Effect of Solution Gas on 1D Steam Rise in Oil Sands
Author: Neil R. Edmunds - Laricina Energy Ltd.
Source: Journal of Canadian Petroleum technology - January 2007, Volume 46, No. 1
Following the UTF Phase A SAGD test, it was reported that: “No significant accumulations [of noncondensible gas] were observed in the steam chamber. This was fortunate because such accumulations appear to be harmful because they inhibit steam transport to the front. “Current simulators incorrectly predict that even small amounts of gas...will accumulate in a steam chamber and substantially impair drainage...more work is required to account for gas transport mechanisms adequately so that reasonable predictions can be made for reservoirs containing greater amounts of gas.”(1) Simulators still do not properly account for known amounts of solution gas in the McMurray formation, in the sense that is not generally possible to produce a reasonable history match of field behavior if the full amount of solution gas is included.(2) The question of how the gas is being removed from the chamber, and therefore what is missing from the simulators, remains unresolved, and is beyond the scope of this paper. Meanwhile, two important points have been passed over. The first is that, even at UTF, some in-significant gas accumulations can be found in the data, as reported for example by Ito.(3) The second is that “reservoirs containing greater amounts of gas”, than UTF/Dover that is, includes most of the current McMurray SAGD projects. In the case of Foster Creek, at the deep end, saturation pressure is 2800 kPa and the solution GOR is estimated to be about 5 times that at Dover. Since the gas removal mechanism(s) is unknown, it has always been possible that it was strong enough under UTF circumstances to prevent “significant” accumulations, but not strong enough to do this in deeper Athabasca reservoirs. However, the difficult nature of the SAGD gas problem, combined with the apparent success of UTF history matching without it, has understandably encouraged a situation where it has become standard practise to ignore solution gas in SAGD simulations.
Standard practice is also to arbitrarily lower absolute and/or relative permeabilities in order to match field performance. In the case of the UTF, there was good correspondance between measured core and history match permeabilites, using “open” relative permeabilites. This correspondance does not seem to apply to deeper, thicker reservoirs such as Surmont or Foster Creek, particularly with respect to steam rise rates.