We thought it appropriate to put out a blog on eVader on May the 4th 🙂
We have been using eVader recently to analyse the impacts of the ‘Beast from the East’, in other words the cold snap and then thaw that hit the UK over the first weekend in March, causing a sudden increase in bursts and causing water supply issues in some regions.
eVader is a network event detection system developed from a deep understanding of water flows in networks and the application of robust statistical processes. We have designed the system to monitor 1000’s of existing network monitoring points, learn what is normal for each point at any time of the year, and then produce an alarm in real time when an abnormal event occurs.
We applied the eVader algorithm to data from around 4000 DMAs in a water company’s water network to see what insights eVader would provide. The figure below shows firstly the eVader output in terms of number of event alarms per hour for 2016 in the top graph. The second graph shows the two week period around the ‘Beast from the East’ for 2018 (note the days in all the graphs align, in order to compare the same response for weekdays and weekends). The third graph shows the air temperature in the region for the same period of 2018.
The week prior to the 3rdMarch 2018 shows the air temperature was below zero for most of the time. On the evening of Friday 2ndMarch, the thaw started. At about the same time there was a sudden increase in abnormal flow events from around the water network. The eVader outputs from the same period 2 years before, show a steady level of events from normal burst and leak breakouts in the system. The analysis shows a rapid breakout of events over the weekend in response to the freeze-thaw, and the speed at which the events were brought back under control. The data from this analysis is currently being analysed to look at the geographical spread of events, and which areas were more affected than others.