IN SWMM manual it is recommended to choose B1=B2=1 for basc representation of transmission losses o a channel.
I think a better way is to note that the groundwater flow equation reduces to Dupuit's formula when B1=B2=2. For transmission losses to and from a channel, choose A1=A2 where A1 = k*B/L where k = hydraulic conductivity (L/T), B = distance from channel to groundwater flow boundary (L), L = length of channel (L). Remember to choose k so that units are compatible with model. Ie if using metric units (Litres per second) choose k = m/s * 1000.
Comments
Robert,
I compiled a document regarding the groundwater flow editor, can I send it to you directly, without posting it publically at this stage?
Regards,
Steven Pells
Yes, you can send me an email to robert.dickinson@gmail.com or robert.dickinson@innovyze.com and I will do my best to answer your question
SWMM5 GW equation:
SWMM5.1 groundwater equation:
Have they really changed or does the documentation have a typo?
(Nice font, btw)
This is the code and it implies that it is A1 - A2 or Term 1 - Term 2 + Term3
// --- compute groundwater component of flow
if ( GW->b1 == 0.0 ) t1 = GW->a1;
else t1 = GW->a1 * pow( (lowerDepth - Hstar)*UCF(LENGTH), GW->b1);
// --- compute surface water component of flow
if ( GW->b2 == 0.0 ) t2 = GW->a2;
else if (Hsw > Hstar)
{
t2 = GW->a2 * pow( (Hsw - Hstar)*UCF(LENGTH), GW->b2);
}
else t2 = 0.0;
// --- compute groundwater/surface water interaction term
t3 = GW->a3 * lowerDepth * Hsw * UCF(LENGTH) * UCF(LENGTH);
// --- compute total groundwater flow
q = (t1 - t2 + t3) / UCF(GWFLOW);
if ( q < 0.0 && GW->a3 != 0.0 ) q = 0.0;
return q;
There is a new groundwater equation in SWMM5
Users can now add their own groundwater outflow equation to
a subcatchment, to be used in place of or in addition to the
standard equation. Similar to treatment functions, the equation
can be any mathematical expression that uses the same ground-
water variables that appear in the standard equation.
Thanks for the reply.
As a check, I just ran a test SWMM model with a catchment of 100 ha, receiving steady precipitation with no evaporation until a steady state was reached. I assumed that the catchment has a creek length of 1000m and catchment width of 1000 m. Once steady state had been reached, the groundwater elevation was 150m and the surface water elevation at the receiving node was 132.4 and receiving node elevation was 130 m. I chose A1=A2=2e-8, representing A1= k.L/(B.As) where k=1e-6 m/s, L = 1000m, B=500m and As=100 (ie 100ha). I chose B1=B2=2 and A3=0. The reported groundwater flow, once steady state was reached, was 0.79 L/s for the entire catchment.
For this problem, Dupuits formula gives q (m3/s per side of channel per m length of channel) =k/(2*B)*((150-130)^2-(132.4-130)^2) = 3.94e-7m/s per side of channel, per m lenght of channel. This gives 0.00079 m3/s for the entire catchment - ie 0.79 L/s.
So this means of representation of groundwater flow appears to be good!
The internal units of all of SWMM 5 are American Standard Units and the conversion to groundwater flow is done at the end of gwater.c
// --- compute total groundwater flow
q = (t1 - t2 + t3) / UCF(GWFLOW);
in which GWFLOW converts from GWFLOW (cfs/ac, cms/ha --> ft/sec) //(5.0.010 - LR)
so the groundwater flow starts out as cms/ha and is converted to ft/sec internally