Nuclear Waste: Management and Technical Problems Continue to Delay
Characterizing Hanford's Tank Waste (Letter Report, 01/26/96,
GAO/RCED-96-56).

GAO reviewed the Department of Energy's (DOE) progress in characterizing
its tank waste at its Hanford site, focusing on: (1) whether it has met
its characterization deadlines; (2) the impediments to meeting
characterization deadlines; and (3) the impact that continued delays
have on the characterization process.

GAO found that: (1) over the past 10 years, DOE has spent over $260
million and made little definitive progress in characterizing the tank
wastes at Hanford; (2) DOE has not been able to meet characterization
deadlines for the 54 tanks with known safety problems; (3) the DOE
contractor has been unable to characterize any of the 177 tanks as ready
for remediation; (4) DOE and its contractor have had problems performing
reliable top-to-bottom samples, gathering sampling data, reconciling
tank contents, and developing an effective tank characterization
management system; (5) disagreement exists as to what kind and how much
information is needed to reliably predict actual waste quantities and
build appropriate treatment facilities; (6) Congress, DOE, and private
contractors need better sampling and characterization information to
reliably predict total program costs; and (7) these uncertainties could
undermine the savings DOE expects to realize by privatizing the tank
waste remediation program.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-96-56
     TITLE:  Nuclear Waste: Management and Technical Problems Continue 
             to Delay Characterizing Hanford's Tank Waste
      DATE:  01/26/96
   SUBJECT:  Radioactive waste disposal
             Environment evaluation
             Reporting requirements
             Tanks (containers)
             Nuclear facilities
             Hazardous substances
             Nuclear waste management
             Safety regulation
             Information gathering operations
IDENTIFIER:  DOE Hanford Tri-Party Agreement
             DOE Tank Waste Remediation System Program
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Secretary of Energy

January 1996

NUCLEAR WASTE - MANAGEMENT AND
TECHNICAL PROBLEMS CONTINUE TO
DELAY CHARACTERIZING HANFORD'S
TANK WASTE

GAO/RCED-96-56

Hanford's Tank Waste Characterization

(302161)


Abbreviations
=============================================================== ABBREV

  EPA -
  DOE -
  DQO> -

Letter
=============================================================== LETTER


B-270726

January 26, 1996

The Honorable Hazel R.  O'Leary
The Secretary of Energy

Dear Madam Secretary: 

The Department of Energy (DOE) has about 55 million gallons of highly
radioactive waste currently stored in 177 underground tanks at its
Hanford site in Washington State.  Some of these tanks have leaked,
and others contain potentially flammable waste.  DOE's program for
developing a safe, permanent way to dispose of the waste in these
tanks is one of its highest priorities.  This cleanup effort is also
one of DOE's most expensive projects; its estimated life-cycle cost
is $36 billion. 

Characterization, the first major step in cleaning up tank wastes,
involves determining the specific contents of the tanks through
sampling and other means.  At Hanford, the characterization program
has experienced a series of delays, leading to numerous revisions in
the schedule for completing the work.\1

The last formal schedule revision took place in January 1994, when
DOE negotiated an agreement with Washington State and the U.S. 
Environmental Protection Agency (EPA) to finish characterization by
September 1999.  Also in January 1994, DOE agreed to implement a
recommendation by the Defense Nuclear Facilities Safety Board to
characterize wastes in the 54 tanks with known or suspected safety
problems by October 1995.  Because of characterization's importance
to the cleanup and safe storage of the wastes, our objectives were to
(1) review DOE's progress in meeting the deadlines, (2) identify any
impediments to making progress, and (3) determine what impact
continued delays could have. 


--------------------
\1 For descriptions of these problems, see Nuclear Waste:  Problems
and Delays With Characterizing Hanford's Single-Shell Tank Waste
(GAO/RCED-91-118, Apr.  23, 1991) and Nuclear Waste:  Hanford Tank
Waste Program Needs Cost, Schedule, and Management Changes
(GAO/RCED-93-99, Mar.  8, 1993). 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

Continuing delays in characterizing Hanford's high-level waste led
DOE to conclude in September 1995 that it could not comply with the
current agreed-upon deadlines.  Westinghouse Hanford Company, DOE's
contractor, has been unable to characterize any of the 177 tanks
sufficiently to declare them ready for remediation.  In addition, DOE
did not meet its commitment to the Defense Nuclear Facilities Safety
Board to characterize the 54 tanks with known safety problems by
October 1995.  Although DOE and Westinghouse believe tank wastes are
stored safely, they are unable to predict when any of the tanks will
be characterized. 

Although many factors have contributed to the continued delays, two
stand out.  First, despite more than 10 years of effort, DOE and
Westinghouse have had difficulty resolving such fundamental questions
as how to take reliable top-to-bottom samples from the tanks, what
types of and how much sampling data to gather, and how to reconcile
conflicting information about a tank's contents.  It is unclear
whether the current characterization effort will produce enough
information to support moving to the next steps of the cleanup
effort.  Second, DOE and Westinghouse have not had effective
management systems for detecting and addressing problems with the
characterization program.  In some cases, managers were unaware of
technical and safety problems; in others, they knew about problems
for lengthy periods before addressing them.  Since February 1995,
Westinghouse has taken some steps to address these sampling and
management problems. 

The limited progress in characterizing the high-level radioactive
wastes may have substantial and far-reaching effects on the $36
billion cleanup program for the tanks.  Delays or inadequacies in
characterization could drive up the cost of other parts of the
program and increase the risk that designers could rely on incomplete
or inadequate information as they develop treatment processes and
facilities. 


