This paper presents an efficient algorithm for estimating the unknown emission rate of radionuclides in the atmosphere following a nuclear accident. The algorithm is based on assimilation of gamma dose rate measured data in a Lagrangian atmospheric dispersion model. Such models are used in the framework of nuclear emergency response systems (ERSs). It is shown that the algorithm is applicable in both deterministic and stochastic modes of operation of the dispersion model. The method is evaluated by computational simulations of a 3-d field experiment on atmospheric dispersion of 41Ar emitted routinely from a research reactor. Available measurements of fluence rate (photons flux) in air are assimilated in the Lagrangian dispersion model DIPCOT and the 41Ar emission rate is estimated. The statistical analysis shows that the model-calculated emission rates agree well with the real ones. In addition the model-predicted fluence rates at the locations of the sensors, which were not used in the data assimilation procedure are in better agreement with the measurements. The first evaluation results of the method presented in this study show that the method performs satisfactorily and therefore it is applicable in nuclear ERSs provided that more comprehensive validation studies will be performed.
1 January 2012 in International Journal of Environment and Pollution
This paper presents an evaluation of an innovative data assimilation method that has been recently developed in NCSR Demokritos for estimating an unknown emission rate of radionuclides in the atmosphere, with real-scale experimental data. The efficient algorithm is based on the assimilation of gamma dose rate measured data in the Lagrangian atmospheric dispersion model DIPCOT and uses variational principles. The DIPCOT model is used in the framework of the nuclear emergency response system (ERS) RODOS. The evaluation is performed by computational simulations of dispersion of Ar-41 that was emitted routinely by the Australian Nuclear Science and Technology Organisation's (ANSTO) previous research reactor, HIFAR, located in Sydney, Australia. In this paper the algorithm is evaluated against a more complicated case than the others used in previous studies: there was only one monitoring station available each day and the site topography is characterised as moderately complex. Overall the estimated release rate approaches the real one to a very satisfactory degree as revealed by the statistical indicators of errors.
1 January 2011 in International Journal of Environment and Pollution
The development of a Data Assimilation algorithm based on variational approach that allows estimation of source emission rate in Lagrangian particle/puff Atmospheric Dispersion Models for non-stationary (wind and source) conditions is presented. The one-dimensional tests including constant and variable in time source emission rate and wind velocity showed that methodology can be extended to take into account variable meteorological conditions. In the 2-dimensional atmospheric dispersion test cases the sensitivity of the developed method on the spatial density of the measurement network was evaluated and an optimal value for the parameter rmod (error of observations proportional to observed concentrations) was estimated.