F I N A L R E P O R T - N O V E M B E R 1 9 9 7

Oak Ridge OperationsU.S. Department of Energy

Type B Accident

Investigation

Board

Report on the

Drum Explosion at

Building C-746-Q,

Paducah Gaseous

Diffusion Plant

DOE/ORO/2061

12/97EASI 704

N O V E M B E R 1 9 9 7

Oak R idge Opera t i on sU.S. Department of Energy

Type B AccidentInvestigation Board Report

on theDrum Explosion at

Building C-746-Q,Paducah Gaseous Diffusion Plant

his report is an independent product ofthe Type B Accident Investigation Board(Board) appointed by James C. Hall,Manager, Oak Ridge Operations.

The Board was appointed to perform aType B Investigation of this accident and toprepare an investigation report in accor-dance with U.S. Department of EnergyOrder 225.1, �Accident Investigations.�

The discussion of facts, as determined bythe Board, and the views expressed in thereport do not assume and are not in-tended to establish the existence of anyduty at law on the part of the U.S. Govern-ment, its employees or agents, contractors,their employees or agents, or subcontrac-tors at any tier, or any other party.

This report neither determines norimplies liability.

TTTTT

INDEPENDENT REPORT

n September 17, 1997, I establisheda Type B Accident Investigation

Board (Board) to investigate thedrum explosion that resulted in the

spill of hazardous/radioactive waste(mixed waste) within Building

C-746-Q located at the PaducahGaseous Diffusion Plant. The Board�sresponsibilities have been completedwith respect to this investigation. The

analysis process; identification ofdirect, contributing, and root causes;

and development of judgments ofneed during the investigation

were done in accordance withU.S. Department of Energy

Order 225.1, �Accident Investiga-tions.� I accept the findings of the

Board and authorize the release ofthis report for general distribution.

James C. HallManager

Oak Ridge Operations

RELEASE AUTHORIZATION

OOOOO

LIST OF TABLES, FIGURES, AND EXHIBITSLIST OF ABBREVIATIONS, ACRONYMS, AND INITIALISMSEXECUTIVE SUMMARY

1.0 INTRODUCTION

1.1 FACILITY DESCRIPTION1.2 SCOPE, PURPOSE, AND METHODOLOGY

2.0 FACTS AND ANALYSIS

2.1 ACCIDENT DESCRIPTION AND CHRONOLOGY

2.1.1 Accident Description2.1.2 Background of Similar Events2.1.3 Chronology of Events2.1.4 Accident Reconstruction and Analysis2.1.5 Emergency Response2.1.6 Personnel Safety2.1.7 Occurrence Reporting

2 .2 MANAGEMENT SYSTEMS AND CONTROLS

2.2.1 Policies and Procedures2.2.2 Roles and Responsibilities2.2.3 LMES Assessment Program

2.3 CONTROLS AND ANALYSIS

2.3.1 Barrier and Control Analysis2.3.2 Change Analysis2.3.3 Causal Factor Analysis

3 .0 CONCLUSIONS AND JUDGMENTS OF NEED

4.0 BOARD SIGNATURES

5.0 LISTING OF BOARD MEMBERS, ADVISORS, AND STAFF

APPENDIX A APPOINTMENT CORRESPONDENCEFOR TYPE B ACCIDENT INVESTIGATION

TABLE OF CONTENTS

i

i ii i iiv

1

1

3

3

3568

1 21 31 4

1 4

1 41 61 9

2 0

2 02 02 0

2 4

2 5

2 6

Table ES-1 Causal Factor AnalysisTable ES-2 Conclusions and Judgements of Need

Table 2-1 Barrier and Control AnalysisTable 2-2 Change AnalysisTable 2-3 Causal Factor Analysis

Table 3-1 Conclusions and Judgments of Need

Figure ES-1 Summary Events and Causal Factors Chart

Figure 1-1 Layout of Building C-746-Q

Figure 2-1 Overpack Configuration Prior to ExplosionFigure 2-2 Summary Events and Casual Factors ChartFigure 2-3 Accident Layout

Exhibit ES-1 Components of HC-1093

Exhibit 1-1 Exterior of Building C-746-Q

Exhibit 2-1 Components of HC-1093Exhibit 2-2 Bottom of 55-gal Polyethylene Drum

LIST OF TABLES, FIGURES, AND EXHIBITS

v iv i i

2 12 22 3

2 4

v

2

47

1 0

i v

2

51 1

i i

AMEF Assistant Manager for Enrichment FacilitiesAMEM Assistant Manager for Environmental Management

CFR Code of Federal Regulations

DOE U.S. Department of Energy

ERD Environmental Restoration DivisionERT Emergency Response TeamESAMS Energy Systems Action Management SystemETTP East Tennessee Technology Park

HAZMAT hazardous materialHAZWOPER Hazardous Waste Operations and

Emergency Response

L literLMES Lockheed Martin Energy Systems, Inc.LMUS Lockheed Martin Utility Services, Inc.

M molar concentration

ORO Oak Ridge OperationsORPS Occurrence Reporting and Processing SystemOSHA Occupational Safety and Health Administration

PCB polychlorinated biphenylPGDP Paducah Gaseous Diffusion PlantPORTS Portsmouth Gaseous Diffusion PlantPPE personal protective equipmentPSS Plant Shift Superintendent

RCRA Resource Conservation and Recovery ActRWP radiological work permit

SME Subject Matter Expert

TRU transuranic

USEC United States Enrichment Corporation

LIST OF ABBREVIATIONS, ACRONYMS, AND INITIALISMS

i i i

IIIIINTRODUCTIONNTRODUCTIONNTRODUCTIONNTRODUCTIONNTRODUCTION

On September 15, 1997, two health physics tech-nicians and two waste management operators dis-covered evidence of a drum explosion. This acci-dent, involving a spill of hazardous/radioactivewaste (mixed waste), occurred in Building C-746-Q at the Paducah Gaseous Diffusion Plant (PGDP),Paducah, Kentucky. On September 17, 1997,James Hall, Manager, Oak Ridge Operations(ORO) established a Type B Accident InvestigationBoard (Board) to investigate this accident in ac-cordance with U.S. Department of Energy (DOE)Order225.1, �Accident Investigations� (seeAppendix AAppendix AAppendix AAppendix AAppendix A).

AAAAACCIDENT DESCRIPTIONCCIDENT DESCRIPTIONCCIDENT DESCRIPTIONCCIDENT DESCRIPTIONCCIDENT DESCRIPTION

The accident was discovered at approximately 9:40a.m. on Monday, September 15, 1997, by twohealth physics technicians and two waste manage-ment operators in Section B of Building C-746-Q.Building C-746-Q is a Resource Conservation andRecovery Act (RCRA) permitted mixed waste stor-age facility. The building was immediately evacu-ated, the Plant Shift Superintendent (PSS) was no-tified, and the Emergency Response Team (ERT)was activated. The initial information relayed tothe PSS at 9:57 a.m. from Building C-746-Q wasthat a drum located in Section B had fallen off apallet into an aisle and spilled its contents. TheAssistant PSS and the ERT arrived at the scene at10:05 a.m. The ERT located a drum lying on itsside (an 85-gal steel drum) and overpacked it in a110-gal, polyethylene-lined, steel drum. The ERTnoticed a large amount of dark brown liquid spilledin and around the dike of acid waste storage areaB-08. Later, the ERT reentered the building to vid-eotape the accident scene. During this entry, the

E X E C U T I V E S U M M A RE X E C U T I V E S U M M A RE X E C U T I V E S U M M A RE X E C U T I V E S U M M A RE X E C U T I V E S U M M A R YYYYY

ERT discovered an open 110-gal, polyethylene-lined, steel drum (HC-1093) without a lid. Furtherinvestigation identified a 55-gal composite steel/polyethylene drum inside HC-1093. (Note: Al-though not known at this time, additional informa-tion discovered later revealed that the accident in-

iv

The accident involved three drumshe accident involved three drumshe accident involved three drumshe accident involved three drumshe accident involved three drumsoverpacked in one configuration.overpacked in one configuration.overpacked in one configuration.overpacked in one configuration.overpacked in one configuration.

Evidence of a drum explosion wasvidence of a drum explosion wasvidence of a drum explosion wasvidence of a drum explosion wasvidence of a drum explosion wasdiscovered the morning ofdiscovered the morning ofdiscovered the morning ofdiscovered the morning ofdiscovered the morning ofSeptember 15, 1997.September 15, 1997.September 15, 1997.September 15, 1997.September 15, 1997.

volved three drums in one configuration. Prior tothe explosion, the container configuration consistedof a 55-gal composite steel/polyethylene drumoverpacked in an 85-gal steel drum that was over-packed in a 110-gal, polyethylene-lined, steeldrum. The 85-gal steel drum was ejected from theconfiguration due to overpressurization. ExhibitExhibitExhibitExhibitExhibitES-1ES-1ES-1ES-1ES-1 shows the components of HC-1093) The darkbrown liquid in and around the diked area of acidwaste storage area B-08 was now foaming. Thebuilding was sealed off, with work limited to stabi-lizing the accident scene.

CACACACACAUSAL FUSAL FUSAL FUSAL FUSAL FAAAAACTORSCTORSCTORSCTORSCTORS

FFFFFigure ES-1igure ES-1igure ES-1igure ES-1igure ES-1 depicts the logical sequence of eventsand causal factors for the accident. It indicates, ina time-sequenced flow, factors that allowed theaccident to occur.

Exhibit ES-1. Components of HC-1093.

v

Figure ES-1. Summary Events and Causal Factors Chart

SUM

MAR

Y E

VEN

TS

AN

D C

AU

SAL

FAC

TO

RS

Causal FCausal FCausal FCausal FCausal Factor Analysisactor Analysisactor Analysisactor Analysisactor Analysis Contributing Causes Contributing Causes Contributing Causes Contributing Causes Contributing Causes DiscussionDiscussionDiscussionDiscussionDiscussion

Procedures Warnings in existing procedures regarding the potential for gas buildupin waste containers were not heeded.

Waste acceptance criteria There was no approved waste acceptance criteria document. The wasteacceptance criteria did not place restrictions on the acceptance of strongoxidizers or address the limitation of polyethylene for long-term storageof strong oxidizers.

Overpack policy or Policies and practices failed to provide guidance on when it ispractice appropriate to repackage as opposed to overpacking a suspect

container.

Occurrence reports Occurrence reports do not always identify the appropriate root causesin past events. There is no single organization or individual responsiblefor tracking or trending information from the occurrence reportingsystem.

Lessons learned There was a failure to properly implement the Lessons Learned Program,and the findings and recommendations in a Lockheed Martin EnergySystems, Inc. (LMES) report (Proposed Neutralization/Pre-treatment forNitric Acid Strip Tank Waste and Other Drummed Lab Wastes Stored inthe Vault 4A Facility, dated September 13, 1991) and a yellow alert(Y-PAD-91-0002, �Polyethylene Reagent Container Failure�) weredisregarded. There is no single organization or individual responsiblefor tracking or trending information from the lessons learned system.

Personnel safety There was a failure to provide hazard analysis and mitigation forexploding drums for hazardous waste operations personnel workingin Building C-746-Q (i.e., there was no safety and health plan).A required safety and health plan, including safety and health permits,must conform to Title 29 of the Code of Federal Regulations (CFR),Part 1910.120, �Hazardous Waste Operations and EmergencyResponse� (HAZWOPER).

DOE oversight DOE failed to perform adequate oversight of environmental and wastemanagement activities at PGDP and did not identify hazards fromexploding waste drums.

Root CauseRoot CauseRoot CauseRoot CauseRoot Cause DiscussionDiscussionDiscussionDiscussionDiscussion

Failure of management LMES management failed to recognize the significance of and/or actcontrol systems appropriately on information regarding the incompatibility between acid

wastes and waste containers.

