Stages [FEA]

A Construction Stage represents one step in the erection sequence. Each stage inherits the deformed, stressed state of its prior stage, activates or deactivates parts of the structure, applies its loads, and passes the resulting state forward.

This sequencing is what makes a staged analysis different from a single all-at-once solve: the structure that carries each load is the structure that existed when that load was applied. A girder erected simply supported and later made continuous carries its self-weight on the simple span forever, regardless of what the finished structure looks like — and only a staged analysis reproduces that.

Stages control the solution and carry the accumulated state. They generate load only through the cases attached to them.

Stage

Prior Stage: The stage whose end state this stage begins from.

Engineering influence. This defines the sequence. The first stage has none; every other stage names its predecessor, forming a chain. The chain is the analysis order, so a wrong prior stage silently reorders construction — applying a load to a structure that does not yet exist, or missing the locked-in forces from a step that should have preceded it.

The chain does not have to be linear in intent: for transient loads such as wind or live load, select the final permanent stage as the prior stage, so the transient case acts on the completed structure and does not feed its effects into subsequent permanent stages.

Construction Day: The day, on the project timeline, that this stage occurs.

Engineering influence. This drives every time-dependent effect. The interval between a member's casting day and the day it is loaded determines its modulus and its creep; the interval between stages determines how much creep and shrinkage accumulate between them. Days that are wrong relative to one another misstate long-term deflection and the redistribution of force between stages — and because creep is strongly nonlinear in age at loading, errors early in the sequence matter more than errors late in it.

Days must increase along the chain. A stage dated before its predecessor is not physically meaningful.

Temperature: The ambient temperature during this stage.

Engineering influence. Used by the time-dependent calculations, where curing temperature accelerates maturity, and as the reference condition for the stage. It is not a thermal load — a temperature change applied to the structure comes from a Temperature Load.

Humidity (%): The relative humidity during this stage.

Engineering influence. A primary input to CEB-FIP creep and shrinkage: drier conditions produce substantially more shrinkage and more creep. The default of 80% suits a temperate exposed environment; an enclosed or arid environment warrants a lower value and produces markedly larger long-term effects. This is a property of the stage, not of the material, which is where engineers often look for it first.

Const. Method: The construction method used for distributing effects within the stage.

Engineering influence. Describes how the work in this stage proceeds, which affects how load and stiffness are introduced across the elements activated in it — whether they all come into service together or progressively. The distinction matters most for long stages activating many elements at once.

Is Active: Whether the stage is included in the analysis.

Engineering influence. An inactive stage is skipped silently, and because stages form a chain, skipping one changes what every later stage inherits — the structure and the locked-in forces they start from. This is more consequential than deactivating an ordinary load case.

Load Type: Classifies the stage's loading for combinations and code checks. It does not change the analysis; it changes which combinations pick the stage up.

Time Dependent

Time Dependent Code: The code basis used for the time-dependent calculations.

Time Dependent Elastic Modulus: Whether the modulus develops with concrete age in this stage.

Concrete Creep Effect: Whether creep is computed in this stage.

Concrete Shrinkage Effect: Whether shrinkage is computed in this stage.

Steel Relaxation Effect: Whether prestressing steel relaxation is computed in this stage.

PT Losses from Structure: Whether elastic-shortening losses are computed in this stage.

Each of these must be enabled both here and on the material to take effect. Enabling it in only one place is the usual reason a staged model shows no creep or shrinkage at all — and the symptom is simply that long-term results equal short-term results, with nothing to indicate why.

Engineering influence, all five. Creep increases deflection over time and redistributes force toward members that creep less, which in a segmental or staged structure changes the final moment diagram substantially compared with the as-built one. Shrinkage produces tension in restrained members and differential strain in composite sections. Relaxation and PT loss both reduce effective prestress over time. Enabling them costs solution time; disabling them understates long-term deflection and, more importantly, misstates the final distribution of force — which is usually the reason for running a staged analysis in the first place.

Creep of Tensile Axial Force: Whether creep is applied to members in axial tension. Defaults to off. Concrete in tension behaves differently from concrete in compression and is often cracked, so applying compression-derived creep to it is not generally appropriate; enable it only where the tensile creep behaviour is genuinely intended.

Nonlinear

Nonlinear: Whether this stage is solved nonlinearly.

Engineering influence. Needed where the stage involves cables, gaps, one-sided members or significant geometric effects — a cable-stayed erection stage analysed linearly ignores the cable behaviour that governs it. Note this is set per stage, so a sequence can solve cheaply where behaviour is linear and nonlinearly only where it matters.

Maximum # of Iterations: The iteration limit within each step of this stage. Raising it helps a slowly-converging stage succeed; it does not improve the accuracy of a converged answer.

Force Tolerance: The out-of-balance force accepted as converged. Loosening it eases convergence while leaving residual out-of-balance force in the state that is carried forward to every subsequent stage — which is why sloppy tolerances are worse in staged analysis than in a single case.

Verification

• Step through the stages and confirm the structure activating at each one matches the erection sequence, using the 3D view.
• Confirm construction days increase along the chain and that the intervals match the programme.
• Check that each stage's prior stage is what you intend, particularly for transient-load stages, which should branch off the final permanent stage rather than sitting in the middle of the chain.
• Compare the final staged result against a single all-at-once analysis. They should differ; if they are identical, the staging is not taking effect.
• Run with time-dependent effects on and off and confirm long-term deflections differ. No difference means the switches are not enabled on both the stage and the material.
• Check that locked-in forces carry forward: a member's force at the start of a stage should equal its force at the end of the prior one.
• Confirm each stage converged rather than hitting its iteration limit.

Common mistakes

• Time-dependent effects enabled on the stage but not the material (or the reverse), so nothing happens and long-term equals short-term.
• A wrong prior stage, silently reordering construction.
• Transient load stages inserted into the permanent chain, so wind or live load effects are inherited by subsequent construction stages.
• Construction days that do not reflect the real programme, misstating creep — with early errors mattering most.
• Humidity left at the default in an environment that is materially drier, understating shrinkage.
• Deactivating a stage and not realising every later stage now inherits a different structure.
• Loosening the force tolerance, carrying residual out-of-balance force through the whole sequence.
• Comparing staged results against a one-shot analysis and assuming the difference is an error. The difference is the point.