When it comes to water distribution systems that use alternative water sources, modelling decisions becomes more complex. It becomes much more than the hydraulic components. This paper explains how we are approaching this work, and how it is applied to solve real-world problems.

Divide and conquer” is a problem-solving strategy. It simplifies a large and complex problem into smaller, simpler problems.

Usually, models take advantage of this concept. In fact, even very complex models will have some simplifications compared to reality. These simplifications are usually well-thought out. The model can still represent the physical processes in an accurate way.

So, what is the issue? We have so many models!

Many models, one problem

A large set of tools, software and models are already available for modelling water systems. We can model very complex water distribution systems (WDSs), taking into account the hydraulics. We can model open channel flows and predict water levels and flooding. We can model hydrological processes of infiltration, evaporation, evapo-transpiration, rainfall-runoff processes and water quality.

We can do it all!

The problem is, we have MANY tools that do different things.

This is normal, because we often want to study a specific aspect. For example, we may want to know if we can transfer water to the users on the top of the hill. We might also want to know if the water distribution system can deal with a fire emergency.

In these cases, we only need to describe:

  • the hydraulic components of the water distribution systems (pipes, pumps, tanks etc.)
  • the initial volume of water in the storages
  • the users’ demands.

In these cases, the rainfall-runoff processes will not have an effect on the water distribution system.

But when alternative water sources (e.g. stormwater) are used, the situation may be different. Stormwater is strongly dependent on rainfall (and infiltration, evaporation, etc.) processes. If the water distribution system uses stormwater, modelling these processes can be important in estimating how much stormwater is available. When alternative water sources are used, the system that needs to be modelled is not limited to the water distribution system anymore. It becomes a much larger system!

Alternative water sources in water distribution systems: What do we need?

Models and software for water distribution systems (WDSs) take into account the hydraulics and the hydraulic components. If you need to operate a system with alternative water sources, you may need to model more aspects.

Through the CRC for Water Sensitive Cities (Subproject C5.1 Intelligent Urban Water Systems), we have worked on the Orange City Council water supply system.

The alternative water supply system is quite complex. It includes two main storages from local catchments, three groundwater bores, several stormwater schemes and a pumping system from an adjacent catchment (37 km away).

  • Hydraulics are needed in order to assess the operating points of the pumps, and hence accurately assess flows delivered, power consumption and energy operating costs.
  • Minimum environmental flows from the catchment, from the nearby river and from the stormwater schemes need to be guaranteed. In general, these environmental flows depend on the inflows in the storages, the current flow in the river and on the time of the year. Hydraulic models cannot be set up to consider these types of inputs.
  • The operation of the system needs to take into account water availability variation in different years, and within the year. This results in the additional issue of integrating short-term and long-term operational strategies. To do this, the duration of the simulation may need to be extended significantly (e.g. one year).  In this case,  evaporation from reservoirs or storages may become an important factor to take into account. In fact, pumping water now into one of the storages does not mean that the same volume will be available in six-months’ time. Evaporation depends on the temperature and on the surface area of the storage. It’s not an easy aspect to model with a typical hydraulic solver.
  • Some of the environmental flows discharged from the upstream catchments are inflows to other storages of the system. This transfer may need to take into account evaporation and infiltration losses. Again, these aspects (infiltration and other losses during open channel flow) cannot be taken into account by hydraulic solvers for water distribution systems.

Pump Operation for Alternative Water Sources (POAWS) algorithm for the integrated modelling of alternative water sources

Many more aspects may need to be integrated in order to model systems that use alternative water sources.

Thanks to the CRC for Water Sensitive Cities project, we (Angela Marchi, Prof Angus Simpson, Prof Martin Lambert and Lisa Blinco) have developed a multi-objective optimisation algorithm. The algorithm optimises the pump operations of systems that use alternative water sources.

We called this toolkit POAWS (Pump Operation for Alternative Water Sources). We have transferred the POAWS toolkit to Samantha McGufficke of Orange City Council and Martin Haege from Geolyse Pty Ltd on the 23rd June 2016. As reported here, the POAWS algorithm will be an important tool for the planning of the operation of their system.


We would like to acknowledge the support from the CRC for Water Sensitive Cities, the Orange City Council and Martin Haege from Geolyse Pty Ltd.