|
Yes. I was thinking, why am I geetting alarmed for a Bank-1 error when bank-2 seems to have more wear?
Possibilities?
1. If you look at the "Actual Cam1-Actual Cam2" graph, you'll notice that the data is skewed towards the positive side. This suggests that cam1 is more likely to have a larger value than cam2. You can also see this in the distribution of points away from zero. I wonder if the Bank1 exhausr chain variocam unit is slow to return to 0 (and needs to be replaced).
2) Why does Cam1 have such a large negative deviation number when we don't see it in the numbers. My suspician is the Porsche may be more sophisticated than we give them credit for.. I suspect that "Deviation" is not just a reference name, but rather a statistical deviation calculated within the DMU. If this is the case, than enough data points far enough from the mean will trigger the alarm. I'll have to check this in the data when I get home.
|