After a few measurements on the vertical boiler, using the cooling curve to estimate heat loss, I achieved a reasonable level of thermal insulation. However, can I turn this into a more methodical approach in the home workshop?
I can electrically heat a block of metal, it would be best if this was copper. Copper would have the correct emissivity and so any radiated heat will be correct. However, the emissivity of copper changes depending on the surface finish:
- Copper heated and covered with thick oxide layer 0.78
- Copper Polished 0.023 – 0.052
Note that a perfect black body has an emissivity of 1.0
This means I probably need to ensure the copper is aged with heat. Also, a first reflective layer might be quite important.
This needs to be a reasonably large surface area if I’m going to get good data with some simple measurements. This also means I need to insulate all of the other surfaces so that heat is only lost through the control surface. That’s a challenge, but I have lots of off-cuts of building foam.
Time to design a test rig. But first, lets look at the background material.
What does maximum boiler insulation look like? When I say “look” I mean in terms of heat loss as this is then something to aim at in terms of design. How much insulation is required? What do those thermal values look like? Are these thermal values achievable?
This was my maximum insulation for the vertical boiler. Reducing the heat loss at 140°C from ~36W (unclad) to ~12W.
The thermal conductivity of the maximum insulation is approximately, kmax = 0.08Wm-1K-1.
The boiler cladding consisted of a number of layers. Working from inside to out they are:
- Teflon sheet – a thin baking sheet wrap
- Medium weight paper wrap
- Polythene – this was a thick clear polythene bag
- Medium weight paper wrap
- Air gap – created using thin wood veneer strips
- Bog oak outer layer
The wood veneer was glued to the outer paper wrap (layer 4) with an aliphatic wood glue. The bog oak strips were then glued to the thin veneer and to each other again using aliphatic wood glue.
The final boiler thermal design lost around 25W. However, the thickness of the insulation was roughly one third that of the maximum design and hence the thermal conductivity was actually better at around 0.06Wm-1K-1.
This appears to be a sensible result when compared to some example Thermal Conductivity (Wm-1K-1) data:
- Expanded Polystyrene = 0.03
- Sheep’s wool = 0.038
- Vertical boiler cladding ~ 0.06
- Wood = 0.13
- Glass wool = 0.040
- Cast iron = 55
- Brass = 110
- Aluminium = 237
- Copper = 398
This insulation is far from practical for the Burrell traction engine as I only have ~1.5mm of depth for insulation and brass outer wrap. My development tool is going to be a block of metal with at least a copper outer surface, insulated in the latest building foam on all surfaces bar the exposed copper. I will instrument the block and electrically heat it. I can then try a number of different insulation materials.
A simple lumped model of the boiler has allowed me to then estimate a number of parameters associated with the boiler.
Looking at the thermal insulation of a boiler and just how thin I could make an insulator I started looking at Multi-Layer Insulation or MLI’s. The effective thermal conductivity of an MLI is in the order of 10-5 Wm-1K-1.