jmg,
I don't intend to reopen the debate of calculating with units versus unitless with you again - let's just say we agree to disagree on that particular issue!
With respect to the basic calculation - the problem first arose in a much more complex worksheet than the one I posted - I just wanted to make the simplest possible demonstration of the nature of the problem.
For the record, the mini-calculation I posted just calculates free-stream air pressure, given by:
p = 1/2*rho*v^2
Density of air (rho) is typically taken as 1.2 kg/m3
For a free-steam velocity of 40, 50 and 60 m/s, the corresponding free-stream air pressures are 960 Pa (0.96 kPa), 1,500 Pa (1.5 kPa) and 2,160 Pa (2.16 kPa) respectively, so the first half of my worksheet (without in-line evaluation) gives the correct solutions. (And yes, these are relatively high wind velocities - but that is what we have to deal with when designing structures for cyclone loading!)
In the second version of the worksheet, performing the in-line evaluation of the velocity vector from base units of m/s to km/hr over-writes the original m/s numeric values in the vector with their km/hr equivalents, but the vector as a whole still carries units of m/s, so the subsequent free-stream pressure calculations are wrong.
Hope this clarifies things. (And yes, I know that if I had worked unitless throughout, and stuck to base SI units everywhere, this problem would not have surfaced at all, but I would rather be able to do my calculations in whatever units work for me and / or the client and / or the source reference I am working from.)