Table ES-1. Causal Factor Analysis

vi

The direct causedirect causedirect causedirect causedirect cause of the accident was the exces-sive buildup of pressure due to chemical reactionswithin the container, resulting in an explosion ofthe container. However, there were also contrib-contrib-contrib-contrib-contrib-uting causesuting causesuting causesuting causesuting causes (causes that, if corrected, would notby themselves have prevented the accident but areimportant enough to be recognized as needingcorrective action) and a root causeroot causeroot causeroot causeroot cause (a fundamen-

tal cause that, if corrected, would prevent recur-rence of this and similar occurrences). Causal fac-tors are identified in TTTTTable ES-1able ES-1able ES-1able ES-1able ES-1, with a short dis-cussion of each cause.

M anagement failed to recognize theanagement failed to recognize theanagement failed to recognize theanagement failed to recognize theanagement failed to recognize thesignificance of previous incidents.significance of previous incidents.significance of previous incidents.significance of previous incidents.significance of previous incidents.

Conclusions and Judgments of NeedConclusions and Judgments of NeedConclusions and Judgments of NeedConclusions and Judgments of NeedConclusions and Judgments of Need

ConclusionsConclusionsConclusionsConclusionsConclusions Judgments of NeedJudgments of NeedJudgments of NeedJudgments of NeedJudgments of Need

LMES has lessons learned LMES management needs to:procedures that could �Adequately implement a Lessons Learned Program.have, if adequately �Assess the criteria for assigning alert levels.implemented, prevented �Ensure the accuracy of occurrence reports and ensure the appropriatethis accident. root cause(s) have been identified.

No documentation exists LMES needs to proceduralize a process for neutralizing corrosive wastesrestricting the storage of prior to long-term storage. Restrictions need to be placed on thestrong oxidizing agents. acceptance of wastes that pose unique hazards when stored. The Waste

Management Division needs to modify waste acceptance criteria(KY/EM-96) accordingly for treatment, storage, and disposal units.

PGDP container LMES needs to:management procedures �Develop guidance to repackage waste as opposed to overpacking thefail to provide guidance container when appropriate.for repackaging wastes �Modify the waste handling procedure (PMWM-1002 IAD) to clarify theas opposed to limitations of polyethylene-lined containers for storage of strongoverpacking. oxidizing agents.

There are no provisions LMES needs to develop a safety and health plan for HAZWOPERfor personnel safety from personnel in Building C-746-Q that conforms to 29 CFR 1910.120exploding drums in (HAZWOPER).Building C-746-Q.

Neither DOE nor LMES DOE needs to track and trend information from the occurrence reportingadequately trends and lessons learned systems, as defined in ORO M 110, Change 2,information from dated May 15, 1997.occurrence reports LMES needs to track and trend information from the occurrenceor lessons learned. reporting and lessons learned systems.

DOE does not perform DOE needs to:adequate oversight. Develop a comprehensive assessment program for PGDP.

Ensure occurrence report information is accurate and complete.

Table ES-2. Conclusions and Judgments of Need

vii

CONCLUSIONS AND JUDGMENTS OF NEED

Based on its investigation and analyses of theresulting findings, the Board determined theconclusions and judgments of need delineatedin TTTTTable ES-2able ES-2able ES-2able ES-2able ES-2.

Neither DOE nor LMES adequatelyeither DOE nor LMES adequatelyeither DOE nor LMES adequatelyeither DOE nor LMES adequatelyeither DOE nor LMES adequatelytrends information from occurrencetrends information from occurrencetrends information from occurrencetrends information from occurrencetrends information from occurrencereports or lessons learned.reports or lessons learned.reports or lessons learned.reports or lessons learned.reports or lessons learned.

1.01.01.01.01.0 INTRODUCTIONINTRODUCTIONINTRODUCTIONINTRODUCTIONINTRODUCTION

On the morning of September 15, 1997, two healthphysics technicians and two waste managementoperators discovered a spill in Section B of Build-ing C-746-Q at the Paducah Gaseous DiffusionPlant (PGDP). The technicians immediately evacu-ated Building C-746-Q and notified the BuildingSupervisor. The Building Supervisor notified thePlant Shift Superintendent (PSS) of the accident at9:57 a.m. The first report to the PSS was that aspill had occurred from a drum that had fallen offa pallet and into an aisle. Later that afternoon, adetermination was made that a drum explosion hadoccurred with a resulting spill of hazardous/radio-active waste (mixed waste).

On September 17, 1997, James Hall, Manager,Oak Ridge Operations (ORO) established a TypeB Accident Investigation Board (Board) to investi-gate this accident in accordance with U.S. Depart-ment of Energy (DOE) Order 225.1, �Accident In-vestigations� (see Appendix AAppendix AAppendix AAppendix AAppendix A).

1 .11.11.11.11.1 FFFFFAAAAACILITY DESCRIPTIONCILITY DESCRIPTIONCILITY DESCRIPTIONCILITY DESCRIPTIONCILITY DESCRIPTION

Contractor activities regarding legacy and/or re-mediation wastes are managed by the local DOEPaducah Site Office with support from DOE-ORO,which is located in Oak Ridge, Tennessee. The fa-cility in which this accident occurred is under pro-grammatic direction of the DOE-ORO Office ofEnvironmental Management.

Utility Services, Inc. (LMUS) is the operating con-tractor for these activities. In the Lease Agreementbetween DOE and USEC, DOE retains the respon-sibility for managing legacy wastes as well as allwastes generated as a result of environmentalmanagement and remediation activities conductedat PGDP. The lease also provides for DOE storageof some USEC-generated mixed waste. LockheedMartin Energy Systems, Inc. (LMES) is the manag-ing and operating contractor for these activities.

Building C-746-Q is a waste storage facility wherethe management, surveillance, storage, and moni-toring of wastes occur. Building C-746-Q is asingle-story, metal-frame building with dimensionsof 272 ft × 178 ft that is situated on a fully dikedconcrete slab.....

1

TYPE B ATYPE B ATYPE B ATYPE B ATYPE B ACCIDENT INVESTIGACCIDENT INVESTIGACCIDENT INVESTIGACCIDENT INVESTIGACCIDENT INVESTIGATION BOTION BOTION BOTION BOTION BOARD REPORARD REPORARD REPORARD REPORARD REPORT ON THE DRUMT ON THE DRUMT ON THE DRUMT ON THE DRUMT ON THE DRUMEXPLEXPLEXPLEXPLEXPLOSION AOSION AOSION AOSION AOSION AT BUILDING CT BUILDING CT BUILDING CT BUILDING CT BUILDING C-746-Q-746-Q-746-Q-746-Q-746-Q,,,,, P P P P PADUCAH GASEOUS DIFFUSIONADUCAH GASEOUS DIFFUSIONADUCAH GASEOUS DIFFUSIONADUCAH GASEOUS DIFFUSIONADUCAH GASEOUS DIFFUSION PLANTPLANTPLANTPLANTPLANT

Building Cuilding Cuilding Cuilding Cuilding C-746-Q is a single-story-746-Q is a single-story-746-Q is a single-story-746-Q is a single-story-746-Q is a single-story,,,,,metal-frame building with dimensionsmetal-frame building with dimensionsmetal-frame building with dimensionsmetal-frame building with dimensionsmetal-frame building with dimensionsof 272 ft x 178 ft that is situated on aof 272 ft x 178 ft that is situated on aof 272 ft x 178 ft that is situated on aof 272 ft x 178 ft that is situated on aof 272 ft x 178 ft that is situated on afully diked concrete slab.fully diked concrete slab.fully diked concrete slab.fully diked concrete slab.fully diked concrete slab.

On July 1, 1993, the United States EnrichmentCorporation (USEC) began managing all uraniumenrichment operations at PGDP. Lockheed Martin

Building Cuilding Cuilding Cuilding Cuilding C-746-Q was classified as-746-Q was classified as-746-Q was classified as-746-Q was classified as-746-Q was classified asNuclear Hazard Category 2 facilityNuclear Hazard Category 2 facilityNuclear Hazard Category 2 facilityNuclear Hazard Category 2 facilityNuclear Hazard Category 2 facility.....

Exhibit 1-1Exhibit 1-1Exhibit 1-1Exhibit 1-1Exhibit 1-1 shows the exterior of Building C-746-Q.The building is shared between LMES (western por-tion) and LMUS (eastern portion, leased from DOE)and is separated by a floor-to-ceiling wall with two6-in. dike risers at each rollup door. The inner wallis of concrete block construction. The outer wallsand roof of the entire building are comprised ofmetal siding. As shown in FFFFFigure 1-1igure 1-1igure 1-1igure 1-1igure 1-1, the westernhalf of Building C-746-Q, where the accident oc-curred, is approximately 172 ft × 178 ft. This halfof the building is used for storage of hazardous/radioactive wastes (mixed wastes), including thosecontaminated with polychlorinated biphenyls(PCBs). The building is a Resource Conservationand Recovery Act (RCRA) Part B permitted facility.DOE has classified the facility as Nuclear HazardCategory 2 due to the quantity of fissile materialstored within the Fissile Storage Area. The wastestorage areas of the building are organized intoliquid and solid sections. The liquid waste storageareas (Section B) are diked with 6- to 8-in. curbs,

N

and ramps are provided for equipment access.There are 14 of these independently diked areasin the building. All diked areas contain palletizeddrums, with the exception of a 4,000-gal tankand the ash receivers.

1 .21.21.21.21.2 SCOPE, PURPOSE, AND METHODOLSCOPE, PURPOSE, AND METHODOLSCOPE, PURPOSE, AND METHODOLSCOPE, PURPOSE, AND METHODOLSCOPE, PURPOSE, AND METHODOLOGOGOGOGOGYYYYY

The Board commenced the investigation on Sep-tember 17, 1997, and submitted the findings tothe Manager, ORO, on October 27, 1997.

The scope of the Board�s investigation was toinclude, but not be limited to, analyzing causalfactors and identifying root causes that resultedin the accident, and determining judgments ofneed to prevent recurrence. The Board conducteda walkthrough and reviewed photographs of theaccident site, reviewed videotapes, reviewedevents surrounding the accident, conducted ex-tensive interviews and document reviews, and

performed analyses to determine the factors thatcontributed to the accident, including any manage-ment system deficiencies.

The purpose of this investigation was to deter-mine the nature, extent, and causation of the acci-dent and any programmatic impact, and to assistin the improvement of policies and practices, withemphasis on safety and waste management ac-tivities and systems. The Board focused on man-agement roles and responsibilities; application oflessons learned from similar type accidents withinDOE (especially those within LMES); and work plan-ning, practices, and procedures. The Board usedthe following methodology:

� Facts relevant to the accident were gathered.

� Various analysis techniques, including eventand causal factor charting, barrier and con-trol analysis, change analysis, and causal fac-tor analysis were used.

2

Exhibit 1-1.Exterior of Building C-746-Q.

Figure 1-1.Layout of Building C-746-Q.

B-07 & B-08Accident Areas

3

2.02.02.02.02.0 FFFFFAAAAACTS AND ANALCTS AND ANALCTS AND ANALCTS AND ANALCTS AND ANALYSISYSISYSISYSISYSIS

This section provides the facts and analysis of theaccident. This section is written with facts as bulletsand analysis as paragraphs.

2 .12.12.12.12.1 AAAAACCIDENT DESCRIPTION ANDCCIDENT DESCRIPTION ANDCCIDENT DESCRIPTION ANDCCIDENT DESCRIPTION ANDCCIDENT DESCRIPTION ANDCHRONOLCHRONOLCHRONOLCHRONOLCHRONOLOGOGOGOGOGYYYYY

The following subsections provide the accidentdescription, background of similar events, chro-nology of events, accident reconstruction and analy-sis, emergency response, personnel safety, and oc-currence reporting.

2.1.12.1.12.1.12.1.12.1.1 Accident Description Accident Description Accident Description Accident Description Accident Description

The accident description information in this sub-section was identified through the investigationprocess. The facts of the accident are presented asfollows:

• The accident was discovered at approximately9:40 a.m. on Monday, September 15, 1997,by two health physics technicians and two wastemanagement operators in Section B of BuildingC-746-Q.

• The building was immediately evacuated, thePSS was notified, and the Emergency ResponseTeam (ERT) was activated.

