I had planned to mount the fuses directly on the pcb but given given the advice received it seems that it would be better to have a in line fuse (at least as far as the gel cell is concerned) as part of the ~300mm long leads that run from the gell cell to the screw terminals on the pcb.
That sounds good. If the cable is pulled out of the screw terminals and shorts, you want the fuse to go.
This is better practice I think because then the leads themselves are fuse protected whereas if the fuse was just on the pcb then nothing 'upstream' from the fuse is protected. There is some small chance that the leads could be chewed by vermin and short out each other or the metal housing.
Or pulled by someone trying to take it apart without unscrewing some parts.
The mains plug pack max output will be around 1amp. What is the engineering rule of thumb for calculating/guestimating the ability of a short to ignite a fire? Obviously a 12V gel cell can supply huge amounts of I for short periods and melt metals. Can a 1amp 15V plug pack supply start a fire?
Current = heat. Heat + something flammable = fire.
1A through a piece of 500mA fuse wire can make it glow red hot, and this could set fire to dry food or straw.
1A through a piece of 10A fuse wire will do nothing. Although 1A is not very likely to make a fire, it just takes one component/wire/connector that is not very good at dissipating heat, and is next to something flammable.
Put some steel wool on top of a 9V battery, and it will burn up as it is so thin. Put cotton wool or paper on this, and a fire will start. Cheap 9V batteries cannot supply 1A, yet can start fires in the right case. However, with suitable wires, PCB tracks that are thicker than hairs and components suitably rated, a fire is unlikely. Note the lead acid battery may leak hydrogen while charging.
A fuse blowing and the equipment not working is not a catastrophic, I'm more concerned with the prospect of a over current situation leading to a fire and secondly trying to limit irreversible damage to the pcb (melted tracks etc.)
Polyfuses are good for things like that - I would be more worried about electronics blowing through overcurrent than a fire starting.
Is it acceptable engineering practice to try obviate the need for a fuse by good design or basic things like (for example) ensuring that vermin can't access cabling or that it can't chafe or should you always work on the basis as my late father advised me "don't ask what will happen IF it fails rather ask what will happen WHEN it fails"?
I think you can avoid the fire problem by stopping anything flammable (straw, food etc) from entering the electronics areas. Blown comonents are another matter, although good design should avoid this. I'd just stick a 10A fuse on the battery (not sure how big the motor is), and a 1A slow blow fuse to the electronics (slow as capacitors may want to fill up).
Andrew