lawrence livermore national laboratory livermore ... · our site. we also have identified numerous...

May 1996

Lawrence

Livermore

National

Laboratory

Science &

Technology Review

Lawrence Liverm

ore National Laboratory

P.O.B

ox 808, L-664Liverm

ore, California 94551

Printed on recycled paper.

Nonprofit O

rg.U

. S. P

ostage

PAID

Livermore, C

AP

ermit N

o. 154

Lightning:Maximizing Safety by Minimizing Effects

Also in this issue:• Groundwater Modeling• Dual-Band Infrared Computed Tomography• Waste Minimization in the Plating Shop



SCIENTIFIC EDITOR

Becky Failor

PUBLICATION EDITOR

Dean Wheatcraft

WRITERS

Arnie Heller, Dale Sprouse, Sue Stull,Katie Walter

ART DIRECTOR

Ray Marazzi

DESIGNER

Ray Marazzi

GRAPHIC ARTIST

Treva Carey

COMPOSITOR

Louisa Cardoza

PROOFREADER

Catherine M. Williams

S&TR is produced by the TechnicalInformation Department as a service for the Director’s Office.

2 The Laboratory in the News

3 Commentary on ES&H Technology

Features4 Mitigating Lightning Hazards

The Laboratory is developing a guidance document for mitigation of lightninghazards to Department of Energy facilities, especially those where nuclear andhigh-explosive materials are handled and stored.

13 Groundwater Modeling: More Cost-Effective Cleanup by DesignInnovative groundwater modeling allows us to more accurately target ourgroundwater cleanup efforts at the Livermore site, reducing the overall scopeand cost of the project.

23 Dual-Band Infrared Computed Tomography: Searching for Hidden DefectsThe dual-band infrared computed tomography technologies developed atLivermore are enabling safe, rapid, quantitative nondestructive inspection andevaluation of manmade structures in three dimensions with 10 times thetemperature sensitivity of single-band systems.

Research Highlight28 Plating Shop Moves to Finish Off Waste

31 Patents

32 Abstracts

S&TR Staff May 1996

LawrenceLivermoreNationalLaboratory

Printed in the United States of America

Available fromNational Technical Information ServiceU.S. Department of Commerce5285 Port Royal RoadSpringfield, Virginia 22161

UCRL-52000-96-5Distribution Category UC-700May 1996

This publication is a continuation ofEnergy & Technology Review.

Page 4

Page 13

3

May 1996

Lawrence

Livermore

National

Laboratory





The Lawrence Livermore National Laboratory, operated by the University of California for theUnited States Department of Energy, was established in 1952 to do research on nuclear weapons andmagnetic fusion energy. Science & Technology Review (formerly Energy & Technology Review) ispublished ten times a year to communicate, to a broad audience, the Laboratory’s scientific and technologicalaccomplishments, particularly in the Laboratory’s core mission areas—global security, energy and theenvironment, and bioscience and biotechnology. The publication’s goal is to help readers understandthese accomplishments and appreciate their value to the individual citizen, the nation, and the world.

Please address any correspondence (including name and address changes) to S&TR, Mail Stop L-664,Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, California 94551, or telephone(510) 422-8961. Our electronic mail address is [email protected].

• •

Cov

er p

hoto

: U.S

. Dep

artm

ent o

f E

nerg

y

About the Review

About the Cover

This month’s S&TR features a report aboutLaboratory work on an awesome, inevitable,unpredictable, and potentially dangerous naturalphenomenon—lightning. The article beginning onp. 4 tells of the development of guidance byLaboratory engineers on how to deal with theeffects of lightning on Department of Energyfacilities, especially those where nuclear and high-explosive materials are handled and stored. Theprinciples provided by Lawrence Livermore’sguidance were recently applied at the DeviceAssembly Facility at the Nevada Test Site todetermine the effectiveness of the facility’slightning protection system. Our lightningguidance is one of many ongoing Laboratoryefforts to improve the safety and health of theworkplace and the environment at large.

Prepared by LLNL under contractNo. W-7405-Eng-48

S&TR is available on the Internet athttp://www.llnl.gov/str/str.html. As referencesbecome available on the Internet, they will beinteractively linked to the footnote references atthe end of each article. If you desire moredetailed information about an article, click onany reference that is in color at the end of thearticle, and you will connect automatically withthe reference.

Electronic Access

We want to know what you think of ourpublication. Please use the enclosed survey formto give us your feedback.

What Do You Think?

2 The Laboratory in the News

Science & Technology Review May 1996

Clementine 2 will have Lab-designed payloadThe Laboratory will provide the integrated payload for

Clementine 2, an asteroid-flyby mission that will extend thelegacy of Clementine 1, the spacecraft that pioneered low-cost,fast-turnaround space flight. Clementine 2 will offer the firstclose-up look at an Earth-crossing asteroid and also launch thefirst manmade objects to come in contact with an asteroid.

The mission calls for flybys of three asteroids over a yearof operation, with a probe launched toward each. The probeswill be autonomous micro-spacecraft that will relay data tothe mother ship right up to the moment they are destroyedon impact.

Air Force Space Command will fund the program throughPhillips Laboratory. The Naval Research Laboratory willbuild the mother ship and handle flight operations.Livermore will provide a multispectral suite of sensors (as itdid for Clementine 1) as well as the probes and will analyzethe mission science data. Clementine 2 will be launched inmid-1998.Contact: Tom Karr (510) 423-3670 ([email protected]).

AVLIS technology nears commercializationThe Atomic Vapor Laser Isotope Separation (AVLIS)

uranium enrichment process, a technology developed at theLaboratory that uses laser beams to enrich uranium for nuclearpower plants, has moved a step closer to commercialization.

In March, the United States Enrichment Corporation(USEC) signed two contracts related to the commercialdevelopment of AVLIS. One, awarded to a team of contractorsheaded by Bechtel National Inc., is for the design of acommercial AVLIS facility. The second, awarded to Babcock& Wilcox’s Naval Nuclear Fuel Design Division, is for thepreliminary design of the AVLIS separator refurbishmentsystem and the uranium management system.

Bechtel’s multiphased contract extends through the year2003; during the first phase, the Bechtel team will employabout 50 people, mainly at USEC’s operations at LawrenceLivermore. The Babcock & Wilcox contract runs for 15 months; about 45 people will be involved in the effort,primarily at the firm’s Lynchburg, Virginia, facility.

The USEC is a government corporation that produces andmarkets uranium enrichment services to more than 60 utilities that own and operate commercial nuclear powerplants in 14 countries, including the United States.Contact: Jim Early (510) 422-6221 ([email protected]).

Astrophysicists, laser scientists eye partnershipThe exploration of uncharted areas for collaboration was

the goal in late February as the Laboratory sponsored itsfirst-ever workshop on laboratory astrophysics experimentswith large lasers.

Mike Campbell, associate director for Lasers, told thegathering that he is committed to making the LawrenceLivermore’s Nova laser available to the broader scientificcommunity. He said he would like 10% of Nova shots to becollaborative basic science experiments.

With access to large lasers, astrophysicists will be able to conduct experiments in hydrodynamics, radiativehydrodynamics, magnetohydrodynamics, and radiativemagnetohydrodynamics.

What started as a workshop for 40 scientists, mostly fromthe national laboratories, attracted 100 scientists representing30 universities, institutes, and laboratories in eight countries.Contact: Bruce Remington (510) 423-2712 ([email protected]).

Wadsworth, Miller receive new Lab positionsJeff Wadsworth and George Miller, both members of

Livermore’s Senior Management Council, are serving in newmanagement positions. The appointments of Wadsworth asDeputy Director for Science and Technology and Miller asAssociate Director for National Security were announcedearlier this year by Director Bruce Tarter.

The AD for Chemistry and Materials Science since 1994,Wadsworth now serves as the key executive responsible forthe quality of science and technology in Livermore’sscientific and technical programs. Among his responsibilitiesare collaborative research with the University of Californiaand oversight of LLNL’s Department of Defense ProgramsOffice and its Office of Industrial Partnerships andCommercialization.

An Associate Director since 1985, Miller moves from theposition of AD for Defense and Nuclear Technologies to hisnew job, which gives him responsibility for developing astrategic and tactical plan that appropriately definesLivermore’s role in national security.Contact: LLNL Media Relations Office (510) 423-3118.

3


ECHNOLOGIES that detect hazards and prevent exposuresand injuries in the workplace have always been key tools of

the Laboratory’s safety and health protection programs. Inaddition, since the 1980s, we have been using technology to helpus develop innovative solutions to complex problems such asgroundwater contamination, legacy waste disposal, and pollution.Now, technology is proving to be a valuable ally in reducing thecost of environment, safety, and health (ES&H) training andcompliance activities. This issue of Science & TechnologyReview presents important examples of Livermore ES&Hsuccesses in our workplace and in a larger environmental context.They represent our continuing effort to make the Lab—and theworld at large—a safe and healthy place to work and live.

Solutions for Workplace HazardsOne major focus of our ES&H research and development

is to supply solutions to unusual workplace hazards. In onesolution—protecting us from natural phenomena such asearthquakes and lightning—Livermore engineers have developeddesign standards for various facilities (see p. 4 of this issue).Another solution is the stainless steel high-efficiency particulate-air (HEPA) filter, which is fireproof and waterproof and madeespecially for use in extreme environments. Because of our needto monitor hazardous environments, we have been involved indeveloping techniques for rapid dosimeter readouts; a realisticphantom, or dummy, for calibrating plutonium lung counters;and a miniature pump-driven air-sampling device worn byworkers to monitor nearby hazardous chemicals.

Our special expertise in evaluating protective clothing andequipment has led to a sophisticated system for testing fire-fighting gear against exposure to direct flame and anexperimental facility for evaluating respirators for their abilityto protect humans under a variety of conditions. This work hasled to the development of a face mask that alerts the firefighterto high carbon monoxide conditions, such as those createdduring wildfires, and techniques for evaluating clothing thatprotects against chemical agents.

Injuries resulting from repetitive motion—like carpal tunnelsyndrome—are the fastest growing cause of lost time at LLNL.To assist in designing keyboards that cause less repetitive strain,we have been developing techniques to analyze thebiomechanical effects of various keyboards.

Cleanup, Prevention, Lower CostsWhen groundwater contamination from solvents was first

discovered at LLNL in 1983, we conducted a study tocharacterize the complex geological structure underneath theLaboratory site. Our new techniques for monitoring our wellsprovide the basis for sophisticated hydrostratigraphic modelsof the groundwater flow and contaminant transport (see S&TR,January–February 1996). These models are now being used insimulation codes (see p. 13) to optimize time—and money—spent to use our smart “pump and treat” process for cleaningup the site. Our work in artificial neural networks andsubsurface contaminant modeling will be a key to futurerestoration efforts.

To help prevent pollution, we have made extensiveinvestments in plating equipment and plating techniques (see p. 28) to eliminate a major waste stream. Now our platingprocesses are safer for workers, use materials that are lesshazardous, and create less waste. We even recycle and reusethe rinse water, which is cleaner than the water coming ontoour site. We also have identified numerous substitutes forozone-depleting solvents, many of which do a better cleaningjob than the original product, and we developed a portablemass spectrometry system to quantify surface contaminationlevels on parts and equipment to provide real-time feedbackduring cleaning operations.

We are using the latest computer technologies to ensurecompliance with various safety and environmentalregulations and to reduce costs. For example, barcodescanners and portable computers are used to track wastecontainers and stored chemicals throughout our facilities;computer-based documentation allows immediate, onlineaccess to manuals and chemical hazard data sheets andpermits supervisors to quickly review the certificationrecords of their employees; and we are finding the WorldWide Web to be a cost-effective and convenient way foremployees to keep up-to-date in their training.