• The initial information relayed to the PSS at 9:57a.m. was that a drum located in Building C-746-Q, Section B, had fallen off a pallet into anaisle and spilled its contents.

• The Assistant PSS and the ERT arrived at thescene at 10:05 a.m.

• The ERT donned Level A personal protectiveequipment (PPE), entered Building C-746-Q,and found a drum lying on the dike curb be-tween acid waste storage areas B-07 and B-08.

of dark brown liquid spilled in and around thedike of acid waste storage area B-08.

• Later in the afternoon, the ERT reentered Build-ing C-746-Q in Level A PPE to videotape thescene to aid in further cleanup and assessmentof the accident.

• During this entry, the ERT found a 110-gal, poly-ethylene-lined, steel drum (HC-1093) in its as-signed position in acid waste storage area B-08with the lid off. The lid was found lying approxi-mately 10 ft away from HC-1093 within acidwaste storage area B-08.

• Found inside HC-1093 was a 55-gal compos-ite steel/polyethylene drum. (Note: Although notknown at this time, additional information dis-covered later revealed the accident involvedthree drums in one configuration. Prior to theexplosion, the configuration consisted of a55-gal composite steel/polyethylene drum over-packed in an 85-gal steel drum that was over-packed in a 110-gal, polyethylene-lined, steeldrum. The 85-gal steel drum was ejected fromthe configuration due to overpressurization.)

• The ERT noted that the dark brown liquid in andaround the dike was now foaming.

• The building was sealed off, with work limitedto stabilizing the accident scene.

• Chemical operators and health physicists reen-tered Building C-746-Q several times over thenext 5 days to neutralize the spilled acid as wellas the remaining acid in the drum. On severaloccasions, Draeger tubes were utilized to iden-tify the types and levels of gases that were evolv-ing from the neutralization reaction. The Draegertubes indicated detectible levels of carbon diox-ide, nitrogen dioxide, ammonia, and otherbyproducts.

The original configuration of the drums prior tothe explosion is shown in FFFFFigure 2-1igure 2-1igure 2-1igure 2-1igure 2-1. A descrip-tion of each container is provided below andshown in Exhibit 2-1Exhibit 2-1Exhibit 2-1Exhibit 2-1Exhibit 2-1.

The 85-gal steel drum was ejectedhe 85-gal steel drum was ejectedhe 85-gal steel drum was ejectedhe 85-gal steel drum was ejectedhe 85-gal steel drum was ejectedfrom the configuration due tofrom the configuration due tofrom the configuration due tofrom the configuration due tofrom the configuration due tooverpressurization.overpressurization.overpressurization.overpressurization.overpressurization.

Initial information reported to the PSSnitial information reported to the PSSnitial information reported to the PSSnitial information reported to the PSSnitial information reported to the PSSwas incorrect.was incorrect.was incorrect.was incorrect.was incorrect.

• The ERT attempted to overpack this drum in an85-gal overpack and were unsuccessful. Theythen requested a 110-gal overpack and placedthe drum inside this 110-gal overpack. Beforeleaving the scene, the ERT noted a large amount

•• 55-gal composite steel/polyethylene drum. The

polyethylene drum was the innermost drum ofthe configuration. There were two bungs locatedon the top of the drum that were in place andtightly closed. The polyethylene drum was brittle.The drum contained a 12- to 15-in. horizontalbreach approximately 1 in. from the top. Onthe bottom, a 24- to 26-in. breach near the sideand a thin 10-in.-diameter area in the polyeth-ylene were found. Within this thin area was one4-in. breach. All that remained of the 55-galsteel drum surrounding the polyethylene drumwas the lid, the ring, and approximately14 to 16 in. of the area nearest the top. Theremains of this steel drum were rusty, corroded,brittle, and thin.

• 85-gal steel drum (first overpack). This drum wasthe middle drum in the configuration. This drumwas ejected from the configuration and foundlying on its side approximately 15 ft away fromits original location in an adjacent diked area.This drum contained a drum lid and ring in placeon the drum. The drum lid was convex, giving

the appearance that it may have withstood pres-sure prior to the explosion. The bottom inch ofthe drum was missing, leaving an open con-tainer. The drum was rusty overall, with a roughand jagged bottom edge.

• 110-gal steel, polyethylene-lined drum (secondand outer overpack). This drum was the outer-most drum in the configuration. This drum wasfound with the drum ring in place on the drum,but the drum lid had been blown off. The drumlid was found approximately 10 ft away on thefloor within the diked area (B-08). The drum lidwas uniformly convex, giving the appearancethat it may have withstood pressure prior to be-ing blown off the drum. The bottom of the drumwas also convex, giving the appearance that itmay have withstood pressure prior to the explo-sion. There were three holes in the bottom ofthis drum believed to have been caused by acidcorrosion after the event. No holes or breacheswere visible in the polyethylene liner; however,after the accident, acid material was leakingfrom this drum.

4

Figure 2-1. Overpack Configuration Prior To Explosion.

THE LIDS OF THE TWOOVERPACKS APPEAREDTO HAVE WITHSTOOD

PRESSURE PRIOR TOTHE EXPLOSION.

2.1.22.1.22.1.22.1.22.1.2 Background of Similar EventsBackground of Similar EventsBackground of Similar EventsBackground of Similar EventsBackground of Similar Events

During the investigation, the Board discovered thatseveral similar events preceded the accident dis-covered in Building C-746-Q. The following para-graphs detail the facts of those events.

• On March 22, 1990, at PGDP, acid waste drumHC-4337 was being transported from Build-ing C-746-B to Building C-746-Q and fell1 ft from a wooden pallet to the floor of Build-ing C-746-Q. The primary container held lowpH hazardous waste and was overpacked ina steel drum. The primary container was lined,but its lid was not lined and was not made ofa material compatible with the low pH waste.The lid was badly deteriorated and leaked

approximately 1 gal of waste onto the con-crete floor and into the steel overpack. Approxi-mately 2 hours after drum HC-4337 wasuprighted, a 1-in. hole in the bottom of thesteel overpack resulted in a 15- to 20-gal spill.

A past corrective action was to place past corrective action was to place past corrective action was to place past corrective action was to place past corrective action was to placethe remaining drums that were knownthe remaining drums that were knownthe remaining drums that were knownthe remaining drums that were knownthe remaining drums that were knownto be incompatible with their overpacksto be incompatible with their overpacksto be incompatible with their overpacksto be incompatible with their overpacksto be incompatible with their overpacksinto compatible storage containers.into compatible storage containers.into compatible storage containers.into compatible storage containers.into compatible storage containers.

An unusual occurrence report, UOR PGDP-90-2-C-746-Q-1, identified the use of incompat-ible containers as the cause of the incident. Acorrective action plan was prepared, and one

5

Exhibit 2-1. Components of HC-1093.

85-gal steel drum

110-gal, polyethy-lene-lined, steel

drum

C O M P O N E N T S O F H C - 1 0 9 3

55-gal compositesteel/polyethylene

drum

Lid and ring for55-gal compositesteel/polyethylene

drum

of the recommended corrective actions wasto �place remaining drums that are known tobe incompatible with their overpacks into com-patible storage containers.�

• On June 5, 1991, at the PGDP Laboratory,Building C-720, a 1-liter (L) polyethylenebottle, which contained, in part, concentratednitric acid, shattered, spilled its contents, andinjured an employee. The event was noted ina yellow alert (Yellow Alert No. Y-PAD-91-0002, �Polyethylene Reagent Container Fail-ure�) that strong oxidizing agents are likely tocause embrittlement and subsequent failureof polyethylene containers. No written re-sponse was required for this yellow alert.

• On July 22, 1991, at the K-25 Site, which isnow known as East Tennessee Technology Park(ETTP), one of four drums containing nitric acidstripping waste generated at ETTP failed andreleased its contents. The drum that failed hadbeen overpacked in an unlined steel drum. Ateam was commissioned to investigate theincident, and its findings were issued on Sep-tember 13, 1991, in an LMES report titledProposed Neutralization/Pre-treatment forNitric Acid Strip Tank Waste and OtherDrummed Lab Wastes Stored in the Vault 4AFacility. This report provides specific recommen-dation g acid wastes that are similar to the acidwastes involved in the accident reported on Mon-day, September 15, 1997, in Building C-746-Qlocated at PGDP. There is no evidence that thereport was shared with sites other than ETTP, al-though PGDP and the Portsmouth Gaseous Diffu-sion Plant (PORTS) had similar operations.

• As recently as July 2, 1997, a Y-12 Plant yel-low alert (Yellow Alert No. Y-1997-OR-LMESY12-0701, �Nitric Acid Causes DrumOver-Pressurization�) was issued. The yellowalert identified the mixing of nitric acid wastewith organic material as the cause. However,the root cause of the accident was a failedpolyethylene-lined container caused by nitricacid. No written response was required forthis yellow alert.

6

2.1.32.1.32.1.32.1.32.1.3 Chronology of EventsChronology of EventsChronology of EventsChronology of EventsChronology of Events

Following are the facts of the events that led to theaccident discovered on September 15, 1997, inBuilding C-746-Q at PGDP. FFFFFigure 2-2igure 2-2igure 2-2igure 2-2igure 2-2 summa-rizes the chronology of significant events and as-sociated causal factors.

� From the meeting minutes titled �Meeting onPGDP TRU Waste,� dated July 23, 1984, ques-tions were again raised about the reliabilityof the nitric acid/radioactive waste drums dur-ing their proposed transfer from Building C-746-B to Building C-400 for solidification. Itwas noted that the inner plastic liner couldfail during transfer.

� Due to concerns with the drum�s integrity,HC-1093 was overpacked in an 85-gal steeldrum in late 1986. The 85-gal steel overpackdid not contain a polyethylene liner.

The integrity of the inner polyethylene/he integrity of the inner polyethylene/he integrity of the inner polyethylene/he integrity of the inner polyethylene/he integrity of the inner polyethylene/steel container had been questionedsteel container had been questionedsteel container had been questionedsteel container had been questionedsteel container had been questionedseveral times beginning in 1983.several times beginning in 1983.several times beginning in 1983.several times beginning in 1983.several times beginning in 1983.

� The container that exploded (HC-1093) con-tained nitric acid/radioactive mixed waste thatwas generated in the 1960s as a result of nickelstripping activities in the Building C-400 nep-tunium recovery process. The wastes from therecovery process were stored in 55-gal com-posite steel/polyethylene containers in Radio-active Warehouse C-746-B, Smelter Area.

� In a PGDP letter and report, dated September26, 1983, titled Drums in C-746-B Radioac-tive Warehouse, a recommendation was issuedfor managing the �lot� of nitric acid/radioac-tive wastes. This document provided the de-tails of an extensive study of the waste in thedrums. A recommendation was made to treatand repackage this waste due to the potentialfailure of the drums. The letter stated that thedrums containing the wastes were of poorquality and should be replaced.

7

Figure ES-1. Summary Events and Causal Factors Chart.

SUM

MAR

Y E

VEN

TS

AN

D C

AU

SAL

FAC

TO

RS

� On January 16, 1990, an attempt was madeto sample and analyze HC-1093 for radio-active and RCRA constituents. The sampletechnician was unable to loosen the bung;therefore, the sample was not obtained untilJanuary 25, 1990. A pH of 1 was not deter-mined until March 21, 1990.

� In May 1990, HC-1093 was transferred fromBuilding C-746-B to Building C-746-Q whenit was determined to contain RCRA hazard-ous waste. HC-1093 was placed in old bay17-01 and later moved to old bay C-03.

� In December 1994, HC-1093 was placed ina 110-gal, polyethylene-lined, steel overpack.This was done in accordance with the correc-tive action plan resulting from the April 6,1990, unusual occurrence report (UORPGDP-90-2-C-746-Q-1).

� In May 1997, HC-1093 was moved from oldbay C-03 to acid waste storage area B-08following completion of floor repairs in B-08(Section B).

