How can I resolve a warning with load history analysis convergence? 

From the program manual, the section "The load history deflection calculation process” describes how the convergence is done.

Load history deflections are calculated using a multi-step process summarized below. For each load history step, the process is performed separately for the instantaneous change in loads at the beginning of the load history step and the sustained changes over the duration of the load history step, in which the loads are assumed to remain constant:

1. Solve cross section forces
2. For each cross section, calculate curvatures including long term effects and prior load history:
    - Gross cross section curvature (using gross section properties)
    - Uncracked cross section curvature (using uncracked transformed section properties)
    - Cracked cross section curvature (using cracked transformed section properties)
    - Creep cross section curvature (takes into account cracking history of the cross section)
3. Using the calculated curvatures and the tension stiffening model, calculate an “average” curvature for each cross section.
4. For each element in the structure, use the average calculated curvatures for the tributary cross sections to set stiffness factors for the element.
5. Re-analyze the structure with the adjusted element stiffnesses and check for convergence. Convergence is measured by the deflection difference between two iterations as measured at a key node.

Repeat 1-5 for each load history step until convergence.

When there is a problem with convergence, a message like this occurs:

WARNING: An error has been found while calculating load history deflections.  The load history deflection solution did not converge.  The expected deflection error is 1.955 mm 

The Load History calc log gives some specifics about the load step and/or iteration where the warning comes.

If you continue, you can still review the Load Step deflections which may give a clue about the problem area (you might see extra large deflection in one area for example).

Adjusting the Calc options to allow a larger convergence tolerance or more iterations is certainly one “solution”, but adding user rebar to prevent cracking may be the better approach (assuming the design strip layout is good).

See Also

RAM Concept Load History Calc Options TN

Structural Product TechNotes And FAQs