By using our specialized skills and working systematicallyon the Laboratory’s specific ES&H requirements, we canimpact the standard of excellence in ES&H for our Laboratoryand in the larger environmental context as well.

Commentary on ES&H Technology

T

Dennis K. FisherAssociate Director, Plant Operations

summer by Michael Kelly, a highschool electronics teacherparticipating in LLNL’s SummerResearch Intern Program. Kellyanalyzed reports of 365 lightning-related occurrences at DOEfacilities since October 1990 andfound that lightning caused a varietyof physical damage and alarmsystem malfunctions. His reportdraws attention to numerous failuresof lightning protection devices,alarm systems, and backupgenerator systems.

Unified Approach Needed

No single document presents aunified approach to lightning safetyand protection, Hasbrouck notes. The

latest version of the U.S. NationalFire Protection AssociationLightning Standard only brieflyreferences surge suppression. TheStandard for Safety LightningProtection Components recentlyissued by Underwriters Laboratoriesprovides only general information.What’s more, says Hasbrouck, theDepartment of Defense Ammunitionand Explosives Safety Standards arevery general and offer neitherguidance nor applicable references.The recently released revision of theInternational Lightning ProtectionStandard is significantly ahead ofU.S. standards.

In response, in 1993 Hasbrouckand fellow engineer Kartik Majumdarproposed developing a guidance

document to help DOE managers inassessing the lightning risksassociated with any facility anddetermining the most effective meansfor mitigating the hazard. That year,the two engineers organized alightning workshop in Floridasponsored by DOE’s Office of RiskAnalysis and Technology. Attendeesfrom DOE and its contractors, theNuclear Regulatory Commission, andother federal government agenciesagreed that a comprehensiveguidance document on lightningprotection was needed as one of aseries on natural phenomena hazardsmitigation.

Released in draft form in 1995, itis anticipated that the “LightningHazard Management Guide for DOE

5


HE awesome sound and light show of a thunderstormhas always been a source of

fear and wonder. At any time some2,000 thunderstorms are in progressaround the globe, causing themajority of forest fires and, in theU.S. alone, hundreds of millions ofdollars in property losses. Lightningis also the leading weather-relatedkiller in the U.S., causing from 100to 200 deaths annually.

Despite these facts, most engineersand architects have at best only arudimentary knowledge of lightningand protection methods, says RichardHasbrouck, LLNL engineer in theDefense Sciences EngineeringDivision within ElectronicsEngineering. A lightning expert,

Hasbrouck is co-author of the draft“Lightning Hazard ManagementGuide” for the Department of Energy.

“Lightning and its associatedeffects are a mystery to manyengineers because these subjects arenot included in most engineeringcurricula,” he says. Department ofEnergy managers whose job it is toassure the mitigation of naturalphenomena hazards (such aslightning) for an operation or facilitymust contend with a hodgepodge ofscientific data related to lightning,commercial products (some ofquestionable worth), unrealisticbuilding codes, folklore, and half-truths. As a result, he says, manyfacilities, instruments, and controlsystems are vulnerable to damage or

lightning-induced upset ormalfunction.

Hasbrouck points out that earliergenerations of electrical andelectronic systems and componentsused vacuum tubes, relays, andanalog control and computationdevices that were intrinsically morerobust against the effects of lightningthan are today’s solid-state,microprocessor-based systems. Briefovervoltages caused by lightning andmanmade transient voltages canimmediately destroy low-power,solid-state components such ascomputer chips or weaken them tothe point that they fail months after alightning event.

DOE facilities’ vulnerability tolightning was underscored last

4


Mitigating Lightning HazardsMitigating Lightning HazardsLawrence Livermore engineers’ investigations of lightning, its hazards, and how to protect against them have led to the development of guidance to aid in dealing with the effects of lightning on DOE facilities, particularly those where nuclear and high-explosive materials are handled and stored. Our guidance document provides risk managers with a unified and graded method for attaining lightning safety.

T

Facilities” will be used by new facilitieswithin DOE, other governmentagencies, and the private sector fordetermining design requirements aswell as for evaluating existing lightningsafety and protection systems.1

“Lightning Hazard ManagementGuide” presents a unified approach tolightning safety and protection thatcombines hazard identification andfacility categorization with a newconcept—the Lightning Safety System.The Lightning Safety System integratesfour lightning safety elements usuallyaddressed as separate topics: a warningsystem, a warning response plan, aprotection system, and a safety systemcertification plan. Hasbrouck notes thatthe extent to which each element isimplemented at a particular site dependsupon the mission of the facility; theelement’s impact upon the safety ofworkers, the public, and theenvironment; and the cost versus benefitof the element’s implementation.

“This flexibility allows managers toapply a graded approach in determiningthe most effective mix of hardware,software, and procedures to solve theirparticular problem,” he says.

Hasbrouck has researched lightningand its effects for more than a decade.He carried out rocket-triggeredlightning tests and was responsible forthe design of the Lightning InvulnerableDevice System to protect nuclearexplosive test device systems fromlightning.2,3 He has also written (andpresented) a tutorial (Lightning—Understanding It and ProtectingSystems from Its Effects4) based onclassic texts, current literature, andLLNL experiments.

The work of Hasbrouck and hiscolleagues is part of a larger LLNLeffort to better understand the effects of lightning. Other members of theLaboratory’s Defense SciencesEngineering Division have long workedwith people in LLNL’s weapons

program and with experts from Sandiaand Los Alamos National Laboratoriesto ensure that nuclear warheads areprotected from lightning (see the box on p. 9). At the same time, LLNLatmospheric investigators have beenworking to determine lightning’scontribution to acid rain. One group isalso studying massive, high-altitude,cloud-to-sky lightning-related eventscalled “sprites.”

Huge Electrical Discharge

Hasbrouck explains that lightning isan electrical discharge of immenseproportions* that accompanies not onlythunderstorms, but also volcaniceruptions, snow and dust storms, andsurface nuclear detonations. At the mid-Northern latitudes, some 80% oflightning occurs within clouds(intracloud). About 20% of all lightningis cloud-to-ground, while an extremelysmall percentage is cloud-to-sky andbetween clouds. (Figure 1)

Cloud-to-ground lightning representsthe greatest threat to people, structures,systems, and components. It can beeither positive or negative. The vastmajority of cloud-to-ground lightning isnegative, that is, it transfers negativecharge to Earth via a channel—thestepped leader—that emanates from thelower portion of a storm cloud andmoves toward the Earth. Once theleader contacts Earth, positive chargemoves back up the negatively chargedchannel, neutralizing it and its source, a negative charge center in the cloud.

During a positive lightning event,on the other hand, a large quantity ofpositive charge is transferred to theEarth. Negative charges move back up the lightning channel to thethunderstorm cloud, temporarily

neutralizing a highly positive-chargedregion within the cloud. Positivelightning occurs much less frequentlythan negative lightning and most oftentoward the end of a thunderstorm,originates in the upper part of athunder cloud rather than in the lowerpart, and can be more severe in itseffects than negative lightning.

A single lightning event, called aflash, typically lasts for many hundredsof milliseconds (see the box on p. 6);intense thunderstorms can produceseveral thousand cloud-to-groundflashes. For each discharge, a tree-likestreamer (or leader) carries chargetoward the ground until the “strikingdistance” (30 to 100 meters) is reached.The oppositely charged return stroketransforms electrostatic potential energyinto electromagnetic energy (radio andlight waves), heat, and acoustic energy(thunder).

As delineated in the draft DOEguide, lightning hazard identificationconsiders the severity of the hazard andits likelihood of occurrence. Theseverity of a flash is defined by the peakamplitude of its return stroke current, itsrate of rise, and the amount of chargetransferred, while the probability of anobject being struck is the product of thelocal ground-flash density times itslightning-attractive area.

The guide recommends thatmanagers combine a timely andcredible threat warning system withsuitable lightning protection methods.The warning system should provide an alert when a lightning threat isidentified and an alarm when lightningis imminent. A suitable plan forresponding to a warning should also be implemented.

Cloud-to-ground lightning occursrandomly, making it impossible to

7


Lightning6


Lightning

Cloud-to-ground lightning is the best understood—and mostdangerous—type of lightning. It comes in two varieties, positiveand negative. Here we will discuss only negative lightning.

As with other types of lightning, negative cloud-to-groundlightning begins when complex meteorological processes, drivenby powerful updrafts, cause a tremendous electrostatic chargeseparation to build up within a thunderstorm cloud. Typically, thebottom portion of the cloud is negatively charged. When voltagelevels of about 50 to 100 million volts are reached, air can nolonger provide insulation, and electrical breakdowns calledintracloud lightning take place within the cloud.

Some 10 to 30 minutes after the onset of intracloud lightning,negative charges called “stepped leaders” emerge from thebottom of the cloud, moving toward the earth in 50-meter-longsteps at speeds of 0.03 to 0.07% of the speed of light (about 100to 200 km/s). (See the illustration below.) The leaders carry thefull voltage of the cloud’s negative charge center and create anionized channel. As the leaders near the Earth, their strongelectric field causes streamers of positively charged ions todevelop at the tips of grounded pointed objects. These objectsmay include pine needles, blades of grass, towers, raised golfclubs, and human heads.

These positively charged streamers flow upward under thestrong influence of the negatively charged stepped leader.When thedistance between a stepped leader’s tip and one of the streamersbecomes small enough (known as the striking distance, from 30 to100 meters), the intervening air breaks down and the leader is

joined to Earth via the streamer. Now a pulse of current known as a“return stroke” ranging from thousands to hundreds of thousands ofamperes moves at one tenth to one third the speed of light (35,000to 100,000 km/s) from Earth through the object from which thestreamer emanated and up the ionized channel to the charge centerwithin the cloud, temporarily neutralizing it. An ionized channelremains in the air, and often, additional negative charges, calleddart leaders, will quickly move down this path, resulting insubsequent return strokes. It is this multiplicity that causes the flashto appear to flicker. After 30 to 60 seconds, the neutralized centerrecharges and is ready to produce another flash.

The return stroke’s extremely high temperature (30,000 kelvin)creates the highly visible lightning channel and instantly turnsmoisture into steam, producing the associated thunder. The entireevent, often consisting of multiple return strokes and typicallylasting up to 1 second, is referred to as a lightning flash.

Most direct damage results from the heavy return strokecurrent that produces a large temperature rise in the resistance ofthe channel through which the charge travels or from arcing at thepoint of attachment. When arcing takes place in a combustible orexplosive environment, fire or an explosion can result. If thelightning current is carried by an enclosed conductor (e.g., withina jacketed cable, through a concrete wall, or beneath a paintedsurface), entrapped moisture is turned into high pressure steam,which can cause the cable or painted object to burst, the wall or atree to explode, or the shoes to be blown off the damp feet of aperson struck by lightning.

Striking Facts about Lightning

Steppedleader

Dart leader

Returnstroke

Subsequent returnstroke

Strikingdistance

30 to 100 meters

Streamer

(a) (b) (c) (d)

Artist’s rendering of a cloud-to-ground lightning flash from (a) development of the negatively charged stepped leader and positively chargedstreamer through (b) the return stroke followed by (c) a dart leader resulting in (d) a subsequent return stroke. * The average electrical discharge of lightning is about 15 coulombs; the highest charge transfer is estimated to be about

350 coulombs. One coulomb is the equivalent to the electric charge of 6.24 ¥ 1018 electrons.

