Pcswmm exercises
Usually you can find the culprit by examining the water levels at each node and flows in each pipe to find the source of the instability (erratic water levels or flows), looking especially for: MTVE or possibly PC-SWMM to see the graphics.Ĭontinuity errors are usually a result of numerical instabilities in one or more pipes(s). I did not generate the extra files needed by MIKE-SWMM. I'll be glad to dispatch the 22 mb output file to you if you want, but Set up in the runoff block, rerun that for 24 hours, and add some inflowĬonstant hydrographs to EXTRAN as Bob recommended at the upper nodes. But I suggest getting a very wet 24 hour period You were right to try 24 hours for a test run a one-year run is too Check for conformance toįigures 2-3 to 2-5 in the manual. Particularly the last two or three nodes. If not, there may be some relatively minor geometry violations. This problem could be alleviated with choosingįollowing Bob and Virginia's ideas might get the error level down to a Lot of transients - I wasn't thinking too clearly when I set INTERĪlso, as Virginia Jensen pointed out, the Courant condition was violatedįor several conduits. Observation is on the basis of 2 hour printout cycle which could miss a More or less normal profiles and discharges. The two branches and the main stem above node HCT9020N appear to have With Bob's suggestion of small constant flows in the upper part of the Q ranging up in the millions of cfs for the final outfall reach and The last 3 reaches show deviant behavior in The reported continuity error was -16000 %. Was 22+ mb and it will rise proportionally to INTER. However, have plenty of disk space this output file INTER should probably be shortened to no more than a half-hour cycle and More detail and an idea on local transients. If you can zoom in on the hydrographs you may get Meaningful simulation on MTVE - or I see your file is structured for I made a one-year run using DELT=5 sec and INTER = 1440, resulting in aĢ-hour printout cycle that should probably be shortened to get a more Mentioned, this really did not exercise the model. That in it ran ok and reproduced your results - 58% error. I agree with Bob Dickinson that the I1 data was missing - after I put YourĮrror will go down if you crank these down although the model may take a
#Pcswmm exercises manual#
The manual recommends bet 0.001-0.005 for both. It looks like yours is set to 1 rightĬheck your simulation tolerances as well (SURTOL and SURJUN) for flow and Typically I have seen this value bet 50-100. Reducing your time step even more (say 1 or 2 sec).Īlso, you may want to check the maximum # of iterations (ITMAX) and Your output file shows thatįor Conduit# HCT9005L, the Courant time step was 5 sec. The model will have a greater degree of stability if youĬhoose a timestep =< the Courant Time Step. Length of the shortest conduit (the limiting factor in your modelīasically). Extran does not like negative slopes so also take a look forĪny neg slopes (HCT9005L) and simply redefine the conduit for a positiveįurther, the courant time step is the time it takes a wave to travel the The length of your simulation in order to "get all of the water out" of Your output file shows that you haveīackflow in a majority of your conduits. Continuity error is computed on a junction by junction basis and thenĪveraged over the entire system.