Citation

Jacobs JA, Mathews MJ, Kleingeld M. J. Fail. Anal. Prev. 2018; 18(4): 927-938.

Copyright

(Copyright © 2018, ASM International, Publisher Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s11668-018-0488-3

PMID

unavailable

Abstract

The de-watering reticulation system of a deep level mine requires a network of pumps to pump water from underground to the surface. These de-watering pumps are exposed to extreme operating conditions, which can result in unacceptably short service lives. It is thus important for mining personnel to use the available tools to improve system operations and maintenance procedures. Current maintenance strategies involve reactive and preventive models, which can lead to failure creep. In order to implement proactive strategies, forecasting models are required. One of such models can be Weibull models, which has proven to be successful for many product failure mechanisms. The Weibull models have been used extensively in the reliability-engineering environment, but have not yet been implemented on a de-watering pump operating in extreme conditions on a deep level mine. A new practical method for predicting pump failure using various Weibull distribution functions was thus developed. The methodology was validated on four pumps, and it was found that the failures of new pumps were successfully predicted based on the Weibull analysis of previously failed pumps. It was also found that a quadratic relationship exists between the characteristic life of a de-watering pump and the operating depth underground. This is significant because all of the factors affecting the service life of a de-watering pump are quantified regarding only the operating depth and the characteristic life.

Language: en

Keywords

Pump failure prediction; Reliability engineering; Weibull analysis