“Lightning is a very-large-amplitudecurrent source,” says Hasbrouck. Thismeans that the same amount of currentwill flow, regardless of whether its pathis of low resistance (a metal flagpole) orhigh resistance (a tree). Much of theenergy contained in a lightning returnstroke is dissipated as heat in whateverpath serves as the current-carryingconductor. A good electrical conductor,e.g., metal structure, will experiencelittle more than minor surface pittingwhere the current enters and exits.Significant damage can result, however,when poor conductors such as a wood-frame building, concrete wall, or treeare struck.

The draft guide presents the“fortress” concept, in which first-levelprotection of a structure is provided bya lightning grounding system. Allelectrically conductive paths thatpenetrate the building (e.g., metallicpipes and vent stacks) are bonded to thelightning grounding system externallyat the point of entry. To protectcomponents housed inside, all electricalconductors pass through transientlimiters (surge arrestors) located insidethe structure as close as possible to thepoint of entry. Also, limiters arerecommended at the power and datainput points of individual systems andcomponents.

Using New Testing Methods

Last year, the document got its firstreal-world application when Hasbrouckengaged LLNL engineer RichardZacharias and Richard Collier, anelecromagnetics consultant from EMA Inc. with experience using swept-radio-frequency testing for facilitylighting studies, to determine theeffectiveness of the lightning protectionsystem of the recently completedDevice Assembly Facility (DAF) at theNevada Test Site. The cavernous DAF

9


Lightning

accurately predict when and where it willstrike. However, lightning announcesitself in several ways. A thunderstorm’scloud-to-ground lightning provides someadvance warning if its visible, audible,and electromagnetic signals are detected.Ideally, a system designed to acquire anddisplay warnings needs to incorporateone or more direct weather observations,National Weather Service reports(including information from the NationalLightning Detection Network), flashdetectors, and electric field sensors.

Strikes Are Inevitable

In the lightning guide, the authorsemphasize that despite nonscientificcommercial claims to the contrary, thecharge in a thunderstorm can bedissipated only by nature’s way—thelightning process. Proper lightningprotection accepts a strike as inevitable,seeks to provide a controlled path forthe current to follow, and minimizesthe development of hazardous potentialdifferences.

8


Lightning

Figure 1. An artist’s depiction of the fourbasic kinds of lightning: (a) cloud-to-skylightning (sprites), (b) cloud-to-groundlightning, (c) intracloud lightning, and (d) intercloud lightning.

(a) Cloud-to-sky(sprites)

(d) Intercloud

(c) Intracloud

(b) Cloud-to-ground

Protecting the Nuclear Stockpile from Lightning

The U.S. military and space programs have long respected the potential of lightning to damage or even destroy vital weapons components as well as aircraft and spacecraft.Lightning caused the annihilation of a World War I arsenal in New Jersey, it almostturned the Apollo-12 launch into a disaster, it has been responsible for several aircraftcrashes, and it led to the destruction of an Atlas-Centaur launch vehicle in 1987, with its$160-million payload.

The vulnerability to lightning of today’s aircraft and spacecraft is greater than in yearspast because critical airborne systems employ vast numbers of solid-state components thatare susceptible to the effects of a lightning strike as well as the associated electromagneticfields. In addition, new designs increasingly substitute composite materials for metallicsurfaces, eliminating what once was, in effect, a flying Faraday cage, that is, an almostcomplete metal enclosure that houses the aircraft’s electrical and electronics systems.

Lawrence Livermore lightning expert Mike Wilson of Defense Sciences EngineeringDivision notes that for people residing in the San Francisco Bay Area, home of typicallyfive lightning storms a year, damage from lightning may seem a far-fetched threat.However, DOE’s Pantex plant, located in Texas, experiences about 60 lightning stormsannually, and the threat of lightning igniting some of the propellants and high explosivesstored at the plant is a real concern. Indeed, DOE considers lightning a particular risk tooperations involving the transport, maintenance, and modification of nuclear devices andtheir associated non-nuclear explosives.

Wilson says that Livermore engineers have been assessing the potential threats fromlightning strikes for more than 15 years as part of the Laboratory’s mission to assure thesafety of nuclear devices. “We’re concerned with all environmental threats to the nuclearstockpile,” he says. “Lightning has been considered by some people to be an awesomeenvironmental threat from which nothing could survive. But that’s not the case. We justneed to understand lightning and protect against it.”

Ensuring that nuclear warheads and their components can withstand a lightning strikefocuses on designing multiple physical barriers that block the transfer of energy from alightning bolt to critical components and materials contained within a nuclear device. Inaddition, LLNL engineers use computer models to mimic the electromagnetic fieldsgenerated by lightning storms that can affect wiring connected to the high explosivesfound in every nuclear device. Other models, based on welding computer codes, simulatethe effects of direct lightning strikes upon metal.

Another area of research is applying statistics to the threat of lightning to sharpen theestimates of the frequency of lightning strikes. The work is similar to Laboratory riskanalyses regarding seismic safety and nuclear power plants.

When underground tests were being planned and conducted at the Nevada Test Site,Laboratory test personnel always kept a watchful eye on the weather. Instrumentsmonitored atmospheric electrification, and the U.S. Weather Service operated a cloud-to-ground lightning locating system. During the late 1980s, the Laboratory Test Programadopted a lightning protection method designed by LLNL engineer Richard Hasbrouckthat took advantage of the fact that the nuclear device system was contained within asteel enclosure. In Hasbrouck’s design, called the Lightning-Invulnerable Device System,the explosive device and associated components reside within a “fortress,” a closed,metallic surface connected to another grounded conductor similar to a Faraday cage. Thedesign was validated through tests that first used simulated lightning at the LightningTransient Research Institute in Miami Beach, Florida, and later rocket-triggered lightningat NASA’s Rocket-Triggered Lightning Facility at the Kennedy Space Center.

The comprehensive lightning appraisal of the Device Assembly Facility at the NevadaTest Site discussed at the left was designed by Laboratory personnel. The appraisalidentified weaknesses in the facility’s lightning protection system through the use of astate-of-the-art swept-radio-frequency testing procedure.

Phot

o by

Tre

va C

arey

11


Lightning

was originally designed to safely andsecurely house nuclear test deviceassembly activities and will be availableto support DOE Defense Programsstockpile stewardship activities. Its roof,rear wall, and much of its side walls areearth-covered.

The concepts presented in the newguidance document were used toevaluate the DAF’s lightning safetysystems. In conducting the study, thereview team focused on the mainreinforced concrete structure, not thefacility’s peripheral outbuildings.

Traditional measurements showedthat the DAF structure exhibited a lowvalue of direct current resistance, notsurprising because it contains a verylarge quantity of interconnected metalin good contact with the Earth.However, was it immune to lightningdamage? To find out, the team went astep further and conducted anelectromagnetic survey, consisting oflow-level radio-frequency testing todetermine lightning’s likely penetrationof the structure. (See the box on p. 11.)

The testing at DAF revealed thatsmall to moderate amounts of lightningcurrent could enter the interior viametallic paths provided by objects suchas the vent stacks and antenna feedlinesthat penetrate the roof top. Thesefindings are helping facility managersevaluate additional protective measuresfor the facility.

The study confirms that cloud-to-ground lightning represents a naturalphenomena hazard in the DAFenvironment and estimates that lightningwill strike some point of the facility aboutonce every 20 years. This figure wasarrived at by multiplying the ground-flashdensity at DAF, which is based on fiveyears of actual NTS lightning strike data(Figure 2) and the lightning-attractivearea of the DAF. If the DAF wereentirely underground, its ground-surfacearea would be the lightning-attractivearea. However, lightning-attractive areaincreases with object height.Consequently, the twelve 39-meter-tallmetal light poles around the DAF’sperimeter are most likely to be hit bylarge-to-severe amplitude strokes (greaterthan about 40 kiloamperes), while small-to-moderate ones are more likely to strikethe structure.

An unexpected outcome of theanalysis showed that a large peak-amplitude stroke to a light pole willproduce essentially the same effectswithin the DAF structure as a smallstroke attached directly to a point on

10


Lightning

��

Vent stack

Blow-out gravel

Soil

Vent

Vent

Personal computer

Vacuum pump

Vacuum pump

Ground straps

Lightning ground system cables

LGS cables

Concrete

Measurement coil

Sense coil Injection coil

Vacuum line

Lightning downconductor

Rebar

Bonding strap

50-W amplifierTest injection cable

Assembly Room

Coaxial cables

Network analyzer

Figure 2. One of several flash-density maps in the area of theDevice Assembly Facility (DAF) atthe Nevada Test Site in southernNevada used to arrive at a strikeprobability estimate. The DAF is atthe center of the map.

Low-level radio-frequency (rf) testing was done on theDevice Assembly Facility (DAF) at the Nevada Test Site totest the effectiveness of the DAF’s lightning protectionsystem. The block diagram below shows an assembly room at DAF outfitted with the rf testing system. The permanentDAF structural elements are in green; the testing systemcomponents are red.

During rf testing, the network analyzer continuously emitsa signal over the range of 10 kHz to 30 MHz, representing(approximately) the rf spectrum of a lightning stroke. Theanalyzer’s output signal travels through coaxial cable to the 50-watt amplifier located on the DAF’s roof. The amplifiedsignal then proceeds to an injection coil and through the testinjection cable to lightning grounding system downconductorsconnected to two of the facility’s metal vent stacks. These ventstacks are interconnected by means of the lightning groundingsystem’s rooftop conductors. Such rooftop penetrations areunintentional electrical conductors, allowing lightning currentsto enter the DAF. Various conductive paths inside the DAF,e.g., bonding and/or mechanical attachments, electrically

connect the vent stacks to the DAF’s steel rebar and thefacility’s lightning grounding system.

As the rf signal travels to the downconductor, it passesthrough a sense coil, which sends a sample of the applied signalback to the network analyzer. Because the injection cable willalter the rf signal, this sample provides the network analyzerwith the characteristics of the signal being applied to the stack.The applied test signal will divide, with some portion flowingon the lightning grounding system conductors and someentering the DAF via the vent stack.

In the configuration shown, the testing system’smeasurement coil detects that portion of the applied signalflowing on the copper vacuum line and sends it to the networkanalyzer. The network analyzer compares this signal’scharacteristics to those of the applied signal sampled by thesense coil. The personal computer is used to analyze andarchive the data. Later, this low-level data is scaled up to levelsassociated with a lightning strike, allowing modeling of worst-cast lightning effects on the facility to determine the adequacyof the facility’s lightning protection system.

Low-Level radio-frequency testing at the Device Assembly Facility

the structure’s roof. The poles,therefore, are expected to effectivelydivert large-to-severe amplitude returnstrokes away from the numerousrooftop points of entry.

Hasbrouck is gratified that the DAFlightning study provided an opportunityto apply the concepts put forth in theguidance document. By employingradio-frequency penetration testing, itwas possible to identify how and howmuch lightning energy would leakthrough “holes” in what the lightningprotection code would have judged tobe a solid facility. He notes that a 1993lightning study of DOE’s Pantex facilityalso recommended that some form ofpenetration radio-frequency testing becarried out in the future.

“Lightning knowledge,” heemphasizes, “is neither archaic norarcane. We cannot prevent lightning,but knowledge of it can help us enhancesafety, protecting us and costly propertyagainst its damaging and potentiallycatastrophic effects.”

Key Words: hazard management, lightning,radio-frequency testing.

References1. R. T. Hasbrouck and K. Majumdar,

Development of a Guidance Documentfor Lightning Protection of DOEFacilities, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-119983 (1995).

2. R. T. Hasbrouck, et al., Final Report ofthe Nuclear Explosive LightningVulnerability Task Force, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-15618 (1984).

