Smithsonian - National Museum of American History, Behring Center
Three Mile Island
Unit 2 nuclear power plant
First looks inside the reactor

Three Mile Island: The Inside Story

Collecting Ultrasonic Echoes at TMI-2

 
Radiation levels in the TMI-2 reactor building in 1980, when it was first entered after the accident, and in 1983, when the core topography survey was done.

Click to enlarge imageFigure 7.1. Radiation levels in the TMI-2 reactor building in 1980, when it was first entered after the accident, and in 1983, when the core topography survey was done.

Learn more

The core topography team arrived at TMI in August 1983. After having choreographed their moves at the mockup in Idaho, they practiced them again on a mockup in the TMI-2 turbine building. August 31 was the first entry into the containment building to set up their equipment. Installation of the apparatus on the work platform took just 45 minutes, after which all personnel left the reactor building.

First a reconnaissance survey was done to establish quickly the shape and dimensions of the cavity and to select optimal operating parameters for the piezoelectric transducers. That first survey provided plenty of cause for worry about the success of the project: the six 10-MHz transducers produced no useful data because the sound waves they emitted proved to be very strongly absorbed by the water in the reactor vessel, loaded as it was with dissolved and suspended matter. And two of the other six transducers, operating at 2.25 MHz, also produced no useful data. Mysteriously, those two transducers were found to be working the next day at the start of the second survey, which provided the principal data for subsequent reductions and displays.

Typical on-the-spot sonar-like range-bearing plot of the data from the horizontal transducer.

Click to enlarge imageFigure 7.2. Typical on-the-spot sonar-like range-bearing plot of the data from the horizontal transducer.

Learn more

 

Surveys were done from the bottom up. The 44-foot (13 m) probe-tipped boom was allowed to descend slowly under its own weight to within about six inches (15 cm) of the floor of the cavity—as determined by the downward-looking transducer. The boom was then stepping-motor-driven upward in 1-inch (2.5 cm) steps. This ensured a more controlled vertical motion in case there were any binding in the control-rod guide tube, whose inside diameter was only 1/8 inch (3 mm) greater than the diameter of the probe. After rising an inch, the probe was rotated through a bit more than 360° in 0.9° steps. At each vertical step a plot was printed of the range-bearing data acquired with the horizontal-pointing transducer, this being the one transducer for which range-bearing data were self-interpreting. A complete survey, with over 50 vertical steps, took somewhat less than four hours.

A similar plot of horizontal transducer data with the outline of the core former (sides), to the same scale, superimposed.

Click to enlarge imageFigure 7.3. A similar plot of horizontal transducer data with the outline of the core former (sides), to the same scale, superimposed.

Learn more

 

Only two surveys were conducted. On a third day the reactor building was reentered to remove the apparatus. “Management policy throughout the cleanup was that research work could not significantly interfere with cleanup work” (ref. 12, ch.5, p.1). And that, despite the fact that much of what the Department of Energy paid for as “research” was the acquiring of data that proved essential for the planning and execution of the cleanup work.

The sonar survey established that the cavity in the reactor’s core was substantially larger than had previously been supposed. This was ascertained immediately with the first preliminary survey. It was presented in dramatic visual form to TMI personnel in the following way: the range-bearing plots of the data from the horizontal transducer were photocopied onto mylar sheets (overhead projector transparencies). These were then stacked, successive plots separated by disks of Lucite, and the whole illuminated from below. (The disks had been prepared in advance for this purpose.) The effect achieved with this improvised 3-D representation of the core void was suggested in the Museum exhibit with a similar construction making use of the original Mylar transparencies.

The August 31, 1983, improvised 3-D representation of the core void as reproduced in the spring 2004 Museum of American History exhibit.

Click to enlarge imageFigure 7.4. The August 31, 1983, improvised 3-D representation of the core void as reproduced in the spring 2004 Museum of American History exhibit.

Learn more

 


Standing out very clearly in the range-bearing plots are 3-foot (0.9 m) lengths of the axial power shaping rods (APSR) hanging down into the cavity. These are the portions of the APSR that were above the core at the time of the accident, and that in the tests of their operability in June, 1982, were driven down into the void. They are all that remained of the APSR after the accident: the 9 feet (2.7 m) that were within the core at the time were melted away.

Among the detailed features revealed by the core topography survey—and to whose revelation particular attention had been given in the design of the survey instruments—was what remained adhering to the upper grid structure, hanging down from it into the cavity.

In one significant respect the topographic maps and topographic model (see Section 8) misrepresent what the sonar survey saw: the walls of the cavity are not smooth and continuous as they appear in maps and model, but are formed largely of broken, hanging fuel rods, as in this still image from an October 1983 video.

Click to enlarge imageFigure 7.5. The walls of the cavity are not smooth and continuous as they appear in maps and model, but are formed largely of broken, hanging fuel rods.

Learn more

 

Such adhering core debris would encumber and complicate the removal of the upper grid, the first step in getting access to the mess in the core. With this in mind, the INEEL team had incorporated upward-looking transducers into the sonar probe. And in working up the data, they again gave particular attention to the presentation of the upward-looking data in visualizable form.