I am rather embarrased to even ask this question as I believe I know the answer is a basic statics question, but need some verification. I am doing a remodel on a one story condo - enlarging the "U" shaped living room / Kitchen to become one large room. The roof consists of a mansard that extends over the exterior 10'-0" plate height bearing wall by 6-1/2 feet of cantilever. There is a return soffit of 2x6's at 24" o.c. with a lath and stucco finish. The mansard roof ends 7.83 feet behind the exterior bearing wall and flattens out to form a flat roof (probably 1/4" in 12" slope) in the center of the home. The high roof framing including the sloping mansard is supported by built-up beams and the interior "U" shaped alcove bearing wall. The roof drops to a 10-foot ceiling in the living room, and an 8-foot ceiling in the Kitchen / Dining room.
The goal is to remove the interior walls that are essentially non-bearing but contain built-up columns supporting beams supporting the two levels of ceilings and the high roof with mansard. The slope of the mansard is about 3.6:12 (it is 8 feet to the bottom edge of the overhang and meets the bearing wall at a plate height of 10-feet).
I am not touching the high roof or the mansards other than to fill in the alcove leading out to the patio and opening the Kitchen/Dining room into the Living room. To do this I will need to introduce one post and approximatly four beams - one of which must be steel because of the loads.
The problem is trying to calculate the reaction of the mansard framing. I have two ways to consider it. The lightest load reaction (at the end opposite the cantilever) occurs if I assume the mansard is just simply supported. I at each end and take the full weight into my analysis. These roofs were designed over 30 years ago and since then they had trouble with creep and with the HOA replacing wood shake with tile. In this house the deflection is not noticible and since I am not addressing it as part of the design problem, then I would end up with a much higher reaction that will transfer to the new framing I am adding.
To make matters worse, there is a hip that I am not working on but will probably result in an uplift at the interior unless I disregard it and consider it simply supported. I am going to be adding about 500 pounds if I treat it as simply supported and in my mind it is better to do this than try to be exact when I have no idea who or how it was originally supported. Based on the slope of the mansard, the hip extends about 9-feet and almost 10-feet back.
I may have answered my own question since at the root of all of this is the basic premise that I am designing a beam and rather than apply a lighter reaction at some point along the new beam, I put a more conservative load on it and make it stiffer. If you were the plan checker, would you be more likely to accept a recreation of the original mansard loading (including any vector load transfers due to the soffit bracing) or allow me to simplify the design by being more conservative. The owner (my client) is more than happy to be on the conservative side of the equation because he knows how poorly these homes were originally constructed.
I think I am getting gun shy lately and am prone to be conservative.
Dennis
California Professional Engineer
Structural Engineering Consultant