One of the Advanced Test Reactor’s (ATR’s) functions is to irradiate and qualify nuclear fuels and materials. Due to the large number of experiment or test positions, the cost, and the limited number of vessel penetrations for instrumentation, in-core instrumentation for most experiments is not feasible. In such instances, modeling of experiment conditions using high-fidelity neutron transport codes can quantify such conditions as fission power density and fissile material burnup during irradiation. Validation of fissile material burnup can only be performed during post-irradiation examination, which typically occurs months—or even years—following irradiation. In most experiments, fission power density and fissile material burnup are directly proportional to the thermal neutron flux in the ATR. Additionally, fast neutrons are born from fission in the ATR core, affording a validation of power distribution within the reactor’s experiment locations. During each irradiation cycle, flux wires installed throughout the ATR can be used to validate computational models and determine an adjusted neutron flux for many of the experiment positions. The flux wires are installed as requested by the experiment sponsors in several of the ATR flux traps and consist of cobalt-aluminum alloy and nickel wires. Both kinds of wire enable measurements of the thermal and fast neutron flux in each experiment position. This paper presents the protocol for validating computational models for experiments using flux wires installed in the experiment positions, as well as the results for flux wires placed in the ATR safety rod guide tubes. The best estimate is typically referred to as the adjusted neutron flux. The calculated unadjusted neutron flux is referred to as the a priori neutron flux. The methods presented here provide the adjusted neutron flux, given both the measured and a prior fluxes. The adjusted flux is compared to the a priori flux to provide a bias in the calculated results and the adjusted results. Two model types are evaluated; an eigenvalue case and fixed-source case. Both conditions demonstrate relatively good agreement. The uncertainty for the adjusted flux ranges from 5% to 6% for all three energy ranges. For the eigenvalue case, the bias between the a priori and the adjusted neutron flux is within the statistical uncertainty in all but two wire pairs. For the fixed-source model, four wire pairs are outside of the uncertainty of the adjusted flux. The bias between the a priori and adjusted fast neutron flux is outside of the statistical range for four wires in the eigenvalue case and nine wires in the fixed-source model. As the differences are not contained to one flux trap, it can be assumed that the biases in the calculated models are attributed to localized effects in modeling. An additional evaluation was performed for the ATF-1 experiment in the ATR “I” positions. The differences between the adjusted and a priori are more pronounced in two of the test positions, indicating that additional model evaluation is needed, in particular in the region near the boundary of the ATR model. It is also noted that the eigenvalue model provides slightly better results in the flux trap positions. The fixed-source model is more computationally efficient though produces less accurate results; the differences in some cases are negligible. The work documented in this paper provides a methodology that extends the validation protocol established at the ATR for flux measurements to validate computational models with limited measurement capability during a cycle.
Read full abstract