Adding value to GIS network data through model validation and verification

Model development, at its most basic level, involves elevating existing information (network data, current and future population data, water usage, rainfall etc) from discrete data sets into a unified and internally consistent database. During this process, the information undergoes rigorous scrutiny. As a result of this analysis, a number of changes to the data often occur. Assuming best practice, these are based either on primary data sets (such as as-built drawings, operator confidence, field monitoring etc) or good engineering judgment and assumption.

Further to the data improvement achieved, a significant review (such as network topology and operation) is undertaken during the validation process. This can identify (amongst other things): unexpected weir settings at CSOs, pumping station operation and capacity, pipe roughness and condition, valve status and PRV settings.

Assuming the changes are correct, the updated network information offers an opportunity to add value to the existing data set. This value can be in the form of confidence in the connectivity, verification of zone/catchment boundaries, identification of erroneous pipe diameters, inverts, conditions and materials, wrongly labelled service connections, nodes located on continuous pipes and duplicated network - to name but a few.

This paper discusses a pilot study undertaken in conjunction with Metrowater Ltd. The study sought to determine the viability of extracting the data improvements from their existing calibrated water distribution models - and installing it into their existing GIS. The pilot study was extremely successful, allowing Metrowater Ltd to add significant value to their existing GIS data, for a fraction of the cost of retrieving the information from the field.