Zarzycki, Edmund (2020) Innovation report : a methodology for estimating gear pump wear-out reliability using pump pressure ripple and an extremely small sample size - the case study of a heavy-duty diesel engine lubrication gear pump. EngD thesis, University of Warwick.

Preview

PDF WRAP_Theses_Zarzycki_2020.pdf - Submitted Version - Requires a PDF viewer. Download (5Mb) | Preview

Request Changes to record.

Abstract

Design for Reliability (DfR) encourages testing products early in the New Product Development (NPD) process to identify and resolve weaknesses quickly. An organisation can then track reliability growth and intervene to ensure the changes in product robustness are in line with a timely release to market. However, for products with long life spans (such as a Heavy-Duty engine (HDE) lubrication gear pump), the evaluation of reliability with an extremely small number of prototype samples is problematic. Budget constraints, product size, and test facilities can limit the possibilities of accurately assessing the initial reliability forming a test planning paradox. The research in this thesis proposes an innovate methodology to minimise this test planning paradox, specific to a gear pump.

The method uses step-stress accelerated degradation testing and Bayesian inference to estimate degradation parameters using only a sample size of two. Post-testing, numerical simulation is used to build a degradation model with larger sample sizes and produce a survival distribution at the quantile of interest. Increasing pump outlet pressure above normal usage accelerates the pump wear and pressure ripple measurements are used to monitor the performance degradation. On inspection, the pumps exhibit erosion on the housing and micro pitting of the gear flanks. The innovative use of a Maximal Overlap Discrete Wavelet Transforms (MODWT) with an Autoregressive Moving Average (ARMA 2,1) extracts a feature from the pressure ripple that provides a stochastic, linear and non-monotone degradation path that is appropriately modelled using a Brownian Motion simulation model. Regression analysis provides a drift and diffusion covariate functional relationship to pump outlet pressure. Given the stress-varying environment of an HDE, Monte Carlo simulations overcome the complexity of replicating vehicle drive cycle and produces a credible reliability estimate validated against a similarly designed high mileage pump.

The application of this original methodology offers the opportunity to minimise the test planning paradox and satisfies populating the reliability growth chart. It is foreseen the method can be adopted for a wide range of positive displacement pumps where is it possible to measure pressure ripple.

Item Type:	Thesis or Dissertation (EngD)
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) T Technology > TJ Mechanical engineering and machinery T Technology > TS Manufactures
Library of Congress Subject Headings (LCSH):	Reliability (Engineering), Service life (Engineering), Gearing, Gear pumps, Product design, Product life cycle
Official Date:	April 2020
Dates:	Date Event April 2020 UNSPECIFIED
Institution:	University of Warwick
Theses Department:	Warwick Manufacturing Group
Thesis Type:	EngD
Publication Status:	Unpublished
Supervisor(s)/Advisor:	Marshall, Jane (Teaching fellow) ; Jones, Jeffrey Alun, 1962-
Sponsors:	Concentric AB ; Engineering and Physical Sciences Research Council ; Warwick Manufacturing Group
Format of File:	pdf
Extent:	xviii, 152, ix leaves : illustrations
Language:	eng

Request changes or add full text files to a record

Repository staff actions (login required)

View Item

Downloads