Technical description

Halcrow and their strategic partner for optimisation studies, Optimatics, carried out a pilot study for Thames Water to demonstrate the advantages of generic algorithms for planning and designing DMAs with rehabilitation schemes.

Details

Since 2003, Halcrow has engaged in a strategic partnership with Optimatics Pty Ltd of Australia, who has pioneered the application of genetic algorithms to the specific problems of water distribution network optimisation.

The OGA (optimatics genetic algorithm) is a mathematical tool which helps planners and engineers make complex design and operational decisions. It feeds into a hydraulic model to analyse a very large number of alternatives to identify the optimal solution. Combined with Halcrow’ extensive experience in UK network analysis, design and management, it can bring measurable efficiencies to the development, reinforcement and expansion of water distribution infrastructure.

The OGA has been successfully applied to a DMA (district metered area) pilot study carried out for Thames Water. The DMA chosen for the study was Maiden Lane, located in the metropolitan area of London, as it has a high leakage rate and was hence selected for wholescale pipe replacement.

This particular DMA serves a population of approximately 6,000 and consists of 24 km of pipe mains.

The main objective of this study was to carry out an optimisation of the distribution system through selective pipe rehabilitation to reduce leakage, provide adequate fire fighting capability and provide a more hydraulically robust supply system. The optimisation incorporated the choice of different pipe replacement techniques, such as slip lining and open cut excavation each with associated costs, different pipe diameters and the option of using more than one pipe in each street and consideration of previously unused pipe routes.

Due to the nature of the OGA formulation process, it has been possible to consider a large number of different scenarios, such as feeding the DMA from different directions or splitting the DMA into two. The solutions presented compared favourably against the Thames Water baseline solution and showed a capital cost saving, whilst maintaining hydraulic requirements. The study cost equated to less than 2% of the potential capital cost savings that were identified.

Thames Water has recently expressed a desire to continue their investigations with Halcrow and Optimatics into further applications for the genetic algorithm software.