Three Mile Island: The Inside Story

INEEL’s Core Topography Survey System

Figure 6.1. This drawing of the upper part of the TMI-2 reactor vessel and work platform atop it shows the 44-foot (13 m) long boom extending down from the work platform above the reactor through the central control rod leadscrew channel to the top of the upper grid structure. Learn more

The “Quick Look” video peek inside the reactor’s core in July and August 1982 necessitated rethinking the cleanup operation. Remediation was going to be much less straightforward than had been generally supposed. In order to carry through that remediation—in order simply to estimate what it was going to take in time, technology, and money to clean up inside the reactor—much more information would be needed about the condition of every structure within the reactor vessel.

That was the basis of the utility’s interest in cooperating in the gathering of more information, though it remained a grudging interest if it did not provide the vividness of video pictures. The Department of Energy and the Nuclear Regulatory Commission saw such information-gathering efforts as essential for understanding what had actually happened inside TMI-2 in those first hours of the accident—which is to say, for understanding the results of an experiment that they could never choose to do, but whose results were of great interest to them in anticipating the consequences of future accidents.

Ultrasonic imaging was a well-established technique in the reactor field, being commonly used to search for defects in materials and construction. It was considered early on to examine damage to the reactor’s fuel rods, but there was no practicable way to put it into play. Likewise, a “sonar” system, i.e., an echo-location system based on ultrasonic pulses, had been considered right from the start as a way to look around inside the reactor vessel. It was recognized that sonar would have an advantage over video if the water filling the reactor vessel were murky—as it was expected to be, and in fact was. But weighing decisively against sonar was the belief that the fuel rod assemblies were still intact, or at least largely intact. In that case the multiplicity and proximity of echoing surfaces would make the sonar data uninterpretable (ref. 12, ch. 5).

The discovery that there was a large void in the core of the reactor made the feasibility, indeed the indispensability, of an echo-location survey beyond question. A team was put together from the Department of Energy’s Idaho National Engineering and Environmental Laboratory. Located on a high-desert reservation almost as large as the state of Rhode Island, this facility was established as the National Reactor Testing Station shortly after World War II (ref. 22). By the early 1980s, it was INEL, “Environmental” not having yet been added to its name (and it remains to this day INEL on the Internet). INEEL was the principal DoE laboratory involved in the analysis and cleanup of the TMI-2 accident, and maintained a field office there for the entire ten years the operation lasted.

The sonar survey team was managed by Michael R. Martin, a metallurgical engineer by training, who had been involved right from the start in the analysis of the damage to the reactor, and who would remain involved in advising and assisting on the cleanup of the core right to the end. The ultrasonic expertise was provided chiefly by Larry S. Beller, who had many years of experience with ultrasonic inspection and imaging of solid materials. But neither he, nor anyone else had built a system to work at the required range or provide images of the desired sort. (Beller was the principal author of the final report of the project, ref. 2, from which details given here of design and operation are taken.) In just four and one half months, the entire data collection system was designed, fabricated, and thoroughly tested, including creation of the software and training of personnel in procedures for setting up and operating the system at TMI-2.

Figure 6.2. At the front and at the center of the core topography survey was the probe. Learn more

INEEL’s core topography survey system was the most technically advanced purpose-built system that had been used for analysis or remediation at TMI-2 up to that time, August 1983. While employing the same work platform and the central control rod access port utilized by “Quick Look,” INEEL’s sonar system—in contrast to that and all subsequent video inspections—was designed to minimize radiation exposure of personnel by remote operation and, further, by using for that remote control only such cabling as was already in place in the reactor building.

Once the mechanical components and interfacing electronics were set up on the work platform above the reactor, all personnel were to leave the containment building and the system was to be operated—for the most part automatically by a Digital Equipment Company minicomputer—from the auxiliary building that housed the reactor’s control room. Stepping motors controlled by that computer effected the rotation and vertical motion of the pulse-echo probe, with the precise vertical position and horizontal orientation of the probe being continually fed back to the computer by digital position encoders. These data were stored electronically, on 8-inch (20 cm) diameter magnetic disks, along with the echo return times for the ultrasonic pulses emitted by each of the twelve piezoelectric transducers at each 0.9-degree rotational step, and each 1-inch (2.5 cm) vertical step.

Although the system was wholly novel, it had to be reliable. Once the measurements began on the scheduled day, they had to go without any serious hitch. Backups could be, and were, at hand for all the components outside the containment building. But the maintainablility of the components inside the containment building was nearly nil—think of fiddling with electronic equipment while wearing the anticontamination clothing shown in Figure 5.1. The system was therefore put together out of components of established reliability and went through a rigorous testing program.

The sonar probe was tested at INEEL in a water-filled pit about the same size as the upper third of the TMI-2 core, into which objects simulating TMI-2 core components had been placed. In order to test system operability using existing cabling at TMI, and under conditions of electrical noise that could exist there, a kilometer of #14 wire was put together by buying up all the heavy-duty extension cord that could be found in eastern Idaho, and then arranging it to provide maximum opportunity for interference. Finally, the work platform above the TMI-2 reactor was mocked up, and installation moves choreographed in complete anticontamination clothing.