John wrote:

Any ideas for identifying which union it might be.

turn all rads off except the tricky one, see which pipe from the
manifold has (at least some) flow warming it?