CAFlood was successfully applied to deliver 120,000 water main burst scenario simulations across London within a week, saving an estimated 27 years computing time if conventional hydraulic models were used. Experimentation was carried out on a range of different sized possible pipe bursts, and a comparison with the speed and accuracy of the industry-standard InfoWorks system was made.
https://doi.org/10.1016/j.proeng.2016.07.409

Processing Times (Minutes) for various pipe diameter bursts, for both single and double precision for CAFloodPro, and single precision for InfoWorks

Thames Water distribution network and 1m LIDAR data coverage, which covers most of greater London area

Shows the varying depths (between 0 and 10.92 meters) and extent at 2 hours of simulation time for a 2,000 Millimeter pipe burst from: (a) InfoWorks output (runtime: 19m47s); (b) CAFloodPro 0.01s minimum time step output (runtime: 3m33s); (c) CAFloodPro 0.03s minimum time step output (runtime: 1m46s); (d) CAFloodPro 1.0s minimum time step output (runtime: 15s);