I have also observed two semicircles in Nyquist plots. How do I relate this with Z'' vs log(frequency)? What might be the origin for these two semicircles in this type of material? How is this related to AC conductivity?
For polycrystals the two relaxations are usually attributed to charge conduction in grain cores and grain boundaries. You can confirm that by comparing the values of capacitance associated to each peak, or through examination of the imaginary part of the electric modulus vs frequency plot. A relaxation related to grain boundaries should not result in a peak in M'' spectra.
Also, the two semicircles in Nyquist plots are exactly the same relaxations you are observing in Z'' vs f plots.
As far as I know impedance spectroscopy is not enough to determine the type of charge carriers. Other techniques are needed in order to do that. You can, however, obtain more details about the conductivity, such as carriers mobility. See for example: