“…For example, in electrostatic analyses, Maxwell's equations are exploited to derive D and E and the error estimated by their comparison using D = E [6]. Although this may provide a boundary of the exact error [5,7], to attain high ÿeld accuracy is very computationally demanding and, in reality, for example, the iteration is aborted before the speciÿed ÿeld accuracy is met, based on the convergence of other quantities e.g. energy [6,8].…”