Reliability, diagnostics and prognostics

 

Maintenance efficiency, and prognostics and health management (PHM) in railway applications

This work was a collaborative research project with Alstom France. We established a method for estimating the effect of corrective and preventive maintenance on the health state of point machines, and to decide when the machines' health indices (constructed using artificial neural network technique) should be reinitialized. Point machines are used for operation of railway turnouts. We are also using the condition-monitoring data to estimate the remaining useful life of these machines in the presence of imperfect maintenance.

 

Modelling failure process and quantifying the effects of multiple types of preventive maintenance for a repairable system

This work was a collaborative research project with Vale Canada Limited. We modelled the failure process of a repairable system, subject to corrective and multiple types of preventive maintenance. We estimated the parameters of the failure process as well as the efficiency of corrective maintenance and various types of preventive maintenance. The proposed methods were applied to a case study of trucks used in a mining site. The paper "Modeling failure and maintenance effects of a system subject to multiple preventive maintenance types", written based on this work received the second-best student paper award of the 2016 Annual Reliability and Maintainability Symposium (RAMS).

 

Reliability and trend analysis of failure data

This work was a collaborative research project with Toronto General hospital. We proposed a method for reliability analysis of complex medical devices with censored and missing failure data. As a case study, we conducted the reliability analysis of a general infusion pump. The paper "Reliability Analysis of Maintenance Data for Complex Medical Devices", written based on this work received the best student paper award of the American College of Clinical Engineering in 2010, and was selected by the Journal of Quality and Reliability Engineering International as the most cited paper in 2011.

Moreover, we proposed new methods for trend analysis of censored failure data for a system whose failure process follows a non-homogeneous Poisson process (NHPP) with a power law intensity function. A novel feature of these methods is that they consider dependent failure histories which are censored by inspection intervals. We used the likelihood ratio test to check for trends in the failure data with censoring. The proposed method was applied for several components of general infusion pump. The paper "Trend Analysis of the Power Law Process with Censored Data" written based on this work received the best student paper award of the Annual Reliability and Maintainability Symposium (RAMS) in 2011.

 

A Bayesian network approach to improve fleet availability analysis

As an important aspect of a fleet management, availability analysis requires modeling uncertainties associated with its two main elements, i.e. reliability and maintainability. We aim to improve prediction and uncertainty analysis of reliability (or failure rate), maintainability (or repair rate), and hence the availability of a fleet of assets, by focusing on common causal factors and rare events, which might impact one or both of the failure and repair rates. To acheive this aim, we proposed a Bayesian network approach with learning as well as analytical features, which allows fleet managers to not just predict the most likely availability of their fleet, but analyze the impact of any observed shift in the influential factors, update prior beliefs about the predictions, and discover the root causes of troubling failure and repair rates.

Connect with us
Contact Us
Mail Contact

RRMR Lab
Dept. of Mechanical & Industrial Engineering
350 Victoria Street, Toronto, Ontario, M5B 2K3, Canada

sharareh@ryerson.ca

Office Contact

RRMR Lab
338A Eric Palin Hall (EPH)
87 Gerrard Street East, Toronto, Ontario, M5B 1G6, Canada

(416) 979-5000 ext 7693