I am learning by myself structural dynamics but confused about the concept of scaled ground motion data. I can't get the reason why the acceleration data should be scaled. I mean, isn't it better to use the original data "as it is" to compute the response of structures so that the results can reflect actual response of a specific structure?

EDIT : Is it the case that when we design a structure, scaled data should be used, while unscaled data can be utilized when we conduct post-seismic-analysis of a structure?

Could someone please summarize the situations when we should use scaled or unscaled data?

  • $\begingroup$ Typically we design for loads never observed at the specific site. How would you infer that a structure can fulfill the design criteria and resist seismic actions with a prescribed failure probability from the verification that it can withstand a measured acceleration ("original data")? Basically, the latter can correspond to any return-period. $\endgroup$
    – rozsasarpi
    Commented Apr 28, 2016 at 8:27
  • $\begingroup$ Hi, @Arpi, thank you for your reply. I have added some edits to my question, could you please take a look at it? $\endgroup$
    – user123
    Commented Apr 28, 2016 at 14:21

1 Answer 1


A ground motion record (accelerogram) is a measurement with high specificity to the fault that caused the earthquake. The recorded peak acceleration, frequencies, duration etc depend on the fault geometry and movement. Of course there are not two identical faults and every time a fault is activated it doesn't produce the same ground motion. Therefore, the value of designing a structure by one "as recorded" ground motion is small, since it is highly unlikely to encounter it in the future. Typically the design with real ground motions requires a set of them (at least 5).

Why scale a ground motion

In an ideal world a structural engineer would have a vast amount of historical ground motions to select from, for a particular site, when designing a new structure. There would be no need to use any records from other parts of the world. Of course this is not happening because severe earthquakes are not an every-day phenomenon and recording ground accelerations started not to far in the past. There are sites where only one or two good records are available. So the solution is to use available records from all over the world, but then the problem is that the seismic risk is not the same all over the planet. Typically, the seismic code provides us with the expected intensity of the seismic excitation, in the form of peak ground acceleration. A solution is to scale an accelerogram to match the code requirements. A problem arises though, because a distrortion in the frequencies is unavoidable. Therefore, it is better to avoid too much scaling, or alternativly to use artificial accelerograms.

Also, scaling of the ground motions is inherently needed for some advanced types of analyses (incremental dynamic analysis).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.