Water Supply Configuration and Water Losses - Nikšić Waterworks Case Study - page 04

Assessment of Water Losses Components

Due to high water losses in the Nikšić distribution system, leak detection activities shall be introduced and maintained over a multi-year period, with an implementation rate that will depend on the available financial resources, among other things. Up until now, detection was carried out in two of the twelve defined consumption zones (DMAs). In both cases, field work was based on application of the balance method, supported by daily water meter readings and occasional use of the minimum night flow method.

Results of activities performed in two consumption zones lead to the conclusion that apparent losses in Nikšić waterworks are significant. Analyses of water meter readings, which were performed on an almost daily basis, led to the following conclusions regarding accuracy of the existing water meters:

  • Reliable water meter reading: 47%,
  • Unreliable water meter reading: 43%,
  • Questionable value of reading: 10%

The results indicated that only about one half of water meter readings provide reliable results. In an attempt to define the non metered, unbilled consumption, an improvised calculation was made. In this calculation, the sum of the numbers of unreliable water meters and water meters with questionable readings was multiplied by the average daily consumption expected during the day:

net q av d (zone 1, nov) = 135 l/con/day;
net q av d (zone 2, dec) = 130 l/con/day

When the resulting value was multiplied by the corresponding number of consumers, unbilled consumption was defined. Although the assumption related to consumption can only be proven through additional research/analysis, the above simplified calculation has proven to be the only realistic method under the circumstances. The goal of this brief analysis was to roughly assess the degree of real consumption increase (known as the coefficient of apparent loss), compared to billed consumption, from the aspect of billing inaccuracies. The value of this degree, for the two investigated consumption zones, amounts to:

K apparent losses = 1.6

The investigated zones included part of the suburban area of the City, which is mostly comprised by household consumers, where the majority of the aforementioned defective water meters and water meters with questionable readings, were linked to individual households (private houses). In this sense, it is assumed that in the central City zones, which are more densely populated with more residential consumers in apartment buildings and flats, a slightly smaller value of coefficient of apparent loss may be expected. Assuming that this ratio, related to the system as a whole, amounts from 1.3 to 1.5, the share of losses in the system is calculated:

K apparent losses = 1.3 ⇒ net Q av yearly (system) =
162 l/s ⇒ P losses = 60 %

K apparent losses = 1.4 ⇒ net Q av yearly (system) =
175 l/s ⇒ P losses = 56 %

K apparent losses = 1.5 ⇒ net Q av yearly (system) =
188 l/s ⇒ P losses = 53 %

The above descriptions mean that, if there were a high degree of water meter accuracy in the system (approximately 90% - it would be unrealistic to expect that all meters in a system with 65,000 consumers will be repaired at the same time), and an overall improvement in the Nikšić billing system (improving accuracy of water metering, elimination of unmetered billed consumption, improvements in data handling and reducing/eliminating billing errors), could reduce the total system losses by 10 – 15 % (compared to the water input volume).

This represents a significant value, because 15% of input under present circumstances amounts to a flow of 60 l/s, or almost half of today's total billed consumption. It can be concluded that the expected outcomes of reducing apparent losses are significant, however the system will still have real losses of no less than 55 %.

Regarding physical (real) water losses, with the exception of a certain number of defective valves, investigations performed in the two consumption zones pointed to the following conclusions:

Losses were detected on a number of secondary branches, but the values of billed consumption (1-2 l/s) and the corresponding measurements conducted during the detection (3–4 l/s), made further detection activities unprofitable; a significant number of minor defects was observed along a stretch of pipeline several kilometers long, which would be difficult to detect (detection of exact location of these defects would require far more time than that provided for implementation of the balance method), with an uncertain outcome and fairly small water saving potential;

In some sections of the main distribution lines of the investigated zones, made from asbestos cement pipes, direct connections to individual consumers were observed. A large number of separate household connections to the main distribution pipeline, under high and variable pressures (both zones are located immediately downstream of the Duklo BPS) may result in a large number of small defects along the main distribution lines and household connections. This was confirmed by measurements since those sections had the greatest difference between the measured flows and billed consumption. The following rehabilitation measures were recommended: replacement of the asbestos pipeline; grouping of consumer connections, which include construction of parallel secondary lines, connected at one point to the main distribution pipeline. The connections would than be made to the new secondary parallel line;

In the first of the investigated zones, the recommended replacements have already been completed, resulting in a reduction of input flow from 8 l/s to 4 l/s, while billed consumption has remained unchanged (about 2 l/s);

The expected outcomes of the described replacements in the second zone (which have not yet been made) are reducing the inlet flow by 6 - 7 l/s (from the current 20 - 21 l/s, to an expected 13 - 14 l/s);

In the second zone, input flow varied considerably with changes to valve openings, located at the entrance point of the subsystem. This variation was up to 7 l/s, during regular supply conditions. The criterion used for the inlet valve operation was to maintain a pressure of 1.8 bars at the critical point in the zone (in the case of zone 2 - the consumer that was located at the highest point). Therefore, implementation of automatic inlet valves, which will maintain the required minimal pressure at the critical point, was recommended. The expected outcome of this measure is the reduction of the zone input flow by 4 – 5 l/s.

Initial results in detection and repair of leaks (real losses) in two investigated zones are encouraging. Draft implementation programs for detection and repair of leaks for the whole network have been prepared, and the foreseen time period needed for implementation of investigations in all parts of the distribution system is several years. After that, results of detection and repair programs will be evaluated and new programs such as the detection and repair program will be drafted and effectively implemented as a regular activity within the waterworks operation.