Early Prediction and Classification of Heat Transfer Degradation in Coolant Channels using Kramers-Moyal Coefficients

The use of in-situ sensing information in advanced reactors is important for informed assessment of reactor operation, maintenance, and safety. To perform an accurate plant risk assessment, such sensing information must be properly analyzed to identify significant physical phenomena. One such phenomena is flow laminarization, which occurs when upward helium flows, such as those in an HTGR coolant channel, are stabilized through flow acceleration due to heating. This flow stabilization can lead to degradation of heat transfer. While the presence of laminarization can be observed through the wall heat transfer coefficient once it has occurred, there are not adequate methods of predicting flow laminarization directly from temporal sensor information, such as velocity and temperature. The fluctuation levels of these signals can provide some insight into the characterization of the flow; however, the direct measurement of the fluctuations often does not provide enough information to predict the degradation of heat transfer due to laminarization. Further probabilistic analysis of these signals through the Kramers-Moyal expansion method can provide some insight into the impact of potential laminarization on degraded heat transfer in reactor coolant channels. Experimentally validated direct numerical simulations are performed to obtain velocity and temperature time series at various locations within a vertical flow channel. The Kramers-Moyal coefficients are calculated at multiple locations along the channel and compared for multiple flow conditions, including cases where laminarization is predicted to occur and where laminarization is not expected to occur. The Kramers-Moyal coefficients are then used to define laminarization criteria, which can be used to predict degraded heat transfer from live sensing data within the reactor. This approach provides a probabilistic framework for predicting reactor behavior from real-time sensing data, improving the operation and maintenance of the reactor.