� On July 28, 1997, HC-1093 and other wastecontainers were analyzed for pH. One of theother waste containers was HC-385. HC-385was located on the same pallet as HC-1093.HC-385 was removed from acid waste stor-age area B-08 when pH analysis showed itscontents to be caustic. To accomplish the re-location of HC-385, the entire pallet (whichalso held HC-1093) was removed from B-08. Later in the same day the pallet, contain-ing HC-1093 but not HC-385, was returnedto B-08.

� On Friday, September 12, 1997, a RCRA in-spection was performed. All storage locationswithin Building C-746-Q were inspected inaccordance with RCRA permit requirements.Inspection requirements include, but are notlimited to, container condition, appearance,etc. All drums were found to be acceptable.

� On September 15, 1997, two health physicstechnicians and two waste management op-erators observed a drum lying on the dike

curb between acid waste storage areas B-07and B-08 in Building C-746-Q. The buildingwas evacuated immediately pending emer-gency response actions.

2.1.42.1.42.1.42.1.42.1.4 Accident Reconstruction and AnalysisAccident Reconstruction and AnalysisAccident Reconstruction and AnalysisAccident Reconstruction and AnalysisAccident Reconstruction and Analysis

The accident was not observed by anyone when itoccurred; therefore, many of the accident recon-struction statements were based upon the Board�stechnical judgment, experience, and analysis of thebest available information. The best available in-formation consisted of sampling and laboratorydata, extensive personnel interviews, pictures of theaccident site, and visual inspection of the accidentsite and materials.

A RCRA inspection of the drumsRCRA inspection of the drumsRCRA inspection of the drumsRCRA inspection of the drumsRCRA inspection of the drumswithin Building Cwithin Building Cwithin Building Cwithin Building Cwithin Building C-146-Q was con--146-Q was con--146-Q was con--146-Q was con--146-Q was con-ducted on Fducted on Fducted on Fducted on Fducted on Fridayridayridayridayriday, September 12, 1997., September 12, 1997., September 12, 1997., September 12, 1997., September 12, 1997.

� The accident involved three drums in one con-figuration. Prior to the explosion, the con-figuration consisted of a 55-gal compositesteel/polyethylene drum overpacked in an85-gal steel drum that was overpacked in a110-gal, polyethylene-lined, steel drum. The85-gal steel drum was ejected from the con-figuration due to overpressurization.

� The 55-gal composite steel/polyethylenedrum (innermost container in the configura-tion) had been in use for approximately 35to 40 years. This drum was used to containwaste from the nitric acid stripping activitiesconducted within Building C-400. The wasteis concentrated with greater than 70% (pHof 0 or 1) nitric acid.

� An LMES daily RCRA inspection of the drumswithin Building C-746-Q was conducted from10:10 a.m. to 11:00 a.m. on Friday, Sep-tember 12, 1997, and the building waslocked at the end of the day. No personnelentered the building until 6:05 a.m. on Mon-day, September 15, 1997.

8

� The accident was discovered at 9:40 a.m. onMonday, September 15, 1997. Personnel con-ducted routine work activities away from theaccident location within Building C-746-Qfrom 6:05 a.m. until the accident was dis-covered at 9:40 a.m..

� The accident occurred in the southwest cor-ner (Section B) of Building C-746-Q as shownin FFFFFigure 2-3igure 2-3igure 2-3igure 2-3igure 2-3. A 110-gal, polyethylene-lined,steel drum (HC-1093) containing a nitric acid/radioactive waste mixture exploded, scatter-ing some of its contents over an inside area(Section B) of the building covering approxi-mately 400 ft2. The nearest wall (west side)as well as the ceiling immediately above thedrum had been splattered with the contentsof the drum. Most of the liquid from the drumwas contained within the diked area (B-08),and the liquid collected/drained to onecorner.

• The explosion did not displace the outer drum(HC-1093) from its original location.HC-1093 was found with the drum ring inplace on the drum, but the drum lid had beenblown off. The drum lid was found approxi-mately 10 ft away on the floor within the dikedarea (B-08). The drum lid was uniformly con-vex, giving the appearance that it may havewithstood pressure prior to being blown offthe drum.

• A 55-gal composite steel/polyethylene drum,containing two bungs in the top, was foundinside HC-1093.

• An 85-gal steel drum was ejected from HC-1093, hit the ceiling of Building C-746-Q,and landed approximately 15 ft away fromHC-1093. The 85-gal steel drum was lyingon its side within an adjacent diked area (B-07) with its lid and ring in place on the drum.The drum lid was convex, giving the appear-ance that it may have withstood pressure priorto the explosion. The bottom inch of the 85-gal drum appeared to have been corrodedby acid.

• A steel drum lid and ring was found lyingadjacent to the 85-gal steel drum. The steeldrum lid and ring were originally thought tobe the lid and ring of the 85-gal steel drum.Closer inspection of the steel drum lid andring showed that the drum lid had two holesin it (bung holes) and that the ring fit the drumlid. Later, it was determined to be the lid andring of the 55-gal composite steel/polyethyl-ene drum.

The concentrated nitric acid stored inhe concentrated nitric acid stored inhe concentrated nitric acid stored inhe concentrated nitric acid stored inhe concentrated nitric acid stored incontact with the polyethylene linercontact with the polyethylene linercontact with the polyethylene linercontact with the polyethylene linercontact with the polyethylene linercaused the polyethylene liner tocaused the polyethylene liner tocaused the polyethylene liner tocaused the polyethylene liner tocaused the polyethylene liner tobecome extremely brittle.become extremely brittle.become extremely brittle.become extremely brittle.become extremely brittle.

The paragraphs below present the analysis of theaccident reconstruction facts.

The concentrated nitric acid stored in contact withthe polyethylene liner over an extended period(from approximately 1960 to 1997) caused thepolyethylene liner to become extremely brittle. Theconcentrated nitric acid slowly seeped through thehairline fractures of the degraded polyethyleneliner. During this time, the continual cycle of heat-ing and cooling of the drum contents due to natu-ral environmental conditions as well as the con-tinual degradation of the polyethylene liner inducedby the concentrated nitric acid caused �wicking�of minute quantities of escaped acid between thesurfaces of the composite steel/polyethylene drum.In this case, wicking means the movement of liq-uids (condensation and concentrated nitric acid)along the outer walls of the polyethylene liner andthe inner walls of the steel drum. The wicking ofthe concentrated nitric acid resulted in an extremelycorroded, thin, and weak 55-gal steel drum. Overtime, more and more hairline cracks developed inthe polyethylene, weakening this liner.

It can be seen in Exhibit 2-2Exhibit 2-2Exhibit 2-2Exhibit 2-2Exhibit 2-2 that the polyethylenedrum is very thin within a 10-in.-diameter bottomarea. The 4-in. breach in the thin and brittle bot-tom of the polyethylene drum could have been

9

A C C I D E N T L A Y O U T

Figure 2-3. Accident Layout.

1 0

1 1

caused by the sampling tool (caliwasa) during thesampling event on July 28, 1997. The samplingevent and the subsequent movements of HC-1093provided enough additional stress to the inner poly-ethylene liner to cause a breach large enough toallow acid to reach a greater surface area of themetal and intensify the acid/metal reaction. The

resulting gases produced from the intensified acid/metal reaction caused the overpressurization of the85-gal steel overpack and HC-1093. Pressure ini-tially built up within the 85-gal steel overpack fromthe acid attack on the inner metal surface of thiscontainer as well as the intense acid attack on theaged and extremely brittle 55-gal steel drum. The55-gal steel drum provided very fine iron particles/powder that reacted with the concentrated nitricacid, causing a rapid buildup of gases.

Nitric acid above concentrations of about 25%(250,000 mg/L) is a strong oxidizer with a historyof unpredictably reactive incidents and a long listof incompatible materials. Nitric acid must be

N itric acid above concentrations ofitric acid above concentrations ofitric acid above concentrations ofitric acid above concentrations ofitric acid above concentrations ofabout 25% (250,000 mg/L) is a strongabout 25% (250,000 mg/L) is a strongabout 25% (250,000 mg/L) is a strongabout 25% (250,000 mg/L) is a strongabout 25% (250,000 mg/L) is a strongoxidizer with a history of unpredictablyoxidizer with a history of unpredictablyoxidizer with a history of unpredictablyoxidizer with a history of unpredictablyoxidizer with a history of unpredictablyreactive incidents and a long list ofreactive incidents and a long list ofreactive incidents and a long list ofreactive incidents and a long list ofreactive incidents and a long list ofincompatible materials.incompatible materials.incompatible materials.incompatible materials.incompatible materials.

Exhibit 2-2. Bottom of 55-gal Polyethylene Drum.

4-In.Breach

10-In. DiameterThin Area

24-26-In.Breach

handled with great care because it is extremelycorrosive and particularly reactive with organics (es-pecially alcohols) and metal powders. Consider-ation should be given not only to the compatibilityof materials with nitric acid, but to controlling ratesof addition, concentration, temperature, agitation,order of addition, transfer routes, storage, andprotection of operating personnel and equipment.The mechanism of nitric acid oxidation of iron,when the concentration is 7 molar concentration(M) or greater (about 400,000 mg/L as nitrate ion),proceeds according to the following set of chemi-cal equations. These chemical equations show theprobable reactions that took place within the85-gal steel overpack.

(1) Fe +6HNO3 Fe(NO3)3 + 3NO2 (gas) + 3H2O

(2) 2NO2 N2O4

It can be seen from equation (1) that 3 moles ofgas (NO2) are generated for every 1 mole of ironpresent; therefore, with the addition of an excessamount of acid, gas buildup is rapid. The mecha-nism of nitric acid oxidation of iron, when the con-centration is between 1 and 2M (about 85,000 mg/L), proceeds according to the following chemicalequation. This chemical reaction could also havetaken place within the 85-gal steel overpack.

(3) 2Fe + 6HNO3 2Fe(NO3)3 + 3H2 (gas)

As the acid/metal reaction continued, the pressurewithin the 85-gal steel overpack increased rapidly.The acid/metal reaction continued primarily withinthe bottom of the 85-gal steel overpack and se-verely weakened the bottom and extreme lower (ap-proximately 1 in.) portion of this drum. As the bot-tom of this drum became corroded by the reac-tion, small holes formed and allowed some of thegases to leak into the 110-gal, polyethylene-lined,steel overpack (HC-1093). This buildup of gasescaused an overpressurization of HC-1093. Thecontinuing rapid buildup of pressure inside the 85-gal steel overpack caused the explosion. The weakpoint of the 85-gal steel overpack was the bottom,where the majority of the acid/metal reaction oc-curred. The release of pressure was concentratedat the bottom of the 85-gal steel overpack, andthe pressure forced the 85-gal steel overpack out

1 2

of the top of HC-1093 with enough force to hit theceiling of Building C-746-Q and be diverted to-wards an adjacent diked area (B-07) within theacid waste storage area. The lid of HC-1093 wasalso blown to the ceiling and landed 10 ft awayfrom its original location.

2.1.52.1.52.1.52.1.52.1.5 Emergency Response Emergency Response Emergency Response Emergency Response Emergency Response

Response by site personnel from discovery of theaccident to completion of emergency responseactivities consisted of the following facts.

• Site personnel were alerted by the healthphysics technicians and immediately evacu-ated Building C-746-Q.

• The Building Supervisor notified the PSS at9:57 a.m. and requested assistance fromthe Fire Services Division and Chemical Op-erations Division. The Assistant PSS, SiteSafety Officer, Incident Commander, andfirefighter personnel responded to the acci-dent scene.

• The Incident Commander established road-blocks to control traffic around Building C-746-Q and established the Incident Com-mand Post in the C-333-A Cylinder Yardnorthwest of Building C-746-Q.

• Two firefighters (the ERT), dressed in LevelA PPE, entered the accident area at 10:30a.m. and overpacked the 85-gal steel drumin a 110-gal, polyethylene-lined, steel over-pack. The firefighters reported a spill ofbrown liquid on the concrete floor. The SiteSafety Officer established a decontamina-tion line for the firefighters as they exitedthe building.