   BACKGROUND
------------------------------------------------------------ Letter :2

The wastes in Hanford's 177 underground tanks are a by-product of
more than 50 years of nuclear weapons production.  (Fig.  1 shows a
typical tank farm under construction.) Storing, managing, and
cleaning up these wastes pose many challenges.  Some tanks, for
example, contain flammable gas or potentially combustible organic
compounds.  Injected into the tanks as liquids, the wastes have
assumed a variety of forms as they have settled and recombined over
the years.  These forms include sludge and a hard "saltcake" that may
have to be pulverized before it can be removed from the tanks. 
Wastes in at least 67 single-shell\2 tanks have leaked or are assumed
to have leaked into the ground as the concrete-and-steel structures
have deteriorated. 

   Figure 1:  Typical Hanford Tank
   Farm Under Construction

   (See figure in printed
   edition.)

DOE's program for addressing these wastes, called the Tank Waste
Remediation System, calls for a series of actions to chemically
characterize the waste in the tanks, remove it from the tanks, and
prepare it for permanent disposal during the several hundred thousand
years in which some of it will remain dangerously radioactive.  This
program is expected to cost $36 billion over its life-cycle. 

Waste characterization is the first major action.  Since fiscal year
1989, the first year for which reliable cost data exist, DOE has
spent about $260 million on characterization.  The purpose of
characterization is to provide sufficient information for safe
storage of the waste in the tanks while awaiting the development of
processes for remediating it, as well as for designing the steps of
the remediation process itself.  These steps include removing the
wastes (retrieval), separating them into low-level and high-level
portions (pretreatment), treating them (vitrification),\3 and
preparing them for permanent disposal. 

DOE began characterizing the waste in Hanford's tanks in 1985.  Since
then, its efforts have repeatedly fallen behind schedule.  DOE's
March 1994 schedule resulted from the renegotiation of an agreement
originally signed in 1989.  This Tri-Party Agreement with Washington
State's Department of Ecology (Ecology) and EPA initially called for
completing the characterization of single-shell tanks by 1998.  DOE
subsequently found itself unable to comply with the characterization
deadline and renegotiated the agreement.  The revised Tri-Party
Agreement calls for characterizing all 177 tanks for retrieval,
pretreatment, treatment, and disposal by September 1999.\4

In a separate agreement to address concerns raised by the Defense
Nuclear Facilities Safety Board, an independent executive-branch
oversight body, DOE agreed to another set of characterization
requirements related primarily to the safe storage of the waste. 
Under this agreement, DOE was to characterize the 54 "watchlist"
tanks--those with known or suspected safety problems, such as
potential flammability--by October 1995 and to sample and assess
safety conditions associated with all 177 tanks by October 1996. 


--------------------
\2 Hanford has 149 single-shell tanks, built between 1943 and 1964,
which have one layer of steel inside the concrete outer wall.  The
remaining 28 double-shell tanks, built between 1968 and 1986, have
two layers of steel. 

\3 Vitrification, DOE's proposed approach for dealing with both the
high-level and low-level radioactive waste, involves converting the
waste into a glasslike product that will hold the waste in place
until its radioactive components deteriorate. 

\4 The Tri-Party Agreement also recognizes that additional
characterization will occur after the wastes are removed from the
tanks and enter pretreatment and treatment processes. 


   CHARACTERIZATION HAS FALLEN
   BEHIND SCHEDULE
------------------------------------------------------------ Letter :3

DOE and Westinghouse have made limited progress in meeting the
agreed-upon deadlines for characterization.  Despite some recent
improvements in sampling capability, Westinghouse remains behind
schedule in taking samples from the tanks.  Its responses to
reporting requirements have consisted mainly of summarizing
previously known information about the tanks' contents.  In September
1995, DOE acknowledged that the Tri-Party Agreement and Safety Board
characterization deadlines could not be met and proposed a two-phased
approach for characterizing tank wastes that would extend
characterization activities well beyond 1999.  As of the time we
completed our work, however, DOE had not formally notified Ecology
that it could not meet the Tri-Party Agreement deadlines. 


      SAMPLING SCHEDULES HAVE NOT
      BEEN MET
---------------------------------------------------------- Letter :3.1

DOE and Westinghouse have made limited progress in meeting the Safety
Board and Tri-Party agreement deadlines (see fig.  2).  The Safety
Board agreement called for taking and analyzing about 216 of 408 core
samples\5 by September 30, 1995.  As of that date, Westinghouse had
completed 42.  The Safety Board agreement also called for taking core
samples from all 54 watchlist tanks by October 1995; by that date,
Westinghouse had obtained core samples from 10 watchlist tanks.  At
Westinghouse's current estimated sampling pace, all 408 core samples
will not be done until 2002--more than 5 years after the agreement's
October 1996 deadline.  Similarly, while the revised Tri-Party
Agreement calls for full characterization of tank wastes by September
1999, DOE's recent planning documents show that at the current
sampling pace, DOE does not expect to meet this requirement until
September 2004. 

   Figure 2:  Deadlines and
   Current Estimates for
   Characterization Work at
   Hanford

   (See figure in printed
   edition.)

Westinghouse has shown some improvement in its ability to take core
samples.  In fiscal year 1994, Westinghouse completed only 3 of 13
planned core samples, but in fiscal year 1995, it completed 39--the
same number it had estimated it would be able to take.  However, the
39 were concentrated in fewer tanks than Westinghouse had planned.\6

Westinghouse currently estimates that it will be able to take about
four core samples per month through March 1996 and five per month
thereafter. 

