Estimating uncertainty in calculated ambient RF field levels for hazard assessments

Just as for measured evaluations, the calculation of ambient levels of radiofrequency (RF) electric and magnetic fields for hazard assessments can entail random uncertainty, the extent of which depends on: how well the method of analysis approximates the governing electromagnetic equations (i.e. Maxwell's equations); how closely the modeled object represents its real counterpart in form and composition, and; computational round off error. While uncertainty estimates are commonly required for RF measurements, they are rarely provided for calculated assessments. In this paper, a process for calculating uncertainty for calculated assessments is proposed. It incorporates a series of validation tests which are used to gauge the uncertainty in the computational technique and round off error. The comparator in the validation tests must be independent of the technique (e.g. measurement of antenna gain) and may vary for different types of analyses (e.g. full wave solutions vs ray tracing methods). Variational targets are set for the validation tests, and the uncertainty estimate for computational uncertainty diminishes as more tests are passed. Once determined in this way, the standard uncertainty for computational technique is then combined in a multiplicative model with standard uncertainties for other influence quantities such as forward power to the antenna and field scattering effects of nearby objects. By using the proposed approach, it will be possible to provide systematic corrections and uncertainty estimates for calculated assessments as they are for RF measurements, thereby allowing meaningful comparisons between them.