Understanding flow to manage the network and reduce leakage

Network Flow Performance - Start



With this tool the user can generate a clear and distinctive display of complex and difficult-to-read flow volume time series, for example, for a particular District Metered Area (DMA) or a supply area. This display provides support in the interpretation of the changes present in the time series in terms of known processes and influences (weather, holidays, etc.) and indications of unknown processes (new leakages, wrong valve positions, customer behaviour).

Underground water distribution networks are occasionally subject to leakages. These are not always easy to determine, nor are the associated losses. Peter van Thienen has developed a method of quickly determining leakage losses: the Comparison of Flow Pattern Distributions (CFPD). Thanks to this innovative method, water companies can make use of the data they already collect to target their efforts in determining the causes of inconsistencies in water balance, as well as gaining more insight into their data and network.

Based on the supply patterns for different periods, the CFPD can characterise and quantify the demand changes and the changes in water losses due to leakage in a specific area.

When there are indications of new leakages and/or wrong valve positions, then further inspections – in the field for instance – can be carried out. The tool therefore constitutes an important source of information to effectively reduce leakage losses, gain insight into customer behaviour and control the distribution process.

 


Tool Expert(s)

Peter van Thienen

Peter van Thienen

Senior scientific researcher, KWR

+31 (0) 30 60 69 602
peter.van.thienen@kwrwater.nl

The analysis of flow volume data using the Comparison of Flow Pattern Distributions (CFPD) method provides insight into customer behaviour, leakages and non-registered network parameters, contributing to effective operational management and leakage reduction.

Discuss Network Flow Performance

To add a comment, please login first.