• The second entry into Building C-746-Q oc-curred at 1:00 p.m. and was performed bytwo firefighters (the ERT) who videotaped

The acid/metal reaction caused a rapidhe acid/metal reaction caused a rapidhe acid/metal reaction caused a rapidhe acid/metal reaction caused a rapidhe acid/metal reaction caused a rapidbuildup of pressure inside the 85-galbuildup of pressure inside the 85-galbuildup of pressure inside the 85-galbuildup of pressure inside the 85-galbuildup of pressure inside the 85-galsteel overpack, leading to the explosion.steel overpack, leading to the explosion.steel overpack, leading to the explosion.steel overpack, leading to the explosion.steel overpack, leading to the explosion.

1 3

the accident scene. The ERT reported thatthe brown liquid spill on the floor was pro-ducing a lighter colored, frothy foam. TheERT conducted air monitoring for nitric acidduring this time. Of the three locationssampled only the area over the spill resultedin a detectable level of 3 ppm nitric acid.The Occupational Safety and Health Admin-istration (OSHA) 8-hour time-weighted av-erage limit is 2 ppm.

• Due to the potential for multiple chemicalreactions, air samples were collected andanalyzed to ensure personnel safety and ap-propriate PPE. This practice continuedthroughout the remaining responseactivities.

the building approximately 80% of normal dutyhours. The building is closed and locked eachday at close of business. The building was un-locked at 6:05 a.m. on September 15, 1997,by a subcontractor chemical engineer and threeenvironmental monitoring samplers. Thechemical engineer and environmental moni-toring samplers were the first personnel to en-ter Building C-746-Q. By the time of discov-ery of the accident at 9:40 a.m., approximately16 people had been in the building. The per-sonnel who had been in the building includedthe chemical engineer, six environmental moni-toring samplers, four health physics techni-cians, one environmental monitoring samplersupervisor, two electricians, and two wastemanagement operators.

• PMWM-1002 IAD, Rev. 1, �On-Site Handlingand Disposal of Waste Materials,� warns ofgas pressure buildup in waste containers.Appendix O of this procedure addressesproper handling of potentially pressurizeddrums, including drums with low pH wastes,and the response of personnel who discoverpressurized drums.

• A Hazardous Material (HAZMAT) Team(chemical operators) was assembled andentered Building C-746-Q at 9:34 p.m. tospray the contaminated area and drums witha soda ash and water mixture. The HAZMATTeam exited the contaminated area at 10:00p.m. to change air packs. The HAZMAT Teamreentered Building C-746-Q at 10:15 p.m.and spread additional soda ash on the floorof the contaminated area. The HAZMATTeam exited the contaminated area at 10:30p.m.

• Bioassay tests were performed on person-nel present when the accident was discov-ered and on first responders to the accident(the ERT). Bioassay test results were nega-tive for all personnel.

2.1.62.1.62.1.62.1.62.1.6 PPPPPersonnel Safetyersonnel Safetyersonnel Safetyersonnel Safetyersonnel Safety

Following are facts concerning the safety of haz-ardous waste operations personnel working inBuilding C-746-Q.

• The Supervisor responsible for Building C-746-Q indicated that personnel are present within

Level A PPE was utilized due to theevel A PPE was utilized due to theevel A PPE was utilized due to theevel A PPE was utilized due to theevel A PPE was utilized due to theunknown conditions.unknown conditions.unknown conditions.unknown conditions.unknown conditions.

No safety and health work permit oro safety and health work permit oro safety and health work permit oro safety and health work permit oro safety and health work permit orsafety and health plan relating to worksafety and health plan relating to worksafety and health plan relating to worksafety and health plan relating to worksafety and health plan relating to workin Building Cin Building Cin Building Cin Building Cin Building C-746-Q was in place at-746-Q was in place at-746-Q was in place at-746-Q was in place at-746-Q was in place atthe time of the accident.the time of the accident.the time of the accident.the time of the accident.the time of the accident.

• Radiological work permit (RWP) 97-ER-003-G, utilized for tours, inspections, and minorhands on activities, and RWP 97-ER-004-G,utilized for minor maintenance and decontami-nation of areas and equipment, were in effectat the time of the accident. Both RWPs expireon December 31, 1997. These RWPs specifyradiological surveys by health physicists fortools, equipment, and materials, and a bodyfrisk for radioactive contamination after anywork is performed in Building C-746-Q. Nosafety and health work permit or safety andhealth plan relating to work in Building C-746-Qwas in place at the time of the accident.

1 4

• Daily RCRA inspections are performed in Build-ing C-746-Q for aisle spacing; container stack-ing, condition, sealing, and labels; safety andemergency equipment; spill containment; gen-eral appearance; unacceptable material; se-curity devices; and cylinder condition. A RCRAinspection was performed between 10:10 a.m.and 11:00 a.m. on September 12, 1997. Allof the RCRA inspection items received an ac-ceptable rating for the last working day priorto the accident. The inspection log did not in-dicate the presence of bulging drums.

The paragraphs below present the analysis of thesafety of hazardous waste operations personnelworking in Building C-746-Q.

Although PMWM-1002 IAD warns of overpres-surization of containers, the procedure does notidentify hazards to personnel involved in routinewaste operations from overpressurized and/or ex-ploding drums or means of mitigating these haz-ards.

The work activity sections of RWP 97-ER-003-G andRWP 97-ER-004-G did not take into account haz-ards from exploding/overpressurized drums and,subsequently, did not specify proper protectiveclothing and equipment for personnel working inthe presence of these hazards. RWPs were not in-tended to protect personnel from hazardous waste(exploding drums, etc.) industrial accidents.

The Board found no evidence of a safety and healthwork permit or safety plan relating to personnelworking in Building C-746-Q. A written, site-spe-cific safety and health plan that identifies hazardsand methods of mitigation is required by 29 CFR1910.120 (HAZWOPER). CP2-EP-EP5031 providesdirection for responding to overpressurized (bulg-ing) drums containing oil, hazardous materials, orhazardous waste. However, hazards to fire servicespersonnel attempting to puncture or open drumsfrom explosions/overpressurization are not identi-fied or mitigated.

2.1.72.1.72.1.72.1.72.1.7 Occurrence Reporting Occurrence Reporting Occurrence Reporting Occurrence Reporting Occurrence Reporting

Occurrence Report ORO-LMES-PGDPENVRES-1997-0008, Waste Drum Rupture Inside BuildingC-746-Q RCRA Waste Storage Facility, was reportedat 9:57 a.m. and categorized as off-normal at 6:45p.m. on Monday, September 15, 1997. The occur-rence was classified as a cross-category item withpotential concerns/issues. The incident category waslater upgraded to an unusual occurrence uponconsideration that Building C-746-Q was a NuclearHazard Category 2 facility. The occurrence was notcategorized within the 2-hour time limit requiredby DOE Order 232.1.

2 .22.22.22.22.2 MANAMANAMANAMANAMANAGEMENT SYSTEMS AND CONTROLSGEMENT SYSTEMS AND CONTROLSGEMENT SYSTEMS AND CONTROLSGEMENT SYSTEMS AND CONTROLSGEMENT SYSTEMS AND CONTROLS

The subsections to follow provide information onthe management systems and controls utilized byLMES to conduct activities at PGDP.

2.2.12.2.12.2.12.2.12.2.1 P P P P Policies and Policies and Policies and Policies and Policies and Proceduresroceduresroceduresroceduresrocedures

Facts pertaining to policies and procedures appliedto this investigation are provided as follows:

• DOE Standard 7501-95 (DOE-STD-7501-95)provides guidance for implementing a LessonsLearned Program.

• A Lessons Learned Program existed on June5, 1991, as evidenced by Yellow Alert No.Y-PAD-91-0002. The current LMES LessonsLearned Program is documented in ProcedureQA-331.

• Title 29 of the Code of Federal Regulations(CFR), Part 1910.120, �Hazardous Waste Op-erations and Emergency Response�(HAZWOPER), requires the development andimplementation of a written safety and healthprogram for employees in hazardous wasteoperations. The program should be designedto identify, evaluate, and control safety andhealth hazards, and provide for emergency re-sponse for hazardous waste operations. Thewritten safety and health plan must be sup-ported by site procedures that implement pro-visions of the plan. CP2-EP-EP5031, Rev. 1,�Oil and Hazardous Materials Spills and Re-leases,� provides direction for responding tooverpressurized (bulging) drums containing oil,hazardous materials, or hazardous waste.

1 5

• LMES Procedure QA-331 identifies criteria forissuing alerts. An appropriate alert level foran uncontained, on-site, hazardous releaserequiring cleanup is subjective and not welldefined.

• LMES Procedure OP-301, �Occurrence Noti-fication and Reporting,� requires investigationand the establishment of the root cause foran accident.

• LMES Procedure PMWM-1002 IAD, �On-SiteHandling and Disposal of Waste Materials,�dated June 1, 1997, provides warnings ofpotential gas buildup in containers. AppendixO of this procedure lists low pH waste solu-tions as those that are likely to generate gasescausing excess pressure within containers.

• Procedure CP2-EP-EP5031, Rev. 1, �Oil andHazardous Material Spills and Releases,� pro-vides direction for responding to releases ofhazardous wastes. Section 6.7, �High RiskDrums,� of this procedure provides directionfor responding to releases of hazardouswastes. This section also provides direction forpersonnel discovering a bulging drum includ-ing the use of Fire Services personnel for punc-turing or opening bulging drums. This proce-dure does not identify hazards to Fire Servicespersonnel who puncture or open a bulging orhigh-risk drum.

• Procedure KY/EM-96 �Waste Acceptance Cri-teria for Treatment, Storage, and DisposalUnits at the Paducah Gaseous Diffusion Plant,�outlines requirements generators must meetprior to transferring waste to PGDP treatment,storage, and disposal units.

• A September 26, 1983 internal memo titled,�Drums in C-746-B Radioactive Warehouse,�noted that the integrity of the drums was ofconcern and recommended repackaging thosedrums.

• Previous yellow alerts (e.g., as described inSection 2.1.2) identified the embrittlement ofpolyethylene-lined containers due to strongoxidizers as the cause of the incident.

• A report prepared in July 22, 1991, at theK-25 Site (now ETTP) titled, Proposed Neutral-ization/Pre-treatment for Nitric Acid Strip TankWaste and Other Drummed Lab Wastes Storedin the Vault 4A Facility, summarized the limi-tations of polyethylene-lined containers for thelong-term storage of strong oxidizing agents.

The paragraphs below present the analysis of thefacts pertaining to the policies and procedures.

LMES Procedure QA-331 implements the LessonsLearned Program that utilizes various levels of alerts(red, yellow, green, and blue). An �uncontainedon-site hazardous release requiring cleanup� wouldjustify a red alert and a managerial response. Thereleased waste resulting from the Building C-746-Q drum explosion did not reach the environment,so it was determined to be contained, and a yellowalert was issued. Evidence shows that the wastewas not contained within the diked area (second-ary containment) since it was found on the wallsand ceiling. So one could argue that the spill wasuncontained. The definition of �uncontained on-site hazardous release requiring cleanup� shouldbe clarified to indicate whether it applies only toreleases that reach the environment or releases notcontained by engineered barriers.

Although numerous similar events have occurredand have been documented and reported in vari-ous forms, LMES management actions to addressthe failure of polyethylene liners have been incon-sistent. ETTP has taken some steps to mitigate thepotential for accidents; however, PGDP continuesto store its inventory in their original containers.The Y-12 Plant has similar problems but possiblyto a lesser extent. LMES needs to reassess its oc-currence reporting and lessons learned systems toensure that the appropriate level of attention isgiven based on the risk involved.

Tracking and trending is notracking and trending is notracking and trending is notracking and trending is notracking and trending is notconducted on lessons learned orconducted on lessons learned orconducted on lessons learned orconducted on lessons learned orconducted on lessons learned orocurrences.ocurrences.ocurrences.ocurrences.ocurrences.

LMES must ensure that the appropriate root causeis identified when reporting occurrences. In pastoccurrence reports the root cause was incorrectlyidentified. The Y-12 Plant Occurrence Report Y-12WASTE-1997-0004 incorrectly identified the rootcause as a drum overpressurization. A yellow alert(Y-1997-OR-LMESY12-0701) issued as a result ofthis accident warned against storing organic ma-terial and nitric acid in closed containers. How-ever, the root cause of the accident was the failureof the polyethylene liner. Because the root causewas incorrectly identified in the occurrence report,the lessons learned from the accident were invalid.

waste management. In its current form, the docu-ment does not have restrictions on receipt and stor-age of strong oxidizing agents.

2.2.22.2.22.2.22.2.22.2.2 Roles and Responsibilities Roles and Responsibilities Roles and Responsibilities Roles and Responsibilities Roles and Responsibilities

The following sections describe roles and respon-sibilities for DOE-ORO and LMES organizationalelements present at the Paducah Site.

2.2.2.12.2.2.12.2.2.12.2.2.12.2.2.1 DOE- DOE- DOE- DOE- DOE-ORO RORO RORO RORO RORO Roles and Roles and Roles and Roles and Roles and Responsibilitiesesponsibilitiesesponsibilitiesesponsibilitiesesponsibilities

The following are the roles and responsibilities fororganizational elements concerned with functionsrelated to the accident.

Oak Ridge Operations

DOE-ORO organizational functions are describedin the ORO Organizational Manual, ORO M 110,Change 2, dated May 15, 1997.

Office of Assistant Manager for Enrichment Facili-ties (AMEF)

This organization is responsible for the develop-ment of plans, procedures, and programs for thedirection of DOE-ORO activities implementingDOE programs at PGDP and PORTS, including,but not limited to, management of environmentalrestoration and waste management activities; man-agement of the OSHA Program; management offacilities not leased to USEC; and associated ac-tivities, including safety and health oversight. Theabove mission is discharged through the Office ofAMEF, Paducah Site Office, and Portsmouth SiteOff ice.

Paducah Site Manager

The Paducah Site Manager performs the follow-ing functions that relate to this accident:

• Provides day-to-day on-site direction and tech-nical oversight of contractor activities in sup-port of environmental restoration and wastemanagement and capital projects.

• Assures that Government and contractor-ex-ecuted functions are carried out in compliance

Care should be taken to correctlyare should be taken to correctlyare should be taken to correctlyare should be taken to correctlyare should be taken to correctlyidentify root causes in occurrenceidentify root causes in occurrenceidentify root causes in occurrenceidentify root causes in occurrenceidentify root causes in occurrencereports.reports.reports.reports.reports.

Tracking and trending is not conducted on lessonslearned or occurrences. Although there have beenseveral similar incidents and lessons learned, LMEScontinued business as usual, issuing lessons learnedand occurrence reports. By using tracking andtrending information from lessons learned andoccurrences, a summary lessons learned or occur-rence report could be written. This summary reportwould consolidate the information from the pastevents and should require a response.

PMWM-1002 IAD, Rev. 1, is the LMES procedurefor managing containers. The procedure does notaddress when to repackage wastes as opposed tooverpacking. The procedure warns of potential gasbuildup in containers resulting from chemical re-actions of low pH solutions. The procedure doesnot address the limitations of polyethylene linersfor storing low pH solutions.

Procedure CP2-EP-EP5031, Rev. 1, fails to requirea hazard analysis prior to puncturing a bulgingdrum. The procedure also suggests utilizing pres-sure relief devices but does not specify when orwhere they are appropriate.

The current waste acceptance criteria at PGDP arein draft. The purpose of the document is to provideguidance to the generator on how to characterize,segregate, and package wastes prior to transfer to

1 6

with DOE Orders and in a manner that pro-tects Government and contractor personnel andthe general public against all environmental,health, and safety hazards arising from per-formance of contract work.

• Implements site environmental restoration pro-grams and activities in coordination with theDOE-ORO Assistant Manager for Environmen-tal Management (AMEM), including remedialinvestigations, alternative evaluations, remedialdesigns, and remedial actions.

• Manages DOE waste management activities atPGDP, including waste processing, storage,transportation, disposal, and minimization.

• Oversees DOE contractor compliance withOSHA, the Clean Air Act; the Clean Water Act;National Environmental Policy Act; Comprehen-sive Environmental Response, Compensation,and Liability Act; RCRA; environmental laws;environmental incident notification and report-ing requirements; and radiological control re-quirements of DOE. Reviews and analyzes com-pliance audits, appraisals, and unusual occur-rence reports; evaluates corrective actions tobe taken by the contractor (LMES) in responseto findings and deficiencies detected; and veri-fies that corrective actions have been fullyimplemented.

• Performs on-site surveillance of DOE activitiesto assure compliance with DOE and other Fed-eral and state environment, safety, and healthrequirements.

Environmental Restoration Division (ERD)

The ERD Director exercises management respon-sibility for the following functions that relate to thisaccident:

• Manages the DOE-ORO Environmental Res-toration Program, including remedial action,transition, and decontamination and decom-missioning programs and waste management

programs at PGDP and PORTS.

• Ensures technical consistency among facil ityenvironmental restoration activities at DOE-ORO sites.

• Coordinates division activities related to theOak Ridge Reservation, PGDP, and PORTS en-vironmental restoration activities with the regu-latory agencies and the public.

• Ensures that contractor-executed functions arecarried out in a manner that protects Federaland contractor personnel and the general pub-lic against environmental, health, and safetyhazards arising from performance of the con-tract work.

• Evaluates prime contractors� performance onenvironmental restoration tasks.

The paragraphs below present the analysis of thefacts pertaining to DOE-ORO roles and responsi-bilities.

The Office of AMEF exercises line responsibility forall DOE functions at PGDP, including environmen-tal and waste management activities; the OSHACompliance Program; and facilities not leased toUSEC and associated activities, including safety andhealth oversight. Continuity of line responsibilityfor DOE-ORO oversight of contractor activities insupport of environmental restoration programs andprojects, waste management programs andprojects, and capital projects is not evident.

The Paducah Site Manager exercises managementresponsibility for day-to-day on-site direction andtechnical oversight of contractor activities in sup-port of environmental restoration, waste manage-

M anagment responsibility for DOEanagment responsibility for DOEanagment responsibility for DOEanagment responsibility for DOEanagment responsibility for DOEprograms at PGDP and PORTS residesprograms at PGDP and PORTS residesprograms at PGDP and PORTS residesprograms at PGDP and PORTS residesprograms at PGDP and PORTS resideswith AMEFwith AMEFwith AMEFwith AMEFwith AMEF.....

The Phe Phe Phe Phe Paducah Site Manager is respon-aducah Site Manager is respon-aducah Site Manager is respon-aducah Site Manager is respon-aducah Site Manager is respon-sible for performing on-site surveillancesible for performing on-site surveillancesible for performing on-site surveillancesible for performing on-site surveillancesible for performing on-site surveillanceof DOE activities.of DOE activities.of DOE activities.of DOE activities.of DOE activities.

ment, and capital projects. The Paducah Site Man-ager is also responsible for performing on-site sur-veillance of DOE activities to assure compliancewith DOE and other Federal and state environment,safety, and health requirements. The Paducah SiteManager implements site environmental restora-

1 7

tion programs and activities in coordination withthe AMEM, including remedial investigations, al-ternative evaluations, remedial designs, and re-medial actions.

Although the Paducah Site Manager is responsiblefor on-site surveillance of DOE activities and tech-nical oversight of designated projects, there is nota comprehensive assessment program in place toidentify threats to facilities and waste operationspersonnel from exploding drums in Building C-746-Q. There were formal, procedure-basedWalkthrough and Surveillance Programs at thePaducah Site Office until mid-1996. At that time,the Paducah Site Office work activities were re-structured to a project management format. Thiswas due to initiation of performance-based con-tracts (incentive task orders) where DOE specifiesthe scope of work and allows the contractor todetermine how the work is done. Each project iden-tifies the environment, safety, and health require-ments and expected performance with less formaloversight by DOE on the actual work being per-formed. The project management structure re-quires each Project Manager to ensure safety over-sight of his/her projects. This can be done by thecontractor (Health and Safety Officer), subcontrac-tor, DOE subcontractor, or the DOE Project Man-ager. DOE interfaces with the project Health andSafety Officer but does not direct the work of indi-vidual subcontractors.

bility to manage the overall Environmental Resto-ration Program. However, an interface documentdelineating the execution of these responsibilitiesdoes not exist.

DOE Facility Representatives and/or Program Man-agers are responsible for ensuring the adequacyof occurrence reports and approving them. Unlikethe Occurrence Reporting and Processing System(ORPS), DOE does not have a formal role in imple-menting the Lessons Learned Program. DOE-OROcan issue alerts but has not to date. DOE is re-sponsible for the collection of data for the purposeof tracking, trending, analyzing, drawing conclu-sions, and disseminating results. However, past oc-currence reports have not always contained suffi-cient information, and tracking and trending hasnot been performed.

2.2.2.2 LMES Roles and Responsibilities2.2.2.2 LMES Roles and Responsibilities2.2.2.2 LMES Roles and Responsibilities2.2.2.2 LMES Roles and Responsibilities2.2.2.2 LMES Roles and Responsibilities

The following are the roles and responsibilities forLMES Facility Managers, LMES Shift Superinten-dents, and LMES Subject Matter Experts (SMEs) inimplementing the Occurrence Reporting and Les-sons Learned Programs.

• LMES Facility Managers have direct line respon-sibility for the operations within a particularbuilding(s). The following are their responsi-bilities in implementing the Occurrence Report-ing and Lessons Learned Programs within theirbuilding(s):

� initiating an occurrence report following an incident;

� determining direct, contributing, and root causes of the occurrence;

� identifying corrective actions and drafting lessons learned;

� searching databases for similar events within their organization;

� reviewing data in Energy Systems ActionManagement System (ESAMS) and ORPSto identify lessons learned or good prac-tices from other facilities; and

� drafting an appropriate alert (red, yellow,green, or blue).

From the above information, the Board determinedthat the implementation of a comprehensive as-sessment program of DOE and contractor activi-ties at PGDP is not evident. The responsibilities forimplementing site environmental restoration pro-grams and activities in coordination with the AMEMis identified as a responsibility for the Paducah SiteManager. The Office of AMEM has the responsi-

The implementation of a compre-he implementation of a compre-he implementation of a compre-he implementation of a compre-he implementation of a compre-hensive assessment program of DOEhensive assessment program of DOEhensive assessment program of DOEhensive assessment program of DOEhensive assessment program of DOEand contractor activities at PGDP isand contractor activities at PGDP isand contractor activities at PGDP isand contractor activities at PGDP isand contractor activities at PGDP isnot evident.not evident.not evident.not evident.not evident.

1 8

• Shift Superintendents transmit the informationfrom the occurrence report to ORPS.

• The LMES SMEs may be part of line manage-ment or may be part of central staff. SMEsreview and approve the alert level providedby line management.

• Senior management within LMES sign off onyellow, green, and red alerts.

The paragraphs below present the analysis of thefacts pertaining to LMES roles and responsibili-ties.

The usefulness of information retrieved from ORPSdepends upon the accuracy of the information thatis in the system. If information is inaccurate, at-tempts to query the system for similar events willbe unsuccessful. For example, attempts to findaccidents that involved the failure of polyethyleneliners would not locate the Y-12 Plant occurrenceor alert because the root cause was misidentifiedin ORPS (e.g., Y-1997-OR-LMESY12-0701).

ORPS and other databases that support the Oc-currence Reporting and Lessons Learned Programsare terminology limited. A query to search for ac-cidents resulting in a �drum explosion� will notidentify drum explosions that were input into ORPSas drum ruptures, drum overpressurizations, drumfailures, etc. Because there is no standard lan-guage, similar accidents may be reported in vari-ous ways. It is unreasonable, however, to estab-lish a common, comprehensive language for oc-currence reporting or alerts.

Facility Managers are not responsible for identify-ing similar events that have occurred outside oftheir building(s), and thus, most do not have ac-cess to ORPS. Therefore, Facility Managers arelimited in their ability to identify similar accidentsand learn from them.

