VisualAnalysis Tutorials: Combined Members Design
This tutorial explains modeling differences and assumptions that VisualAnalysis makes when analyzing combined and split members. Below are a series of identical beams laid out to provide a side by side comparison of combined vs. split member design results. The top beam is a single member, the middle beam is split into five members, and the bottom beam is comprised of five combined members. The system is loaded with a uniform dead and live loads. Each beam span is 20 feet and was given its own design group pictured below.
VisualAnalysis uses the end points of an element as an initial determination of its braced length. This means that splitting and combining members can have large affects on a models design results. Engineers need to use good judgment and intuition to ensure that their model is producing the desired behavior.
In the example above, VisualAnalysis treats all combined members as if their braced length is 20 feet and all split members are treated as if their braced length is 4 feet. The model was analyzed and each group was designed. Results for each design group are shown below in the report.
The top single beam and bottom combined beams, having much longer braced lengths, yielded larger sections than did the split members. Recognize that the split members are not actually braced every four feet so the design results are not accurate.
Situations arise when members must be modeled as split or combined. In these cases, the members braced length can then be controlled under the parameters for the design group.
Beam bracing is specified for the top and bottom flanges of the beam. It can be entered as a specific length, continuously braced, unbraced, or braced at fractional points along the beam.
Let us now model the split beams as completely unbraced. Bracing parameters of the model must now be changed to achieve proper and efficient design results. This can be done without having to change split or combined members. For design group 2, define the bracing as a "Set L" of 20 feet. With this amendment and one design iteration, we arrive with the results of all three design groups pictured in the lower left report.
Results become almost identical. Ambiguities in the depicted results stem from the Cb coefficient. This variable is calculated based on the moment rate of change across the member. Although we have changed the unbraced length of the members of Design Group 2 to 20 feet, the software is unable to determine where the Cb calculation should start, end, and on what size of element the calculation should be made over. It is because of this that the Cb value for each individual element of a split member is calculated.
It may be more efficient to model split members as combined.
Splitting or combining an element can affect VisualAnalysis design results in situations where deflection limits control. For the sake of easier comparison, all previous design groups were changed to a W14x48 section.
Member deflection limits can be controlled through the design group parameters by using choosing either span ratios or actual deflections.
The "Member Span Ratio" and "Member Deflection" limits measure an elements deflection in relation to its end points. This means that the dmember value shown on the left is used in deflection limit calculations. The "Total Span Ratio" and "Total Deflection" limits use the overall deflection that a member experiences in relation to the global structure. This means that the dtotal value is used in deflection limit calculations.
Let us assume that these beams must meet an L/400 "D + L" deflection limitation. When all three design groups are set using a "Member Span Ratio" limit VisualAnalysis yielded the following results.
Notice that using a "Member Span Ratio" deflection limit for split members does not give proper deflection criteria. The "member deflection" limit also would be inadequate. In this situation the dtotal value (as seen in the previous picture) needs to be taken into account for correct deflection limits.
Below are the design results with the middle span's deflection limit changed to a "Total Span Ratio". Each individual member is a fifth of the length of the entire span so to achieve the same deflection limitation, a value of L/80 is required.
Note that this method takes into account deflection of the endpoints of the members as well as the deflection of the members with respect to their endpoints. This means that axial deflection of modeled columns will be included in the deflection calculations. Axial deflection should be negligible, so it shouldn't cause a significant error.