3. Richard T. Hasbrouck, et al., FinalReport for the Device Systems LightningProject, Lawrence Livermore National

Laboratory, Livermore, CA, UCRL-21068 (1988).

4. R. T. Hasbrouck, Lightning—Understanding It and Protecting Systemsfrom Its Effects, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-53925 (1989).

For further information contact Richard Hasbrouck (510) 422-1256([email protected]) or Kartik Majumdar (510) 422-5060([email protected]). For information onthe Device Assembly Facility study,contact Richard Zacharias (510) 423-0714([email protected]).

12


Lightning

RICHARD HASBROUCK holds a B.S. in electrical engineeringand is a Registered Professional Engineer in California. Hejoined the Laboratory in 1968 and is currently a seniorelectronics engineer. During his career at Lawrence Livermore,he has supported various projects in the test program.Involvement in nuclear explosive safety studies led to his studyof lightning safety, culminating in the lightning hazardmanagement concept. He has written numerous papers and

reports on this subject and is co-author of the draft “Lightning Hazard ManagementGuide for DOE Facilities” (1995). Hasbrouck is a consultant and co-director forengineering of the National Lightning Safety Institute.

In a collateral Laboratory assignment, he is the aviation project officerresponsible for the aviation safety interface between Lawrence Livermore and theDOE on Laboratory projects related to aviation, and he was general manager of theLaboratory’s F-27F aircraft operation. Prior to joining the Laboratory, he designed,tested, and fielded electro-optical, instrument servomechanisms for astronauttraining simulators produced by Farrand Optical Co. in New York for NASA’sProject Apollo.

About the Engineer

13

Groundwater Modeling:More Cost-EffectiveCleanup by Design

Groundwater Modeling:More Cost-EffectiveCleanup by DesignComputer modeling is proving its usefulness as cleanup ofcontaminated groundwaterproceeds at the Livermore site. Modeling is an extremelyeffective tool for deciding whereand how groundwater remediationefforts should be directed. Our models arebeing made available to others for moreefficient remediation around the country.


ROUNDWATER modeling usesmathematical methods to helpscientists “see” what is happening

underground, to make up for what wecannot see with our own eyes. Thediscipline of groundwater modeling hasbeen around for at least 25 years, butwith the powerful desktop computersand advanced software available today,computational modeling is an easierand more effective task than it used tobe. Evaluation processes that used totake days or even many weeks can nowbe done in minutes and often with ahigher degree of accuracy.

G We have developed several newsoftware tools that can be used bygroundwater remediation plannersanywhere. MapIt, for example, can reada variety of one-, two-, and three-dimensional data sources and will allowremediation planners to rapidly produceinput files for the various simulationcodes. With MapIt, we have reducedthe time needed to regrid and executenew three-dimensionalconceptualizations from months tohours. In the past, a different “codepreparation” program was required foreach groundwater simulation code.

Another tool is PLANET, an easy-to-use, point-and-click, drag-and-dropprogram that replaces laborious, manualoperation of modeling codes to evaluatealternative remediation scenarios(Figure 1). Using these and other newlydeveloped tools, groundwater scientistsor engineers at Lawrence LivermoreNational Laboratory and elsewhere canquickly prepare and simulate robustthree-dimensional conceptual models ofour site.

We now have the ability to simulategroundwater flow and transport in alarge number of possible configurations

Scientists at Lawrence Livermorehave been developing and usingadvanced numerical modelingtechniques for decades because thedesign of nuclear weapons requiresextensive modeling prior to testing. Itmade sense, when groundwatercontamination was discovered, for ourin-house researchers to continuemodeling, albeit this time with a verydifferent goal.

Why We Model

We have known since 1983 thatthere are contaminants in groundwaterbeneath the Livermore site. (See the boxon top right.) But we can only see thesoil and water beneath the Laboratory in small, drilled samples of soil and inwater samples taken from monitorwells. Because we cannot take coresamples of the Laboratory’s entiresubsurface or cover the site withmonitor wells, there are large gaps inour data.

How have we determined the bestway to clean up this uncertainenvironment? And how long it willtake? As described in the box on thebottom right, groundwater modelingallows scientists to develop a picture ofthe workings of an otherwise invisiblesubsurface. Using computersimulations of the flow and movementof groundwater contaminants, we canevaluate the migration rates and pathsof contaminants in groundwater andsoil, assess potential health risks, selectextraction and injection well locations,and optimize hydraulic control andcontaminant removal in ourremediation designs.

Modeled simulations of groundwaterbehavior have served as an effectivedecision-support tool for Laboratoryscientists and engineers in our efforts toclean up contaminated groundwater atthe Livermore site as quickly andinexpensively as possible. Modelingprovides a means of rapidly retrieving

15


Groundwater Modeling

of a collection of wells (known as a wellfield) used to extract contaminatedgroundwater from the subsurface. Thisability is a key to finding an optimalremedial system design or designs.Conventional methods of formaloptimization can practically consideronly a few hundred simulations ofpossible well field configurations,whereas thousands or even millions ofpossible configurations are needed tofind the most effective designs. As aresult, the use of conventional formaloptimization methods seems impracticaland in need of a new approach.

New methods enable us to quicklyevaluate millions of prospectiveengineering designs and optimize

remediation pumping strategies. Thesemethods use artificial neural network(ANN) technology to process a muchsmaller set of simulations, repeatedly, forany and all configurations. ANNs, whosedevelopment was inspired by studies ofthe human brain, can be “trained” topredict the cost, extracted mass, andcontainment information that a modelsimulation normally generates. ANNspeed is remarkable—the technology canevaluate thousands of well configurationsper second. With ANNs to repeatedlypredict outcomes for a particular wellfield, a genetic algorithm, inspired byevolutionary concepts such as naturalselection, directs the search for the bestwell fields to meet remediation goals.

14



NO

RT

H

Volatile organic compound plume concentrations in parts per billion

5

10

100

500

10,000

Injection well

Extraction well

Irrigation well

Scale: Meters0 1000500

Initial condition After 10 years

After 50 yearsAfter 30 years

What Is a Model?

A conceptual groundwater model begins as a collection of information about the system’s material properties. That information then becomes a mathematicaldescription of an existing groundwater system, coded in a programming language,together with a quantification of the system’s boundary conditions, parameters, andinternal sources and sinks of groundwater and contaminants. Sometimes confusingly,the computer code used to simulate the groundwater system is also referred to as agroundwater model. So the word “model” is sometimes used for the description of the system under study and sometimes for the computer code that generated thedescription.

Groundwater modeling is a process of mathematically analyzing the mechanismsand controls of groundwater systems and the policies, actions, and designs that mayaffect these systems. Models help researchers understand subsurface fluid flow andfluid-related mass-transfer and transformation processes. They are also of use inanalyzing the responses of subsurface systems to variations in both existing andpotential new stresses, e.g., pumping out contaminated groundwater or returningtreated groundwater to the subsurface. Modeling is an effective tool for screeningalternative remediation technologies and strategies; the resulting “what-if” simulationsare helpful for finding the most efficient, cost-effective methods for remediatinggroundwater contamination. By looking at many variables in various combinations,modeling sometimes point the way to nonintuitive ways that complex systems respondto stress and may, thus, lead to new well field designs and remediation strategies.

Groundwater Contamination at the Laboratory

The contamination at the Livermore site consists of widely distributed plumes ofseveral volatile organic compounds (VOCs), tritium, fuel hydrocarbons, and someheavy metals. Some of this contamination has traveled in groundwater off site,especially on the south and west sides of the Livermore site. Contamination wasdiscovered in 1983, and Livermore was declared a Superfund site in 1987.*

The standard method for handling groundwater contamination is to extract thecontaminated groundwater and treat it. Known as “pump and treat,” this method isvery slow. Laboratory scientists have therefore seized the initiative to develop fast,inexpensive, innovative site restoration technologies that can also be appliedelsewhere and to demonstrate these technologies at a Superfund site. Restorationactivities at the Laboratory to date have centered on a smart pump-and-treatphilosophy that includes detailed characterization, validated modeling, phasedimplementation of remediation, directed extraction, and adaptive, time-managedpumping. The Laboratory has also used underground steam stripping, electricalheating, and vacuum extraction methods to treat the gasoline contamination. Othertechnologies being studied include abiotic, microbial, and thermal oxidation.

In late 1995, environmental regulatory agencies declared soil cleanup above thewater table to be complete at the site of an underground gasoline spill. This is the firstformal regulatory closure of a nonexcavation cleanup activity at the Livermore sitesince cleanup began in 1988.

* Site 300, an area east of the Livermore site that has been used since 1953 for testing non-nuclear,high-explosives compounds, was also declared a Superfund site in 1990. However, all discussionsin this article are related to groundwater modeling and remediation at the Livermore site.

Figure 1. PLANET, graphical userinterface software we have developed,can be used by groundwaterremediation planners at any site toexamine various contaminationextraction and injection scenarios.This two-dimensional simulation of apossible cleanup scenario for theLivermore site shows howconcentrations of volatile organiccompounds would decline over time.

wells, by dynamic steam stripping, andby soil vapor extraction above the watertable (Figure 2).

But even with all this characterizationinformation to apply to the conceptualmodel, data are still insufficient togenerate a highly resolved conceptualmodel for groundwater remediationbecause the subsurface is not uniform.For instance, a core sample taken at point A indicates soils with hydraulicconductivity, diffusion, and absorptionproperties that are very similar to those of a core sample taken at point B, 30 meters away. The soils between thosetwo points may or may not be similar.Soils have been laid down andrearranged unevenly over millennia, andtiny fingers and braids of different soiltypes intermingle. Each type of soil hasdifferent, nonuniform properties thatgovern how quickly or slowlycontaminants can travel through that soil.

A time-honored method of making up for missing data is to interpolate.Another method of filling in these datagaps is inverse modeling, a mathematicalmethod of backing into missing data,which is used in many scientificdisciplines where a lack of detailed datais a problem. Measured patterns ofgroundwater flow and contaminantstransmitted through the Earth’ssubsurface can be used to back-calculatesoil properties with a variety of inversesolution methods.

The conceptual model,which relies in part oninterpolated data, alsoincorporateshydrostratigraphicanalysis, a processdescribed in theJanuary/February 1996 issueof Science & Technology Review.1Hydrostratigraphic analysis integrateschemical, hydraulic, and geologic dataand includes exhaustive trend analysesperformed to produce a map of

subsurface connections that indicateswhere contaminants can and cannottravel. Hydrostratigraphic analysis wasborrowed from the oil and gas industrywhere it is used to determine the bestway to exploit underground reserves. Itis fairly new to groundwater modelingbut has proven to be an effective tool inthe modeling process.

This composite of actual andinterpolated or back-calculated data isused to create the conceptual model ina mathematical description, ormodeling code, that simulates what ishappening in the subsurface. But the

conceptual model is still not finalbecause many valid interpolations arepossible. So the next step is to test themodel and calibrate it. As new fielddata become available, they arecompared with the modeling codesimulations. If the simulations do notmatch reality, then the conceptualmodeling process continues untilsimulated and field data come intoagreement. This calibration part of themodeling process, known as circularmodeling, is essential to develop thebest picture possible of how thesubsurface behaves (Figure 3).

17



and analyzing a variety of data, and asmodeling methods have become moresophisticated, models have been thekey to refining the scope of the cleanupwork. Cleanup efforts can be targetedmore precisely, thereby reducing thescope of the project and saving bothtime and money.

