Proportional hazards modelling for the analysis of reliability field data

Walker, E., 1989. Proportional hazards modelling for the analysis of reliability field data. MPhil, Nottingham Trent University.

10183544.pdf - Published version

Download (28MB) | Preview


This thesis reviews the theory of the Proportional Hazards Model for use in analysing reliability field data. Since, much work in reliability is concerned with providing statistical models for the lifelengths or interfailure times of equipment, and hence provide predictions for their performance in the field, reliability field data might be considered most appropriate on which to base ones model.

Proportional Hazards Modelling has great potential in the context of analysing reliability field data, since it assesses significant affects on the lifelengths or interfailure times of equipment due to both internal and external covariates.

Three applications of PHM to the analysis of reliability field data, are presented. The applications illustrate that there is no unique form for reliability data. The form of the data is influenced by the system structure, the system's deployment, repair and maintenance regimes, and the data collection procedures. It is for these reasons that Proportional Hazards Modelling can not be applied in a black box fashion. This thesis demonstrates the questions that require resolving before fitting a model and the difficulties involved in extracting data from poor collection processes.

Four commonly used graphical diagnostic procedures, which assess the fit of the model and indicate whether or not the model's assumptions are violated are highlighted. These procedures are discussed in detail and extended. The additions to the forms applied previously in the literature ease the visual inspection of the plots.

An exploratory data analysis approach to Proportional Hazards Modelling of reliability field data is advised.

Item Type: Thesis
Creators: Walker, E.
Date: 1989
ISBN: 9781369317084
Divisions: Schools > School of Science and Technology
Record created by: Jeremy Silvester
Date Added: 01 Oct 2020 09:46
Last Modified: 20 Sep 2023 09:52

Actions (login required)

Edit View Edit View


Views per month over past year


Downloads per month over past year