Cleaner coal: how small do you like your reactors?
Coal will remain a staple fuel for electricity generation despite the carbon dioxide it emits. However, research at Imperial College London into gasification and high-pressure reactors is helping to increase efficiency.
I once introduced our solid fuel research at Imperial College as "the second oldest profession". I was, of course, referring to our legendary ancestor, the charcoal maker. Few people now realise that even King Henry VIII's gun foundries ran on charcoal, causing much deforestation in southern England. It's no accident that Sweden, with its vast supply of timber, became the next great manufacturer of steels and guns.
When coal arrived on the scene, it stole the show. With its greater energy density, it was cheaper to transport. The UK will probably need both coal and biomass for a long time to come. Yet, for a decade and more, government ministers virtuously ignored the coal option for power generation. It seems advisors were telling them that burning coal produces carbon dioxide. Then, suddenly, in April 2009, four new coal-fired power stations were announced.
The announcing minister did not mention which technologies had been selected, which he probably should have. Instead, he talked of pilot schemes for the magic antidote - 'carbon capture and storage' - for which there is no technology on the market. 'Experts' estimate it would cost a mere two pence per kWh. It comes to between £7bn and £12bn annually, depending on the rate of UK capacity utilisation. It makes nuclear decommissioning look like a bargain.
After the current economic trough, however, step-increases in energy consumption are projected for developed economies, as well as for developing and transitional ones. There are fascinating contradictions at work within the current debate.
Whatever else happens, one way to conserve resources and reduce emissions is to come up with more efficient energy technologies. That is where energy engineering enters everybody's life. Furthermore, it does not require genius to see that coal and biomass will hold their own in fuel mixes of the future.
We will need to investigate new fuels, new processes and new reaction conditions. Such tests are less expensive when carried out on smaller scales than the industrial plant level. Pilot-plants are meant to be smaller models of actual process equipment. However, a pilot-scale high-pressure fluidised-bed, say, could easily occupy several adjacent large sheds and require teams of engineers manning the beast for around-the-clock operation. Expensive! That is why most pilot-plants see very limited action. Indeed, some are constructed for just one experimental campaign, and the capabilities of pilot rigs are soon dated. Each new campaign is usually preceded by a costly refit.
Experiments at bench scale are quicker and cost far less. At Imperial College, we have taken solid fuel testing to a new level of compact design. Small-scale experiments do not always translate into exact numbers for plant design, but they will give you trends and help you select what experiments to run at pilot scale. It allows running fewer costly pilot tests. Bench-scale rigs can also be used for trouble shooting, revealing why things go wrong at pilot level and even at plant scale.
However, before considering the midget reactors, it's important to know what it takes to run experiments on solid fuels.
Laboratory studies require an ability to characterise the underlying behaviour of the fuel, with as little reference as possible to sample configuration or reactor design. This is no different from requiring the result of any measurement to be entirely independent of the method of measurement.
In practice, we make do with results as independent as possible from the method of measurement. Nevertheless, the importance of striving to minimise the effect of experimental design on our measurements cannot be overstated. Basically, we subdivide sample particles as finely as is practicable and we try to assess the behaviour of sample particles - as much as possible - in isolation from one another.
Most high-pressure fluidised-bed designs involve a reactor body fitted into a furnace; insulation is wrapped around it and the assembly then placed in a pressure casing, made of low-alloy steel. The insulation ensures that the pressure-casing is not exposed to high temperatures. It usually ends up as a large piece of equipment.
We went the other way. We used the reactor body as the pressure vessel. The body is small, machined to 48mm outside diameter, 34mm inside diameter and just over 50cm long. This much alloy (Incolloy 800HT) will cost about £100. It has remarkable performance with a creep rupture limit of 1,000 hours at 1,000°C and 40 bars.
Then we went a bit further. By attaching electrodes to the top and bottom, we used the reactor body as the resistance heater. No furnace. No insulation. No pressure casing. Plus, the flanges rest on two half-moon rings set in grooves machined into the body. Hence, no welding. And, for safety, it fits snugly into a small box constructed of 1/8in mild steel.
This toy-sized reactor has served as platform for designing several different types of test. The initial design was for batch experiments, featuring a bubbling fluidised bed. Some 50-100mg of sample was forced into the fluidised bed at 30 bar, through a tiny water-cooled high-pressure probe. Developing the probe used all the capabilities of our very versatile workshop. Yields were repeatable to within ± 2 per cent; gases and solids could be recovered for characterisation.
At the time, British Coal was developing the Air Blown Gasification Cycle (ABGC). It wanted to know why 20 per cent of its feed carbon turned up as unreactive char. We ran tests with different particle sizes. It turns out that, when carbons are held at 1,000°C for longer than 10 seconds, they begin to anneal. They become stable - unreactive. Larger particles (3mm diameter) fed into the pilot reactor could not be consumed in 10 seconds. So they baked and lingered.