During fiscal years 1994 through 1995, Westinghouse also used other
types of samples besides core samples to augment its understanding of
tank wastes, completing 177 of 210 planned samples.\7 These other
methods yield results that are generally considered less
comprehensive than core samples, because they usually do not involve
a top-to-bottom sampling of the waste.  However, the results of these
samples can supplement what is learned from core samples because they
can provide information on tank vapors and liquids that core samples
may not provide.  Even with these other efforts, to date no tank has
yet been sufficiently characterized either to meet the Safety Board's
sampling requirements or to support any of the subsequent steps in
the waste treatment process.  The Director of DOE's characterization
division said that he was unable to estimate when the
characterization of any tank would be completed. 

DOE and Westinghouse officials believe that characterization
difficulties are not affecting the safe storage of tank wastes. 
According to the assistant manager of DOE's tank waste remediation
program, Westinghouse has placed controls on tank farm operations to
reduce the risk of an unintentional release.  The controls include
using sparkless equipment and avoiding certain types of drilling
procedures.  However, these controls have made it more difficult for
DOE to maintain its desired sampling rate. 

DOE and Westinghouse characterization managers have acknowledged that
they will not meet the Safety Board or Tri-Party Agreement deadlines. 
They have prepared a draft revision of the original implementation
plan agreed to with the Safety Board.  This draft, which does not
specify a completion date, is discussed in more detail in the next
section.  DOE and Westinghouse also acknowledged the need to propose
changes to the Tri-Party Agreement, but at the time we completed our
work, they had not yet submitted a formal proposal to Ecology or EPA. 
However, Ecology has informed DOE and Westinghouse that in the
state's view, the inability to obtain adequate samples does not
provide sufficient grounds for renegotiating the characterization
milestones of the Tri-Party Agreement.  Ecology expects DOE to
formulate a plan to compensate for past inadequacies and to meet
commitments under the current Tri-Party Agreement schedule. 


--------------------
\5 A core sample is a top-to-bottom sample of the tank waste using a
hollow-core drill bit about 1 inch in diameter. 

\6 Westinghouse planned to take core samples from 30 tanks in fiscal
year 1995.  However, it took the samples from only 16 of the 30 tanks
because it needed to obtain more complete information from the tanks
sampled. 

\7 These other types of samples are vapor, auger, and grab samples. 
Vapor samples involve extracting gases from the space at the top of
the tanks.  Auger samples, considered a cost-effective method of
obtaining samples from tanks with less than 23 inches of waste, are
obtained using a large, steel, hand-turned drill bit to sample 15
inches of the waste.  Grab samples are taken using a bottle attached
to a cable that is lowered into the surface of soft or liquid waste. 


      CHARACTERIZATION REPORTS
      MAINLY SUMMARIZE INFORMATION
      ALREADY KNOWN
---------------------------------------------------------- Letter :3.2

The revised Tri-Party Agreement requires DOE to submit, for Ecology's
approval, characterization reports on a certain number of tanks each
year through 1999.\8 All 23 of the reports submitted through fiscal
year 1994 were based mostly on existing "historical" data about the
wastes, supplemented with limited sampling results.  The 30 reports
submitted on September 30, 1995, were also based primarily on
historical information, although they contained some results of
samples taken since May 1989. 

Reaction to the value of these reports is mixed.  DOE accepted the
reports, but Ecology, the body that must approve the reports, has
criticized their contents, including Westinghouse's extensive
reliance on historical information.\9 Rather than approve the 23
reports submitted in fiscal years 1993 and 1994, Ecology received
them on the condition that Westinghouse develop additional
characterization data and resubmit the reports for approval within 2
years.  In connection with the fiscal year 1995 reports, Ecology's
characterization team leader said that 25 of the 30 reports were
inadequate because they contained mainly historical information,
modeling results that had not been verified, and limited analytical
results.  He considered the results of the analyses of the most
recent samples insufficient because they were limited to determining
whether tank wastes were being maintained in a safe condition and did
not contribute to any remediation step. 


--------------------
\8 The agreement calls for reports on 3 tanks for fiscal year 1993,
20 in fiscal year 1994, 30 in 1995, 40 in 1996 and 1997, 30 in 1998,
and 14 in 1999. 

\9 Because the reports were submitted as required, the director of
DOE's characterization division told us that he authorized an
incentive payment to Westinghouse.  For fiscal year 1995, DOE
provided Westinghouse with a payment of $843,333 for submitting the
required 30 reports by September 30, 1995.  Under the terms of DOE's
draft performance criteria, Westinghouse can earn $2 million for
submitting the required 40 reports in fiscal year 1996 if the reports
meet DOE's requirements. 


   CHARACTERIZATION PROGRESS HAS
   BEEN IMPEDED BY TECHNICAL
   DIFFICULTIES AND LACK OF
   EFFECTIVE MANAGEMENT CONTROLS
------------------------------------------------------------ Letter :4

While many factors have contributed to the slow pace of the
characterization effort, the primary reasons for slow progress are
that (1) DOE and Westinghouse have not yet determined how to
successfully draw reliable samples and characterize the waste and (2)
managerial weaknesses with the characterization program and other
aspects of the tank farms have exacerbated delays and contributed to
operational inefficiencies. 


      CHARACTERIZATION APPROACH
      CONTINUES TO CHANGE
---------------------------------------------------------- Letter :4.1

DOE and Westinghouse have been unable to develop and implement a
characterization approach that has been successful in meeting the
requirements of the Tri-Party Agreement or their commitments to the
Safety Board.  Three different approaches have been attempted since
the Tri-Party Agreement was signed in 1989, and in each case, the
approach has not been successfully implemented.  As a result, DOE and
Westinghouse are still having difficulty answering fundamental
methodological questions that have existed since the characterization
program began:  how to take reliable samples, what types of and how
much sampling data to gather, and how to reliably predict the waste
constituents on the basis of sampling and other data. 

Limited progress in taking samples has led DOE and Westinghouse to
search for another characterization approach that would satisfy the
requirements of the Tri-Party Agreement as well as their commitments
to the Safety Board.  The latest approach, proposed in September
1995, aims to evaluate the program on a continuing basis while taking
samples and collecting data, rather than following a fixed schedule. 
DOE now views sampling as an iterative process in which the total
number of samples needed depends on the results of the initial
samples taken.  This viewpoint is in contrast to DOE's earlier
strategy that required a fixed number of samples. 

DOE's latest characterization strategy has two phases:  Phase one
concentrates primarily on ensuring safe storage while demonstrating
the approach's overall validity.  More specifically, phase one
includes

  reducing the number of core samples from 408 to 109, which would be
     done by grouping tanks believed to contain similar wastes and
     taking core samples from the 25 to 30 tanks considered
     representative of the various groups;

  supplementing data obtained from the reduced number of core samples
     with historical data, temperature and moisture measurements, and
     data obtained from other types of samples (such as vapor and
     auger samples); and

  using computer models to analyze the various data in order to
     predict the tanks' contents and evaluate the risk of potential
     combustion. 

Westinghouse characterization project officials acknowledged that if
assumptions about the validity of this approach are not successfully
demonstrated during fiscal years 1996 and 1997, more core samples and
other types of samples may be needed in phase one. 

Phase two of Westinghouse's approach would focus primarily on the
characterization of tank wastes to support the treatment and disposal
steps.  The time frames, funding requirements, and sampling
strategies for phase two are currently undefined.  After evaluating
the results of phase one, Westinghouse plans to formulate and
implement phase two beginning in fiscal year 1997. 


      ADEQUACY OF PROPOSED
      APPROACH IS UNCERTAIN
---------------------------------------------------------- Letter :4.2

The September 1995 characterization proposal, while more in line with
what DOE can realistically expect to accomplish in the next several
years, has generated concerns about whether it will provide
sufficient characterization information to proceed with remediation
efforts.  These concerns have been expressed by regulators, advisory
bodies, and other persons involved with the remediation effort. 
Among the main concerns about the adequacy of the proposed approach
are whether historical and sampling data can be reconciled, whether
the computer models will reliably predict actual quantities of
specific tank wastes, and whether the information being developed
will be thorough and accurate enough to proceed with the various
steps in retrieving and treating the wastes. 


         DISPARITIES EXIST BETWEEN
         HISTORICAL DATA AND
         SAMPLING INFORMATION
-------------------------------------------------------- Letter :4.2.1

Having accurate knowledge of the amounts of waste components is
important in choosing pretreatment and treatment technologies and
designing facilities.  Westinghouse's new approach would rely to a
significant extent on historical data for purposes of
characterization.  However, as far back as October 1991, DOE
recognized that its historical data were incomplete and unreliable. 
The Safety Board has expressed concern that historical data are not
complete, reliable, or representative because inadequate operational
controls have resulted in limited information about (1) the specific
types of waste placed in tanks and (2) the chemical processes
occurring in the tanks.  DOE and Westinghouse have acknowledged that
such disparities exist and plan to determine how to resolve them in
phase one of their proposed approach. 


         RELIABILITY OF COMPUTER
         MODELS IS UNTESTED
-------------------------------------------------------- Letter :4.2.2

Computer models are key components in Westinghouse's approach to
predict tank waste constituents and evaluate potential safety
problems.  However, their reliability is largely untested. 
Westinghouse plans to test the models' reliability during a
"demonstration" period in fiscal years 1996 and 1997, when some tank
wastes will be characterized and compared to the models' predictions. 
The Safety Board, among others, has raised concerns about
reliability.  The Board's technical staff concluded that "Significant
portions of the new strategy are based on simplified models and
simulants that may not adequately represent tank wastes."

The limited information available to date shows examples of
substantial differences between the models' projections and the data
obtained through core samples.  In September 1995, Westinghouse
reported the results of 144 possible comparisons between the models'
projections and the core sample results in 12 tanks.  These
comparisons included waste constituents, such as chromium and
phosphate, that are important in determining the volume of glass
needed for the vitrification process.  The core sample results were
at least three times higher or lower than the models' projections in
about 25 percent of the comparisons.  In addition, for another 20
percent of the comparisons, the sample results showed the presence of
such constituents as cesium 137, phosphate, and total organic
carbons, while the models' projections indicated that these
constituents did not exist in the tanks.\10


--------------------
\10 The remaining comparisons differed from the models' results by a
factor of less than three. 


         SUFFICIENCY OF
         CHARACTERIZATION
         INFORMATION IS UNCERTAIN
-------------------------------------------------------- Letter :4.2.3

Considerable uncertainty exists about the characterization
information needed to design methods and facilities for cleanup. 
Westinghouse is currently developing criteria known as data quality
objectives; these criteria specify what information will be needed
for each step in remediating tank wastes.  Although Westinghouse's
approach is based on obtaining sufficient data from a reduced number
of core samples, in 1992 Westinghouse told us that between 2 and 14
core samples may be needed to adequately characterize a single tank. 

We asked an independent nuclear engineering consultant to review
Westinghouse's approach to determine if it will produce sufficient
information about the waste to meet DOE's objectives.\11 His review
raised concerns about the reliability of the information that will be
developed for most of the tanks under Westinghouse's proposed
approach.  The consultant concluded that the proposed approach
involving limited sampling may yield adequate characterization
information for about 31 single-shell tanks and several double-shell
tanks believed to contain waste that is relatively homogenous, but he
concluded that the approach may be considerably less reliable for
approximately 135 other tanks.  For those tanks, the consultant
concluded that more core samples than originally planned, rather than
fewer, may be needed to reconcile disparities between the tanks'
waste contents as derived from sample analyses and the tanks' waste
contents as deduced from historical data. 


--------------------
\11 For specific technical assistance in performing our review, we
used the services of Dr.  George W.  Hinman, an independent nuclear
engineering consultant, who is currently Director, Office of Applied
Energy Studies, Washington State University.  Dr.  Hinman has worked
for about 40 years in the nuclear energy field in industry,
government, and academia. 


      MANAGERIAL WEAKNESSES HAVE
      CONTRIBUTED TO OPERATIONAL
      INEFFICIENCIES
---------------------------------------------------------- Letter :4.3

The technical complexities associated with characterizing tank wastes
highlight the need for an effective management system for detecting
and addressing problems.  However, such a system has been lacking in
the characterization program.  Instead, technical and safety problems
have gone uncorrected for considerable periods, either because
managers were unaware of the problems or because they were slow to
take action on problems they knew about.  In an April 1995 letter to
the Safety Board, the Secretary of Energy acknowledged that numerous
problems affecting the characterization program have been caused by
ineffective management. 


         LACK OF AWARENESS LIMITED
         TIMELY ACTION
-------------------------------------------------------- Letter :4.3.1

We found instances involving operational and safety-related problems
in which Westinghouse or DOE managers were initially unaware of
circumstances that caused delays or increased safety risks.  For
example: 

  Before being placed into service, a new rotary-mode sampling truck
     was inspected in July 1994, 3 months later than planned. 
     Westinghouse inspectors reported that welds on the truck did not
     meet design code requirements.  Consequently, the truck was
     inoperable for an additional 3 months; its unavailability
     contributed to Westinghouse's obtaining considerably fewer
     rotary-core samples in 1994 than originally planned.  The
     report's authors stated that Westinghouse management lacked
     commitment in identifying and tracking such deficiencies. 

  Westinghouse managers were not aware that workers were operating
     the push-mode core sampler without an operable instrument called
     a bottom detector to prevent damaging or drilling through the
     bottom of a tank.  When this practice was reported in February
     1995, the report stated that the workers involved did not have
     adequate knowledge of safe sampling procedures.  As a result,
     push-mode sampling was halted for more than 2 weeks while safety
     procedures were reevaluated and workers received additional
     training, according to Westinghouse's deputy operations manager
     for characterization. 


         ACTIONS WERE OFTEN SLOW
         EVEN WHEN MANAGEMENT WAS
         AWARE OF PROBLEMS
-------------------------------------------------------- Letter :4.3.2

We also found instances in which management was aware of problems
with characterization--or with tank farm maintenance activities
affecting characterization--but was slow to address them.  For
example: 

  In 1993, virtually all sampling activity was suspended for more
     than 6 months following a safety violation in which a
     maintenance worker contaminated himself and others while using
     unapproved procedures to unclog a blocked drain.  This incident,
     referred to as the "rock-on-a-rope" occurrence because of the
     extremely primitive methods used, was the culmination of a
     series of incidents that indicated deficiencies in operations at
     the Hanford site, including inadequate procedures and
     personnel's lack of awareness of important technical procedures. 
     DOE took limited actions on these incidents until this
     substantial event occurred. 

  If wind speeds exceed 15 miles per hour--a common occurrence at
     Hanford--core samples cannot be taken unless sampling equipment
     and operations are protected from the wind.  Although sampling
     delays associated with Hanford's windy conditions have been
     apparent for years, no solutions were advanced until January
     1995, after the Safety Board had suggested many times for more
     than a year that a wind barrier be fabricated. 

These examples of management's ineffectiveness are supported in
several broader studies.  In October 1990, for example, the Safety
Board issued a statement concluding that management's attention to
the characterization effort was inadequate.  More than 2 years later,
an internal DOE review found that there were "significant weaknesses
in the safe control, adequate management, and technical
implementation of field, laboratory, and supporting project
activities." In January 1995, about 2 years later, DOE acknowledged
to the Safety Board that confusion still existed over who, at the
management level, was responsible for managing and coordinating
various characterization activities. 

In a 1995 letter to the Secretary of Energy, the Safety Board placed
the management issue at the center of the lack of progress.  It said
the principal source of difficulty was

     "the loose and ineffective structure of the technical and
     administrative organizations assigned to characterization of the
     waste tanks.  That has caused numerous delays for relatively
     trivial reasons that could have been readily overcome by a
     strong and determined manager with sufficient authority.  .  . 
     ."


   CHARACTERIZATION DELAYS COULD
   INCREASE COSTS AND ADD TO
   CLEANUP UNCERTAINTY
------------------------------------------------------------ Letter :5

DOE's problems in keeping characterization on schedule affect more
than just compliance with various agreements.  Other problems include
potential cost increases for the characterization effort, inefficient
use of a laboratory, and uncertainty about carrying out other aspects
of the remediation program. 


      FINAL COST OF
      CHARACTERIZATION IS UNKNOWN
---------------------------------------------------------- Letter :5.1

Delays in characterizing tank wastes raise the likelihood that DOE's
most recent estimates of total characterization costs are
understated.  Since fiscal year 1989, the earliest date that reliable
cost data were available, DOE has spent about $260 million on
characterization, and in August 1995, it estimated that it would need
to spend at least $569 million more through fiscal year 1999 to meet
the Safety Board and Tri-Party agreements. 

Characterization work beyond fiscal year 1999 will most likely
include sampling the majority of the 177 tanks to obtain information
to support retrieval, pretreatment, treatment, and disposal of the
wastes.  The amount of additional program funds needed after fiscal
year 1999 to support these characterization activities has not been
estimated.  The director of DOE's characterization division said that
DOE will not know for several years what these program costs are
likely to be. 


