MrShorty's quite right about the need to look for the underlying physics in solving your problem more precisely. Excel's TREND and FORECAST functions are straight line fits to data and so are not too good with non-linear relationships. The way forward is to transform the data to a coordinate set that results in a (close to) linear relationship then apply the least squares fit.
It is always useful to look at the goodness of fit (e.g. PEARSON) to see if the relationship holds well - graphing is also useful.
A simple transform for the gas equation is to change the temperature from Celsius to Kelvin^-1 - i.e. divide 1 by C+273. This gives a very good fit. Assuming the new value is in cell A21, the following formulae work:
=FORECAST(1/(273+A21),$B$4:$B$18,1/(($A$4:$A$18)+273))
=TREND($B$4:$B$18,1/(($A$4:$A$18)+273),1/(273+A21))
The arguments are a little complex but they allow the coordinate transform without the need to calculate a new column.
Bookmarks