Estimation of Water Balance and Water Losses in Water Utilities – Experiences from the Belgrade Waterworks - page 4


Places where real losses from distribution pipes occur can be divided in two categories, depending on a flow magnitude (WSA/WCA, 1994):

  • A burst is place where a significant quantity of water flows out of the damaged or cracked pipe where a flow rate of leakage is over 500 l/h under pressure of 50 m;
  • Leaks (at valves, connections, hydrants, etc.) are place where water flows out of the pipe with a flow rate of less than 500 l/h under pressure of 50 m.

Measurements of night flow revealed that average normal water use in households by night amounts to 1.7 l/connection/h in the UK, or 0.6 l/capita/h in Germany (WSA/WCA, 1994). It can be assumed that night water consumption in households is mainly independent from the pressure in the distribution network, since water is used predominantly by devices which operate on a fixed water volume principle (flushing of toilets, washing machines, etc.).

Water flow rate through orifices or cracks depends on the size of opening and water pressure. Theoretically, flow through openings (leakage) should be proportional to square root of pressure. However, numerous field measurements revealed that the relationship between leakage and pressure is represented in the following equation (Thornton, 2005):


where: Q1 is leakage under water pressure p1 inside pipe, Qo – is leakage under water pressure po inside the pipe. N1 is determined by field measurements of minimal night flows and real losses in the DMA. Investigations reveal that N1 is usually in range from 0.5 ro 1.5, but can go up to 2.5 (Thornton, 2005). N1 depends on the pipe material, operating conditions, type of pipe damages and other factors.

Numerous field investigations on different DMAs in the UK were collected and analyzed (WSA/WCA, 1994). Investigations include measurements of minimal night flows and estimation of net night flow under different pressure in DMAs. Statistical analyses of data on pressures and net night flows were performed and a dimensionless parameter Leakage Index (LI) was introduced which represents ratio of net night flow at pressure p and net night pressure at referent pressure. Further statistical analyses (WSA/WCA, 1994) led to establishment of the following relationship between LI and the average pressure in studied area NP:



LI enables estimation of net night flow under different average pressures in DMA and estimation of total daily volume of leakage by using measured net night flows. For more accurate estimations of daily volumes of leakage, night water losses shall be multiplied by a factor that would represent pressure variations during the day. This factor can be calculated by using relationship for LI given above (WSA/WCA, 1994).