Frontier optimisation of strategic water company targets

SoLow

SoLow (Strategic Optimisation of Leakage Options for Water Resources) is our planning tool that we developed to identify the most economical mix of leakage management options to achieve targets at Water Resource Zone (WRZ) level. It accounts for the interaction of options and the potential efficiency gains one option might bring to another and baseline activities.

The environmental impact and implications of other supply and demand interventions on the water available for use are also considered. The tool has been adopted by 11 water companies, as an enhancement to traditional approaches used across the industry.

It determines the most cost-effective solutions based on the inputs provided using a optimisation algorithm, both for scenarios which require a specific level of leakage and those looking to find the sustainable economic level of leakage. SoLow provides a high-level summary of these leakage management scenarios, and the accompanying detailed investment information required at a company and resource zone level over the whole life of the intervention. 

The tool uses baseline information such as network characteristics, natural rate of rise, background leakage, and property forecasts to inform investment, and accounts for adjustments in these appropriately in response to chosen interventions. The tool has been developed to optimise wider demand strategies such as water efficiency impacts on consumption for both households and non household customers to meet the needs of future water demand reduction.

Quality Assurance

Quality assurance and confidence assessments on data consider the risk of achieving leakage management savings. This provides within AMP strategies to track and target leakage management activities to achieve annual targets, adjusting leakage management options based on currently applied leakage levels and DMA characteristics. It ensures schemes are distinct in that the same DMA will not be targeted with overlapping interventions, in order to appropriately account for the benefits associated with leakage management activities. Alongside this, machine learning is used to adjust benefits and to enable the model to learn and update from both historical and current data provided on leakage management option effectiveness in the field. 

Scenarios are further assessed for their risk based on natural rate of rise variability, accounting for best- and worst-case scenarios to achieve targets. This enables the tool to become a record of how an intervention was predicted to perform and how well it did perform. Providing a platform for re-optimisation helping to justify the adjustment of leakage management strategies DMA by DMA in response to performance, supporting leakage management practitioners in the delivery of short- and long-term targets.