To confirm this story, we turned to another rig, a 'wire-mesh reactor' that required less sample and gave better time resolution.
Engineering science comes alive when new technology is being developed for commercial scale. We get to hear about all that goes wrong and we find the reasons and, sometimes, the solutions.
Our next problem was to explain the baffling distribution of ammonia concentrations in the gas product from the pilot-scale ABGC reactor. Being fed at 200kg per hour, it was neither cheap nor quick to change operating conditions. At bench scale, however, we could match reaction conditions quickly and easily. However, to simulate the spouted bed, we needed continuous sample injection. That had never quite worked previously, at such small scale. My colleague Nigel Paterson came up with a design for injecting about 3g per minute, which I was convinced would never work. It has not failed in over ten years of operation.
Using British Coal data, we identified suspect conditions and went to work. As expected, the ammonia in the product gas diminished at higher temperatures. Similarly, changing the coal/air ratio changed the product gas composition. But one unexpected result was to see how doubling the steam input (not a problem in our little reactor) would double the ammonia content of the product stream. During the same set of experiments, we managed to track rates of cyanide formation as well.
In another project, we determined extents of trace element emissions during the gasification of sewage sludge pellets. Above 900°C, the depletion of barium, lead and zinc in bed solids was accompanied by enrichment of fines, collected in a downstream filter. Mercury just flies away and must be caught separately.
Some sludges from the north of England contained inordinately high amounts of zinc. Apparently, zinc oxide is used as white pigment in facial creams. The water companies told us "Northern girls use more makeup". Who knows?
Our little fluidised-bed is still busy. We recently completed work making liquid fuels from waste plastics and the rig is now being used for oxy-fuel firing experiments, burning coal in a mixture of oxygen and carbon dioxide. We are trying to develop aspects of the missing carbon capture technology.
What happens when you need 2,000°C at similar pressures? The conditions are those of entrained flow gasifiers, currently very popular in China. Or, you may need to know the effect of heating rate on your fuels. In that case, we need to go smaller still.
Several laboratories around the world have developed their own version of this reactor. Many of the features described below, however, are unique to the rigs we developed at Imperial College, and proved immensely versatile. They were originally designed for contacting coals with hydrogen and optimised for liquid yield determinations, at up to 70-80 bars and 850°C. Time-temperature programming was a little more involved when we first started. We had to make our own analogue/digital converters. These days, £80 will buy you a multi-channel programmable card. When we were asked to compare reactivities of coals under power plant combustion conditions, the temperature range was extended to 1,500°C.
As part of the ABGC project, we were also asked to investigate differences between the gasification reactivities of a set of 'world' coals. At that stage, we added a steam injection capability. Since the wire-mesh cell casing is 'cold', we preheated the gas flow path within the cell before injecting the steam, to avoid condensation.
As new experimental research moves away from the UK shores, we are being asked to support novel technologies in places as far removed as China and South Africa. To support the new two-stage gasifier designed by Thermal Power Research Institute in Xi'an (Shaanxi Province, China), the range of the wire-mesh reactor was extended to 2,000°C, operating at up to 40 bar, using steam injection. This allowed us to report on the reactivities of Chinese coals under entrained gasification conditions,
before their new pilot plant started up.
Another application involved examining the behaviour of coal particles injected into channels near the base of blast furnaces (tuyeres) and raceway cavities. The procedure improves the economics of
iron-making by requiring less coke, an expensive feedstock, the manufacture of which leads to much pollution. At high coal-injection rates, however, operational problems become apparent.
The work was carried out at up to 2,000°C and 6 bars. The instrument was modified so the sweep gas composition could be altered sequentially, to simulate the stages that injectant coal particles 'see' in tuyeres, raceways and the furnace. The initial stream of nitrogen (pyrolysis) was followed by a 20-100ms burst of oxygen-enriched air, before switching over to CO2 to gasify the residue.
We found that much of the coal does not combust in the tuyeres and raceways. The time available in the blast was short and air-fuel contact within the plume of pyrolysing and combusting fuel was poor. If the coal is not all consumed by gasification on its way up the furnace, it turns up as dust in the top-gas - not desirable!
The most recent extension of wire-mesh reactor capabilities was undertaken during a 'zero-emission carbon' project. It required contacting samples of coal (and wood) with mixtures of hydrogen and
steam at up to 950°C, under 90 bar pressure. The data gave us a wide ranging envelope for selecting operating conditions for the actual process.
We are now investigating whether some heavy oil upgrading methods used in oil refineries can be overhauled. The wire-mesh reactor lends itself to such work without refit. Samples of heavy oils are 'painted' onto the mesh. Char residues obtained at different temperatures and heating rates will be correlated with initial sample layer thickness on the mesh. It may lead to new ideas for plants in Canada, to improve their seemingly wasteful methods of upgrading feedstock from the Athabasca tar sands
Professor Rafael Kandiyoti is senior research fellow in the Department of Chemical Engineering at Imperial College London.