Christchurch Implements World Class Smart Wastewater Network

Christchurch was one of the earliest cities in NZ to have extensive flow and overflow monitoring throughout their wastewater network, providing both data and insight for planning and operational requirements, was well as alarms, duration and quantity for overflow event reporting.

Traditionally a coarse system of overflow confirmation was used where sand was placed on the overflow weir and washed away during an event. This method was low tech, low on detail and required manual labour. In time, this was migrated to a network of thin-plate overflow weirs and pressure sensors. However, the installation of non-return valves on the wastewater overflow pipes throughout the city rendered use of weir ratings incorrect due to weir drowning, overstating the volume of wastewater lost to the environment.

A unique approach was required for overflow quantification, and lead Mott MacDonald to utilise a primary depth and conductivity sensor in the overflow manhole, coupled with pressure, ultrasonic and velocity sensors installed in the overflow pipe. Significant reduction in stated overflow volumes have been the result of the improved system as it is truly measuring the volumes of overflow.

However, data collection is only part of the solution for active monitoring. The balance focussed on innovative approaches to exacting specifications, requiring upgrading Mott MacDonald’s Moata analytics platform to support a commitment to high data availability, assured times for ‘observation-to-availability-online' of data, timely alarms in all weather conditions and much more frequent reporting from gauges than is typically found throughout New Zealand. This timeliness allows control and intervention in near real-time, moving overflow monitoring past a ‘post event’ reporting system to one that encompasses the best qualities of traditional SCADA, the modern technologies and availability of web browser delivery and ability to overlay increasing smarter data analytics.

Future directions include short-time period rainfall forecasting, driving the prediction of flow and overflow events using machine learning trained on physical models and informed by the long-term data we have now collected.