Pushover Cases [FEA]

A Pushover case is a nonlinear static procedure: the structure is pushed by an incrementally increasing lateral load or displacement until a target is reached, producing a capacity curve of base shear against displacement. It shows the sequence in which members yield, the ductility available before collapse, and the lateral strength — none of which a linear analysis can reveal.

Pushover controls the solution and reads the model's nonlinear ingredients (plastic hinges, material nonlinearity, geometric stiffness). It generates load through the pattern it pushes with.

A pushover case is only meaningful if the model actually contains nonlinearity — plastic hinges on the members expected to yield, or inelastic line elements. Pushing a purely elastic model produces a straight line that reveals nothing about capacity.

Control

Reference Node: The node whose displacement is monitored and controlled — conventionally a point at the top of the structure, at the deck level or the centre of mass.

Engineering influence. This node defines the horizontal axis of the capacity curve, so the whole result is expressed relative to it. Choosing a node that is not representative — one on a locally flexible member, or one that barely moves in the pushed direction — produces a capacity curve that misrepresents the structure's global behaviour, and in displacement control it can prevent the analysis progressing at all.

Reference Direction: The direction the reference displacement is measured in.

Control CoorSys: An optional coordinate system for the reference direction, so a skewed structure can be pushed along its own axes rather than the global ones.

Engineering influence. For a skewed or curved bridge the governing direction is usually not a global axis; pushing along global X on a skewed bridge pushes partly across the bents and produces a mixed response that is hard to interpret. Several directions are normally checked.

Reference Displacement: The target displacement the analysis pushes to.

Engineering influence. This sets how far the capacity curve extends. Too small a target stops the analysis before the structure yields significantly, producing a curve that looks linear and uninformative; too large a target pushes far past collapse, where results are meaningless but harmless. It should reach at least the code's target displacement for the seismic demand being assessed.

Max # of Steps: The maximum number of increments used to reach the target.

Engineering influence. More steps resolve the capacity curve more finely and follow the changing stiffness through each yielding event more accurately, at proportionally more run time. Too few steps jump over the formation of hinges, so the curve misses the knee where yielding begins and the sequence of hinge formation is lost. If a pushover fails to converge, increasing steps is usually the first remedy.

Initial Load Factor: The load factor applied at the first step.

Load Factor Increment: How much the load factor increases each step.

Engineering influence, both. These govern force-controlled stepping. A large increment reaches the target quickly but can step across a yielding event and fail to converge; a small one is robust and slow. Note that a purely force-controlled analysis cannot pass a peak in the capacity curve — once the structure begins to soften, no increase in load is possible and the analysis stops. Capturing post-peak, softening behaviour requires displacement control.

Displacement-Controlled Steps: Whether stepping is controlled by displacement rather than by load.

Engineering influence. This is the parameter that determines whether you can see the descending branch. Displacement control imposes the movement and reports the force required, so it follows the capacity curve through and past the peak into softening — which is where ductility and collapse behaviour are assessed. Force control cannot do this and terminates at the peak. For any pushover intended to assess ductility, displacement control is the correct choice.

Nonlinear

Nonlinear: Whether the case is solved nonlinearly.

Engineering influence. Effectively mandatory for a meaningful pushover — with it off, hinges never form, materials never yield, and the "capacity curve" is a straight line whose slope is just the elastic stiffness.

# of Iterations: The iteration limit within each step. Raising it allows a slowly-converging step near a yielding event to succeed; it does not make a converged answer more accurate.

Force Tolerance: The out-of-balance force below which a step is accepted as converged.

Engineering influence. Defines what counts as equilibrium. Loosening it eases convergence at the cost of accepting residual out-of-balance force — which in a pushover is particularly misleading, because the residual can masquerade as capacity the structure does not have. Set it relative to the base shear magnitudes in your model.

Settings

Structure Group: The part of the structure active for this case. Leave empty for the whole model.

Is Active: Whether the case is solved. Inactive cases are skipped silently.

Verification

• Plot the capacity curve. It should show an initial elastic slope, a knee as hinges form, and a plateau or descending branch. A straight line means nothing yielded — check that hinges exist and that Nonlinear is on.
• Confirm the initial elastic slope matches the structure's elastic lateral stiffness from a linear case.
• Check the sequence of hinge formation against engineering expectation; hinges should form where the design intends them, typically at column tops and bases rather than in the deck.
• Confirm the analysis reached the target displacement rather than stopping early on non-convergence.
• Re-run with more steps and confirm the curve is stable.
• Compare base shear at yield against a hand plastic-mechanism calculation.
• Check that the reference node's displacement is representative by comparing it with other points on the structure.

Common mistakes

• No hinges or nonlinear members in the model, producing a straight-line "capacity" curve.
• Force control used to assess ductility, so the analysis stops at the peak and never shows the softening branch.
• A target displacement too small to reach yielding, making the structure look elastic.
• A poorly chosen reference node — locally flexible, or barely moving in the pushed direction.
• Too few steps, stepping over hinge formation and losing the sequence.
• Loosening the force tolerance to force convergence, accepting residual out-of-balance force as capacity.
• Pushing only along global axes on a skewed or curved bridge, missing the governing direction.
• Reading results past the point where the analysis stopped converging and treating them as valid.