I'll have another go!
The 20tpd is 20 tonnes of feedstock, VBC in this case. From p.24 of the JORC (14/8/12 ann.), the coal is 63.5% and 7.5% ash. That leaves 29% of 20 tonnes to make stuff from, which is 5.8 tonnes out of the 20.
The output, from various Thermaquatica and company announcements is between 98.5% and 99.5% water, so let's take it at 99% water and 1% product. The density of the coal is about the same as water. So if the output is 1% stuff from 5.8 tonnes of input, the output is 5.8 tonnes divided by 1%, which equals 580 tonnes of output per day. 574.2 tonnes of water, 5.8 tonnes of organic product. One tonne of water is 1,000 litres, so every day the pilot plant puts out 580,000 litres, which is about 15 back yard swimming pools.
If it is working as a continuous process, which has been indicated as one possible option, that is 580,000/24, or 24,000 litres per hour or 400 litres/minute. P.5 of the last quarterly states that the "immediate or primary product as it exits the reactor has an immediate use, without further treatment." Which means that the PDU, with a feed of 5kg per day, or one four thousandth the rate of the proposed pilot plant, should have produced 6 litres per hour.
But 580,000 litres per day, for an application rate of 10lt/ha, treats 58,000 ha of wheat. With 20 million ha of grain (wheat, barley, maize etc) in Aus, the pilot plant will produce enough for one application for the whole of Australia in 345 days. But they'll have to store it somewhere, and stop it going mouldy whilst in storage; maybe just being without light will stop the mould growing.
However, what you have to realise is that this process uses a lot of energy. The water and coal need to be heated up to ~300C for the reaction to take place. When you boil a kettle you heat 2 litres of water through 80C. The pilot plant requires 580,000 litres to be heated by 280C, three and a half times the temperature rise of the kettle, and 290,000 of the volume. That's just over a million 2lt kettles. Every day.
The process also requires the output to be quenched, that is, to be cooled rapidly, to stop the reaction from continuing.
http://thermaquatica.com/content/uploads/OHD-Clearwater-Preprint-Draft.pdf
They don't say what temperature the product needs to be cooled to, but quenching involves a lot of water for rapid cooling. This also removes a lot of high grade heat in exchange for low grade heat. Which means that not much, maybe 30%, of the heat can be recovered for heating the feedstock. Be interesting to see what details we get in the PFS of the water and energy balances.
Hope that puts it in easier to conceptualise terms.
15 back yard swimming pools of water required each day, to produce 15 swimming pools of product.
Energy to boil one million kettles each day (less really, depending on heat recovery).
Output sufficient to treat the whole annual Aussie grain acreage once per year.
(20min delay)