝑖𝑖𝑗𝑗 = 𝑥𝑥, the FPT beyond the last inspection 𝑀𝑀𝑖𝑖𝑖𝑖~ℐ𝒢𝒢�(𝐿𝐿 − 𝑥𝑥) 𝜂𝜂𝑖𝑖⁄ , (𝐿𝐿 − 𝑥𝑥)2/𝜎𝜎2 �, and the density function is given byIf component (𝑖𝑖, 𝑗𝑗) fails before any degradation measure istaken, i.e., 𝐿𝐿𝑖𝑖𝑖𝑖 = 0 , the likelihood is merely contributed by 𝑀𝑀𝑖𝑖𝑖𝑖 = 𝑌𝑌𝑖𝑖,𝑖𝑖.
Q10. What is the estimated degradation drift under the cases where the number of working components is 3?
The estimated degradation drift under the cases where the number of working components is 3, 2 and 1 are 0.0202, 0.0471 and 0.0774, respectively.
Q11. What is the simplest way to simulate the failure threshold?
By assuming the failure threshold to be 0.4, the authors use the following parameter setting to simulate the data: 𝜽𝜽 = (𝛽𝛽0, 𝛽𝛽1, 𝜎𝜎)′ = (0.1, 1, 0.25)′ .
Q12. How can the authors solve the optimization problem for load-sharing systems?
as mentioned before, various management problems related to system reliability can be formulated for such systems, and to solve the optimization problem robustly by considering parameter uncertainty should be very useful for decision makers in various industries.
Q13. What is the effect of smaller inspection intervals on the estimation accuracy?
more test systems enhance the estimation accuracy by reducing the uncertainty significantly, while smaller inspection intervals help more to reduce the bias of reliability inferences.
Q14. How do the authors compute the initial estimates of 0 and 1?
As with the link function defined in (2), wecompute the initial estimates of 𝛽𝛽0 and 𝛽𝛽1 by fitting a linear regression model as follows:where 𝟏𝟏𝐽𝐽 is the 𝐽𝐽 -dimensional column vector with all elements equal to 1.
Q15. How can the reliability function be evaluated?
In Appendix B, the authors use an approximation-based simulation method to generate samples of the failure time, then the reliability function can be evaluated non-parametrically via simulated life data.
Q16. What is the simplest way to model the degradation of a system?
Under the assumption that the components in the systems are of the same type and the workload on each component is equal, it is reasonable to imply that each surviving component is suffering from an equal damage that leads to degradation growth at an arbitrary time.
Q17. What is the way to estimate the MLE of a sample?
Under comparatively large samples, a multivariate normal (MVN) distribution provides a satisfactory approximation for the joint distribution of the MLEs.
Q18. What is the simplest way to model the degradation of components in a loadsharing system?
As stated in Kong and Ye [7], the authors can resort to establishing a link function to connect the workload and degradation model parameters for degrading components in loadsharing systems.
Q19. What is the effect of NN on the estimates of 0 and 2?
This implies that, regarding the accuracy of the estimates of 𝛽𝛽0 and 𝛽𝛽2, the influence of Δ𝜏𝜏 gets larger when Δ𝜏𝜏 is relatively large, whereas 𝑁𝑁 always puts a significant effectTR-2017-587R29on all the three estimates.
Q20. What is the result of the proposed guessing method?
The result implies that the proposed guessing method accelerates the estimation procedure via providing initial points close to the MLEs.