David Starkey sells tanks that might be adapted: "The smallest [insulated] milk tank fitted to a milk lorry produced is 9000l, [costing] £2750--£3500" as of 2008-05.
10W loss is impractical needing many metres of foam or vacuum insulating panes. However, instead of a 5t tank losing 10W (10% loss of stored heat over 6 months), why not use a 10t tank losing ~100W (~50% loss of stored heat over 6 months)? The extra energy to put in the tank is readily available in the summer and free, and I need only ~10--20% of the insulation, ie something nearer 50cm than 5m+. The downside is that we'd definitely need to dig down to accommodate a large enough tank as we just don't have lots of ground space. But we might not have to dig much.
Additional summer excess might be used to heat up the mass around/under the tank to reduce heat leakage and to act as a further heat store in its own right.
System Extension
I want to ensure that any storage tank contains some spare heat-exchanger coils at various heights to allow for possible future expansion, eg:
a seasonal thermal store as summer heatsink and winter heatsource
greywater heat recovery
2010 on: Solar PV plus Wind
I found it hard to find credible sophisticated suppliers of solar thermal, so instead went with more solar PV which has the advantages of familiarity and being able to export any excess to the grid rather than 'waste' them.
As of April 2010 the business now has installed enough PV to offset:
Consumption of electricity by my office networking, laptop, etc.
UK Internet-based services such as the main server in co-lo.
UK business travel and related implied direct emissions.
The next immediate aim in 2010/2011 is to try to find wind CapEx projects to cover the emissions from co-lo servers abroad (currently US/India/Australia).
As of 2010-09-17 the company purchased a stake in [defunct http://www.unitywind.com] Unity Wind that should have avoided about 5tCO2 per year if the project had gone ahead, which should have been comfortably enough to run three overseas servers at a local grid intensity of up to 1kgCO2e/kWh (eg mainly coal) and a consumption of up to 190W (~4.5kWh/d or 1664kWh/y) each. This CapEx could have been good for 20+ years once the project was running. But UW's equipment was vandalised and planning permission refused.
By comparison, the h2ope Stockport Hydro scheme hopes to avert 87.5tCO2/y for £250,000, ie 0.35kg/£/y, or about 29x more expensive than UnityWind per tCO2/y but still maybe 4x to 5x cheaper than solar PV on the roof. A small portion of long-term company reserves has gone to this project.
The company is working on a super-insulation experiment (with aerogel) as another route to help others offset/avoid emissions.
In early 2011 the company put a small amount of long-term reserves into Ecotricity's EcoBonds, and into the Ovesco/Lewes community solar project.
Also in early 2011 the company paid for a domestic PV installation in Kingston-upon-Thames.