2. Eating through the surface of a drum will release heat. That heat could also cause the lead azide to ignite.
Ok, and?
If the plan is to detonate it... why not modify the plan to be "prepare for it to detonate then try and handle it (robotically) in a non-detonative fashion".
Doing experimental chemistry on 59 barrels with a robot seems to spread the explosion risk over several days. Also assumes you can source a mobile robot suited for doing chemistry in the next few days and train the operator up how to use it.
Even now the only reason for "the next few days" seems to be their unwillingness to simply blow warm air at the barrels to keep them warm. An entirely artificial constraint.