Diagnostics of Regular Performance of Water Distribution Networks and its Relation to the Network Reliability - Pattern Recognition for Reliability Assessment of Water Distribution Networks - page 184

More in details, the optimised layout of the network sn73 in Figure 7.7 (right) shows clearly
the lower values of ADF avg and NBI compared to the overdesigned network on the left,
despite the fact that the optimised network supplies the lower demand. This fact will also be
the biggest contributor to the comparable values of NPI , PBI and I n for the two networks,
while the lower NRT value of the optimised layout results from the lower total volume.
Figure 7.8 depicts the situation where the network of bigger capacity (on the right) supplies
also the bigger demand. Some discrepancy exists in the fact that this network has the lower
NRT values and yet the higher NBI and ADF avg , the latter also being an impact of the two
connections to the source.
The most difficult to compare are the results in Figure 7.9. Both networks supply the same
demand with the same source head and the one the right is three times bigger in terms of the
number of pipes. This is visible in the difference between the values of NRT , but one would
expect substantial difference in the other measures, too. Still, the values of NPI and I n look
slightly lower on the right diagram although the value of NBI is expectedly higher; the reason
for this has to be found in lower elevations of the smaller network.
The entire networks look fairly to very reliable from the perspective of ADF avg , which gives
almost horizontal curve in all the diagrams. The initial impression is that the substantial
increase of the pipe diameters and subsequent increase of all other reliability measures does
little to improve this index, as it already has quite high value. To get more insight into the
correlation between ADF avg and other reliability measures, the similar diagrams as for the
growth of demand have been produced in Figures 7.10 to 7.12. The conclusions while
observing these figures can be summarised in the following bullets:
- Regardless the level of demand, all three diagrams show logical response as far the values
of ADF avg : the network sn73 is more reliable than its optimised version (in Figure 7.10),
as well as are the networks with larger number of pipes, shown in Figures 7.11 and 7.12.
- The higher values of PBI and NPI for optimised layout in Figure 7.10 are the
consequence of lower demands i.e. the higher nodal pressures.
- The different form of the curves for the same index results from the presentation of that
index in the different range of ADF avg . After continuous increase of pipe diameters, each
network enters the state where this action has little implication for the increase of ADF avg .
It is specifically visible in the results of Figure 7.11.
- Further extrapolation of the index curves of the optimised network in Figure 7.10, by
increasing the pipe diameters, would eventually yield the shape ending with the saturation
zone of ADF avg . How big this zone will be also depends on the head at the source(s).
Equally, the extrapolation of the index curves representing non-optimised networks, by
reducing their pipe diameters, would bring them closer to the curves representing
optimised networks. Consequently, the anticipated shape of a curve covering the entire
range, from optimised diameters to the diameters resulting in no further increase of
ADF avg , would be a kind of S-shape. The results shown in Figure 7.8 represent the very
end of that curve, zoomed-in.
- As discussed in relation with Figure 7.9, the higher values of NPI shown in Figure 7.12
for ngt01-50 results from lower nodal elevations of this network.
As was the case with the demand growth analyses, the statistical analysis of the correlation
between the reliability measures has been done for the diameter increase in all ten networks,
by applying the Pearson's test. The results are shown in Table 7.3.