The direct and indirect hydro-dynamic loads in the suppression pool of the BWR fleet were developed well before 2000. These loads were based on scaled experiments and numerical solutions using one dimensional models. The analysis was cumbersome and conservatisms were added in multiple steps.
As these loads spread throughout the containment, as global vibrations, they are generally a part of all structural verification inside the containment. This fact has made it hard and unpractical to challenge and revise these loads, as a change could lead to significant re-work. A consequence of this is that loads in the suppression pools have seldom been revisited, regardless if they cause local or global vibrations. This is problematic when new equipment is needed. The design of this equipment suffers from the very conservative loads.
Experience has shown that these loads can be challenged and refined using updated techniques, leading to significantly lower loads. This was realized during the modernization of Swedish plant Oskarshamn Unit 2 with Mark II containment, where loads following pool swell proved to be particularly challenging and it was decided to investigate possibilities to reduce the conservative loads.
Due to the large scale of the condensation pool coupled with the transient and small scaled condensation of steam at the drywell vent pipe nozzle full CFD resolution is not feasible. Instead lumped models in GOTHIC was used to increase resolution from the 1D approach of normal containment analysis, to a resolution that can account for the features of the condensation pool. This showed that the pool swell was less uniform than initially thought, leading to fewer objects affected as well as lower loads on objects that suffered from these loads. A full CFD analysis was then used to resolve phenomena working on even shorter time scales leading to a complete rework of all local loads.
The loads addressed using updated codes and modelling techniques was pool swell impact (PSI), pool swell drag (PSD) and local drag loads due to pressure relief valve opening, LV/SRV.
The current work shows that using updated modelling techniques and aligning results with previous analysis and documentation, it is possible to reduce loads for some events in the suppression pool without violating safety for the power plant. The results from the GOTHIC model shows that the cylindrical shape of the pool will create an uneven velocity distribution radially at pool swell resulting in much smaller loads at the outer boundary.
CFD analysis of the LV/SRV event shows that the loads are reduced in comparison with previous methodology and this is mainly because of shadow effects. The calculations also verifies assumptions used in the previous methodology.
Effect of Mechanical Ventilation on Accidental Hydrogen Releases—Large-Scale Experiments
