Assessment of apparent losses due to meter inaccuracy – a comparative approach

Abstract

The empirical method for the determination of apparent water losses, using the assessment of consumption patterns and the laboratory testing of water meters, is compared against alternative methods of comparative billing and meter change analyses for one of the largest water utilities in South Africa. Using the empirical method, apparent losses are estimated to have an average value of 12% of the billed volume with a range from 9.4% to 14.6% that is dependent on meter size ratios. This overlaps with the estimates, of 8%−10% for a utility with direct feed good quality water but high meter age (> 10 yr) and low accuracy, currently proposed in some studies and used by industry. The estimate from the comparative billing analysis method is 14% and it is sensitive to how the data is processed and analysed. The meter change method yielded an estimate of 4.7% for only a subset of the data. Both results of the alternative methods are in line with previous studies, with the comparative billing analysis performing better, but requiring further refinement for better accuracy and repeatability. The empirical method remains the gold standard in assessing apparent water losses but is undoubtedly very laborious, expensive and out of reach of the budget of many utilities. The development and validation of alternative methodologies therefore holds great promise but these are substantially reliant on comprehensive meter information and credible billing datasets which are rarely available in most utilities in Africa.

Keywords: apparent water losses, water metering, metering errors, water loss management, non-revenue water (NRW)