Designing A Connection

by Don Friedman on May 9, 2016

A large percentage of any college course on steel design teaches students how to design connections of various types, most commonly beam-to-beam and beam-to-column connections. The topic of connection design is so ubiquitous that the AISC created a steel-connection sculpture that has been built on many college campuses as a teaching aid, showing multiple connection types. For example, this sculpture was built at Ohio State by the John Edwards Steel Company:

So, how many engineers does it take to design a connection? Most recently in our office, five, and it’s not because we’re especially dumb.

Design is almost always taught to both architecture and engineering students with new-building examples because they are simpler. Working in existing buildings involves dealing with constraints that make some ordinary construction methods difficult or impossible. Our recent example involved a change to the penthouse of a steel-frame apartment house constructed in the 1920s. If we simply looked at the frame – if we treated the problem as an abstraction of steel design only – all we needed to do was install one beam and remove another. But the existing building is not simply a frame. It is, rather, an assembly of many materials and systems that houses hundreds of people.

The constraints on our design included an existing roof that could not be penetrated (the recently-applied membrane is still in warranty), masonry walls that could not be penetrated (facade repairs have recently been performed), limits on steel size that can be carried in the freight elevator, and limits on noise and dust. That list is challenging enough, but there are ripples outward from each constraint. For example, leaving the masonry intact meant that the new beam could not simply run straight across at the desired height, right below the roof slab. It had to be slanted, or kinked, or in some other fashion change its height. But slanted or kinked beams develop horizontal thrust under gravity loads – don’t blame me, blame physics – and therefore need more complicated end connections.

After some brainstorming – which in an open-plan office means everyone around joins in, which is how we end up with five people talking about one connection – we think we have a solution. Our beam has become a small frame, so the loads are delivered and the connections are made at the floor level rather than the ceiling. It requires a bit more steel, but it can be constructed without violating any of the constraints, and that makes it good renovation design.

Previous post:

Next post: