Thanks for the post bear.
If ICE is the consideration, three separate design approaches would likely produce three anomalous solutions to the goal. Redundancy works well if in a "fallback" position, so I like your anomalous heat solution.
You also raised a valid point vis-a-vis the statistical credibility of such a design. Placing confidence in the divergence as you put it. (sorry to keep quoting the deleted post.) Here I think the data collected by Northwest can come in handy. Airbus has a baseline to compare against any data that may be generated from experiments, if they so chose. Who knows, even the beans may be convinced.