Determining the influence of population variation on compliance with radiofrequency exposure limits

How best to model humans for accurate SAR estimates remains an open question. Much work has been done over the years, attempting to quantify which and how various anatomic parameters such as tissue morphology, size, dielectric properties and relative location, affect energy absorption; and how to incorporate those parameters into models that provide conservative SAR estimates protecting all individuals. Since the available computational models are either highly complex realistic models or simple single tissue models, this issue is not easily explored, since the identified parameters are not easily varied within the models.

Attempts have been made in this thesis to overcome these limitations by creating a unique alternative model of the human head termed Geometry Head, incorporating a reduced set of tissues in a semi-homogeneous, simplified geometry for which the key parameters may be varied parametrically. Data was gathered from literature regarding what variations exist in human populations for tissue parameters, and used to populate the model. The computational model was validated using simplified physical models of human heads, custom made for this purpose. The Geometry Head model was used to systematically test what relationships exist between several anatomic variations in adult humans and consequent energy absorption due to RF exposure at a single frequency of 900 MHz, namely: cranial thickness; skin thickness; head size; and dielectric properties of tissues.

Overall, results suggest that SAR predictions could not have been anticipated using a priori reasoning, as location and magnitude of SAR maxima were seen to alter in unexpected and irregular ways with variations in tissue parameters. A compromise model like the one proposed here, which provides a point of diminishing returns between the high complexity of the multi-tissue models and the single-tissue homogeneous model, allows increased granularity for determination of SAR maxima in sensitive tissues.

Results of this work provide some inroads into establishing the level of complexity necessary for simulating human heads for radiofrequency exposure compliance calculations. Relationships between magnitude and location of regions of high SAR in the head and variables listed above have been quantified and qualified in this thesis.

The model proposed here is available for public use upon request. It can easily be amended to incorporate more or fewer tissues with varying levels of complexity; augmented to better resemble anatomy of children's heads; or used in its current state to further explore relationships between anatomic variations and absorption at the same frequency used here, or other RF frequencies.