Hi David
You are correct that using adequate pre-heats, post heats and slow cooling will give you good results most of the time. The problem comes when you are welding something that is either impractical to pre-heat adequately or you do not have the necessary facilities to do this.
To explain my approach, I will give you a run-down on the reasoning:
When welding cast iron, you have two potential areas for problems to develop. The weld metal itself, and the HAZ. When using a filler that will form a hard brittle weld metal (e.g. E7018) you have to take steps to reduce the cooling rate significantly. This can really only be done with very high pre-heats. As mentioned before, we are talking of around 500°C or so. You can however get around this problem by using a filler that will not form a hard weld. (e.g. ENi-CI filler) This does however bring us to our second problem area, namely the HAZ.
Due to the high carbon content of the cast iron, the HAZ will always form martensite and carbides, unless a very high pre-heat is once again used. Once again, if this is not practical, a different approach is required. Seldom will a "low temperature" pre-heat give much benefits in terms of reducing the HAZ hardness. What it will however do is form a wider HAZ, because with an increase in pre-heat, a greater area around the weld will move above the lower critical temperature during welding, resulting in the martensitic phase change upon cooling. (And carbide precipitation) The wider the HAZ, the bigger your chances of getting a crack forming. The other alternative is to accept that the HAZ will be hard and brittle, but to try and reduce the size of the HAZ. This will ultimately lead to a lower probability of cracking. This is the approach that I think is best here.
The only other problem that remains are the cooling stresses that are greater, the faster the cooling rate. To overcome these, you need to deposit small amounts of weld metal at a time. This means small string beads of a short length deposited at a time, and then "staggering" the welds. I believe that the job being described here would typically lend itself to this approach, because of it's ornamental nature.
Hope this makes sense.
Regards
Niekie Jooste
Fabristruct Solutions