For example, in 1987 when theLaboratory was placed on the Superfundlist, little modeling had yet been done.Our early simulations using a very simplemodel indicated that if we took no actionto remediate the contamination, volatileorganic compounds would spreadthroughout the Livermore basin in 500 to800 years and become sufficiently dilutedas to no longer be a problem. However,this “no action case” precluded manyuses of the groundwater in the Livermorebasin during that time period.Subsequent modeling efforts, which haveimproved in their physical realism overtime and are based on remediating thecontamination using the latest cleanuptechnologies, have reduced the timeframe considerably, as shown in Table 1.Today, using state-of-the-art, three-dimensional modeling, a risk-basedapproach to groundwater cleanup, andaccelerated cleanup of source regions, wecan greatly reduce the estimate forsuccessful remediation of contamination.

Conceptual Modeling

The first step in the modelingprocess is to develop a conceptualmodel, which is the initialrepresentation of the subsurface,including both the saturated andunsaturated groundwater zones. Thisconceptual model incorporates all fielddata and laboratory measurements—a huge quantity of material. AtLivermore, hundreds of monitor wellsboth on site and off provide informationabout groundwater elevations andcontaminant concentrations,distribution, motion, and naturaldegradation. Thousands of core samplesprovide geologic data. We haveextensive geophysical logging andseismic information, data ongeochemical properties and reactionsunderground, and information aboutquantities of contaminants removed bygroundwater pumping at extraction

16



Table 1. Estimated cleanup time decreases as the realism of our conceptual models improves*

Simple 2-D 3-D“Tank” Model CFEST Model CFEST Model

Date 1990 1992–1994 1995–1996

Time to MCLs 80 years 75 years Approximately 30 years(5 ppb)

Time to risk-based 35 years 40 years 10 to 15 yearsremediation (25 ppbassumed for this example)

*Applies to remediation of volatile organic compounds in distal contaminant plumes whose source ofcontamination has been controlled.

Notes:MCLs = Maximum Contaminant Levels. These are the maximum levels currently allowed by regulatory agencies.ppb = parts per billion. CFEST is a finite-element, flow-and-transport model used in the remediation industry.

Figure 2. This visualization of core sample data represents part of the first step indeveloping a conceptual model. Subsurface connectivities, which show howgroundwater moves, have not yet been developed in this step.

180

180

160

140

120100

80

60

160

140

120100

80

60

40 60 80

100 120 140 16020

20

20

20 4060 80 100

120

140

100

120

100

6040

100120

140

100

160

20

80

120

10080

80

6040

100120

140

20

80

Treatment Facility A Area

180180

NORTH

LLNL

LLNL

LLNL 2020

��

Figure 3. (a) The three-dimensional conceptual modelof the hydrostratigraphic unitsaround Treatment Facility A inthe southwest corner of theLivermore site shows theunderground streambedsthrough which groundwatermoves. The three-dimensionalmodel presents considerablymore information about thesubsurface than does (b) thetwo-dimensional model of theLivermore basin.

(a)

(b)

behavior of flow and contaminantmigration processes; remediationoptimization and risk analysis issues aremore visible; the importance ofdegradation and reaction processesismore evident; and cleanup times andcapture zones can be estimated.

But because our world has threedimensions, two-dimensional modelshave their limitations. In two-dimensional models, vertical variationsin contaminant concentrations and soilhydraulic properties are averaged out,potentially overestimating the timenecessary to clean up a site (Figure 5). Ingroundwater remediation, time equalsmoney, so overestimating time alsomeans running up the cost of the cleanupwith untargeted well field designs andperhaps wasting taxpayer dollars.

Three-dimensional conceptualmodels are traditionally time-consuming and costly to implementmanually in groundwater modelingcodes, but with new software tools, wehave made considerable advances intheir use. Because three-dimensionalmodels are more representative of thereal world, we can incorporate three-dimensional hydrostratigraphicrepresentations and more effectivelyassess cleanup strategies, costs, andimpacts. Wells can be targeted forcleanup of specific volatile organiccompounds (VOCs). Three-dimensionalmodeling can also be used to developtargeted implementation strategies forsuch new technologies as dynamicunderground stripping, abioticreduction, and microbial filters.

Laboratory Innovations

The Laboratory has initiated anumber of significant developments ingroundwater modeling. For example,we use a variety of simulation modeling

codes, each of which has its own“mesh” system for accepting suchconceptual model data as conductivity,thickness, and hydraulic head. Themesh frequently must be modified toimprove accuracy in simulations, tomodify the conceptual models, and toresolve remediation-stressed conditions.Manually translating, or “mapping,”conceptual model data to a particularmesh was a very tedious, time-consuming process and one that wasprone to error, even for two-dimensional models. With three-dimensional models, mapping ontomillions of mesh nodes would be almostimpossible without electronic tools.While some mapping tools did exist,they were designed to be used with aspecific modeling code and allowed forlittle geological complexity.

As a result, Lawrence Livermore isdeveloping MapIt, which can map allconceptual model data onto any meshso that it can be used by virtually any

19



Forecasting the Future

Groundwater modeling is very usefulfor estimating what will happen to thesubsurface in the future (Figure 4). Whatwill the contaminant plumes look likeand where will they be in 10, 30, or 100 years if we do nothing? If we domake efforts to remediate thecontamination, how will the undergroundenvironment respond to manmadechanges in the subsurface? What willhappen to the contaminant plumes whengroundwater is pumped out? What willhappen to the surrounding soils andgroundwater? Is it beneficial for thecleanup effort to return treatedgroundwater to the subsurface? Will fuelhydrocarbons degrade naturally withoutposing unacceptable risks to Laboratorypersonnel and the public? Where shouldadditional wells be located and whatshould be the rates of extraction andinjection? And most important: how longwill it take to clean up the groundwaterto meet regulatory requirements?

Modeled simulations with thecalibrated conceptual model can be veryeffective in helping to answer thesequestions. The more accurate thecalibrated conceptual model, the moreaccurate the model forecast simulationswill be.

From 1-D Models to 3-D

Simple questions about thesubsurface can often be answered using dimensionless or one-dimensionalconceptual models. Two-dimensionalconceptual models, on the other hand,provide a better framework fororganizing and relating geologic,hydrogeologic, and chemicalinformation. Two-dimensional modelsgive scientists, the public, and regulatorsa more realistic picture of the general

18



Figure 4. We used the three-dimensional CFESTmodeling code to forecastthe decline in concentrationsof perchloroethylene (PCE),a volatile organic compound,beneath Treatment Facility A(TFA) in the southwestcorner of the Livermore site.The decline shown in thesimulations at the right forthe years 1995, 2009, and2026 assumes that thesource of contamination forthis distal contaminant plumehas been controlled and that“smart” pump-and-treatcleanup methods continuefor the duration. The PCEmass removed in these threedimensional simulationsagrees closely withmeasured PCE removal.

12

8

Cum

ulat

ive

tota

l V

OC

mas

s re

mov

ed, k

g

4

0Jan90

Jul90

Jan91

Jul91

Jan92

Jul92

Jan93

Jul93

Jan94

Jul94

Jan95

Jul95

Cumulative total VOCmass removed, kg

Two-dimensionalCFEST results

Month

Figure 5. A two-dimensional simulation usingthe CFEST modeling code underestimatedthe total volatile organic compound (VOC)mass that would be removed at TreatmentFacility B, which is on the west side of theLivermore site. The three-dimensionalsimulation predicted that by April 1995, wewould have removed approximately 3.25 kg ofVOC mass whereas we had in fact removedover 10 kg from the actual three-dimensionalsubsurface.

Modeling on the Internet

To facilitate widespread access toour remediation data, the Laboratory’sEnvironmental Restoration Divisionnow has a home page on the WorldWide Web that provides authorizedusers access to a wealth of accumulateddata. Not only can users view anddownload static documents, images,and product and technology overviews,but they also have access to up-to-dateproject status information, statisticalprocessing capabilities, databaseinformation, and estimating tools. Attheir desktop computer, they can view

HotMap, which provides a variety ofinformation on any well or combinationof wells at the Livermore site, or theycan use PLANET, MapIt, or PDEase, apartial differential equation solver forinverse modeling (Figure 8). For thefirst time, Laboratory scientists andfederal and state regulators have quick,timely access to data in a form that isuseful to them.

Looking Ahead

Three-dimensional groundwatermodeling is proving its usefulness on aregular basis, and it is still in its infancy.

21



modeling code (Figure 6). Twomodeling codes that use different meshsystems can now be used forsimulations using the same conceptualmodel data, and they can then becompared in more meaningful ways.MapIt includes a “feature” databasethat provides for considerable geologiccomplexity. MapIt is easy to use andallows the user to view and manipulatethe features in the conceptual model aswell as the graphical representation ofthe simulation model. Another benefitof MapIt is its use of an electronicallyencoded, time-stamped conceptualmodel, which encourages the use ofconsistent conceptual models across allmodeling efforts.

We have also developed PLANET(Pump Layout and Evaluation Tool), aninteractive software package that givesenvironmental remediation planners atany site the ability to quickly evaluate alarge number of remediation scenariosas part of an overall strategy ofhydraulic management. PLANETprovides a simple, site-map-orientedinterface to industry-standard flow-and-transport modeling codes. WithinPLANET, a series of chemical transport

simulations can be interactivelydesigned and displayed

to examine the

migration of existing contaminantdistributions in the saturated zone,under natural conditions, as well asunder various proposed remediationscenarios. PLANET can be adapted foruse with either two-dimensional orthree-dimensional modeling codes.

Artificial neural networks allow us tomake use of optimization techniquesthat were not possible in the pastbecause the evaluation of thousands ofalternatives was so time consuming.ANNs’ ability to evaluate thousands ofdesigns per second compares to the 3 to4 hours that a modeler normallyrequires to simulate a single design at acommon Unix workstation not equippedwith ANNs. In a small-scale trial of theANN approach, we analyzed 28 potential extraction and injectionwell locations that had been selected byhydrogeologists as capable ofcontaining and cleaning up thegroundwater contamination atLivermore within 50 years (Figure 7).

We were trying to identify the leastexpensive subset of these 28 locationsthat would prevent the spread ofcontaminants and remove as much asthe full 28-location strategy. After usingANNs to evaluate 4 million of over 268 million possible subsets, we foundacceptable strategies that involved asfew as 8 to 13 locations and yet metcontainment and contaminant-removalgoals. Installation, maintenance, andtreatment costs associated with thesestrategies were less than 35% of thosefor the 28-location strategy. We nowapply the ANN approach to well fieldscontaining up to 268 potential locationsand search for the optimal answers to awide variety of management questions.

20



Figure 6. The huge number ofnodes on the mesh for the three-dimensional conceptual model ofthe Livermore basin illustrates theusefulness of MapIt for mappingconceptual model data to a mesh.Mapping soil conductivity,thickness, hydraulic head, andother data by hand to those nodeswould take months if it were donemanually. With MapIt, a newthree-dimensionalconceptualization of asite can be createdin just a few hours.

NO

RT

H

Gre

envi

lle R

oad

Vas

co R

oad

Patterson Pass Road

Extraction well

Monitor well

Injection well

Contaminant plume

Arroyo

LLNL boundary

Scale: Meters

0 500 1000East Ave.

LLNL

10

10

10

50

100

50

500

250

150

10

10

10

50

100

50

500

250

150

Number of locations

Contaminant contained?

