Hydraulic Reliability Diagram and Network Buffer Index as Indicators of Water Distribution Network Resilience - Pattern Recognition for Reliability Assessment of Water Distribution Networks - page 122

head found in Table 2 of Prasad and Park (2004) was used to calculate the nodal elevations,
which became (in msl): node 1 = 190, node 2 = 160, node 3 = 170, node 4 = 165, node 5 =
160, node 6 = 175, node 7 = 170. The same head of 210 msl was used at the source, and
uniform k-value of 0.5 mm was assigned to all the pipes, which is an approximation of the
originally used Hazen-Williams roughness factor.

Figure 5.22 Case network from Todini (2000) - pipes: L(m), nodes: Q(l/s)

Eight different configurations of diameters were randomly taken over from Table 3 of the
same reference and the results were compared with the calculations done in this research;
these are shown in Table 5.3.

Table 5.3 Results comparison with Prasad and Park (2004)

Network
Number

Pipe Diameter (mm), from Prasad and Park (2004) Prasad and Park Trifunović

1 2 3 4 5 6 7 8 I r I n I r I n NBI

1 609.6 609.6 609.6 25.4 609.6 25.4 609.6 609.6 0.9002 0.6223 0.8919 0.6168 0.0029

2 609.6 609.6 609.6 25.4 609.6 25.4 609.6 558.8 0.8999 0.6141 0.8917 0.6087 0.0029

6 609.6 609.6 609.6 609.6 609.6 203.2 609.6 609.6 0.9038 0.8007 0.8955 0.7938 0.6110

9 609.6 609.6 609.6 609.6 609.6 101.6 609.6 609.6 0.9036 0.7749 0.8955 0.7682 0.5030

11 609.6 609.6 609.6 609.6 609.6 609.6 609.6 609.6 0.9038 0.9038 0.8957 0.8955 0.6520

15 609.6 609.6 609.6 609.6 609.6 406.4 609.6 609.6 0.9037 0.8523 0.8956 0.8446 0.6506

19 609.6 609.6 558.8 609.6 609.6 609.6 609.6 609.6 0.8989 0.8927 0.8904 0.8842 0.6541

20 609.6 609.6 609.6 558.8 609.6 609.6 609.6 609.6 0.9037 0.8923 0.8956 0.8841 0.6503

The table confirms the assumption of taking the minimum required head H * as the sum of
nodal ground elevation and PDD threshold pressure. Minor differences with the results of
Prasad and Park come likely from the conversion of the roughness factor. The table also
shows that the NBI values are consistently lower than those of I r and I n , which is easy to
explain; the source of supply is connected to the rest of the system with single pipe. NBI is
significantly lower in case of the network configurations 1 and 2, for the same reason as in
case of the networks in Figure 5.1b-d: pipes 4 and 6 are too small and the entire system
functions actually as branched network. This again shows that NBI captures the impact of
network connectivity better than the resilience indices.