18 September 2005

Bitumen Upgrading

As I previously talked about in my post on the use of nuclear steam generation for the Alberta tar sands, they require a large energy input in the form of steam in order to separate the heavy oil from the dirt. The heat for steam is currently almost exclusively generated by natural gas. There are two basic methods of extracting bitumen from tar sands: the surface layers can be open pit mined. Deeper layers can be separated in-situ, by pumping large quantities of steam underground and then extracting the warmed bitumen.

In either case, the bitumen needs to be upgraded (mostly through hydrogenation) . Typically this would be done by reforming -- splitting the hydrogen off methane and then transforming the bitumen into synthetic light crude. Syncrude is currently burning off 1.35 Mbtu of natural gas per barrel of light crude they produce. This raises a crucial question for the basis of nuclear steam: how much natural gas is needed for separation, and how much is used for upgrading?

The natural gas used for upgrading cannot be replaced by nuclear generated steam. Hydrogen could be produced from nuclear power, either thermally or through electrolysis, but this is prohibitively expensive compared to steam reforming. I was eventually able to find this link, which on page twelve states that mining consumes 0.25 - 0.30 MBtu of natural gas for extraction and 0.15 - 0.45 MBtu for upgrading. In-situ is far worse at 1.0 - 1.2 MBtu per bbl with the need for later upgrading:


The wide variation in the inputs for upgrading are a result of the wide variation in the quality of bitumen deposits. This sort of puts the damper on the idea of using nuclear for steam production. Nuclear steam isn't highly mobile, so it's not very useful for in-situ production, where major cost savings could be made. However, in the mining case it's clear that the separation is only about half of the natural gas inputs. It might be possible to use nuclear power to preheat a stream of methane used for reforming, but it probably won't make the nuclear steam plant significantly more economical.

The problem remains that the steam inputs from bitumen are quite small compared to the scale of a nuke plant. Even at five million barrels a day production from mining, only about 1.3 GJ of steam could be used per day. That's only about 55 MW of heat production which represents about a 25 MW electric output -- i.e. nothing. One lesson here is that nuclear plants produce massive quantities of waste heat that goes unused.

There is another possibility for upgrading to natural gas. Syngas, produced from gasified coal, consists of Carbon Monoxide and Hydrogen gas. The other option is simply to wait for the price of natural gas to rise above the cost of hydrogen generated by nuclear power. Neither is likely to significantly offset the natural gas consumption used to render tar sands into light sweet crude oil.


Engineer-Poet said...

Back a few decades, small-scale nuclear power was investigated for military purposes.  The SL-1 test reactor (with its rather inadequate control and safety systems) was an unfortunate example, but it was in something like the right power range:  about 3 MW thermal.

It would definitely be possible to build a small reactor suitable for a mining site, even a mobile one.  But how big would it have to be to run the upgrader too?  Envelope time....

Assuming that all the energy in the gas winds up as hydrogen (not accurate but generous), .45 million BTU of hydrogen per barrel is 474 MJ/bbl.  If this hydrogen is made by electrolysis at 75% efficiency, that's 633 MJ/bbl of electricity.  500,000 barrels/day would require 3.66 GW of power - definitely a lot.

Robert McLeod said...

When I go home I can use Allprops to figure the actual hydrogen efficiency of steam reforming. The next step would be to research thermal hydrogen production from nuclear heat, since that's a better path than electrolysis.

Engineer-Poet said...

BOTE:  If US fuel demand is 13 mmbbl/day and it takes 474 MJ of electricity to upgrade 1 bbl using electricity, the power required to hydrogenate the motor fuel would be 71 GW.  That's 39% of the electricity required to replace the fuel directly.