The intent of the Lessons Learned Program is tocollect and disseminate positive and negative in-formation with other organizations and sites. How-ever, since no single organization or individualwithin LMES is responsible for tracking and trend-ing lessons learned, the collection and dissemi-

nation of like events is not performed, and thesuccess of the program is hindered.

2.2.32.2.32.2.32.2.32.2.3 LMES Assessment P LMES Assessment P LMES Assessment P LMES Assessment P LMES Assessment Programrogramrogramrogramrogram

The following are facts of assessment activitiesperformed by LMES at the Paducah Site.

• Assessment activities performed by LMES aredescribed in Procedure PMQA-1050, Rev. 1,�Integrated Management Assessment Pro-gram.� The purpose of this program is to pro-vide a system of assessments to identify defi-ciencies in performance relative to laws, regu-lations, and DOE Orders and requirements;provide corrective actions for each deficiency;and track corrective actions to closure.

• The assessment activities performed by LMESconsist of organization self-assessments, aPerformance Evaluation Program, complianceperformance measurement system assessment,Facility Excellence Program, readiness reviews,evaluations of Performance MeasurementTeam implementation, management reviews,occurrence reporting system, performance in-dicators and trend analysis, lessons learned/alerts, root cause analysis and issues manage-ment, compliance with contractual require-ments, and management review of assessmentreports.

• Schedules for assessments are developed semi-annually, and deficiencies are identified fromactivities being reviewed. Corrective actionplans and schedules are developed to correctidentified deficiencies. Corrective actions aretracked until closure and reported to DOEmonthly.

The paragraph below presents the analysis of thefacts pertaining to the LMES assessment activities.

Deficiencies in the Lessons Learned Program androot cause analysis are identified in Section 2.2.1of this report. Assessment activities for Building C-746-Q include walkthroughs, technical audits,management assessments, RCRA inspections,compliance audits and surveillances. However,

1 9

these assessment activities did not review lessonslearned reports, occurrence reports, and othertechnical reports that identify problems with nitricacid/hazardous wastes stored in polyethylene/steelcontainers prior to the start of the assessments.As a result of these deficiencies, assessment ac-tivities did not detect problems withoverpressurized drums in Building C-746-Q untilHC-1093 exploded.

2 .32.32.32.32.3 CONTROLS AND ANALCONTROLS AND ANALCONTROLS AND ANALCONTROLS AND ANALCONTROLS AND ANALYSISYSISYSISYSISYSIS

The subsections to follow provide the barrier andcontrol analysis, change analysis, and causal fac-tor analysis.

2.3.12.3.12.3.12.3.12.3.1 Barrier and Control Analysis Barrier and Control Analysis Barrier and Control Analysis Barrier and Control Analysis Barrier and Control Analysis

Barriers and controls generally can be classifiedinto one of four categories: (1) engineering de-sign, which is the most effective means of elimi-nating hazards from a work environment;(2) safety devices, which are placed on individualpieces of equipment, around whole processes, oron employees as a secondary protective measureto ensure work safety; (3) warning devices, whichare placed on individual pieces of equipment orprocesses that cannot feasibly be designed toeliminate every hazard; (4) procedural controls,which are written, implemented, and enforced inwork environments where additional measures arerequired to eliminate the effects of an existing orpotential hazard that cannot be reduced throughdesign or the use of safety and warning devices.

Three types of barriers were present in BuildingC-746-Q prior to the accident. Engineering de-sign barriers, safety devices (PPE), and procedural(administrative) controls were present, and theyall failed to prevent the accident. However, doorlocks and radiation protection barriers kept un-authorized personnel from the area where theaccident occurred. TTTTTable 2-1able 2-1able 2-1able 2-1able 2-1 presents a detailedbarrier and control analysis.

2.3.22.3.22.3.22.3.22.3.2 Change Analysis Change Analysis Change Analysis Change Analysis Change Analysis

Change analysis considers failures in barriers fromplanned or unplanned changes in a system that

disturb normal operations. TTTTTable 2-2able 2-2able 2-2able 2-2able 2-2 shows detailsof the change analysis performed by the Board.

2.3.32.3.32.3.32.3.32.3.3 Causal F Causal F Causal F Causal F Causal Factor Analysisactor Analysisactor Analysisactor Analysisactor Analysis

The direct causedirect causedirect causedirect causedirect cause of the accident was the excessivebuildup of pressure due to chemical reactions withinthe container, resulting in an explosion of the con-tainer; however, there were also causal factors, i.e.,contributing causes and a root cause. Contribut-Contribut-Contribut-Contribut-Contribut-ing causesing causesing causesing causesing causes are causes that, if corrected, would notby themselves have prevented the accident but areimportant enough to be recognized as needingcorrective action to improve the quality of the pro-cess. Root causesRoot causesRoot causesRoot causesRoot causes are the fundamental causes andassociated corrective actions that, if corrected,would prevent recurrence of an event or adverseaction. The causal factor analysis presented in TTTTTableableableableable2-32-32-32-32-3 uses techniques from Management and Over-sight Risk Tree based root cause analysis and theInstitute of Nuclear Power Operations�s Good Prac-tice OE-907, �Root Cause Analysis.�

2 0

Procedure OP-301, To ensure investigation and identifi- Barrier failed because:�Occurrence Notification cation of the root cause following an �Not all occurrence reports identifyand Reporting� occurrence. the appropriate root causes.

�Tracking and trending were not performed.

Procedure PMWM-1002 To outline requirements for segre- Barrier failed because:IAD, �On-Site Handling gating, collecting, storing, treating, �Did not address limitations on using polyeth-and Disposal of and disposing of hazardous wastes. ylene-lined containers for the long-termWaste Material� storage of strong oxidizing agents.

�Did not address specific criteria for overpacking vs. repackaging.�Did not adequately address when pressure relief devices are appropriate or allowable.

Waste management To provide controls for segregating, Barrier failed because:control system collecting, storing, treating, and �There were no approved waste acceptance

disposing of hazardous wastes. criteria that provided requirements for waste treatment, storage, or disposal.�Disregarded findings and recommendations from prior events.�Disregarded warnings in PMWM-1002 IAD regarding buildup of gas pressure in waste containers.

Overpack To provide storage and containment Barrier failed because:drum HC-1093 of nitric acid mixed wastes. �Nitric acid degraded the polyethylene liner

and reacted with the primary container and the steel drum overpack.

Overpack policy/ To provide containment of primary Barrier failed because:practice to use container of acid mixed wastes. �Policy was not followed in late 1986.12-mil polyethyleneliner for overpacks

Procedure QA-331, To outline roles and responsibilities Barrier failed because:�Lessons Learned and provide implementation �Management failed to recognize similaritiesProgram� guidance. of past alerts.

�Tracking and trending were not performed.

Building C-746-Q To provide containment for Barrier failed because:concrete floor dikes hazardous waste spills on �Although this barrier contained the mixed

the floor of Building C-746-Q. wastes that spilled within the limits of the dikes, the mixed wastes exploded beyond the limits of acid waste storage area B-08.

29 CFR 1910.120 To provide a site-specific safety and Barrier failed because:(HAZWOPER) Safety health plan to identify, evaluate, �There was not a site-specific safety andand Health Program and control safety and health health plan or work permit for Building

hazards to hazardous waste C-746-Q that provided for hazard identifica-operations. tion and mitigation for employees exposed

to hazards from overpressurized drums.

Barrier and Control AnalysisBarrier and Control AnalysisBarrier and Control AnalysisBarrier and Control AnalysisBarrier and Control Analysis

Barrier Purpose Performance

2 1

Table 2-1. Barrier and Control Analysis

Steel/polyethyleneprimary containers areutilized for long-termstorage of nitric acid/radioactive wastes.

Planned/Normal Present Difference Analysis

Reliability and integrityof steel/polyethyleneprimary containerquestioned in 1983,1984, and 1991 butcontinued to be used.

Steel/polyethyleneprimary containers arenot suitable for long-term storage of nitricacid/radioact ivewastes.

Gas buildup fromsteel/acid reaction issufficient to causeexplosion in 110-galoverpack.

85-gal overpackshould consist of 12-mil polyethylene linerinside 85-gal steeldrum.

Normal policy/prac-tices not followed, andpolyethylene liner wasnot used.

85-gal steel overpackwas quickly penetratedby nitric acid whenleak developed inprimary container.

Polyethylene linerinside steel drumwould have extendedcontainer life of 85-galoverpack.

Primary containersafely contains acidwastes, and there is nogas buildup in contain-ers and overpack.

Primary containerleaked acid into 85-galsteel overpack causinga rapid developmentof gas.

Low pH acid wastesreacted with steel/polyethylene drum andcaused gas buildup.Low pH acid wastesreacted with steel/polyethylene drum andcaused gas buildup.

Failure to determinelong-term reliability ofcontainers leads topolicy/practice ofoverpacking nitric acid/radioactive wastes.

Procedure PMWM-1002 IAD warns of gasbuildup in wastecontainer requiringinstallation of ap-proved pressure reliefdevice.

Building C-746-Qwaste containers donot have pressure reliefdevices.

There are no safe-guards installed onwaste containers toprevent additionalexplosions.

Potential for wastecontainer explosion inBuilding C-746-Q notmit igated.

Personnel safety forBuilding C-746-Qrequires hazard identi-fication and mitigationaccording to 29 CFR1910.120(HAZWOPER).

Building C-746-Q isunsafe for personnelengaged in hazardouswaste operations dueto potential for explod-ing container.

Potential for explosionof waste containerscaused Building C-746-Q to be unsafefor normal activities.

Safety in BuildingC-746-Q must beenhanced due topotential for containerexplosion.

Lessons LearnedProgram provides fordisseminating, identify-ing, and utilizingpositive and negativeoperating experiences.

Lessons learned notimplemented and notutilized where previousnitric acid/radioactivewaste drum storage isconcerned.

LMES management didnot utilize lessonslearned information torepackage nitric acid/radioactive wastes incompatible containers.

The accident couldhave been preventedby using previousinformation fromlessons learned.

Occurrences requiretimely identification,categorization, notifica-tion, evaluation, correc-tions, and reporting.

Occurrence reportingsystem improperlycategorized the occur-rence and did notprovide timely notifica-tion.

Categorization of eventwas updated to un-usual, and notificationdid not meet the 2-hour time limit.

Corrections to occur-rence report wererequired.

2 2

Change or Difference Change or Difference Change or Difference Change or Difference Change or Difference Analysis Analysis Analysis Analysis Analysis

TTTTTable 2-2. Change Analysisable 2-2. Change Analysisable 2-2. Change Analysisable 2-2. Change Analysisable 2-2. Change Analysis

Procedures

Waste acceptance criteria

Overpack policy or practice

Occurrence reports

Lessons learned

Personnel safety

DOE oversight

Root CauseRoot CauseRoot CauseRoot CauseRoot Cause

Failure of managementcontrol systems

Contributing CausesContributing CausesContributing CausesContributing CausesContributing Causes Discussion Discussion Discussion Discussion Discussion

TTTTTable 2-3. Causal Fable 2-3. Causal Fable 2-3. Causal Fable 2-3. Causal Fable 2-3. Causal Factor Analysisactor Analysisactor Analysisactor Analysisactor Analysis

Warnings in existing procedures regarding the potential for gasbuildup in waste containers were not heeded.

There was no approved waste acceptance criteria document. Thewaste acceptance criteria do not place restrictions on the accep-tance of strong oxidizers or address the limitations of polyethylenefor long-term storage of strong oxidizers.

Policies and practices failed to provide guidance on when it isappropriate to repackage as opposed to overpacking a suspectcontainer.

Occurrence reports do not always identify the appropriate rootcauses in past events. There is no single organization or individualresponsible for tracking or trending information from the occur-rence reporting system.

There was a failure to properly implement the Lessons LearnedProgram, and the findings and recommendations in an LMESreport (Proposed Neutralization/Pre-treatment for Nitric Acid StripTank Waste and Other Drummed Lab Wastes Stored in the Vault4A Facility, dated September 13, 1991) and a yellow alert (Y-PAD-91-0002, �Polyethylene Reagent Container Failure�) weredisregarded. There is no single organization or individual respon-sible for tracking or trending information from the lessons learnedsystem.

