Seismosoft | Earthquake Engineering Solutions

Hi Seismosoft Team,

1. How do we determine an appropriate time-step for non-linear time history analysis ? I will be employing IDA for an assessment of the collapse capacity.

I've heard many thumb rules from people and they usually relate the analysis time-step to the time-step at which the ground motion record is sampled. E.g., one rule is to use one-tenth of the time-step at which the ground motion is sampled.
However, I want to have a rational explanation for the selection of analysis time-step, which is dependent on the integration scheme, rather than thumb rules in order to have a sound justification.

2. Does the analysis time-step has to do anything with the period of vibration of the structure ? If yes, can you please let me know how ?

Thanking you in anticipation.
Muhammad Rashid

Hi Rashid. While this question is not really a SeismoSoft question, but more of a structural dynamics theory question, I will offer up the following:

Seismospect adopts a default maximum time step of 0.02 times the natural period being investigated for response spectrum generation.
Seismosignal adopts a default maximum time step of 0.02 times the natural period being investigated for response spectrum generation.

You will find values such as 0.05T and 0.04T in the literature and in textbooks.

One approach is to use the smaller of (a) the accelerograms time step or (b) 0.02T.

I have never seen a recommendation of 0.1 times the accelerograms time step.

Perhaps this information will prove useful to you Rashid.

Hi Tim,

Thank you for your kind response. I was asking this question from the perspective of the integration schemes implemented in Seismostruct.

Best regards,
Rashid

The default integration scheme (i.e. HHT) with the default parameters is the superior option, both in terms of stability and accuracy.
You do not need to change these settings.
Seismosoft Support