Make a flash curve of the bottoms product and make corresponding reboiler duty and relief rate calcs for each vapourisation step. For a narrow boiling mixture I wouldn't bother tho'.
What is optimistic about the assumption of the normal bottoms composition during relief? Doesn't the manual heat balance method result in normal reboiler duty / latent heat of top tray liquid for your case?
The "relief Systems Handbook" of Parry gives criteria for thermal relief. Furthermore there are some articles that give guidance ("What you should know about liquid thermal expansion", HP November 2008, and "Decide whether to use thermal relief valves", CEP December 1993).
If you're doing a project for AstraZeneca you should use their standard as long as it is more stringent than applicable industry standards / design codes.
From you question you seemed to have the concepts mixed up. I'm sure in the standard a statement is made (or was made) that the vapour relief required is equal to the total rate of heat absorption divided by the latent heat of vapourisation, or W = Q/dH.
Equation A-7 is for vessels which do not contain liquid while W = Q/dH is for vessels which do. The A-7 equation is still part of the latest version (2008).
