Surface and deep well pump monitoring via efficiency metrics

Presents the results of some collaborative work carried between the Control Systems Centre (CSC) at Sunderland Polytechnic and the Sunderland and South Shields Water Company (SSSWCo.). The SSSWCo. supplies a large population in the North West of England using three types of raw water sources which include bore holes, reservoirs and river abstraction. Water abstracted from bore holes requires little treatment other than disinfection but incurs high a pumping cost, while the other sources require much more complex treatment at specialised treatment works where chemical cost represent the larger part of the operation costs. The company owns and operates three treatment works at Wearhead, Mosswood and Lumley. Wearhead and Mosswood are reservoir sources while the Lumley plant takes water from the river Wear. The results of experiments carried out to establish the feasibility of using existing instrumentation on both surface and deep well pumps to determine their performance metrics is presented here. The fundamental idea behind the approach relies on the use of linear regression relationships and recursive estimation algorithms to highlight unmeasured characteristics in the metering data