Barnabas

Hi to All!

I hope some of you will find this problem interesting. I’m trying to simulate water abstraction resulted by mining activity in a fractured rock aquifer, with the use of an equivalent porous medium model built up in FEFLOW, (so no discrete features for fractures). I had no problem with the steady-state analyses, the model successfully reproduced the hydraulic heads, which had been measured back in time before the mining activity started. I used the so calculated hydraulic head distribution as initial condition for the transient calculations.

For transient calculations, I’ve represented water abstraction as Dirichlet-type of boundary condition, set for the nodes of the mine tunnel network, giving the measured drawdown time-series as time-varying hydraulic heads. The mining activity lasted 30 years long, 1800-2500 m3/d water was pumped out of the tunnels annually, which caused a maximum 500 m drawdown above the tunnels.  My problem is that transient  calculations will give either a much smaller drawdown for that amount of water abstraction, set as boundary constraints for the nodes of the tunnel network, or without the constraints it will give  a much larger pumping-rate (about  14500 m3/d ) causing though the right drawdown  at the tunnel nodes.  Since the aquifer is confined, I thought specific storage would be the parameter, the amount of abstracted water could be influenced by, but setting its value really small: 10e-06 1/m (which is about the value for a rock without any fractures) instead the default value: 10e-04, the amount of abstracted water decreased only to 13000 m3/d. After that I tried with a specific storage value of 10e-10 but the decrease was even less.

My first question: should I not be able to minimize the aquifers water-loss caused by drawdown,  setting the value of specific storage very close to zero?
My other question: is there a way to reduce the aquifers water loss caused by drawdown without significantly changing the value of the hydraulic conductivity?

Any help or advise is very appreciated.