Percentage of contaminant removed

Cost relative to baseline

8 9

Yes Yes

97% 98%

26% 31%

13

Yes

105%

34%

10

1010101010

1010

1010

1010

100100 100100

5050

5050 10101010

1010

1001005050

(a) (b) (c)

(a) (b) (c)

Figure 7. We use artificial neuralnetworks (ANNs) and genetic algorithmsto greatly accelerate the identification ofoptimal groundwater cleanup strategies.The larger figure represents the initialdistribution of volatile organic compounds(VOCs) and the locations of 28 extractionand injection wells. The smaller threefigures are the top-ranked patterns foundafter evaluating 4 million designs. Thecontours show VOC concentrationsremaining after 50 years. Percentages inthe table use the cost and performance ofthe full 28-well pattern as a point ofreference. These evaluations are basedon two-dimensional simulations ofgroundwater behavior.

Advancing computer technology andour continued creativity are key to thefurther advancement of groundwatermodeling at the Laboratory for beneficialuse everywhere. It is interesting to notethat the personal computers that can be

bought off the shelf today are aspowerful as the Cray 1 supercomputersof less than 10 years ago. As computersbecome ever more powerful, ourmodeling capabilities can only expandfor effective global applications.

Key Words: artificial neural network(ANN), genetic algorithm, groundwatercontamination, groundwater remediation,hydrostratigraphic analysis, inversemodeling, volatile organic compound(VOC).

References1. “Groundwater Cleanup Using

Hydrostratigraphic Analysis,” Science &Technology Review, UCRL-52000-96-1/2 (January/February 1996), pp. 6-15.

For further information contact Robert J. Gelinas (510) 423-2267([email protected]).Information on flow and transportmodeling is also available from Peter McKereghan of Weiss Associates(510) 450-6151. Virginia Johnson (510)423-2005 ([email protected]) is aLaboratory expert on ANNs and geneticalgorithm optimization.

22



ROBERT J. GELINAS joined the Laboratory’s TheoreticalPhysics Division in 1966 as an applied physicist. He received hisPh.D. (1965) and an M.S.E. (1961) in Nuclear Engineering and aB.S.E. (1960) in Chemical Engineering from the University ofMichigan. He is currently the Environmental Transport GroupLeader in the Environmental Restoration Division of theEnvironmental Protection Department. His experience includesleading groups of scientists and engineers in energy, defense,

environmental, and high-power laser projects at both LLNL and in commercialR&D, where he was principal scientist and manager of the Science ApplicationsInternational Corporation facility in Pleasanton, California, from 1975 to 1985.Gelinas has published extensively in the fields of reactive fluid flow and transport,with applications to radiative weapons systems, atmospheric and subsurfaceenvironments, high-average-power and Nova-class lasers, hydrocarbons, nuclearsystems, and nuclear weapons effects.

Figure 8. From theHotMap menu, anauthorized user islinked to an array ofinformation aboutthe Livermore site.

About the Scientist

23

Dual-Band Infrared ComputedTomography: Searching forHidden Defects

Dual-Band Infrared ComputedTomography: Searching forHidden Defects

(a) (b) (c) (d)

Dual-band infrared computed tomography systemsdeveloped at Lawrence Livermore are providing highlysensitive and accurate three-dimensional nondestructiveinspection and evaluation of manmade structures in avariety of applications inside and outside the Laboratory.

(Above) The dual-band infrared (DBIR) laboratory that theFederal Highway Administration is currently using for bridgeinspection is a converted mobile home. The mobile DBIRlaboratory’s cameras, mounted about 4 meters (13 ft) above theroadway, scan the reinforced-concrete bridge deck for defectscalled delaminations. (Below) Delaminations are seen at thecenter of (a) the 8- to 12-micrometer (µm) longwave thermalinfrared image and (b) the 3- to 5-µm shortwave image, both ofwhich also show clutter. Clutter is identified in (c) the spectraldifference map and later removed from (a) and (b) to create (d), which clearly shows only the delamination (minus clutter) asthe bright yellow area with anomalous heat flow at the center ofthis image where temperatures are about 2°C warmer at noonand 0.4°C cooler at midnight.


ICTURESQUE bridges such as the recently famous ones inMadison County, Iowa, have secured a fond place in Americanhearts. But the fact is that many of them and their less attractive

fellows are badly in need of repair. Federal highway officialsestimate that 20% of the country’s half-million two-lane bridges arestructurally deficient.

Beginning in April 1996 and continuing through the summer, aconverted motor home will roam the highways in several states,testing a system developed at Lawrence Livermore that can pinpointthe flaws in these well-traversed and rapidly aging bridges. Fundedby the Federal Highway Administration (FHWA), this dual-bandinfrared (DBIR) system uses a technique known as dual-bandinfrared computed tomography (DBIR-CT), which locates defects inmaterials by sensing time-dependent temperature differences. (Seethe box on p. 25 and the images below.) Our system promises tomake bridge inspections more reliable, faster, and safer. It also has awide range of nondestructive inspection and evaluation applications,including unmasking metal corrosion in aircraft skins, assessingstructural damage in reinforced concrete buildings, analyzing theintegrity of containers of radioactive waste, and identifying corrosionin exposed petrochemical pipelines.

P

25


Dual-Band Infrared Computed Tomography

Finding flaws in the bridges’ concreteand asphalt roadbeds and repairing themis expensive—an estimated $3 billionannually. In addition to the dollar outlay,there is the emotional cost: theaggravation that motorists endure whilehighway crews go about the slow,meticulous task of pinpointing problems.Currently, bridge crews must close lanesto traffic while they conduct visualinspections—looking for potholes, ruststains, cracks, or broken pavement. Todetect hidden flaws, they manually drag achain across the deck, listening forauditory differences that indicate areas ofpossible cracking—it is a little likerapping your knuckles along drywall,hunting for a secure anchor point. Thechain technique, however, is slow,disruptive, unreliable, and costly andraises serious safety concerns. Also, itworks only with concrete, not asphalt-covered roadways.

Road-Testing the Technology

The Laboratory’s DBIR systemprovides reliable bridge inspectionswhile minimizing lane closures. It isdesigned to eliminate the need to shutdown lanes for routine inspections orhave road crews risk possible injury asthey dodge traffic. With the mobile unit,highway crews can conduct theirexaminations as they ride across abridge. The Laboratory’s system alsowill let inspectors peer beneath aroadbed’s surface to hidden troublespots before they get out of hand. Forexample, if defects like rebar corrosionare identified before they becomeserious, maintenance crews couldperform less costly repairs.Delaminations (concrete layerseparations beneath the surface) can bediscovered before they lead to possiblyhazardous potholes. The FHWA teamwill test the prototype DBIR system thissummer by using it to examine bridges

and then compare the data collectedwith the defects road crews actuallyuncover and repair.

Steve Chase, the FHWA engineer incharge of nondestructive evaluationtechnology, calls the system, whichwas developed at the Laboratory, “anevolutionary development in bridgedeck inspection.” While traditionalinfrared imaging systems work at onlyone wavelength, the Laboratory’ssystem collects images at two separateranges of wavelengths (3 to 5 and 8 to 12 micrometers), allowing greaterprecision in computer calculations oftemperatures. The dual-band systemcompensates for the influence ofsurface contamination on materials andsurface compositional differences, bothof which can skew readings intraditional, single-band systems.

The bridge inspection DBIR systemtaps into the Laboratory’s well-honedexpertise in nondestructive inspectionand evaluation, developed over theyears in support of the nation’s nuclearweapons program. The system alsocapitalizes on ideas proposed by theLaboratory for detection of buried landmines during the Gulf War.

In fiscal year 1994, the Laboratorybegan adapting the DBIR technology to meet the Federal HighwayAdministration’s needs. Feasibility testswere conducted with tower-mountedinfrared cameras, which overlookedasphalt-covered and exposed concreteslabs that served as a surrogate bridgedeck. (See the illustrations in the box on p. 25.)

The feasibility tests were used tooptimize the DBIR system response tothermal differences between normal anddefective concrete structures and toclarify interpretation of corrosion insteel reinforcements that causes hiddendelaminations. These hidden flaws aretypically masked by clutter from oil,grease, paint, patches, shadows, rocks,

wood, plastic, metal, or concretecomposition variations.

With good results from the feasibilitystudy, the Livermore team last yearmounted DBIR cameras on a telescopingmast located at the front of a motor homethat had been converted for field tests.The mobile DBIR Bridge DeckLaboratory successfully completed itsfirst road test in November 1995 on theGrass Valley Creek Bridge near theNorthern California town of Weaverville.Scientists were able to view 1-meter-longsections of roadbed approximately1 meter (3 ft) long and one lane (3 metersor 10 ft) wide while traveling at 40 kilometers per hour (25 mph). VIEWcomputer codes developed at Livermoresped processing and analysis of theimages, which had been recorded andstored on a high-speed hard disk.

Livermore researchers are awaitingresults of the spring and summer 1996 shakedown tests by the HighwayAdministration before planning thenext phase of their DBIR activities.The FHWA’s Chase is optimisticabout the technology’s future: “If itproves to be a valuable technology,and I think it will,” he says, “there’spotential for commercialization by aprivate company.”

Aging Aircraft Applications

According to the Laboratory’s Nancy Del Grande, a principal scientistfor DBIR inspection capabilitiesdevelopment, the technology hasnumerous applications in addition toFHWA bridge inspections. Thetechnology has applications inLaboratory programs—identifyingthermal stress and damage to opticalcomponents used for the NationalIgnition Facility, measuring emissivityand radiative heat transfer for StockpileSurveillance, and characterizing thethermal efficiency of the uranium

24



The use of dual-band infrared computed tomography (DBIR-CT) imaging as an inspection tool is based on the knowledge thatflawed and corroded areas of a manmade material or structureheat and cool differently than do areas with no defects. DBIR-CTuses two thermal infrared bands to provide time sequences ofhigh-contrast images called temperature maps of naturally orflash-heated materials and structures. When these maps areprocessed using computer codes developed at LLNL, they unfoldthe location and amount of hidden defect and corrosion damage.DBIR-CT enables researchers to analyze heat flow patterns at 10 times the sensitivity of single-band systems, to imagestructural defects in three dimensions, to differentiate surface andsubsurface clutter from corrosion damage, and to quantify thatdamage so that major damage in need of immediate attention canbe differentiated from minor defects.

The DBIR inspection and evaluation process works like this(see the figure below): First the material or structure is heatedeither naturally by the sun or artificially using pulsed lasers,quartz lamps, or flash lamps. Bridge decks, for example, arenaturally heated; airplane fuselages are heated by flash lamps.The material or structure is scanned by the DBIR-CT systemusing two ranges of infrared wavelengths—3 to 5 and 8 to 12 micrometers. The resulting time-sequence temperature maps

Dual-Band Infrared Computed Tomography: How It Works(Figures c–1 and c–2 below) at each wavelength show heat-flowanomalies that could be caused by defects or corrosion or byclutter such as paint or junk metal on the road surface (in ourexample below) or subsurface globs of excess epoxy within anairplane skin.

To weed out clutter, the temperature maps are processed usingthe computed tomography capabilities provided by the VIEWcomputer code developed at LLNL. The resulting DBIR image-ratio patterns on the temperature map (Figure c–3) showingtemperature variations from both surface-only features and fromsubsurface defects are compared to DBIR image-ratio patternsrelated to “emissivity noise” (Figure c–4) that show only surfaceclutter. The clutter from unseen concrete chemical differencesstands out clearly on the emissivity-noise map, and when theemissivity-noise patterns are subtracted from the temperature mapthat shows both emissivity noise and defects, the resultingtemperature map (Figure c–5) shows only the flaws—theirlocation, size, shape, depth, and severity.

The DBIR system’s capabilities shown in the results oflaboratory tests (see the figures below) have been confirmed byfield tests (see the images bottom of p. 23) and are being put touse in further highway tests by the Federal HighwayAdministration that began in April 1996.