      ANALYTICAL LABORATORY WAS
      NOT EFFICIENTLY USED
---------------------------------------------------------- Letter :5.2

Delays in sampling the wastes affected utilization of the 222-S
analytical laboratory, a facility Westinghouse operates at Hanford. 
This laboratory, which was recently expanded to deal with the
expected volume of incoming samples, has a staff of more than 143
full-time-equivalent positions and a budget of nearly $12 million.\12

Westinghouse anticipated that about two-thirds of the laboratory's
capacity would be needed to analyze the samples.  However, during
fiscal year 1995, the lower-than-expected volume of samples required
only 25 percent of the total capacity available.  Laboratory
personnel used an additional 20 percent of capacity during the year
in various activities, such as developing internal procedures and
process controls.  Consequently, more than half of the laboratory's
capacity went unutilized.  According to the analytical laboratory's
performance documents, sample analysis has increased to 53 percent of
total capacity in the first 2 months of fiscal year 1996 because more
samples had recently been obtained. 


--------------------
\12 We previously reported on the need for additional laboratory
space in Nuclear Waste:  Hanford Tank Waste Program Needs Cost,
Schedule, and Management Changes (GAO/RCED-93-99, Mar.  8, 1993). 


      REMAINING REMEDIATION STEPS
      MAY BE DELAYED AND COST MORE
---------------------------------------------------------- Letter :5.3

The most significant effect of the delays may be on the rest of the
steps in the remediation process--retrieval, pretreatment, treatment,
and disposal.  These other steps of the process depend on the
adequacy and quality of characterization information.  In particular,
data on the quantities and chemical properties of such waste
components as chromium, phosphate, cesium 137, zirconium, and
plutonium are important for determining the most efficient
pretreatment technologies and the design of treatment facilities. 
Insufficient and untimely characterization information could either
delay the construction of those facilities or cause the construction
to proceed without sufficient information, increasing the risk of
costly errors. 

DOE and Westinghouse do not share the view that limited
characterization information could jeopardize the success of
subsequent steps in the remediation process.  Despite
slower-than-expected progress on characterization, DOE plans to begin
testing equipment for retrieving tank wastes in fiscal year 1996 and
to begin designing treatment facilities the following year. 
According to the assistant manager of DOE's tank waste remediation
program, DOE believes that (1) current knowledge is sufficient to
proceed with the initial facility design and retrieval of selected
wastes and (2) additional characterization information will be
available before the design reaches a critical phase.  Ecology's
position is that sufficient data to begin designing treatment
facilities currently exist, but if DOE is unable to characterize tank
wastes at its expected rate, the lack of characterization information
could ultimately jeopardize the success of the remediation program. 

Rather than constructing and operating its own facilities to treat
the tank waste, DOE is considering privatization as an alternative
approach.  Under such an approach, a company or a consortium of
companies from the private sector would finance, design, build, and
operate pretreatment and treatment facilities and deliver the
finished product--in this case, vitrified waste encased in stainless
steel containers--to DOE for a fee.  DOE expects this approach to
save billions of dollars because the potential for innovation in the
private marketplace could lead to greater efficiencies and improved
performance.  Authorization to pursue this approach was obtained from
the Secretary of Energy in late September 1995.  In November 1995,
DOE drafted a request for proposals to be reviewed by interested
parties and expects to issue this request in February 1996. 

Whether or not DOE moves ahead with privatization, it will be
responsible for providing the characterization information necessary
to proceed with remediating the tank wastes.  DOE's current strategy
of proceeding with limited characterization information could
increase the risk that facilities may not perform as needed and/or
may need costly modifications to perform safely and efficiently.  For
example, DOE has conflicting data on the quantities of key waste
constituents, such as chromium and phosphate, that affect the quality
and durability of the vitrified glass product in which the waste will
be immobilized.  The independent nuclear engineering consultant we
asked to review the program believes that without more reliable
information on these and other elements, proceeding to construct the
facilities is risky because they (1) may be built with insufficient
capacity to process wastes containing greater-than-expected
quantities of certain components or (2) may be built with excess
capacity, resulting in needless expense.  In our previous work on
DOE's waste-processing facilities, we found that DOE had experienced
major start-up problems, cost increases, and schedule delays caused
in part by a "fast track" approach where construction began before
major technical uncertainties were resolved.\13


--------------------
\13 For more information on this issue, see Nuclear Waste:  Defense
Waste Processing Facility--Cost, Schedule, and Technical Issues
(GAO/RCED-92-183, June 17, 1992). 


   DOE AND WESTINGHOUSE HAVE TAKEN
   SOME ACTIONS TO IMPROVE
   CHARACTERIZATION
------------------------------------------------------------ Letter :6

In its performance evaluation of Westinghouse for the period from
October 1994 through March 1995, DOE pointed out substantial problems
with Westinghouse's management of the characterization effort.  The
evaluation stated as follows: 

     "[Westinghouse] has been unsuccessful in demonstrating tangible
     results on waste characterization; total program performance is
     considered a significant deficiency for the evaluation period. 
     Limited progress has been demonstrated on field sampling
     improvements, technical basis development, or program
     optimization through process and productivity improvements. 
     Adverse cost and schedule performance during the evaluation
     period indicate a strong probability to exceed available program
     funds without corrective action."

DOE and Westinghouse have begun some efforts to bring greater
management control to the characterization effort, which had four
different DOE managers between August 1994 and July 1995.  In
February 1995, DOE changed the effort from a "program" to a "project"
with a single manager.  The Westinghouse project manager reports
directly to the Westinghouse vice president for tank waste
remediation.  DOE and Westinghouse also clarified the lines of
reporting accountability within the program and increased the amount
of time that managers spend in the field observing characterization
activities. 

Other accomplishments that DOE and Westinghouse reported over the
past 8 months include the completion of a variety of technical
documents outlining DOE's new characterization strategy;
characterization criteria, called data quality objectives, describing
tank waste safety, disposal, and historical data requirements;
historical tank content estimate reports for all tanks; and 42 vapor
tank sampling and characterization reports to address noxious vapor
concerns.  To enhance their capability to sample tank wastes, DOE and
Westinghouse deployed three rotary-mode core sampling trucks and
placed into operation an X-ray imaging device to provide real-time
data on the recovery of core samples.  These changes hold some
promise for improvement, but it is too early to tell if they fully
address the operational deficiencies and management weaknesses that
have plagued the characterization program to date. 

For the period during which many of these improvements were made
(April 1995 through September 1995), DOE gave Westinghouse a
"satisfactory" performance rating for its characterization project
achievements.  DOE's evaluation praised Westinghouse for increasing
management's attention to program issues and developing and issuing
"sound technical basis" documents to the Defense Nuclear Facilities
Safety Board.  However, the evaluation also stated: 

     "Information provided by the contractor through the Data Quality
     Objective (DQO) process has been insufficient to determine when
     a tank is fully characterized and [when] the need for further
     sampling is no longer required.  .  .  .  The root of this
     problem is a lack of adequate discipline in the definition of
     characterization needs and objectives, and the subsequent
     operations executed to accomplish those needs/objectives.  The
     result has been excessive cost due to inefficient sampling and
     an inadequacy of data required to meet DQOs."


   CONCLUSIONS
------------------------------------------------------------ Letter :7

After more than 10 years and about $260 million invested in trying to
characterize the tank wastes at Hanford, little definitive progress
has occurred.  Disagreement still exists over how much and what kind
of characterization data are needed to reliably predict actual
quantities of waste constituents and build appropriate treatment
facilities.  Inadequate management attention has impeded solutions to
these problems.  DOE's current proposal raises questions about
whether enough characterization information will be available to
build effective facilities for retrieving the wastes and preparing
them for permanent disposal.  If the information proves to be
inadequate, further technical problems and cost overruns are likely,
jeopardizing the success of the overall program and increasing the
potential that funds may be used unwisely. 

All parties, including DOE, Westinghouse, potential private
contractors, and the Congress, need further assurance that the
characterization program has a sound technical foundation.  Answers
are needed to such questions as how much sampling and what kinds of
sampling methods are sufficient to reliably characterize a tank's
contents; how to reconcile disparities between existing data on tank
contents and actual waste sample data; and how much characterization
information is needed before the design and construction of
pretreatment and treatment facilities should begin.  Without this
information, it will be difficult to reliably predict when the
overall program will be done or how much it will cost.  Furthermore,
these uncertainties could undermine the savings DOE expects to
realize by privatizing the tank waste remediation program. 


   RECOMMENDATIONS
------------------------------------------------------------ Letter :8

To ensure that Hanford's tank waste characterization program will
provide a sound foundation for designing and building waste treatment
facilities, we recommend that the Secretary of Energy commission an
independent review of the characterization program, using an
organization such as the National Academy of Sciences, to resolve
questions about the technical adequacy of Hanford's characterization
strategy.  The review should focus on determining (1) how much and
what kind of information is sufficient to reliably characterize the
tank wastes and predict the quantities and conditions of the waste
constituents and (2) the amount and quality of characterization
information needed for DOE to proceed with the design and
construction of waste treatment facilities. 

To ensure that funds for the overall tank waste remediation program
are spent as wisely as possible, we recommend that the Secretary of
Energy defer funding the construction of pretreatment and treatment
facilities until (1) the technical adequacy of the characterization
program has been confirmed or established and (2) sufficient waste
characterization information is available to reliably define the
requirements of those facilities. 


   AGENCY COMMENTS AND OUR
   EVALUATION
------------------------------------------------------------ Letter :9

We provided a draft of this report to DOE, the Westinghouse Hanford
Company, and the Washington State Department of Ecology for their
review and comment.  We discussed the report with officials from DOE,
Westinghouse, and Ecology, including the assistant manager for DOE's
Tank Waste Remediation System, the director of DOE's characterization
division, and the director of DOE's safety division; the director of
Westinghouse's tank waste characterization project and the ecology
coordinator for Westinghouse's Tank Waste Remediation System. 
Overall, the officials agreed that the report was accurate and
factual; however, DOE, Westinghouse, and Ecology disagreed with
several aspects of the report, including the tone and substance of
our conclusions and recommendations.  In addition, DOE, Westinghouse,
and Ecology provided annotated comments on technical aspects of the
draft.  We have incorporated those comments where appropriate. 

DOE said that while the report accurately describes past difficulties
with the characterization program, the report does not adequately
recognize the performance improvements accomplished since February
1995.  DOE provided a list of accomplishments that included
developing program strategy documents, completing tank safety
analyses and tank characterization reports, performing laboratory
testing of wastes, acquiring new drilling equipment, and increasing
the number of samples taken.  In determining DOE's progress, we used
the criteria in the agreements DOE has signed with Washington State,
EPA, and the Safety Board.  We also developed and compared data on
waste samples planned and accomplished, since sample data are
essential for completing characterization.  In addition, we
documented other activities DOE accomplished that could help in
meeting characterization objectives and included many of them in our
report.  We believe that we have adequately emphasized that DOE
recently has made more progress in taking and analyzing samples and
in completing other characterization activities. 

