You could use the unitary circle where the length of a 90deg angle equals half Pi.
For instance, 2x diameter bend radius would have a length of (2 x diameter bend radius x Pi / 2). Units remain unchanged.

Not going to be much different.

5D maybe, but not 1.5 to 2D

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

Use Smacna guide to duct design and interpolate if needed.

Does that apply to copper pipework as well nuuvox000? I wondered about using the inverse square law on the 1D and 1.5D fittings from ASHRAE to approximate equivalent lengths for 2D.

My apologies, somehow I missed the part about it being copper; I thought you were talking about air through duct.

No worries :)

I believe the pressure drop of the custom fitting bends can only be measured through testing, not calculations. In practice, I think you can just use the worst case because it's likely not worth your time to differentiate between the 1.5 and 2.0, especially if you're using a safety factor.

I think you are over-thinking this. You will see that different publications show different K-values and other data and in real-life many fittings get made in a shop and vary from the text-book fitting. Especially the more complex ones. So there is quite some variety. Imagine a Y, or bullhorn etc. Also note, that the tests are made with long duct up and down-stream. In real life you have multiple fittings close to each other and they influence each other. For design, just use the higher pressure drop to be conservative.

Also allow for some field-changes, like routing around a structural element. And your plans may show a 90°, and the contractor will use an assembly of 45° or some other deviation since the sprinkler guy decided to install his pipes before the ducts.

Cameron Hydraulic Data provides a table of "Resistance coefficient K" to calculate the equivalent length of pipe for 90 degree bends of different r/d (bend radius / pipe diameter).

Unfortunately for you, the r/d values are integers, and there is a big difference between 1 and 2.

The K value then gets plugged into a formula that includes the "Friction Factor" from the Moody diagram, which in turn is dependent on pipe roughness and the flow Reynolds number.

How many of those inputs do you know with any certainty? (Hint: Probably none)

Unless pipe bends are a dominant feature it's really not worth the effort of going through the work when all the inputs are guesses.

Thanks all for your input. I'll use worst case for now as it seems like there's a lot I'm missing to try calculate pressure drop at a higher bend radius.