2018 | ANNUAL REPORT 73 pinch in the sensor’s wire that cleared up after temperature cycling. The signal follows the reactor temperature well, for most of the test. Some intermittent signal loss was observed over the last few reactor cycles.This is evidence that the coil was the component that failed, as any other compo- nent failing would not allow for a recovery. Figure 5 shows the normalized delay time and TC temperature of the three-segment Inconel sensor. As with the single-segment sensor, for most of the irradiation, the signal closely matches the reactor temperature. Some anomalous behavior can be observed during the early part of the irradiation, seen as an opposing response between the first and second segments (as one signal increases, the other decreases proportionately). It is unclear at this point if this is a physical phenom- enon or an artifact of the signal- acquisition process.This sensor also failed intermittently before finally failing after ~7000 hours. Figure 6 shows the normalized delay time and TC temperature of the single-segment titanium sensor. This sensor performed well through almost 3000 hours before failure. Unlike the Inconel sensors, there was observed a slow decrease in measured delay time.The likely explanation of this drift is fast-neutron damage causing a slow increase in the elastic modulus of the titanium.The effect appears to have saturated by the last operational reactor cycle, but this behavior may make titanium a poor material for UTs. AGR 5/6/7 Also based on the results of the ULTRA irradiation, one UT, using molybdenum as the sensing wave- guide, was included in the Advanced Graphite Reactor (AGR)-5/6/7 irradiation test at INL’s Advanced Test Reactor (ATR).This UT has performed very well through the early irradiation.This thermocouple was calibrated prior to installa- tion in the experiment. Figure 7 shows the temperatures measured Figure 7 AGR 5/6/7 UT results