A proposed model for the correlation of power ramp test results

The trend toward extended burnup operation in the world's light water reactors (LWRs) has recently received more attention than considerations for reliability at conventional burnups. Although the performance level of LWR fuel rods has been rather high, the few remaining threats to fuel rod integrity should not be overlooked. Pellet cladding interaction (PCI) or stress corrosion cracking (SCC) is one of the remaining threats. The use of operating guidelines that precondition the fuel rods and/or reduce stresses during a power escalation has been beneficial in reducing the incidents of fuel failure. The guidelines for boiling water reactors (BWRs) have been more restrictive than those for pressurized water reactors (PWRs). The incentive co reduce or eliminate operating restrictions is, therefore, higher for BWRs. But the PCI/SCC issue has not disappeared completely from the PWR. PCI models that have been offered over the last five-to-ten years include, for example, the Canadian Fuelograms where a minimum power level and a minimum power change are necessary conditions for PCI/SCC. Another PCI model, recently modified, was called FOSHO for power shock. With the early version of this model, Scandpower predicted PCI solely on the basis of power change. Another approach to predicting PCI evolved from the simplified treatments in publications where data from test reactor experiments are plotted on graphs of power vs. burnup and thresholds are drawn on the basis of these results. A threshold, for example, has been treated as a power level below which failures by PCI or PCI/SCC do not occur. If the ramp experiments that lead to a conclusion on thresholds for a given design all start from a common power level, however, a peak power and a change in power cannot be distinguished. Most investigators now recognize that chemistry plays a role in the failure of fuel rods by PCI. Thus stress corrosion cracking is being treated explicitly in more recent models. Although the actual nature of the chemical attack is not known, the chemical influence needs to be accounted for in the predictions of PCI/SCC. Predictions based on mechanical considerations alone have not been universally successful. For understandable reasons, the modeling of SCC continues to be difficult. The data are showing considerable stochastic variation, and two samples of common design may respond differently to a common ramp in power. Difficulties in the interpretation of the relevant statistics would be minimized if larger numbers of fuel rods were tested. Tests cannot be conducted in large numbers, since operators of power reactors seek to avoid failures and single rod tests are very expensive. Another source of complication in modeling SCC has been the lack of adequate characterization of early tests and the absence of detailed power histories. The industry is rapidly approaching a turning point in the above areas as programs initiated in the mid-seventies are yielding larger amounts of data upon which conclusions can be based. This paper is intended to offer a model that addresses PCI/SCC in power ramp tests, explains many (but not necessarily all) of the observations, and provides a basis on which continued investigations can build. The model correlates direct empirical results, iu a manner consistent with known physical phenomena, to predict the results of power ramp experiments. The model does not apply to the spectrum of transients which can occur in power reactors, since explicit modelling of ramp hold times and ramp rate effects are needed and this remains beyond the state of the art. A general case based on a modern PWR fuel rod, as represented by PWR current design, is used to describe the model. The considerations necessary to apply the technique to other designs, subjected to other conditions, are only discussed qualitatively. A comparison of the model to representative power ramp test data is shown; however, a more extensive evaluation is recommended before adopting the model for other than developmental work