IMHO it is not normally necessary to go to these lengths modelling a dummy leg. You have a whole lot of approximations in your calculation already, since wall thickness and thus weight and stiffness throughout may vary by +/-12.5% dependent upon material and code. In addition, every restraint has flexibility which we nearly always omit.
Since attaching a dummy leg will probably destroy any flexibility in an elbow, ignore the elbow. Model a direction change and to be conservative, apply the correct sifs to both the attached elements. Either attach the dummy leg and support ,or simply support the intersection point. That depends on temperature mostly.
Complex modelling is certainly justified sometimes, but by no means always.