Dear Ahlam Khalil ,

The duration of the recordings must at least last for few cycles of the lowest oscillation frequency. Assuming that the lowest frequency oscillation is f, then the satisfactory observation time must ne at least few periods of T=1/f

In this way you can easily estimate the periodicity of the vibration.

As the observation time increases the frequency of the oscillation will be more sharply located.

Best wishes

Dear Abdelhalim abdelnaby Zekry

thank you very much for your answer.

but I want to know if there is a relationship between the duration of recordings and the period of soil or sea waves or any another kind of seismic waves.

thank you again.

Hi dear Ahlam Khalil

As far as I know, there shouldn't be any rational relationship between the exact (full length) recording and the period of the soil. However, the significant duration that we choose from acceleration spectra based on "normalized arias intensity" might have something to do with a period of the soil. (I am NOT sure, please check)

Dear Ahlam Khalil

According to the SESAME HV user guidelines, they recommend the following condition to be fulfilled : f0 > 10 / lw. for example if the expected f0 is 0.2 Hz so the Minimum value for lw is 50 and Recommended minimum record duration is 30 [min]. see table 1 in the attached file at page 18

Dear Ahlam Khalil

I do agree with Hazem Badreldin it need at least 30 minutes for duration of recording. The longer the measurement the better.

According to the SESAME HV user guidelines, for reliable data processing you must have 10 cycles which depend on the desired frequency and the window length. knowing that during processing the programs consider only the stable windows, so in order to increase the number of these windows you must make an acquisition with a longer duration (more than 30 min) which will allow you to acquire a lot of samples so automatically increase the number of cycles.

Concerning the seismometer, the longer the acquisition time, better is the sensor stabilization.

it is required to do Stationarity test. The stationarity test is a property of time series which states that the value of the variable doesn't change with time.