This would probably give an over-stiff model but would have the advantages of being a linear contact model, of not having to prepare the model in any way and being computationally inexpensive. Using this approach will maintain contact between the back faces. So, we could just connect the back faces using bonded surface to surfaces contact as shown in figure 2. If we imagine our flange assembly, the bolts exist just to hold the two flange back faces together.
Indeed, in such circumstances it is not unusual to produce a meshed model that is too large to be able to be handled, in a reasonable amount of time, by the computational resources available. When we try to mesh such assemblies, we can often end up with very high element counts, resulting in many system equations and therefore excessive run times. Many engineering structures are assemblies that are held together using bolted connections and it is quite normal for design engineers to build CAD models that contain many nut, bolt and washer sub-assemblies within a wider full assembly model. Do you model the whole bolt, with washers and nuts, possible even the thread? Or do you use some level of abstraction, some representation of the bolt, which captures the behaviour of the bolted connection, but allows a much more efficient solution? We discuss some of the possible approaches below.
Working with bolted assemblies is one area where this trade off is very apparent. It can take many years of experience of both product knowledge and simulation techniques to understand where to compromise the simulation model to enable it to be solved in the computational resource available while maintaining a reasonable level of accuracy. One of the difficulties for engineers that are new to the simulation world, is that there are often many decisions that need to be made by the engineer, that trades off the accuracy of the simulation versus computation efficiency.
0 kommentar(er)
