Again many thanks for sharing the details!
One thought comes to my mind. What if the excess heat was caused by a local overheating of the heater? Then we can get very high temperatures just before the heater will fail in some spots. Does the heater always failed in each experiment after excess heat was observed?
In one of my experiment I have also melted the heater to the state when part of it was basically in the fluid state. Actually I think that it happened because excess heat occured, but I was unable to measure temperature in the center/core of the reactor. So it is important to check implication. If the heater failed because of excess heat inside the tube or failing heater caused excess heat. From your photos I think that it was the first option so there was real excess heat.
I am looking forward for the experiments with AC. It would be nice to find if there is any difference between applying AC or DC to the heater.