Spurious exponentiality observed when incorrectly fitting a distribution to nonstationary data

Ascher, Harold; Hansen, Christian K.

doi:10.1109/24.756089

Cited by 44 publications

(27 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Asher and Hansen [2] point out the dangers of incorrectly identifying a non-exponential distribution as exponential. They suggest validating all data, understanding the nature of the data, performing a trend test after visually examining the data and, if there is no evidence of a trend , conducting an exponential goodness-of-fit test on the data before pronouncing it as 'exponential'.…”

Section: Introductionmentioning

confidence: 99%

On misinterpretation of the asymptotic property of system time‐between‐failures

Chambal

Keats

2001

Quality & Reliability Eng

View full text Add to dashboard Cite

SUMMARYIf a system, not necessarily one in series, is composed of components having various time-to-failure distributions and components are replaced good-as-new as they fail, then the system time-between-failure distribution tends toward the exponential. Many practicing reliability engineers, incorrectly invoking this property, model their systems with an exponential time-to-failure. We show, under two conditions, using a hypothetical fleet of vehicles, the severity of this error. Modeling time-to-failure as exponential results in gross over-sparing and high unavailability costs as well.

show abstract

Section: Introductionmentioning

confidence: 99%

On misinterpretation of the asymptotic property of system time‐between‐failures

Chambal

Keats

2001

Quality & Reliability Eng

View full text Add to dashboard Cite

show abstract

“…The procedure followed in determining the repairable system model is based on that suggested in [1] and [10]. The first step involves a graphical technique of plotting cumulative number of failures versus cumulative operating time .…”

Section: Methodsmentioning

confidence: 99%

Modeling failure data of overhead distribution systems

Stillman

2000

IEEE Trans. Power Delivery

View full text Add to dashboard Cite

An overhead power line can be regarded as a collection of components each of which contribute in a particular way to its reliability. A overhead distribution system is a collection of such lines. A repairable system is a system which after failing, can be restored to an operational, but not a new state by repairing or replacing one or more components. Thus work treats the distribution system as a holistic entity with the aim of measuring the pattern of failures as an indication of reliability growth or deterioration. Comparison is made using case studies of an urban and a rural system based on trend analysis, the Laplace statistic and confidence limits. It is argued that although the pattern of failure of a distribution system may emulate a stationary point process, this can be misleading in the assessment of reliability and the impact of maintenance policies.

show abstract

“…The in-depth use of this conceptual basis for practical cases can be found in Ascher and Hansen (1998) and Saldanha et al (2001).…”

Section: Introductionmentioning

confidence: 99%

Rate of occurrence of failures based on a nonhomogeneous Poisson process: an ozone analyzer case study

Xavier¹,

Azevedo

Sousa

et al. 2012

Environ Monit Assess

View full text Add to dashboard Cite

Atmospheric pollutant monitoring constitutes a primordial activity in public policies concerning air quality. In São Paulo State, Brazil, the São Paulo State Environment Company (CETESB) maintains an automatic network which continuously monitors CO, SO(2), NO(x), O(3), and particulate matter concentrations in the air. The monitoring process accuracy is a fundamental condition for the actions to be taken by CETESB. As one of the support systems, a preventive maintenance program for the different analyzers used is part of the data quality strategy. Knowledge of the behavior of analyzer failure times could help optimize the program. To achieve this goal, the failure times of an ozone analyzer-considered a repairable system-were modeled by means of the nonhomogeneous Poisson process. The rate of occurrence of failures (ROCOF) was estimated for the intervals 0-70,800 h and 0-88,320 h, in which six and seven failures were observed, respectively. The results showed that the ROCOF estimate is influenced by the choice of the observation period, t(0) = 70,800 h and t(7) = 88,320 h in the cases analyzed. Identification of preventive maintenance actions, mainly when parts replacement occurs in the last interval of observation, is highlighted, justifying the alteration in the behavior of the inter-arrival times. The performance of a follow-up on each analyzer is recommended in order to record the impact of the performed preventive maintenance program on the enhancement of its useful life.

show abstract

Spurious exponentiality observed when incorrectly fitting a distribution to nonstationary data

Cited by 44 publications

References 11 publications

On misinterpretation of the asymptotic property of system time‐between‐failures

On misinterpretation of the asymptotic property of system time‐between‐failures

Modeling failure data of overhead distribution systems

Rate of occurrence of failures based on a nonhomogeneous Poisson process: an ozone analyzer case study

Contact Info

Product

Resources

About