As the weather warms in my part of the world the family and I spend more time outside. After having dinner outdoors one night the children petitioned my wife the following day to let them eat breakfast outside, too. These are the days that help us spend a lot more time together working and playing: Bonding.
Bonding seems almost natural to some of us. But it doesn't always come so easy for those of us using Simulation. Despite the default global contact condition of Bonded, why won't certain bodies and faces bond??? The reason exists because of the different nature of the elements in-use. Solid elements are three-dimensional. Shells are two-dimensional. And Beams one-dimensional. And bonding, along with other contact conditions, needs to be applied from a source to a target; the latter always being a face.
For example, here's a simple structure that incorporates solid brackets, SolidWorks beams, and a planar surface (meshed with shells). This is what we call a "mixed mesh."
The brackets and surface body are modeled atop the weldment. In otherwords, they are touching faces. But, when ran with the Direct Sparse solver, one is greeted with the following: 
Are the restraints inadequate? No. This slightly mis-leading error is an indication that bonding is not being respected. The solution is an oft-repeated mantra around my office: Local bonded contacts will set you free.
Ironic, isn't it?
Local contacts trump contacts between components and anything set Globally. In this case, however, the software cannot resolve the apparent contact. So local bonding is needed. How and where?
A contact pair is created between the face of each body and the beam in play.
The brackets are self-explanatory: each bottom face of a bracket touching/weled to the beam needs to have a local bonded contact. (It is possible to make only one contact for both these brackets, setting the target/face set 2 to both bracket faces.... while this will work it is poor documentation and I've seen cases where the bonding still fails so I do not recommend it)
The shell crosses over two beam bodies. Again, I could satisfy this with one local bonded contact but instead chose two: one beam per the shell face.
Once the local bonding is satisfied you should get something like this: 
See? Bonding is good for the family and good for simulation.
As a side note, I want to make you aware of an option whilst creating your weldment members in SolidWorks. Please be sure to clear this checkbox option so that your joints are calculated correctly in your Simulation study: 