A

D

E

C

B5

4

1

2

3

1.8

m

1.8 m

– =

(c–3) Temperature mapshowing surface clutter andsubsurface defects.

(c–4) Emissivity-noise mapshowing surface clutter only.

(c–1) 8- to 12-µm long-wavelength temperature map.

(a) 1.8-m (6-ft) square concrete test slab 18.4 cm (7.5 in.) thick with(b) five surface clutter sites (A–E) along the perimeter and fiveStyrofoam delamination sites (1–5) 5.1 cm below the surface.

(c–2) 3- to 5-µm short-wavelength temperature map.

(c–5) Temperature map showingsubsurface defects only.

A

B

C

D

E

techniques to look for fuselage defects.However, in order to determine theextent of corrosion damage, thefuselage must be taken apart.

Livermore’s DBIR detectionmethods do not require aircraftdisassembly. When applied to aircraftstudies, DBIR can show the extent ofmetal corrosion near the aircraft skin’ssurface or deep within it. (Figure 1.) Itcan characterize the type of defectsinvolved—such as gaps or areas withpoor adhesive bonds with or withoutmetal loss from corrosion. Also, it candifferentiate between corrosion andconditions that may be mistaken forcorrosion, such as fabrication ripples,surface roughness, and uneven sealantin a lap joint.

While the DBIR system for bridgeinspections relies on natural heatsources, the system for aircraftinspections uses flashlamps and thusrelies more on the computedtomography (CT) aspects of thetechnology than do DBIR bridgeinspections. The metal skin is heatedwith uniform thermal pulses eachlasting a few milliseconds. Theresulting surface temperature changes,which vary with location and time, arethen imaged by the dual-band scannerfor analysis by computer. Hotterreadings indicate areas of potentialcorrosion. A patent on the Laboratory’sactive (flash-heated) and passive(diurnally-heated) DBIR processes forimaging anomalous structural heatflows was issued in August 1995.

Livermore researchers haveconducted several DBIR-CT tests oncommercial aircraft. A demonstration atBoeing in early 1995 used a uniformpulsed-heat source to stimulate infraredimages of hidden defects in an aircraftfuselage. The DBIR camera and imageprocessing system produced

temperature, thermal inertia, andcooling-rate maps. In combination,these maps characterized the defect site,size, depth, thickness, and type. LLNLresearchers are able to quantify thepercent metal loss from corrosion abovea threshold of 5%, with overalluncertainties of 3%.

The Laboratory team’s goal is toproduce a single corrosion defect mapthat eliminates clutter from excesssealants, ripples, and surface features.Such a map could be incorporated in acommercial DBIR scanner, making iteasier to assess damage. The Laboratoryhas worked out a Cooperative Researchand Development Agreement withBales Scientific Incorporated tocommercialize DBIR corrosioninspection technology.

A recent round-robin investigation ofnondestructive investigation equipmentused to detect hidden corrosion on U.S.

Air Force aircraft indicated that falsedetection of corrosion results in costly, unnecessary, and destructiveexploratory maintenance, said Del Grande. “We expect our technologyto cut the cost of destructive exploratorymaintenance in half by eliminatingclutter,” she said.

Key Words: dual-band infrared computedtomography, nondestructive inspection andevaluation.

For further information contact Nancy Del Grande (510) 422-1010([email protected]). Other key Laboratory researchers on theDBIR projects are Phil Durbin, ProjectManager, Federal Highway Administrationproject (510) 422-7940 ([email protected])and Ken Dolan, project manager, FederalAviation Administration project (510) 422-7971 ([email protected]).

27



spin-forming process for our AdvancedDevelopment and ProductionTechnology. Applications in areasoutside Lawrence Livermore programsinclude detecting corrosive thinning andpitting within exposed petrochemicalpipelines, assessing structural damage inreinforced concrete buildings, andanalyzing the steel wall thickness and

integrity of radioactive waste containers.Since early 1992, Del Grande and herassociates have developed DBIR toquantify metal corrosion in aircraftfuselages, with funding from the Federal Aviation Administration’sTechnical Center.

Currently, aircraft inspectors usevisual, ultrasound, and electronic

26



Figure 1. (a) Phil Durbin uses theDBIR system scanner, flashlamp,and spectral hood to look for metalloss from corrosion within the skin ofan airplane. (b) What the DBIRsystem sees over time afterflashlamp heating. Early- and late-time thermal inertia, temperature,and cooling-rate maps distinguish,through the power of computedtomography, subsurface clutter(shallow sealant excess) from deepcorrosion (metal loss) in lappedmetal splices.

0 to 5% metal loss

40 to 50% metal loss

(b–1) Airplane skintemperature mapsshowing defects

and clutter.

(b–2) Timegramcooling-rate maps

of defects andclutter.

(b–3) Early-timethermal inertia maps

showing shallowclutter in red.

(b–4) Later-timethermal inertial

maps showing deepdefects in red.

Shallow sealant excess Deep corrosionin lap splice

Temperature maps (left to right) compare (b–1) airplane skin temperatures at two locations, (b–2) timegram cooling rates atthose same locations, (b–3) early-time thermal inertias (0.1 to 0.5 seconds) after flash heating at the locations, and (b–4)late-time thermal inertias (1 to 3 seconds after heating) for (top row) noncorroded or barely corroded (0 to 5% metal loss)aircraft skins and (bottom row) heavily corroded (40 to 50% metal loss).

NANCY DEL GRANDE (née Kerr) received her A.B. in physicsfrom Mount Holyoke College in 1955 and her M.S. in physicsfrom Stanford University in 1957. She joined the Laboratory’sTest Program in 1959. She became the first woman to design andconduct an experiment using x-ray spectroscopy to measure thetemperatures of nuclear devices stored underground at theNevada Test Site. She pioneered the technology transfer of x-raytemperature methods for nuclear device diagnostics to the dual-band infrared (DBIR) precise airborne temperature measurement

method for detecting underground objects and applied it to depict deep aquifers atthe Long Valley, California, geothermal resource area and to locate buried landmines at the Yuma Proving Grounds.

In 1992, she transferred to the Nondestructive Evaluation Section of theLaboratory’s Engineering Sciences Division where she has been principal investigatorand principal scientist for DBIR imaging projects to detect corrosion in aging aircraft(1992–1995) and delaminations in bridge decks (1993–present). She is the author orco-author of over 50 publications on x-ray spectroscopy and infrared physics and hasbeen issued two patents, one for a technology to identify anomalous terrestrial heatflows from buried and obscured objects, and another a system to image anomalousstructural heat flows from corrosion within aircraft skins and bridge decks.

About the Scientist

(a)

(b)

29

Science & Technology

Waste Minimization

system, the facility now recycles spent solutions of nitric,sulfuric, and hydrochloric acids with a recovery efficiency ofabout 80%.

In addition, the facility is now using ferric sulfate materialto clean aluminum parts as a substitute for a process that usedconcentrated acids. The new solution lasts much longer anddoes not cause a problematic insoluble aluminum fluoride filmthat formed with the acid process.

Also, the facility substituted oxalic acid for sulfuric acidwhen anodizing aluminum. Because oxalic formation is muchless corrosive than sulfuric acid, equipment, floors, and tankslast much longer. As an added bonus, anodized aluminumcoatings with oxalic acid are much smoother.

Another approach to minimize waste has been to extend thelife of materials. A case in point is electroless nickel solutions,widely used to coat a variety of surfaces by immersion in a bathcontaining a chemical reducing agent. The solutions have atendency to spontaneously plate out on the tank and associatedequipment, adding to the plating cost. Also, expensive treatmentis required before the spent solution can be disposed of. Inresponse, Steffani and his technicians incorporated anelectrodialysis process that reduces both chemical purchases anddisposal costs. The technology separates dissolved solids fromthe nickel ions, making it possible to reuse the tanks many timeswithout disposal. The electrodialysis operation is simple andruns virtually unattended.

Substituting New TechnologiesOne of the most important materials substitutions was made

for hexavalent chromium plating, until recently the standardchromium plating process. This process generates airemissions, effluent rinse water, and solutions that are toxic andsuspected of being carcinogens. Facility people estimated thatit would cost about $25,000 to demonstrate compliance with

Material discharged to municipal sewers must meet strictguidelines, and wastes that must be trucked off site cost atleast 75¢ per liter ($3 per gallon) for transport, treatment, anddisposal. Standard degreasing equipment was replaced with avariety of aqueous cleaning processes including soak andelectrolytic cleaning and ultrasonic cleaning (mentionedabove). These were connected to a rinse water recyclingsystem capable of recycling 2,840 liters (750 gallons) a day.The recycled water is further purified by an ion exchange unitfor operations other than simple rinsing.

As a result of these actions, the facility is saving over11.4 million liters (3 million gallons) of water per year and nolonger sends any water to the City of Livermore’s sewersystem. What’s more, the shop’s recycled water has fewermetallic ions than the water supplied to city residents.Nominated by the City of Livermore, the Laboratory receiveda 1993 award in recognition of these accomplishments fromthe California Water Pollution Control Association.

Dealing with AcidsMinimizing the use and waste of inorganic acids has been

one of the facility’s largest concerns. Such acids arecommonly used in shops to remove surface blemishes,produce bright surfaces, strip unwanted metals and coatings,and prepare metal surfaces to receive other coatings. Theacids are also used as electrolytes for coatings produced byelectrolytic oxidation such as anodizing. Eventually the acidsbecome unusable or weakened because of contamination withmetals or the conversion of hydrogen ions into hydrogen gas.The acid then becomes waste and a new acid batch is started.

The facility installed diffusion dialysis equipment torecover acids and separate the metal component for recoveryand sale to refineries. The equipment uses a new kind ofmembrane that can withstand low pH solutions. Using this

Kelye Allen of Livermore’s platingshop plates (left) a metal bellowscomponent to be used in theLaboratory’s lasers program with an environmentally friendlynickel–tungsten–boron deposit,which replaces hexavalentchromium, a plating material that is a source of toxic and possiblycarcinogenic air and waterpollutants. On the right, she rinsesoff excess nickel–tungsten–boronmaterial with water that is treatedand reused.

28


One of the two dozenradiofrequency cavitiescoated with copper by the Laboratory’s Chemical andElectrochemical Processes Facility as part of our work on theStanford Linear Accelerator Center’s “B-factory.” The cavitiesare 60 cm in diameter and 35.6 cm tall.

HEN physicists at Stanford Linear Accelerator Center’s“B-factory” last year needed to coat two dozen 35.6-cm-

tall cavities with pure copper and ultraclean some 50,000 otherparts for a set of unprecedented high-energy experiments, theyturned to specialists at the Lawrence Livermore Chemical andElectrochemical Processes Facility.

The facility is operated by the Metal Finishing Group ofMechanical Engineering’s Manufacturing and MaterialsEngineering Division. Over the past three decades, the facilityhas earned a reputation as one of the top-flight metal finishingcenters in the nation, using chemical and electrochemicalprocesses for wide-ranging assignments from LLNL researchprograms, other national laboratories, and internationalresearch agencies such as CERN (the European high-energyphysics research organization).

During the past few years, the facility has also made aname for itself by embracing environmentally consciousmanufacturing principles. Working in stages, the shop hasadopted scores of improvements that included recycling strongacids, substituting Earth-friendlier materials, and eliminatingcyanide in its operations. The advances have been made aspart of a larger, Laboratory-wide effort to encourage pollutionprevention and waste minimization activities.

