Determining average volumes of residential water consumption in flow rate interval categories

Abstract

Water utilities install and maintain domestic water meters but so far in some utilities customer meters are replaced on a run-to-fail basis. In some countries there is not yet a mandatory requirement to replace water meters. Most Australian utilities are required to have meter replacement programs, but there is still limited understanding of meter accuracy performance with in-situ age, particularly when considering its starting or minimum registration level (Qs). As a consequence, a fleet of water meters can become less accurate (i.e. under or over register flow) and also the minimum registration of the meter can increase, thereby resulting in higher volumes of unaccounted for water. To remedy this situation, an understanding of water consumption patterns, particularly at the lower flow rate levels, needs to be investigated in order to understand the effect that an aged fleet of meters with higher Qs will have on unaccounted for water for the water business; such unaccounted water can account for a significant sum of lost revenue so may be case to replace meters earlier. To obtain the necessary data for this flow data categorisation task, standard council residential water meters needed to be replaced with Actaris CTS-5 water meters (new meters as replaced in 2010-11). These smart meters measure flow to a resolution of 72 pulses/L or a pulse every 0.0138L. The smart meters were connected to Aegis Data Cell series R-CZ21002 data loggers. The loggers were programmed to record pulse counts at five second intervals and recorded data are sent on a weekly basis via GSM/GPRS network to an assigned server system. Each logger was wired to a meter, labelled and activated prior to installation. According to metrological characteristic of a new Actaris CTS-5 meter, the Qs (i.e., minimum flow rate) is 2L/hour (or 0.033L/min). Theoretically, water with a flow rate less than the Qs flow rate of a meter cannot be measured. The aim of this technical report is to understand the average volume of water used by a typical residential household in each water flow rate category (i.e. < 5 L/hr; 6-10 L/hr). This data will assist with the modelling of unaccounted for water in a typical fleet of aged meters in Melbourne, and elsewhere. High resolution water data obtained from a fleet of 400+ smart meters in South East Queensland (SEQ) and Melbourne were utilised for this problem.