DOE and Westinghouse were concerned that our report could be
interpreted to mean that because of problems with the
characterization program, DOE could not ensure that the tanks are
safe.  DOE and Westinghouse officials believe that the tanks are safe
because of the controls they have put in place over tank farm
operations, including sampling activities.  These controls are
designed to prevent harmful releases due to such conditions as high
temperature and/or flammability risks.  We did not evaluate the level
of safety associated with the tank wastes and did not intend to imply
that the tanks are unsafe because of deficiencies in the
characterization program.  We revised our report to clarify that DOE
believes the tank wastes are being safely stored.  However, our
report does explain that the Safety Board has directed DOE to conduct
additional characterization of tank wastes to ensure that they are
safely stored.  Until those characterization activities are complete
and the waste constituents are better understood, DOE has placed
controls over the tanks to provide an added level of assurance. 

While DOE officials agreed with the value of having an outside
technical review of the characterization program, they noted that DOE
had recently begun such an effort with one of its contractors,
Pacific Northwest National Laboratory.  Specifically, DOE is planning
to fund a study team to help resolve critical uncertainties related
to the safety of the tank wastes.  The team will also develop an
approach to integrate data needs to support all aspects of the tank
waste program, from current operations to treatment and disposal of
the wastes.  This proposal was drafted in December 1995, after we
completed our field work.  On the basis of the initial documentation
DOE provided, it appears the study team will focus on resolving
uncertainties related to tank safety rather than evaluate the
viability of the characterization program.  We continue to believe
that a group, independent of ongoing tank waste and related DOE
activities, needs to address the technical feasibility of the
characterization strategy. 

DOE and Westinghouse disagreed with our view that difficulties in
characterizing the wastes could affect the remaining steps in the
disposal program, including the design and construction of facilities
to treat the wastes.  DOE said that at sampling rates achieved since
March 1995, it expects characterization of the wastes to be complete
by 2004, which DOE believes is adequate to support the disposal
program.  In addition, DOE said that enough information exists now to
proceed with the design and construction of treatment facilities. 
DOE expressed concern that any deferral of funding for remediation
could make it difficult to keep the project moving forward to
accomplish waste treatment and disposal.  Ecology officials shared
similar concerns. 

We believe our report accurately describes the potential effect that
characterization difficulties could have on the remaining steps in
the remediation program.  For example, even DOE's latest schedule for
completing characterization could be in jeopardy.  First, DOE is
unsure if Westinghouse can maintain a core sampling rate of four to
five samples per month.  Recent additional controls placed on the
tanks and other sampling problems may make this sampling rate
difficult to achieve.  Second, DOE's projected completion of
characterization by 2004 is based on a characterization approach that
has not been validated.  Third, quantities of certain waste
constituents need to be determined to minimize uncertainties that
affect the construction of treatment facilities.  On the basis of our
discussions with DOE and Ecology, we agree that facility design
activities could proceed as characterization work continues, and we
have modified our report accordingly.  However, on the basis of this
report and our previous work, we continue to believe that
construction of treatment facilities should not be funded until the
technical adequacy of DOE's characterization strategy is confirmed or
established by independent sources and sufficient waste
characterization information is available to reliably define the
requirements of those facilities. 


   SCOPE AND METHODOLOGY
----------------------------------------------------------- Letter :10

Most of our work was performed at DOE's Hanford site in Washington
State.  To determine DOE's progress in meeting the tank waste
characterization commitments, we reviewed tank waste characterization
milestones that DOE committed to with the Washington State Department
of Ecology and the Environmental Protection Agency in the Tri-Party
Agreement and with the Defense Nuclear Facilities Safety Board.  We
compared these commitments with DOE's actual sampling results through
September 1995, the latest month for which data were available. 

To identify impediments to progress and determine what impact these
impediments could have, we reviewed tank characterization reports,
engineering studies, characterization technical basis documents,
characterization project strategy documents, and other materials.  We
reviewed Ecology and Safety Board reports and correspondence with DOE
on concerns associated with the characterization program, and we
reviewed program documents detailing program improvements in the
management of the characterization project and in Westinghouse's
sampling capability.  We also reviewed DOE's and Westinghouse's cost
estimates of the characterization program and the Tank Waste
Remediation System. 

We supplemented our reviews of reports and other documentation by
interviewing DOE and Westinghouse officials, including the assistant
manager for DOE's Tank Waste Remediation System, the director of
DOE's characterization division, Westinghouse's vice president for
tank waste remediation, and various others with program
responsibilities.  We also interviewed officials from oversight
agencies, including Ecology's Tank Waste Remediation System
coordinator and characterization team leader, and Defense Nuclear
Facilities Safety Board members and their staff. 

We conducted our work from May 1995 through January 1996 in
accordance with generally accepted government auditing standards. 


--------------------------------------------------------- Letter :10.1

As you know, 31 U.S.C.  720 requires the head of a federal agency to
submit a written statement of the actions taken on our
recommendations to the Senate Committee on Governmental Affairs and
the House Committee on Government Reform and Oversight not later than
60 days after the date of this letter and to the House and Senate
Committees on Appropriations with the agency's first request for
appropriations made more than 60 days after the date of this letter. 

We are sending copies of this report to appropriate congressional
committees and other interested parties.  We will also make copies
available to others on request. 

Please call me at (202) 512-3841 if you or your staff have any
questions.  Major contributors to this report are listed in appendix
I. 

Sincerely yours,

Victor S.  Rezendes
Director, Energy, Resources, and
 Science Issues


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix I

RESOURCES, COMMUNITY, AND ECONOMIC
DEVELOPMENT DIVISION, WASHINGTON,
D.C. 

Bernice Steinhardt, Associate Director
James Noel, Assistant Director

SAN FRANCISCO/SEATTLE FIELD OFFICE

William R.  Swick, Core Group Manager
Thomas C.  Perry, Evaluator-in-Charge
Robert J.  Bresky, Staff Evaluator
Drummond E.  Kahn, Staff Evaluator
Stanley G.  Stenersen, Senior Evaluator


*** End of document. ***