TRANSMITTED TO M. WHITAKER 02/08/96

DEFENSE NUCLEAR FACILITIES SAFETY BOARD

                                                  November 27, 1995

MEMORANDUM FOR:          G. W. Cunningham, Technical Director

COPIES:             Board Members

FROM:                    David C. Lowe

SUBJECT:            Savannah River Site (SRS)   In-Tank Precipitation (ITP)
                    Review of Cycle 1 Testing (November 8, 1995)

1. Purpose:  This trip report documents a review by the Defense Nuclear
   Facilities Safety Board's (Board) technical staff (D. Lowe and J. Roarty) on
   November 8, 1995, regarding In-Tank Precipitation (ITP) flammability issue
   resolution program.

2. Summary:  The Radioactive Operations Commissioning Test Program (ROCTP) is
   in progress at ITP.  The large benzene release rates are again observed, but
   are now attributed to the mixing action of the four submerged mixing pumps. 
   There is no observable release during washing operations indicating that
   "trapped" benzene is not the benzene release mechanism.  These conclusions
   are preliminary, but it appears that the turbulent shear during pump mixing
   of  the waste solution produces a uniform benzene release from the liquid. 
   Westinghouse Savannah River Company (WSRC) is conducting additional in-tank
   tests and plans to conduct laboratory experiments to determine the benzene
   release mechanism.

3. Background:  In 1983, a test was conducted in Tank 48 to determine the
   effectiveness of the sodium tetraphenylborate (NaTPB) precipitation process. 
   Coincident with the introduction of wash water, four pumps were started to
   mix the solution.  A significant release of benzene was observed which was
   postulated to be due to the release of benzene "trapped" in the excess NaTPB
   solids as they were dissolved.  This "trapped" benzene phenomena were also
   observed in some University of Florida experiments, therefore, the 1983 test
   observations were attributed to "trapped" benzene.  However, recent
   laboratory tests have failed to confirm this hypothesis,  as the observed
   benzene releases are 2-3 orders of magnitude less than the values observed
   in the 1983 test.

4. Discussion:

   a. Radioactive Operations Commissioning Test Program: The test program is
      in progress at ITP with Batch 1 complete.  The tests conducted during
      Batch 1 have yielded some unanticipated results.  The large benzene
      release rates are again observed, but are now attributed to the mixing
      action of the four submerged mixing pumps (i.e., release rate less for
      one pump operating than four pumps operating, and when pump speed is
      decreased the release rate is correspondingly decreased).  There is no
      observable release during washing operations indicating that "trapped"
      benzene is not the benzene release mechanism.  These conclusions are
      preliminary, but it appears that the turbulent shear during pump mixing
      of  the waste solution produces a uniform benzene release from the
      liquid surface.  WSRC is conducting additional in-tank tests and plans
      to conduct laboratory experiments to determine the benzene release
      mechanism.

      The two instrument poles located at geometrically opposite ends of the
      tank are indicating uniform vapor space temperature and concentration
      profiles.  This implies that the tank headspace is well-mixed.  However,
      during these tests the lowest sample point was 21 inches above the
      surface.  In future tests, the instrument poles will be lowered to
      within five inches in order to determine if a benzene rich layer is
      formed near the surface.

      The benzene release rates are greater than expected during non-wash
      operations, but WSRC stated that they are still operating within their
      safety envelope.  The potential safety implications of these new
      findings are being studied by WSRC.  Additional testing is required to
      determine the mechanism and ensure that controls are adequate.  If the
      benzene release mechanism turns out to be pump operation, then
      additional controls may be required (e.g., interlock to shutdown pumps
      on loss of ventilation and purging, additional controls to prevent
      inadvertent pump operation, etc.) and some controls may be relaxed
      (e.g., controls on water additions).

   b. Benzene Generation Rate Test Results: The experimental program is
      ongoing at Georgia Institute of Technology with 4 of 10 barrel tests
      complete.  There is no evidence of the "trapped" benzene phenomena during
      recent Savannah River Technology Center (SRTC) bench-scale tests,
      Georgia Institute of Technology barrel-size tests, or full-scale testing
      at ITP.  WSRC is adjusting the barrel experimental program as additional
      information is obtained from the in-tank testing.

   c. Analytical Effort: Little progress has been made in the 3-D analytical
      code effort.  The code is still not predicting the dynamic response of
      purge flow.  An extension of the contract on code development is
      underway as this effort affords the only means for investigating local
      benzene stratification (vertical or radial) away from the two instrument
      locations.