On Nuclear Reactors as Refinery Process Heaters

Chemical engineers hear this from time to time, how great it would be if only the oil refineries would install small nuclear reactors (presumably modular) to provide the process heat to the different refinery units.  The nuclear reactors would replace the existing furnaces and heaters that burn hydrocarbons.  The selling points were: almost zero operating cost for the nuclear reactors, and zero pollution from the heaters.   In earlier days, that meant real pollution such as sulfur oxides, nitrogen oxides, and sometimes particulate matter (PM) if heavy oil was a fuel being burned.   More recently, the nuclear advocates have added carbon dioxide to the list of pollutants that would be avoided.

Chemical engineers devote a considerable amount of time to researching and engineering ways to improve the process.  These ways typically include reducing operating costs, improving yield, reducing feedstock required, reducing process upsets, and others.   The small nuclear reactors would, supposedly, reduce the operating costs.

I first ran across this in my process engineering days in 1982, while working at an oil company that owned three refineries.

We were less than impressed with the capabilities, the cost, and the severe limitations of small nuclear reactors to provide process heat.

The first limitation was the temperature that could be provided by the reactors.  That turned out to be much, much lower than most of those that the fired heaters achieved.  As just two examples, and these are the two largest heaters in most refineries, the atmospheric crude unit heater had an inlet temperature of approximately 540 degrees F, and an outlet temperature at least 100 degrees hotter.   The second example is in the same range, a catalytic reformer first reactor heater has an inlet temperature of approximately 850 degrees F, and an outlet temperature of a bit more than 900 degrees F.   These two fired heaters, the crude unit and reforming unit, were the largest in heat load (Btu/h) and thus were the best candidates for the nuclear reactors.

The nuclear proponents then explained that the reactor produced a hot water stream at approximately 615 degrees F.  That hot water would be used in a heat exchanger to heat the oil, or whatever else we needed heating.   We were a bit puzzled.  Even if a heat exchanger had superb design, the hottest we could bring a process stream to would be no more than 550 to 560 degrees F.    So, we asked why the outlet temperature was so low in a nuclear reactor.

It turns out that the nuclear reactor has hundreds of fuel rods that contain the uranium that actually undergoes fission.  The fuel rods are made of a special metal, zircaloy, that must not be allowed to exceed a certain temperature.  If the zircaloy overheats, the water around the fuel rods begins to react with the zircaloy metal.   The oxygen in the water combines with the metal alloy, and the hydrogen forms a gas.   So, we could not obtain the hot water at any higher temperature.

Then, we asked about the cost of a nuclear reactor, even knowing the entire exercise was pointless.    For a crude unit fired heater at 300 million Btu/h fired duty, or approximately 80 MW of process heat, the cost was estimated back then at $200 to $250 million.

We never got around to the questions of temperature control, how long for startup and shutdown.

We said Thank you very much.   And the nuclear proponents left.


Roger Sowell

Copyright © 2017 by Roger Sowell, all rights reserved



One thought on “On Nuclear Reactors as Refinery Process Heaters

  1. One possible alternative for the nuclear power plant’s output temperature being too low for most refinery uses was suggested by a colleague. He proposed to use an electric heater to obtain the temperatures required. The electric heater would be powered by electricity produced in the nuclear power plant.

    I considered that proposal, and pointed out a few drawbacks. First, the reactor portion of the nuclear plant would necessarily be 3 times larger in size, and substantially more in cost. We figured the total investment for larger reactor, steam generator system, and power generating turbines would easily quadruple the initial cost. That put the cost to build the nuclear power plant in the range of $800 million to $1 billion.

    There was no way a refinery would elect to do that when they could purchase power from the grid, or even better, burn natural gas in their own combined-cycle gas turbine power plant that also produced process steam.

    — Roger Sowell


Comments are closed.