The paper deals with the creation of a differential pressure model with artificial neural networks (ANN). Particular, model development and verification tests are considered. One of the main features in reactor safety research is the safe heat dissipation from the reactor core and the reactor containment of light-water reactors. In the case of loss of coolant accident (LOCA) the possibility of the entry of isolation material into the reactor containment and the building sump of the reactor containment and into the associated systems to the residual heat exhaust is a serious problem. This can lead to a handicap of the system functions. To ensure the residual heat exhaust it is necessary the emergency cooling systems to put in operation which transport the water from the sump to the condensation chamber and directly to the reactor pressure vessel. A high allocation of the sieves with fractionated isolation material, in the sump can lead to a blockage of the strainers, inadmissibly increase of differential pressure, build-up at the sieves and to malfunctioning pumps. Hence, the scaling and retention of fractionated isolation material in the building sump of the reactor containment must be estimated. This allows the potential plant status in case of incidents to be assessed. The differential pressure is the essential parameter for the assessment of allocation of the strainers. For modelling we use artificial neural networks. To build up the ANN, the available experimental data are used to train the ANN.
- Nuclear Engineering Division
Using Artificial Neural Networks in the Reactor Safety Research
- Views Icon Views
- Share Icon Share
- Search Site
Kratzsch, A, Ka¨stner, W, & Hampel, R. "Using Artificial Neural Networks in the Reactor Safety Research." Proceedings of the 16th International Conference on Nuclear Engineering. Volume 2: Fuel Cycle and High Level Waste Management; Computational Fluid Dynamics, Neutronics Methods and Coupled Codes; Student Paper Competition. Orlando, Florida, USA. May 11–15, 2008. pp. 891-897. ASME. https://doi.org/10.1115/ICONE16-48611
Download citation file: