Integrated system for population dose calculation and decision making on protection measures in case of an accident with air emissions in a nuclear power plant

Primož Mlakar,Marija Zlata Božnar,Boštjan Grašič,Borut Breznik
DOI: https://doi.org/10.1016/j.scitotenv.2019.02.309
2019-05-20
Abstract:The accidents in Chernobyl and Fukushima remind us that nuclear power plants should continuously invest resources in improving safety and in risk management. This paper presents the methodology for developing a measuring and modelling system with a high degree of automation, which enables predicting the effects of the spreading of radionuclides from the nuclear power plant to the atmosphere. The end result is the calculated population doses in the event of an accidental release, which is an essential piece of information needed by first responders to take proper action. The key challenge addressed by this methodology is how to build a system so that its operation is maximally automated, ongoing and in real time. Moreover, in a way that "fresh", normalized results for the hypothetically most probable types of emissions are always available to operators. The principle that normalized, fresh results are always automatically available to operators is the only real assurance that they will almost surely be available in the event of an accident and panic. This way, we can avoid performing complex model calculations at the operator's request when the accident is already taking place. The methodology divides the building of the system into key modules, which are substantiated and described. The theoretical section is followed by a description of implementation on the example of the Measuring and Modelling System at the Krško Nuclear Power Plant (in Slovenia). The system has been tested in regular nuclear emergency exercises and rated excellent by IAEA inspections; it has been operating automatically, continuously and in real time for many years. The availability of automatic results is counted for the last two years. Measurements and diagnostic modelling results were available for more than 96% and forecasts were available in more than 91% of all half-hour intervals.
What problem does this paper attempt to address?