An investigation of apparent, intermittent fuel failure in TRIGA fuel

Full text: The 1 Mw TRIGA Reactor is cooled by natural convection and 16N and 41Ar which were released from the pool caused an undesirably high background in the gas detector used for fission product monitorings. It was necessary to convert the gas detector which was very sensitive to 16N and 41Ar to a particulate monitor. The monitor chosen was a commercially available moving tape continuous air monitor. It was placed adjacent to the pool wall where air under the reactor bridge could be sampled at the surface of the pool water. This monitoring system proved to be adequate under normal conditions. However, on several occasions, sharp increases in activity of short, duration were noted on the monitor recorder. These brief high levels of activity were similar to those described by other facilities which had experienced intermittent fuel element leakage and the integrity of the reactor fuel was suspected. The filter paper from the monitor was removed and its gamma spectrum obtained using a 30 cc Ge(Li) detector. This analysis revealed only the naturally occurring airborne isotopes, and it was concluded that the recorder increases in activity were not from fuel leakage. It was suspected that the activity increases were due to the intermittent release of radioactive gases from the pool water at higher temperatures. A series of experiments was performed which demonstrated the following: 1. At pool temperatures up to 80oF, the reading of the CAM remained constant. From 80 to 95oF, the reading increased by nearly a factor of 3 and became more erratic. 2. With the air pump off, the CAM response did not change. 3. A gas monitor set up in series with the CAM exhibited nearly an identical response and its spectrum showed only the existence of 41Ar and 16N. 4. After reactor shutdown, the CAM reading decayed with a half-life of 1.4 hr. It was concluded that at the higher pool temperatures large bubbles containing 41Ar escaped to the pool surface and the 41Ar was seen through the shield of the CAM. The response of the monitor from the 41Ar resembled responses reported in the literature from intermittent fuel element leakage. The problem was solved with the installation of the permanent fission product monitor which transports the air to the monitor located outside the reactor bay which eliminates the16N and 41Ar background. This monitor is capable of detecting <1 mg of 235U placed in a flux of 1 x 1013 n/(cm2 sec) and is insensitive to 41Ar fluctuations near the pool. (author)