View Full Version : Flow Velocity

April 9, 2015, 01:58 PM
My Model isn't replicating the measured flow velocities. I know there are some false readings within the measured data, but for the most part, the monitor was performing as required. My DWF and WWF aren't in too bad of shape. I was curious what inputs I might manipulate to get my model to react similar to the measured data. I have bumped up the N value for my pipes (as it is an old system, with mostly vitrified clay sewer). The pipe sizes and grades are accurate to the best of my knowledge(based on as-constructed information). I'm sure it is something obvious that my rookie modeling senses aren't picking up. Please see the attached graph.

April 13, 2015, 05:27 AM
Hi Jimmy67,

(assuming you have no user-defined controls or pumps nearby, and your recorded data is accurate)
From your velocity plot, it looks like you're overpredicting a baseflow or perhaps a tradeflow. Also usually, you would be using Colebrook White to represent roughness for pipes, rather than Mannings (manning's was designed for non-constricted flow). As a starting point i would consider removing the baseflow/tradeflow and convert to Colebrook White roughness. After removing the baseflow velocity, you may want to play further with roughness to get a better match (typical ks (colebrook white roughness) values for VC pipes range from 0.03 - 0.15 (new) and 0.3-3.0 (old) from Butler and Davies (2011)).

Also your recorded data seems to be displaying a significant slow response component - given the scale of the falling limb (approx 1 week for the main storm event) - this looks to be most probably groundwater infiltration. You may want to consider a way to represent this in your model, as the loss of capacity for that length of time after a storm even can pose significant flood risk and operational issues!

Good luck!