|To estimate friction introduced by plastering strips inside of laboratory tunnels, an equation named after Reidar Birkeland is used at NTNU. In this thesis, four different geometries with three different strip distances are simulated numerically using OpenFOAM to test the accuracy of the equation. The calculated friction factors from 36 runs are analyzed. Other factors, including how length to development and secondary currents interact with the strip roughness, are also looked at.
With a 24 percent average overestimation of the friction factor lambda, the Birkeland equation is decidedly out of its domain of accuracy in these cases. To ensure reliable head loss measurements, length to development was investigated. Starting out with a uniform velocity profile gave values between 2 and 3.5 of C=lambda*L/D, with L being length to development and D being four times the hydraulic radius, for most geometries and flow velocities. An implicit assumption made by Birkeland was that the strips act as isolated roughness elements, with one not being affected by the former. This was found to be true for all cases tested.