Environmental Engineering Reference
In-Depth Information
Under what conditions is the increase of investment costs more affective for improvement of
the reliability than the increase of operation and maintenance costs? Reliability analysed on
the level of whole network, or significant part of it, will be much more efficiently maintained
by investing into additional pipe capacity rather than by installing additional pumping
capacity. The solutions that favour additional pumping on account of reduced pipe diameters
hide calamities with potentially serious implications for service levels.
Furthermore, the following conclusions address the research hypothesis made.
Based on a given topography, network layout and demand scenario, there is a unique
reliability/resilience footprint that can be described as a function of ADF. This footprint
reflects the network buffer. Although, it could not have been mathematically proven that the
footprint obtained by HRD is a unique one, this diagram has offered consistently logical and
transparent results, leading to the wide range of NBI-values, easy to be interpreted. The
overall feeling from numerous analyses is that the combination of HRD and NBI offers the
most complete picture of the network reliability/resilience.
Reliability measures derived using demand-driven hydraulic models are less accurate than
those derived by the use of pressure-driven demand models. As elaborated above, they are
less accurate in general, not offering direct translation to the loss of demand, as well as they
are less consistent in more extreme range of pressure drops.
Increasing the connectivity between the pipes improves network reliability in general. For a
given supply scheme, there is an optimal network geometry that can be described by a 'shape
index', which can be correlated to other reliability measures. This hypothesis could not have
been proven. The connectivity measures proposed in this research showed less consistency
than those of NodeXL. They appear to be rather coarse and reflect the difference in
connectivity only in case of visibly different network layouts. The simpler network grid index
( NGI ) and the average value of network connectivity factor ( NCF avg ) performed often better
than the more complex network connectivity index ( NCIx ) or the network shape index ( NSIx ).
Also, having the most of these values within the range between 0 and 1, regardless the
number of nodes and links, gives an impression that all these measures would work better as
parameters of more complex formula including the network resistance and supplying heads.
Getting a lower value of the connectivity index for much larger network can be confusing at a
time.
It is possible to make a quick reliability snapshot of a system by looking at typical hydraulic
indicators. There is a clear implication from the interrelation between the pressures,
flows/velocities and network resistance. The quick snapshot is obtained by the simulations
done by NEDRA package but the implication of the interaction between the main hydraulic
parameters is not always clear. The trends of single parameter change in the single network
are mostly obvious but none of the indicators has shown 100% clear correlation with the
ADF, regardless the layout.
The networks with generally higher pressures, despite potential for increased leakage, have
more of a buffer to maintain the minimum service level during a single event of the
component failure. This has been mostly indicated by the two resilience indices from the
literature but not always reflected in the values of ADF avg or NBI . Clearly, the network
reliability is influenced by the combination of geometric and hydraulic properties, rather than
by one of these groups alone.
Search WWH ::




Custom Search