Introduction - Pattern Recognition for Reliability Assessment of Water Distribution Networks - page 2

1.1 WATER DISTRIBUTION MODELS

Water is life. Water distribution networks live as humans: they are born, grow, get old, may
suffer from 'stress', 'high cholesterol', 'blood pressure', 'haemorrhage', 'heart attack' or else.
However, they rarely 'die' although they are nearly dead when they have been poorly
designed and/or constructed, or little has been done about their operation and maintenance.
On contrary, some of their users can die as a result of such an abysmal situation.

A lot has been written about the relevance of potable water for public health. Researchers in
the field of water distribution are continuously concerned with performance improvement of
distribution networks, analysing specifically the 'diseases' related to water demand and
leakage reduction, corrosion growth, water hammer impacts, pump failures or else. Drawing
parallels with medicine may therefore not be so ridiculous; it is almost that the average life
expectancy of water users could be brought into a proportion with the average lifetime of the
network components. Some futuristic research topic could aim at possible correlations
between the condition of distribution networks and medical records of the population
supplied from them. It is not quite clear how feasible such a research could be, but it is very
clear that computer models would be playing essential role in it.

Hardly any field of civil engineering has benefited so early from the development of PC
computer technology, as it did the field of water distribution. Hydraulic modelling software
launched massively in the developed countries in early eighties, has been speedily introduced
all over the world. Water quality modelling applications that followed with the delay of some
10-15 years are nowadays equally available in practice.

Such breakthrough allowed a single water distribution expert to analyse dozens of design and
operational scenarios for the same time that would be required by a dozen of experts to
analyse a single one in the era of manual calculations and hydraulic tables and diagrams,
being daily practice just a couple of decades ago. The trend accelerated significantly by the
end of the millennium, and the challenges of water distribution in the 21 st century go even
beyond optimisation of design and operation of selected layouts. New possibilities are opened
to look deeper into the mechanisms of corrosion, sediment transport and other phenomena
that are affecting maintenance practices and eventually play role in the overall ageing of the
system.
Based on what has been achieved only within the last decade, it is fair to believe that a model
that could simulate full behaviour of distribution network throughout a longer period of time
than just a few days is not necessarily a dream. Apart from readily available results showing
the network hydraulic performance, such model could suggest how the network should be
managed throughout its entire design period, namely:
1. What are the most effective maintenance practices?
2. How to deal with renovation and expansion of the network?
3. How to improve the network reliability?

The prerequisites for having such a model, very likely sorted according to the degree of
research complexity, are:
- To have complete picture of complex mechanisms that take place in pipes, namely the

growth patterns of various deposits, corrosion, biofilms, etc.