If these were my units, I'd be looking to resample both of them - if it can be done safely. Then you can see what the latest values are, what variation there is in each gas, and beging to see if one of them has a naturally high ethane level. It may be naturally high, or there may be early...
Start with the data and find a decent ERM - otherwsie you'll be chasing processes (functionality) and incurring larger and larger expense to make it work in the future. A good database design at the outset at least lets you build the functionality on a solid supportable base.
After some useful discussion - the need for trends, fault history etc, a number of questions still remain...
Why do you think the ethane should be ~10?
How many DGA results do you have for this unit, and were they all analysed at one lab?
What is the load history on this unit - full 60 MVA or...
