Floorbeam Transverse LL Analysis - Factors Affecting Runtime
Date | November 2025 |
Product | AASHTOWare BrDR |
Version | All versions |
Description | Floorbeam Transverse LL Analysis -- Factors Affecting Runtime |
Fixed | N/A |
Floorbeam Transverse LL Analysis – Factors Affecting Runtime
Floorbeam Transverse LL Analysis – Factors Affecting Runtime
The performance of a transverse live load analysis depends on multiple parameters including total travelway width, transverse loading parameters, number of vehicles, and selected outputs. The purpose of this article is to provide insights into the impact of analysis parameters on runtime as well as suggestions for reducing analysis runtimes while still obtaining reasonably accurate results. In this article, the transverse LL analysis of a 105.5’-long floorbeam as part of a Truss-Floorbeam-Stringer (TFS) model is discussed, but a similar process is used for 3D FEM transverse LL analysis as well.
When the travelway width exceeds 70 ft, the warning message shown below is displayed in the Analysis Log pane. The actual amount of “processing time” is unknown and highly dependent on analysis parameters used.
If the travelway width is less than approximately 72 ft, transverse analysis runtimes are generally reasonable when default transverse loading increment input parameters of 2’ and 4’ are used.
However, for travelway widths exceeding 72 ft, runtimes can increase exponentially with additional width.
Defining More than One Travelway:
Separate travelways are not recognized by the transverse LL loading algorithm. The algorithm will always create a single travelway extending from the left-most edge to the right-most edge of all travelways entered. This behavior is discussed in the ‘AASHTO LRFD/LRFR Superstructure Method of Solution Manual’, as shown in the excerpt below:
Three travelway definition scenarios are shown in the graphic below. All three will result in the same transverse LL analysis runtime and floorbeam actions (moments, shears, displacements and rotations). For Scenarios 2 & 3, the program will use one combined travelway equivalent to Scenario 1.
Effect of Overall Travelway Width & Transverse Loading Parameters on Runtime:
Analysis runtime is significantly affected by the overall width of the travelway and the ‘Transverse loading’ analysis parameters entered in the ‘Analysis’ tab of the superstructure definition window. The default transverse analysis parameters are ‘2’ for the ‘Vehicle increment in lane’ and ‘4’ for the ‘Lane increment’. These default values will generally produce accurate analysis results with reasonable execution times for travelway widths less than 72-ft (6 lanes of traffic).
Using a value greater than ‘2’ for ‘Vehicle increment in lane’ will result in the 10-ft uniform load portion not being shifted within a 12-ft wide traffic lane. Using a value greater than ‘12’ for ‘Lane increment’ will result in gaps created in subsequent rounds of lane placement within the travelway and could lead to inaccurate analysis results.
Analysis results are more accurate with smaller transverse loading parameters, but at the expense of runtime, which can increase substantially for a wider travelway.
The accuracy of the analysis and the runtime is directly related to the number of unique transverse position IDs generated. Each transverse position ID represents the transverse location of one or more traffic lanes along the length of the floorbeam.
For a floorbeam transverse LL analysis, an intermediate output file named ‘Span LL Summary.txt’ is generated. It can be found under the ‘Live Load’ subfolder for each floorbeam analyzed.
Below is a screen capture from the beginning of the ‘Span LL Summary.txt’ file showing the building of nodal loads applied to the floorbeam FE model for each transverse position ID.
Below is a screen capture from the body of the ‘Span LL Summary.txt’ file showing the computed action result for the individual position IDs. The excerpt below is for shear on the right side of FE model Node 1.
The table below shows the typical impact of travelway width and transverse loading parameters on analysis runtime. The runtimes given in the table are for an LFR analysis of a single EV2 vehicle. The computer used was a Dell i7 @ 2.3 GHz 11th gen with 64 GB RAM and 2 - 1TB SSDs laptop. Runtimes will vary by computer specs.
As can be seen, the number of unique transverse position IDs generated for evaluation, as well as runtimes, increase exponentially for travelway widths over 72 ft with the most significant impact for the default transverse loading parameters of 2’ and 4’. When the transverse loading parameters are increased, the number of transverse position IDs and runtimes are reduced substantially using transverse loading parameters of ‘3’ and ‘6’.
A question mark ‘?’ for the runtime indicates the analysis was terminated after approximately 30 minutes, which was enough time to know how many transverse position IDs were generated.
For a transverse floorbeam analysis, the reason for the significant runtimes is the fact the program computes and tracks a maximum and minimum value for six different actions for each node in the floorbeam FE model for each transverse position ID.
In the floorbeam example used here, the simply-supported floorbeam has 29 FE nodes -- 11 nodes are generated for 1/10 points, 14 nodes for the stringers (1 at each stringer), 2 nodes are generated at floorbeam cross-section changes (top and bottom flange transitions), and 2 nodes are created at user-defined POIs.
The transverse live load analysis will result in the following approximate number of actions being considered for one floorbeam per vehicle: 2X for max & min, 6X for each separate action (moment-left, moment-right, shear-left, shear-right, Y displacement & Z rotation), multiplied by the number of FE nodes, and then multiplied by the number of transverse position IDs. For the 101.75’-wide travelway case with transverse loading parameters of 2’ & 4’, the number of action values computed is approximately:
2 max & min x 6 actions x 29 FE nodes x 594,364 position IDs = 206,839,000 values
Effect of Fractional Travelway Width:
Another consideration is the effect of a travelway width that results in fractional number of lanes. This aspect is shown in the table below for the 80.875’-wide travelway.
The number of transverse position IDs generated for transverse loading parameters of 2’ & 4’ is 42,408. This number is 2X the number of transverse position IDs for an 80’-wide travelway. So, a travelway width only 0.875’ wider results in the doubling of transverse position IDs and double the runtime.
The reason this happens is that the transverse loading algorithm will place 12’-wide lanes progressing left-to-right and then also right-to-left. A fractional travelway width can result in a small offset distance between the lane positions between starting from the left side of the cross-section versus starting from the right side.
Effect of Number of Vehicles Evaluated:
The runtime can also be affected by the number of LL vehicles included in the analysis. The runtime will be at least linearly related to the number of vehicles (i.e., the runtime for two vehicles will be at least 2X that for one vehicle), but could be non-linear for very wide travelways with smaller transverse loading parameters that result in a large number of transverse pattern IDs being evaluated as well as the use of more vehicles. Including multiple vehicles in an analysis requires additional virtual memory usage. The performance of virtual memory degrades exponentially with the amount used.
Effect of Analysis Output Options:
Analysis runtimes will increase if analysis output options are selected because these output files can be large and require significant disk I/O processing.
Unselecting the analysis output options will reduce runtime. Keep in mind, though, these intermediate reports can provide valuable information and insight into analysis results.
Accuracy of Analysis Results:
Fewer transverse position IDs generated and evaluated may lead to reduced accuracy of the analysis results. There is a trade-off between runtime and accuracy, and the question becomes, is the additional runtime worth the increased accuracy? Shown below are LL moment and shear plots for the 105.5’-long simply-supported floorbeam for the EV2 vehicle using two different transverse loading parameters for the widest travelway width of 101.75’. For the analysis that used 2’ & 12’ for the transverse loading parameters (13,120 transverse position IDs), the maximum moment is 3.7% higher than the analysis that used 3’ & 6’ for these parameters (3,192 transverse position IDs). For the shear analysis, though, the maximum end shear is the same at 192.19 kips, but the intermediate shears along the length of the floorbeam vary with the analysis that used 2’ & 12’ parameters producing values up to 5% larger.