Stability monitoring of a natural-circulation-cooled boiling water reactor

Methods for monitoring the stability of a boiling water reactor (BWR) are discussed. Surveillance of BWR stability is of importance as problems were encountered in several large reactors. Moreover, surveying stability allows plant owners to operate at high power with acceptable stability margins. The results of experiments performed on the Dodewaard BWR (the Netherlands) are reported. This type reactor is cooled by natural circulation, a cooling principle that is also being considered for new reactor designs. The stability of this reactor was studied both with deterministic methods and by noise analysis. Three types of stability are distinguished and were investigated separately: reactor-kinetic stability, thermal-hydraulic stability and total-plant stability. It is shown that the Dodewaard reactor has very large stability margins. A simple yet reliable stability criterion is introduced. It can be derived on-line from thhe noise signal of ex-vessel neutron detectors during normal operation. The sensitivity of neutron detectors to in-core flux perturbations - reflected in the field-of-view of the detector - was calculated in order to insure proper stability surveillance. A novel technique is presented which enables the determination of variations of the in-core coolant velocity by noise correlation. The velocity measured was interpreted on the basis of experiments performed on the air/water flow in a model of a BWR coolant channel. It appeared from this analysis that the velocity measured was much higher than the volume-averaged water and air velocities and the volumetric flux. The applicability of the above-mentioned technique to monitoring of local channel-flow stability was tested. It was observed that stability effects on the coolant velocity are masked by other effects originating from the local flow pattern. Experimental and theoretical studies show a shorter effective fuel time constant in a BWR than was assumed. (author). 118 refs.; 73 figs.; 21 tabs