Impacts of Node Connectivity on Reliability of Water Distribution Networks - Pattern Recognition for Reliability Assessment of Water Distribution Networks - page 139

The results of all calculations assessing the configuration of networks L1 to L7 are
summarised in Table 6.6 ( NCF avg = 2.36).

Table 6.6 Configuration assessment of network layouts L1 to L7

No. conn. NGI NCF1 NCF2 NCF3 NCI1 NCI2 NCI3 NSI1 NSI2 NSI3

L1 0.571 2.18 2.18 2.18 0.923 0.923 0.923 0.527 0.527 0.527

L2 0.571 2.27 2.27 2.27 0.962 0.962 0.962 0.549 0.549 0.549

L3 0.571 1.82 1.82 1.82 0.769 0.769 0.769 0.440 0.440 0.440

L4 0.571 2.36 2.36 2.36 1.000 1.000 1.000 0.571 0.571 0.571

L5 0.571 1.45 2 1.09 0.613 0.846 0.461 0.351 0.484 0.264

L6 0.571 1.18 1.64 1 0.499 0.694 0.423 0.285 0.396 0.242

L7 0.571 1.09 1.46 0.91 0.461 0.618 0.385 0.264 0.353 0.220

6.3.3 Network Diagnostics Tool

To be able to check the correlations of the proposed indices and factors with network
reliability, the network diagnostics tool (NDT) has been developed in C++ programming
language by using EPANET toolkit functions, which can process unlimited number of large
networks (each up to 10,000 nodes/links) in single simulation run, assess their geometry and
hydraulic performance, eventually draw the energy/pressure balance and correlate these to the
loss of demand under failure conditions.

The input for the programme is a network file or group of files created in *.inp format in
EPANET, the latter including a text file listing the total number of INP-files and their names.
In addition, a uniform PDD threshold pressure and exponent are specified for PDD
calculations in which the loss of demand is calculated with or without pipe failure.
Furthermore, each network is assessed on the following list of parameters:
1. number of nodes and links, ( n and m , respectively),
2. number of basic loops and complex loops ( l and L n , respectively),
3. the network grid index ( NGI ),
4. the network connectivity factor and index ( NCF, NCI ), in all three versions,
5. he network shape index ( NSI ), in all three versions,
6. total pipe length and volume ( L tot and V tot , respectively),
7. total demand in demand driven mode of hydraulic simulation (DD),
8. minimum pressure and corresponding node ID (in DD mode),
9. maximum pressure and corresponding node ID (in DD mode),
10. the resilience index of Todini ( I r ) in DD mode,
11. the network resilience of Prasad and Park ( I n ) in DD mode,
12. the network buffer index proposed in Chapter 5 ( NBI ) in PDD mode,
13. the average available demand fraction ( ADF avg ) calculated by failing the pipes
consecutively,
14. the minimum available demand fraction ( ADF min ) and corresponding pipe ID causing the
highest loss of demand.

In the single network calculation, additional table will include complete results of PDD
hydraulic analysis per pipe, which includes the calculation of ADF j . The pipes will be
counted based on the loss of demand they cause, in 10 categories, at increments of 10%.
Furthermore, it is possible to do multiple calculations on a single network using fixed or