DEV Community

Domonique Luchin
Domonique Luchin

Posted on

Setting boundary conditions in RISA-3D: what the manual does not tell you

After six years of designing pipe racks and equipment platforms in the oil and gas industry, I've learned that RISA-3D's boundary condition setup can make or break your structural analysis. The manual gives you the basics, but it doesn't tell you about the gotchas that cost hours of debugging.

Here's what I wish someone had taught me when I started.

The foundation modeling trap

Most engineers set foundation boundary conditions like this:

Node 1: Fx=Fixed, Fy=Fixed, Fz=Fixed, Mx=Fixed, My=Fixed, Mz=Fixed
Enter fullscreen mode Exit fullscreen mode

This works for simple cases. But when you're modeling a 200-foot pipe rack with 40 foundations, fully fixed conditions create artificial stress concentrations that don't exist in reality.

Real foundations have some flexibility. Even concrete pads on good soil rotate slightly under load.

I model most spread footings like this instead:

Node 1: Fx=Fixed, Fy=Fixed, Fz=Fixed, Mx=Spring, My=Spring, Mz=Fixed
Enter fullscreen mode Exit fullscreen mode

The spring constants depend on your soil conditions and foundation size. For a typical 8'x8' concrete pad on medium clay, I use:

  • Mx spring: 50,000 kip-ft/rad
  • My spring: 50,000 kip-ft/rad

This reduces unrealistic moment concentrations at column bases by 30-40% in my experience.

The pin connection mistake everyone makes

RISA-3D defaults new members to "Fixed-Fixed" end conditions. You need to manually change these for realistic behavior.

Bracing connections are almost never fully fixed in real structures. They're typically bolted with 2-4 bolts that can't transfer significant moments.

For diagonal bracing, I use:

  • Start: Pinned
  • End: Pinned

For horizontal bracing between pipe rack bents:

  • Start: Fixed (welded to column)
  • End: Pinned (bolted connection)

This change alone fixed convergence issues I was having on a recent platform design with 150+ diagonal braces.

Temperature effects that the manual ignores

The RISA manual shows you how to apply thermal loads. It doesn't explain how boundary conditions interact with temperature changes.

I learned this the hard way on a 300-foot pipe rack in Texas. Summer temperatures reach 105°F, winter can hit 20°F. That's an 85°F swing.

With all foundations fully fixed, thermal expansion created massive fictitious forces. Some columns showed 200+ kip compression just from thermal effects.

The fix: release one end of long structures for thermal movement.

For pipe racks longer than 150 feet, I model one end with sliding supports:

Expansion end: Fx=Free, Fy=Fixed, Fz=Fixed, Mx=Spring, My=Spring, Mz=Fixed
Fixed end: Fx=Fixed, Fy=Fixed, Fz=Fixed, Mx=Spring, My=Spring, Mz=Fixed
Enter fullscreen mode Exit fullscreen mode

This matches real construction where expansion joints or slotted bolt holes accommodate thermal movement.

Load path debugging with boundary reactions

When your model won't converge or gives weird results, check the boundary reactions first.

In RISA-3D, go to Results → Joint Reactions → Boundary Reactions.

Look for these red flags:

  1. Uplift at interior supports: Usually means your load combinations are wrong or you have sign errors
  2. Massive horizontal reactions: Often indicates restrained thermal expansion
  3. Unbalanced reactions: Total reactions should equal applied loads

Last month I had a model where one foundation showed 500 kip uplift. Turned out I had accidentally applied wind loads as downward instead of lateral. The boundary reactions caught this immediately.

The integration order trap

This one's subtle but important for dynamic analysis.

RISA-3D uses different integration orders for different member types. The default works for most static analysis, but dynamic problems need consistent integration.

For seismic analysis of equipment platforms, I set all members to the same integration order:

  1. Select all members (Ctrl+A)
  2. Properties → Member → Advanced
  3. Set Integration Order to 3 for all members

This ensures consistent mass distribution and mode shapes.

Soil-structure interaction reality check

The manual treats soil springs as simple linear springs. Real soil behavior is more complex.

For equipment platforms on pile foundations, I model the soil-structure interaction in stages:

Stage 1: Fully fixed foundations for initial sizing
Stage 2: Replace with nonlinear springs based on pile analysis
Stage 3: Apply working loads and check serviceability

The spring constants come from separate pile analysis in programs like LPILE. But even approximate springs give more realistic results than fully fixed conditions.

Quick validation checklist

Before finalizing any model, I run through this boundary condition checklist:

  • [ ] Foundation springs match soil conditions
  • [ ] Pin connections used where appropriate
  • [ ] Thermal expansion accommodated for long structures
  • [ ] Boundary reactions are reasonable
  • [ ] No artificial stress concentrations at supports

You can catch 80% of modeling errors by questioning whether your boundary conditions match the actual construction.

Your turn: next time you set up a RISA model, try the foundation springs instead of fully fixed conditions. Compare the column base moments between the two approaches. You'll see the difference immediately.

What boundary condition issues have you run into? Drop your questions in the comments.

Top comments (0)