The defueling of the damaged reactor at Three Mile Island Unit 2 (TMI-2), is conducted under 9.15 m (30 ft) of water for radiation shielding and nuclear criticality control. Adequate reactor coolant water clarity is necessary in this effort. After the start of active defueling, the reactor coolant water clarity rapidly deteriorated due to the presence of suspended colloids and a microbiological growth, which the originally designed filtration system could not adequately remove. Therefore, an alternate filtration technique was required. Deep-bed filtration was chosen and tested as a potential alternate method. The deep-bed testing program consisted of three distinct phases:

  1. small-scale single-element testing using reactor coolant and demineralized water
  2. full-scale hydraulic testing using demineralized water
  3. canister loading tests using a mockup knockout canister.
The test data from each phase are presented with an analysis of the results. Since the design objectives of turbidity values <1 ntu and a cumulative filter flow >378 500ℓ (100 000 gal) were not achieved, deep-bed filtration technology was not an acceptable substitute for the original filters. Certain deep-bed loadings did provide filtration with limited cumulative filter flow; however, based on the test results, the improvement in water clarity would have been inadequate for TMI-2 needs. In general, the results of the deep-bed testing program indicate that the loadings that provided effective filtration did so at the expense of reduced run time and cumulative flow through the filter, whereas the loadings that provided higher cumulative filter flow and run time sacrificed filtration effectiveness. The results further indicate that the deep-bed filter failed to provide effective filtration at higher levels of suspended solids in the influent reactor coolant. To ultimately solve the TMI-2 water clarity problems, a melamine formaldehyde coagulant enhanced by a diatomaceous earth bodyfeed was utilized.