Reposting my comments for the benefit of non-team guys...
Smokeless CAN and DOES self-ignite from degradation. It usually is fine if left alone but once it starts degrading, opening the container lets in air which accelerated the decomposition. That orange gas is NO2 and NO3. The temperatures were not high enough to liberate the oxygen and cause a run-away. But once the lid was popped, the extra air pushed it over the top. Boom.
And since it was also self-heating, the burn rate of this normally slow propellent was greatly enhanced.
...
The problem starts when the cellulose is first nitrated. The cellulose, usually a processed wood pulp that has undergone the Kraft process to remove most lignin (nature's "glue"). The process uses concentrated nitric and sulfuric acids and takes some time. The pulp doesn't dissolve, the OHs' (hydroxyls) H (hydrogen) are replaced with NO2 BUT the cellulose does not dissolve, remaining much like the original pulp form. This gives the acids places to "hide". And because the hydroxyl terminals are now replaced, the resulting nitrocellulose becomes hydrophobic. This traps the acid, using coarse pulp can add to greater trapping effect. Naturally, this raw material is well washed but even the best of washing will not remove all acids. These trapped acids are neutralized in further processing after the nitrocellulose is dissolved in acetone/ethyl acetate or the proprietary solvents (Cellulosolves TM) by addition of alkaline materials...remember the gas tube clogging issue with early Ball Powders in the M16? That was from calcium carbonate, added for the very purpose of acid neutralization. Better stabilizers/preservatives are now used, mostly organic amines. The powder used in early M193 ammo had been made from reclaimed propellents, a common practice. Only this reclaimed powder/propellent already had calcium carbonate, leading to this problem.
Powder oxidizes naturally, that solvent smell is residual solvents which slowly displace oxygen from the internal voids of the powder. Ball Powder and sphericals are formed under water, the solvents partition into the water, hardening instantly. But if left too long in the water, some of the organic amines can be leached out along with too much solvent, leading to a dry grain that has poor stability. Once the powder starts to emit the orange vapor, the rusty look, it cannot be salvaged, the acids produced start a chain reaction. Usually, the limited oxygen in things like loaded ammo prevents run-away. But in this case, just opening the container let in enough. And because of the mass, it automatically ignited.
Now, most powders are "artifically aged" for testing using a heated, oxygenated "cell". Exposure to acids is also done in some cases. But sometimes, the artifical aging process can fail to identify a bad lot. Or improper storage after the powder has left the manufacturer can cause premature failure.
Yes, they could have made other mistakes, like not adding any stabilizers although this is highly unlikely. A sample of the raw nitrocellulose is ground with water, then titrated with a base to determine the acid number, a measure of the residual acid