My multi-bolt wood connection will surely have ductile behavior; I only need to check yield modes III and IV, right?
An engineer would be negligent to assume that all bolted connection geometry in wood design experiences ductility and only select yield modes can be checked. While there are two fundamental mechanisms of deformation (i.e. wood bearing (crushing), and plastic hinge formation in fastener), the latter does not occur nor guarantee load redistribution to the engineer in a multi-fastener connection, and requires intentional detailing to elicit such behavior. All five of the yield modes must be checked to determine which one is controlling, and there is a multitude of geometrical effects that can determine which one controls your design specifically.
The NDS Yield Mode Equations are based upon the European Yield Model (EYM) framework. Which is empirically built around the two assumptions noted above, of fastener strength attainment:
- Compressive strength of the wood main member is surpassed (Modes I or II).
- One (Mode III) or two (Mode IV) plastic hinges form in the fastener.
While this failure model is plastic, the ductile behavior assumed only develops if the geometry allows said plastic hinges to form in the fastener(s). Meaning that if the geometry minimizes the chance for the fastener to yield, the controlling behavior becomes crushing of the wood main, or side member(s). When bearing on the wood (crushing) controls the design, the connection behavior becomes brittle, and the connection can begin to split locally or lose capacity suddenly.
Design choices engineers make that might negate the ductile behavior of their connection can take form of the following:
- Thin side members relative to fastener diameter.
- Shortened edge distances or end distances.
- Low dowel bearing strength relative to fastener yield strength.
- Fastener penetration too short to develop plastic hinges.
- Multiple fastener rows violating row spacing requirements.
A connection can be designed safely for either controlling behavior, this is not meant to say that all connections must be controlled by yield modes III or IV. Simply that the engineer must design for all yield modes, knowing which one controls their connection, not assuming that any one controls without verification.