Survey of recent applications of reactor noise techniques at Argonne National Laboratory

Abstract

Recent applications of reactor noise techniques to the improvement of subcritical reactivity and β eff measurements in fast critical assemblies and to fast power reactor surveillance by neutron/acoustic noise cross correlation are reviewed. The Inverse Kinetics Rod Drop (IKRD) technique has been the primary subcriticality measurement technique applied at ANL because of its relative insensitivity to counter efficiency and prompt lifetime changes and the fact that a near-critical reference is not needed. Estimates of the random errors in IKRD reactivity measurements resulting from noise effects have been developed by simulation studies and theoretical analysis. Relative standard deviations of 1–20% are indicated by these methods for initial Δk from −0.0005 to −0.035, respectively, under conditions typical of the Fast Test Reactor (FTR) mockup experiments in the ZPR-9 Fast Critical Facility. Space-dependent kinetics calculations have been applied to the IKRD technique to study systematic errors that are functions of detector location as well as changes in kinetics parameters and effective source strength in the course of a rod drop. In various situations, systematic errors from these causes may be as low as several parts per hundred for detectors located outside the axial reflector due to cancellation of errors. For other detector locations, the principal error components might be in the same direction, thereby amplifying the resultant error. Noise measurements have been used as cross checks on IKRD results. The break frequency derived from cross power spectral density or polarity correlation has been found not to be a good reactivity estimator in fast power reactor critical assemblies presumably because wide disparities between neutron lifetimes in the core and reflector produce deviations from point kinetics in the range of the break frequency. Various two-detector mid-frequency polarity coherence function methods have produced good reactivity estimates, although these methods depend on near-critical reference measurements and are perturbed by differences in detector efficiency between the reference and measured systems. Correction methods to account for detector efficiency and β eff changes have been developed. The accuracy of fast critical experiment measurements of Δk k are directly affected by the accuracy with which β eff is known. Recent conflicting experimental data on β i , λ i has been the incentive for use of a noise-based integral β eff measurement technique. The technique is a modified detector noise variance-to-mean method using sampling intervals from 8 to 200 sec; β eff has been measured to a precision of ±2.9% in the FTR mockup critical assembly. Exploratory work on the use of neutron/acoustic noise cross correlation as a tool for surveillance of fast power reactor systems has been carried out using the Experimental Breeder Reactor II. Cross correlation of neutron signals from in-core detectors with acoustic signals from accelerometers mounted on various mechanical components connected into the core structure has been studied up to several thousand hertz. Flow-induced subassembly vibrations at about 10 hertz have been observed.