Yes, optical ray tracing is a simple simulation method. Good results can be obtained if the near field coupling of the antennas can be neglected - so it is well suited for "higher frequencies" where the wavelength is much smaller than the dimensions of the chamber.
It was developed in 1946 for the use in the VHF and UHF range. The advantages are the simple design (a balun is not required) and the large bandwidth. Due to the symmetrical construction the H-plane pattern is circular. Today antennas for a frequency range 1 to 18 GHz are sold. They are called "omnidirectional" but they have many peaks and nulls in the E-plane pattern.
For accurate NSA measurements you need to have also a proper calibration for the different positions and polarizations. That means your antennas should be calibrated at horizontal polarization at 1 and 2 m height and vertical polarization at 1 and 1.5m. We determinate the dual antenna factor DAF for the pair of antenna, and the results have to be used as reference for your NSA measurements.
We are accredited for antenna calibration up to 20 GHz. In the higher frequency range we are able to determinate the antennafactor. Therefore we will issue a test report, where no uncertainty is stated. For a price-information please request an official offer.
We are able to calibrate as follows: Measurement of the radiated RF - Fieldstrength of a comb generator in the frequency range 30 - 1000MHz. The system is set up typical at 1 m height above groundplane. The measurement distance is 3 or 10m. The frequency spacing is 10 MHz. If an other one is needed, please let us know. We recommend to measure at both polarisations. In the moment we are not accredited for this service, we are going to issue an ISO-calibration certificate, with a statement of the measurement uncertainty.
For accurate NSA measurements for a distance of 3 m we don´t recommend to use the results, which have been calibrated at 10 m distance: When you change the distance, also the coupling between groundplane and the two antennas yields differently. We made an investigation and found out that the results can be up to +/- 2 dB. So the total uncertainty increases and the value is too big to validate measurements sites in an accurate way (and fulfill the 4 dB criteria). So we recommend to calibrate the pair of antennas also at the matching distance, in your case at 3 m.
Our standard uncertainty in the frequency range 1 Hz – 150 MHz is 15%. The calibration is done in a TEM-cell. If a lower uncertainty is needed, we can do this in this way: At 50Hz we can perform a single frequency calibration. We use a very accurate plate capacitor and a special coil for generating E- and H-fields. Then we achieve an uncertainty lower than 10% for generating the field. This kind of generating works only narrow banded and can´t be used at higher frequencies (>10kHz).
A project dealing with this topic was started in the year 2000. The output was published 2004 in CISPR 16-1-4 Amd. 1 Ed. 1.0. Additional information can be found at the IEC homepage (you have to be a member of a national standardization committee to download working documents): Link to IEC