The second nonstationary random process mathematical model, which is based on data, is the a
modification to the first nonstationary random process mathematical model. To build a model of this kind, a
nonstationary random process mathematical model is set up through fast Fourier transformation and Herbert
transformation based on available original records, and random samples are produced by numerical
simulation. As an example, wind speed samples are simulated using the second model based on the
observed wind speed records.