Generalization of the variance-covariance method for microdosimetric measurements. III. Numerical simulation to assess the applicability of the formulae.

Numerical simulations in terms of microdosimetric data for 5-MeV neutron are utilized to test the twin detector method of microdosimetric measurements in its new, extended form. Three different conditions of measurement are considered: (I) constant dose increments per measurement, (II) fluctuating dose increments per measurement, and (III) fluctuating dose increments combined with slow changes of the dose-rate ratio for the two detectors. Under each of the conditions, large numbers of measurement series are simulated, and the estimation formulae for the mean and the two subsequent moments of the dose-weighted single-event distribution, d(y), are applied. The estimation formulae exist in three different forms: uncorrected formulae that apply to condition I, corrected formulae that are valid also for condition II, and fully corrected formulae that remain applicable under condition III. The results of the simulations indicate the acceptable ranges of doses per measurement interval. It can be seen that all three parameters are obtainable through the twin detector method, and that the fully corrected formulae for the two lower moments are generally applicable. The determination of the third moment is possible only under limited conditions, and the fully corrected formula is not useful in this case, while the corrected formula has some applicability even under condition III.