📄 Why are Processing Times longer when selecting “Response node = dynamic load node”

Why are Processing Times longer when unchecking “Response node = dynamic load node”

⚠️

Under Construction


The longer processing time when the "Response node = dynamically loaded node" option is unticked is expected behaviour of the MasterFrame Dynamic Analysis module.

Why the Analysis Takes So Long

When you untick the "Response node = dynamically loaded node" option, the software must perform a dramatically higher number of calculations to determine the response factors. Here's why:

  • Combinatorial Calculations Typically, the "Response node = dynamically loaded node" option assumes that the worst response factor at a given point occurs when the dynamic excitation is applied at that same point. This simplifies the calculation by measuring the response only at the loaded node.
  • Comprehensive Cross-Referencing However, when this option is unticked, the software is instructed to calculate the response at every selected response node for the dynamic excitation applied at every selected loaded node. For example, if you have 50 loaded nodes and 100 response nodes, the software performs 50 x 100 = 5,000 combinations of calculations. This combinatorial increase in operations is what leads to the substantial increase in analysis time.
  • Time Step and Frequency Range: For each of these numerous combinations, the analysis also cycles through the entire defined frequency range (e.g., 100 different frequencies) and performs calculations for each time step in the analysis period.
  • Finite Element Surfaces: If your model includes Finite Element (FE) surfaces, which are common for detailed floor vibration analysis, this further contributes to longer analysis times due to the higher number of degrees of freedom involved. For large models with FE surfaces, analysis times can range from minutes to hours or even days, as you've experienced.

This comprehensive calculation is necessary for specific scenarios, such as assessing the impact of footfall in a corridor on a sensitive area like an operating theatre, where the excitation and response locations are distinct.

How to Speed Up Analysis Without Compromising Accuracy (Significantly)

While some trade-offs might be necessary, here are several strategies to reduce analysis time without unduly compromising the accuracy of your results:

  • Strategic Node Selection:

    • Focus on Relevant Areas: When unticking "Response node = dynamically loaded node", only select the nodes that are truly relevant as loaded nodes and response nodes. Avoid selecting the entire structure as both loaded and response nodes, as this leads to an unmanageably high number of permutations.
    • Define Specific Regions: For example, if analyzing a corridor's effect on an adjacent room, explicitly select only the nodes in the corridor as loaded nodes and only the nodes in the room as response nodes. This significantly reduces the total number of node combinations that the software needs to evaluate.

  • Optimize Modal Analysis (Natural Frequencies):

    • Use "Frequencies in a Range": Instead of computing all natural frequencies (0-1000Hz), which is generally not recommended for large models, specify a targeted frequency range that is relevant to the type of dynamic loading you are simulating. For footfall analysis, a suggested upper range is typically 30-40Hz (3-4 times the 4th harmonic of walking frequency). This limits the number of mode shapes the software needs to calculate, directly reducing analysis time.
    • Careful Mode Selection: While the default is to use all mode shapes identified in the modal analysis (which is recommended for general accuracy), if you can confidently identify and select only the specific mode shapes that significantly contribute to the vertical motions or response you are interested in, this can further reduce analysis time. However, this requires careful engineering judgment.

  • Refine Time Step (for Force Profile, Base Acceleration, or Transient Analysis):

    • Balance Resolution and Time: The time step (interval) directly impacts the number of calculations. A smaller time step provides better resolution but increases analysis time. Adjust the time step to be as large as possible while still adequately capturing the structural response, particularly periodic motions. For initial runs, you might use a coarser time step and then refine it for critical areas or specific nodes.

  • Consider Model Simplification (where appropriate):

    • Analyze Sections: If feasible, consider analyzing individual floors or smaller parts of the structure that are of primary interest, rather than the entire multi-storey frame, as analyzing full multi-storey models is "computationally hungry" for dynamic analysis.
    • 2D/Grillage Analysis: For planar structures or grillages, performing a 2D analysis instead of a full 3D analysis can reduce analysis time by eliminating zero stiffness terms and reducing degrees of freedom.

  • Hardware and Software Considerations:

    • Up-to-Date Software: Ensure you are using the latest version of MasterSeries software. A major improvement to the solver engine in version 2015 significantly reduced dynamic analysis times and memory consumption.
    • System Resources: Dynamic analysis is memory and CPU-intensive. A computer with a faster processor and ample RAM will naturally complete analyses quicker.

By carefully applying these strategies, you can significantly reduce the analysis time for dynamic models, even when investigating scenarios where the load application and response measurement points differ, while maintaining a reasonable level of accuracy for your engineering design.