Hi,

When i am doing a WRC evaluation using CAESAR, I am being asked to input direction cosines for nozzle and vessel. My vessel is a vertical vessel and hence cosine in Y is 1. My nozzle is on top of dished which is also in vertical direction. Hence I am making my nozzle cosine also 1. But I get an error "Directional Cosine of the Vessel and Nozzle Should Be Orthogonal". What should I do? my nozzle is not orthogonal to vessel.

You say WRC. Is this WRC 107 (or the similar WRC 537) or WRC 297? WRC 297 does not address attachments to spherical shells. Your head is more of a spherical connection rather than a connection to a cylinder.
You say evaluation. These WRC documents do not evaluate stress, they provide a means of calculating (estimating?) stress.
Do you wish to model the relative flexibility of this connection? While both of these documents calculate stress, only WRC 297 provides estimates on shell flexibility. If you choose the WRC 297 nozzle flexibility calculation the CAESAR II "flexible nozzle" input requires the orientation of the (supposedly cylindrical) shell. Some would use a different reference vector for the vessel just to get the model to run. But this will improperly apply some of the stiffness terms. I would suggest using the "custom" nozzle definition so that you can correctly define the flexible (and rigid) directions.

I am Using WRC 107 module. Yes, as you mentioned its a connection to dished end. I want to confirm if the dished end nozzle is over stressed or not due to the loads transmitted by pipe. I hope I make sense.

For attachments to spherical shells the vessel orientation is arbitrary. For example, a vertical nozzle off the top head could have the vessel orientation as X or Z. The axial symmetry of the shell allows this - as long as you are in the center of the head.

Thank you.

Fortunately the nozzle lies in the center of head. What if I have more than one nozzle. If I do WRC 107 calculations is it valid for nozzles which are offset from center of vessel/head ?

An oft-asked question. That depends...
There is much uncertainty in trying to push these algorithms away from their intended use. If you have the time, search this Forum.
Certainly there is FEA software built specifically for these connections (e.g., NozzlePRO by Paulin Research Group)