Mining operations in the high mountains increasingly rely on water being pumped up from sea level for their process needs. Future process plants may be able to use sea water in which case raw sea water with a higher density than desalinated water will be pumped up. Pumping large volumes of water is usually done by multiple, constant speed, high pressure, multi-stage, centrifugal pumps operating in parallel.
Such pumps operate from a common suction line and discharge into a common manifold. As a result intake and discharge pressures as well as the flow rates are correlated, which complicates the evaluation of individual pump’s efficiency. Such correlations will also impact the operating point of each individual pump and will introduce dependencies between individual pump performances.
Correct treatment of such dependencies and correlations is required for predictive maintenance and efficiency improvement. In a model study of a five-pump pump station, we use simple measures of performance (head, power and efficiency) in non-dimensional form based on non-dimensional figures of merit for an individual model pump. Parallel operation of five pumps, three conforming to specifications, one with a reduced static head, and one with increased internal friction, can be modelled easily using a simple empirical model.
In order to identify underperforming pumps in parallel operation of a number of pumps, it is required to be able to compute efficiency on a single pump basis. When single pump operation in a station is rare, such single pump data can only be obtained by simultaneous measurement of discharge head and discharge flow for each pump individually.
To compliment this work it is planned to directly measure and track individual pump efficiency by installing sonar flow meters on individual pump discharges on an existing pipeline, and to report on this work in the near future.