Lionheart Minerals - a Postscript and recommendation:
I had removed the body before letting loose with the hairdryer, and left it in a stable condition with one side back to straight, one very slightly bowed out.
The next evening I checked the sides had not moved, and then reassembled the body to the chassis.
When admiring my handiwork, I discovered the slightly bowed out side was still slightly bowed out, but the straight side was now slightly bowed inwards!
So I removed the body once more, only to find the bowed out side was still the same, but the bowed in side had reverted to straight again.
It looks very much like the body may have been moulded nice and straight, but the chassis plate - when chassis rail sides are clipped onto the outer faces - is very slightly too wide for the recess moulded into the body underside. So when the thing is assembled the chassis exerts an outward force on the side below floor level, thereby causing the upper sides to bow inwards!
I tested this theory by inserting the die-cast chassis plate into the underfloor recess without the plastic chassis rails in place, and twisting a screwdriver in the resultant gap between plate and body moulding - this caused the straight side to bow in. Further twisting caused the bowed out side to go straight.
On the other hand, the more likely scenario is that by heating and bending the body sides without the chassis in place, the gap for the chassis plate could have closed up too tight
So a file was taken to the die-cast plate to remove paint and a small amount of metal, and the lower part of the four door side stanchions were eased a little where they protrude below floor rail and meet the chassis side rails.
Reassembly proved a success.
The recommendation here must be to LEAVE THE
BODY attached to the chassis and simply apply heat and muscle to straighten sides.
I'll have a look at the wooden bodied coke wagons next when time allows. Not sure the above cause will apply to them.