There was a failure to provide hazard analysis and mitigation forexploding drums for hazardous waste operations personnelworking in Building C-746-Q (i.e., there was no safety and healthplan). A required safety and health plan, including safety andhealth permits, must conform to 29 CFR, Part 1910.120(HAZWOPER).

DOE failed to perform adequate oversight of environmental andwaste management activities at PGDP and did not identify haz-ards from exploding waste drums.

Discussion Discussion Discussion Discussion Discussion

LMES management failed to recognize the significance and/or actappropriately on information regarding the incompatibility be-tween acid wastes and waste containers.

2 3

Causal FCausal FCausal FCausal FCausal Factor Analysisactor Analysisactor Analysisactor Analysisactor Analysis

2 4

3.03.03.03.03.0 CONCLCONCLCONCLCONCLCONCLUSIONS AND JUDGMENTS OF NEEDUSIONS AND JUDGMENTS OF NEEDUSIONS AND JUDGMENTS OF NEEDUSIONS AND JUDGMENTS OF NEEDUSIONS AND JUDGMENTS OF NEED

This section of the report identifies the conclusionsand judgments of need, as determined by theBoard, by using the accident analysis methods de-scribed in Section 2.3. Conclusions of the Boardconsider significant facts and pertinent analyticalresults. Judgments of need are managerial andprocedural controls believed necessary to prevent

or mitigate the probability or severity of a recur-rence. They flow from the conclusions and causalfactors and are directed at guiding managers indeveloping follow-up actions. TTTTTable 3-1able 3-1able 3-1able 3-1able 3-1 identifiesthe conclusions and corresponding judgments ofneed identified by the Board.

TTTTTable 3-1. Conclusions and Judgments of Needable 3-1. Conclusions and Judgments of Needable 3-1. Conclusions and Judgments of Needable 3-1. Conclusions and Judgments of Needable 3-1. Conclusions and Judgments of Need

Conclusions and Judgments of NeedConclusions and Judgments of NeedConclusions and Judgments of NeedConclusions and Judgments of NeedConclusions and Judgments of Need

Conclusions Conclusions Conclusions Conclusions Conclusions

LMES has lessons learnedprocedures that couldhave, if adequatelyimplemented, preventedthis accident.

No documentation existsrestricting the storage ofstrong oxidizing agents.

PGDP container manage-ment procedures fail toprovide guidance forrepackaging wastes asopposed to overpacking.

There are no provisions forpersonnel safety fromexploding drums in Build-ing C-746-Q.

Neither DOE nor LMESadequately tracks or trendsinformation from occur-rence reports or lessonslearned.

DOE does not adequatelyperform oversight.

Judgments of Need Judgments of Need Judgments of Need Judgments of Need Judgments of Need

LMES management needs to:�Adequately implement a Lessons Learned Program.�Assess the criteria for assigning alert levels.�Ensure the accuracy of occurrence reports and ensure the appropriate root cause(s) have been identified.

LMES needs to proceduralize a process for neutralizing corrosivewastes prior to long-term storage. Restrictions need to be placed onthe acceptance of wastes that pose unique hazards when stored. TheWaste Management Division needs to modify its waste acceptancecriteria (KY/EM-96) accordingly for treatment, storage, and disposalunits.

LMES needs to:�Develop guidance to repackage waste as opposed to overpacking the container when appropriate.�Modify the waste handling procedure (PMWM-1002 IAD) to clarify the limitations of polyethylene-lined containers for storage of strong oxidizing agents.

LMES needs to develop a safety and health plan for HAZWOPERpersonnel in Building C-746-Q that conforms to 29 CFR 1910.120(HAZWOPER).

DOE-ORO needs to track and trend information from the occurrencereporting and lessons learned systems, as defined by ORO M 110,Change 2, dated May 15, 1997.LMES needs to track and trend information from the occurrencereporting and lessons learned systems.

DOE needs to:�Develop a comprehensive oversight program for PGDP.�Ensure occurrence report information is accurate and complete.

B OB OB OB OB O A R D S I G N AA R D S I G N AA R D S I G N AA R D S I G N AA R D S I G N A T U R E ST U R E ST U R E ST U R E ST U R E S

2 5

4.04.04.04.04.0 BOBOBOBOBOARD SIGNAARD SIGNAARD SIGNAARD SIGNAARD SIGNATURESTURESTURESTURESTURES

______________________________________________ __________________David R. Allen DateDOE Accident Investigation Board ChairpersonU.S. Department of Energy, Oak Ridge Operations

______________________________________________ __________________James S. Campbell DateDOE Accident Investigation Board MemberU.S. Department of Energy, Oak Ridge Operations

______________________________________________ __________________Brian C. DeMonia DateDOE Accident Investigation Board MemberU.S. Department of Energy, Oak Ridge Operations

______________________________________________ __________________William M. (Mike) Arendale DateDOE Accident Investigation Board MemberU.S. Department of Energy, Oak Ridge Operations

2 6

L I S T I N G O F B OL I S T I N G O F B OL I S T I N G O F B OL I S T I N G O F B OL I S T I N G O F B O A R D M E M B E R S ,A R D M E M B E R S ,A R D M E M B E R S ,A R D M E M B E R S ,A R D M E M B E R S ,A DA DA DA DA D V I S O R S , A N D S TV I S O R S , A N D S TV I S O R S , A N D S TV I S O R S , A N D S TV I S O R S , A N D S T A F FA F FA F FA F FA F F

5.0 LISTING OF BOARD MEMBERS, ADVISORS, AND STAFF

Board Chairperson David R. Allen, DOE-ORO

Board Member James S. Campbell, DOE-ORO

Board Member Brian C. DeMonia, DOE-ORO

Board Member William M. (Mike) Arendale, DOE-ORO

Advisor Gregory A. Bazzell, DOE Paducah Site Office

Technical Editor Kimberlee A. Davis, PAI Oak Ridge Office

Administrative Support Melissa J. Howell, EASI, DOE Paducah Site Office

Administrative Support Teresa G. Fields, EASI, DOE Paducah Site Office

Graphic Design Sharon P. Partin, EASI, DOE Oak Ridge Office

2 6

APPOINTMENT

CORRESPONDENCE

FOR TYPE B

ACCIDENT

INVESTIGATION

A P P E N D I X A

(4/93)

United States Government Department of Energy Oak Ridge Operations Office

memorandumDATE: September 16, 1997

REPLY TOATTN OF: SE-32:Mullins

SUBJECT: TYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRATYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRATYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRATYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRATYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRAWASTEWASTEWASTEWASTEWASTE STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC.,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC.,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC.,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC.,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC.,PADUCAH,PADUCAH,PADUCAH,PADUCAH,PADUCAH, KENTUCKYKENTUCKYKENTUCKYKENTUCKYKENTUCKY

TO: David R. Allen, Technical Support Division, SE-32

You are hereby appointed Chairman of the Investigation Board to investigate the subjectincident which was discovered on September 15, 1997. Although the incident does not meetinvestigation requirements as defined by DOE Order 225. 1, the similarity to other incidentswithin Lockheed Martin Energy Systems, Inc. (LMES), at Paducah and Oak Ridge cause megreat concern.

You are to perform a Type B investigation of this incident and to prepare an investigationreport. The report shall conform to the requirements detailed in DOE Order 225.1 and DOEG 225.1-1, Implementation Guide for Use with DOE 225. 1, Accident Investigations. Thescope of the investigation is to include, but is not limited to, analyzing causal factors andidentifying root causes which resulted in the incident, and determining judgments of need toprevent recurrence. The Board will also focus on management roles and responsibilities,application of lessons learned from similar type accidents within the Department (especiallythose within LMES), and work planning, practices and procedures. If additional resourcesare required to assist you in completing this task, please let me know and it will be provided.You and members of the Board are relieved of your other duties until this assignment iscompleted.

The following employees have been appointed to serve as members of the Board:

Brian C. DeMonia, Waste Management and Technology Development, MemberJames S. Campbell, Operations Division, Trained InvestigatorW. Mike Arendale, Nuclear Safety Division, Member

The Board will provide my office and Robert Poe, Assistant Manager for Environment,Safety, and Quality, with periodic reports on the status of the investigation and not includeany findings or arrive at any premature conclusions until an analysis of all the causal factorshave been completed.

Draft copies of the report should be provided to LMES and appropriate ORO staff for factualaccuracy review.

David R. Allen - 2- September 16, 1997

The final draft of the investigation report should be provided to me by October 20, 1997. Discussions of theinvestigation and copies of the draft report will be controlled until I authorize release of the final report.

James C. HallManager

cc:T. J. O’Toole, EH-1, HQ, 7A-097/FORSA. L. Alm, EM-1, HQ, 5A-014/FORSG. S. Podonsky, EH-4, HQ, C-303/GTND. Vernon, EH-2 1, HQ/GTNJimmy C. Massey, LMES, 761 Veterans Ave., Kevil, KY 42053S. A. Polston, LMUS, MS-100, PaducahJimmie Hodges, EF-22, PaducahSteve Wyatt, M-4, ORR. W. Poe, SE-30, ORR. R. Nelson, EW-90, ORJ. W. Parks, EF-20, ORW. T. Cooper, EH-24, OR Steve Wyatt, M-4, OR

Department of Energy

Oak Ridge Operations OfficeP.O. Box 2001

Oak Ridge, Tennessee 37831-

September 16, 1997

Dr. Robert I. Van HookPresidentLockheed Martin Energy Systems, Inc.P.O. Box 2009Oak Ridge, TN 37831-8001

Dear Dr. Van Hook:

TYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRA WASTETYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRA WASTETYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRA WASTETYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRA WASTETYPE B INVESTIGATION - RUPTURED WASTE DRUM, C-746-Q RCRA WASTESTORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC., PADUCAH,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC., PADUCAH,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC., PADUCAH,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC., PADUCAH,STORAGE FACILITY, LOCKHEED MARTIN ENERGY SYSTEMS, INC., PADUCAH,KENTUCKYKENTUCKYKENTUCKYKENTUCKYKENTUCKY

As a result of the subject incident which was discovered on September 15, 1997, I am direct-ing that a Type B investigation be conducted. Although the incident does not meet investi-gation requirements as defined by DOE Order 225.1, the similarity to other incidents withinLockheed Martin Energy Systems, Inc. (LMES), at Paducah and Oak Ridge cause me greatconcern.

The investigation will be performed by the following individuals:

David R. Allen, Technical Support Division, ChairmanBrian C. DeMonia, Waste Management and Technology Development, MemberJames S. Campbell, Operations Division, Trained InvestigatorW. Mike Arendale, Nuclear Safety Division, Member

The scope of the Board’s investigation will include, but is not limited to, analyzing causalfactors and identifying root causes which resulted in the incident, and determining judg-ments of need to prevent recurrence. The investigation will be conducted in accordance withDOE Order 225.1. The Board will also focus on management roles and responsibilities,application of lessons learned from similar type accidents within the Department (especiallythose within LMES), and work planning, practices, and procedures.

Dr. Robert I. Van Hook -2- September 16, 1997

Draft copies of the investigation report will be provided for factual accuracy review. The final draftreport of the investigation should be provided to my office by October 20, 1997.

Sincerely,

James C. HallManager

cc:T. J. O’Toole, EH-1, HQ, 7A-097/FORSA. L. Alm, EM-1, HQ, 5A-014/FORSG. S. Podonsky, EH-4, HQ, C-303/GTND. Vernon, EH-21, HQ/GTNJimmy C. Massey, LMES, 761 Veterans Ave., Kevil, KY 42053S. A. Polston, LMES, MS-100, PaducahJimmie Hodges, EF-22, PaducahSteve Wyatt, M-4, ORR. W. Poe, SE-30, ORR. R. Nelson, EW-90, ORJ. W. Parks, EF-20, ORW. T. Cooper, EH-24, OR