Quantifying Transient Uncertainty in the BEAVRS Benchmark Using Time Series Analysis Methods

摘要

The BEAVRS benchmark has been instrumental in showing the efficacy of high fidelity modeling tools to model realistic PWR models. Recent work has been focused in quantifying the uncertainty in areas of data measurement, data processing, and simulation tools. Time series analysis methods were investigated as a means to calculate transient detector uncertainty data, and we expect to find that this uncertainty should be consistent with measurement and post-processing uncertainty that were calculated at each burnup step. For linear models, errors were on the order of 0.6% to 0.8%. We also use simulation codes to calculate RMS error to more accurately model reaction rates, and expect to find that such modeling tools will yield more accurate results than linear regression models. Future work involves looking at stationary models instead of non-stationary models as added references for fitting available data. In addition, we hope to cross reference CASMO/Simulate with other codes to verify that our results are consistent across multiple software platforms.