Courant Condition

I'm new to the stormwater modeling community. I'm currently attempting to build my first model which consists entirely of ditches connected by culverts in places where there is some man-made obstruction to the ditch. I'm having some difficulty with my "courant condition" for some of the open channels (ditches) in my model. I started with a time step of 60 seconds, which produced a courant condition warning for just about every ditch. I've had to reduce the time step to 1 second in order to prevent this warning for all ditches.

Can someone explain what the repercussions of doing this may be or perhaps suggest some other solutions to the problem. Thanks in advance.

Jeff
JThompson@HOBBSUPCHURCH.COM

Dear Jeff,

If you have solved the problem, just reducing the time step, consider yourself very lucky!

Regards

Giovanni
giovanni_diconza@LIBERO.IT

Jeff -

Try setting NEQUAL=1 on the B2 line. The program then automatically "lengthens" the channel that the program sees while cutting its Manning's n an equivalent amount (see Eq. 2-6 in the Extran manual and accompanying discussion) to comply with the Courant requirement.

Actually, 60 seconds is somewhat on the long side for EXTRAN, in my experience; I usually start out with 5 seconds and NEQUAL=1, and have sometimes had to go to 1 sec. That usually eliminates the Courant problem.

Good luck

Ron Kilmartin
ronkilmartin@ATTBI.COM

Jeff,

I would like to add a bit to Ron's comment. The NEQUAL=1 does work very well to eliminate CFL problems in links but if you use it with a 60 second time step and large natural channels you are probably increasing the volume of your system by a tremendous amount. A general guideline (which is in the SWMM 4 output) is that you should not increase the full volume of your system by more than 5 to 10 percent.

There are other sources of instability in Extran besides the link CFL condition that you might acerbate by using a large time step and the NEQUAL=1 option. The longer natural channels might:

(1) cause instabilities at the nodes as the area of the nodes change during the Extran "HEAD" calculations,

(2) The smaller roughnesses required in the creation of the equivalent conduits may cause instabilities as the HGL moves vertically (Ron has described this as "Area Shock" in past discussions).

(3) the continuity equation for the node may experience problems in the beginning because of the larger node volume.

Regards,
Bob Dickinson
DickinsonRE@CDM.COM