The operation’s environmental efforts have been sosuccessful that the facility has decreased its discharge of waterto the Livermore sewer system from 11.4 million liters(3 million gallons) to zero! And whereas in 1991 the facilitywas producing 227,100 liters (60,000 gallons) of chemicalwaste to be trucked off site, it now produces only 3,785 liters(1,000 gallons) of this waste annually, a dramatic reduction.

As a result of these accomplishments, plating shoppersonnel are sharing lessons learned with both Department ofEnergy centers and private firms on how to respond totightened environmental regulations and waste disposal costsby minimizing wastes and substituting better procedures.

The products of chemical and electrochemical finishing arefound throughout modern society, from galvanized nails toshiny chrome-plated faucets, from automobile trim to the goldand silver electrical contacts in computers and telephones.Such finished parts can be stripped and replated if they tarnish,scratch, or wear out. However, heavy metals, toxic chemicals,and large volumes of waste (much of it hazardous) for manyyears seemed an inescapable part of the metal finishingbusiness at the Laboratory and elsewhere.

But that assumption changed for Livermore in the late1980s, according to Jack Dini, Materials Engineering andMechanics section leader, when the Chemical andElectrochemical Processes Facility began to scrutinize itsoperations with an eye on minimizing waste and preventingpollution. The in-depth look was prompted by stifferregulations, higher costs to dispose of wastes, a growingenvironmental consciousness, and the appearance of newtechniques in the marketplace that used fewer hazardousmaterials and produced less hazardous wastes. The result wasmodification of many operations over several years, but mostimportantly, says Dini, the adoption of a different mindsetcentered on incorporating environmental consciousness inevery facet of the facility’s operations.

“Today we have a safer and cleaner facility, produce muchless pollution, and have maintained the quality of partsprocessed through our operation,” Dini says. “Being in theforefront of both metal finishing technology and wasteminimization is not only possible, but they go hand in hand.”

What’s more, the environmental changes have producedsizable cost savings. Today the facility is saving more than$500,000 per year. According to Chris Steffani, shopsupervisor, many of the savings resulted from relativelysimple material substitutions. For example, costly chlorinatedcleaning compounds were replaced with much cheaper, andmore environmentally kind, materials. For more thoroughcleaning, high frequency sound waves remove very fineparticulates, and the residual water is recycled for anotherround of cleaning chores.

Because wastewater typically represents the largest wastestream, shop personnel have given it very high priority.

Research Highlight

Plating Shop Moves to Finish Off WastePlating Shop Moves to Finish Off Waste

W

31


30


tougher Bay Area Air Quality Management Districtregulations on emissions from hexavalent chromium and thatcompliance costs would mount in future years.

As a result, the facility adopted a compound composed of59.5% nickel, 39.5% tungsten, and 1% boron as a substitutefor hexavalent chromium. The new material has excellentwear, corrosion resistance, and mechanical properties. It alsoposes less of an environmental risk and reduces energycosts.(See the photos on p. 29.)

Much of the effort to adopt more environmentally friendlyprocedures and materials has centered on cyanide. Because ofthe hazardous nature of cyanide, extensive safety precautionsmust be incorporated when manufacturing the electroplatingchemicals, transporting them to user sites, using theelectroplating process, and disposing of wastes. For example,if cyanide-based solutions become too acidic, large amounts ofpoisonous cyanide gas are liberated. The electroplatingindustry as a whole has suffered many accidents—and a fewdeaths—from cyanide use. In addition, cyanide use costsmore—disposal costs for LLNL wastes containing cyanidesare about $1.50 per liter ($6 per gallon) compared to75¢ per liter ($3 per gallon) for noncyanide wastes.

Over the past several years, the Livermore facility foundsubstitutes for almost every process involving cyanide. Forexample, copper cyanide plating was replaced with a copperpyrophosphate process. The substitution works quite well andproduces parts as good as those obtained with the coppercyanide process.

Until recently, silver was the remaining metal still relyingon cyanide formulations for electrodeposition of lasting andrelatively thick deposits. In fact, silver–cyanide platingsolutions are the most highly concentrated cyanide solutionsused in the plating industry.

In the spring of 1995, LLNL entered into a $2-millionCooperative Research and Development Agreement(CRADA) with Technic Inc. of Providence, Rhode Island,with the goal of providing industry with an environmentallybenign alternative to the silver–cyanide plating process fordepositing thicknesses greater than 250 micrometers.Laboratory researchers have obtained stress and hardness dataand have analyzed the structure of the deposit withmetallography and x-ray diffraction. The results show that theelectroplating industry for the first time can confidently platesilver with a process that uses no cyanide.

The CRADA effort has been so successful that theLaboratory has nominated our partner for a President’s GreenChemistry Challenge Award, a program sponsored by avoluntary partnership of industry, the American ChemicalSociety, the Council for Chemical Research, and the U.S.

Environmental Protection Agency. This awards programrecognizes fundamental breakthroughs in chemistry that areuseful to industry and accomplish pollution prevention throughsource reduction.

Helping Small FirmsAnother successful collaboration is with the Northern

California Association of Metal Finishers. This partnership,sponsored by the LLNL’s Small Business Initiative,established a Model Metal Finishing Facility at theLaboratory’s Chemical and Electrochemical Processes Shop toassist regional businesses in acquiring and implementingchemical processing technology and providing wasteminimization consultation.

Steffani handles calls daily from small metal finishers withquestions on minimizing waste or about technical processes.Many calls result in immediate solutions. Other calls involve arequest to tour the Laboratory facility or to use the modelfacility for a few days to work on a new process. Steffani hastoured more than 70 small electroplating facilities and helpedmany of them set up waste minimization programs of their own.

A new partnership is with the U.S. EnvironmentalProtection Agency’s Common Sense Initiatives, a programaimed at enhancing interaction between business and the EPA.In a $300,000 project supported jointly by EPA and LLNL’sSmall Business Office, the Chemical and ElectrochemicalProcesses Shop is selecting four to five projects aimed athelping small electroplating businesses establish wasteminimization and pollution prevention programs. Dini notesthat a third of the nation’s small electroplating firms have goneout of business in recent years largely because of the burden ofmeeting environmental regulations.

To date, the shop’s largest partnership is with StanfordLinear Accelerator Center (SLAC), Lawrence BerkeleyNational Laboratory, and other LLNL programs to design andbuild the so-called “B-factory” (named after the elusive B-meson subatomic particle) at SLAC. (See the photo on p. 28.) The shop is receiving $750,000 for ultracleaning some50,000 parts that make up 1,400 ion pumps and for coppercoating 300 meters of connecting pipe and two dozen, 60-cm-diameter, 35.6-cm-tall radio-frequency cavities that will keepelectrons traveling at nearly the speed of light.

Key Words: metal finishing—chemical and electrochemicalprocesses, electroplating; pollution prevention; waste minimization.

For further information contact Jack Dini (510) 422-8342 ([email protected]) or Chris Steffani (510) 423-1780 ([email protected]).

Waste MinimizationEach month in this space we report on the patents issued to and/orthe awards received by Laboratory employees. Our goal is toshowcase the distinguished scientific and technical achievements ofour employees as well as to indicate the scale and scope of the workdone at the Laboratory.

Patents

Patent issued to

Daniel D. DietrichRobert F. Keville

Thomas E. McEwan

Ger Van den EnghRichard J. Esposito

Robert J. ContoliniSteven T. MayerLisa A. Tarte

Troy W. Barbee, Jr.Gary W. JohnsonDennis W. O’Brien

Margaret L. CarmanKenneth J. JacksonRichard B. KnappJohn P. KnezovichNilesh N. ShahRobert T. Taylor

Anthony M. McCarthy

Patent title, number, and date of issue

Electron Source for a Mini Ion TrapMass Spectrometer

U.S. Patent 5,477,046December 19, 1995

High Speed Sampler andDemultiplexer

U.S. Patent 5,479,120,December 26, 1995

Multiple Sort Flow Cytometer

U.S. Patent 5,483,469January 9, 1996

Removal of Field and Embedded Metalby Spin Spray Etching


High Performance Capacitors UsingNano-Structure Multilayer MaterialsFabrication


Methods for Microbial Filtration ofFluids


Silicon on Insulator with Active BuriedRegions


Summary of disclosure

An integrated electron source and mass analyzer/detector with miniatureion cyclotron resonance mass spectrometer (MS) having low powerconsumption and ion trap MS that Fourier analyzes the ion cyclotronresonance signals induced in the trap electrodes. This portable, low-power MS with integrated sensors and electronics can detectenvironmental pollutants and illicit substances.

A demultiplexer based on a plurality of banks of samplers (each bankcomprising transmission lines for transmitting an input signal) and strobesignal and sampling gates at respective positions for sampling the inputsignal in response to the strobe signal. Strobe control circuitry is coupledto the plurality of banks and supplies a sequence of bank strobe signalsto the strobe transmission lines.

A flow cytometer with means for deflecting and focusing charged dropletsinto multiple streams. A pair of oppositely charged plates disposed oneach side of a droplet flow with a respective ground plane for each plateproduce a curved and focused electric field between the plates to moreaccurately focus deflections of the charged droplets.

A process for uniformly removing metal from a substrate surface andabove metal-containing trenches formed in the substrate for producing anessentially planar surface across the substrate and the trenches. Thesurface is rotated while directing an etchant onto the surface.

The fabrication of high energy–density capacitors by nano-engineered,multilayer synthesis technologies, whereby the materials, thickness oflayers, interfacial quality, and conductor configuration can be preciselycontrolled. Magnetron sputtering of very thin conductive and dielectriclayers is used.

A method for purifying contaminated subsurface groundwater bycontacting the contaminated subsurface groundwater with resting statemethanotrophic or heterotrophic microorganisms that produce longlifetime contaminant-degrading enzymes. The micro-organisms arederived from surface cultures and are injected into the ground to act as abiofilter. The contaminants include organic or metallic materials andradionuclides.

A method for introducing patterned buried components in a silicon layerand forming contacts for the buried components in the thin silicon layerafter fabrication of the silicon-on-insulator (SOI). This process can beaccomplished by using excimer laser-doping techniques during theformation of the SOI and after the SOI has been fabricated. This methodapplies to SOI wafers made from a silicon wafer bonded to a substratesuch as glass or an oxidized silicon wafer.


32


Groundwater Modeling: More Cost-Effective Cleanup by Design

Mathematical modeling allows groundwater remediationplanners at Lawrence Livermore National Laboratory toview the soils and pathways in the site’s subsurface wherecontaminated groundwater moves. To improve our pictureof the subsurface, we have developed innovative approachesto groundwater modeling and a number of programs thatfacilitate our use of industry-standard simulation codes. Weare applying artificial neural network (ANN) technology toevaluate and optimize “smart” pump-and-treat remediation.We have developed PLANET, a graphical user interface forrapid evaluation of pump-and-treat scenarios, and MapIt, asoftware tool that can read almost any type of data sourceand quickly produce input files for a host of simulationcodes. Evaluation and data input tasks that used to takeweeks and months have been reduced to hours, minutes, oreven seconds. These tools and technologies can be used byremediation planners at any site in need of groundwatercleanup. The ability to forecast a groundwater remediationproject to its closure allows planners to reliably reduce timeand cost as the project progresses.■ Contact: Robert J. Gelinas (510) 423-2267 ([email protected]).

Mitigating Lightning Hazards

A new draft document provides guidance for assessingand mitigating the effects of lightning hazards on aDepartment of Energy (or any other) facility. Written by twoLawrence Livermore engineers, the document combineslightning hazard identification and facility categorizationwith a new concept, the Lightning Safety System, to helpdispel the confusion and mystery surrounding lightning andits effects. The guidance is of particular interest to DOEfacilities storing or handling nuclear and high-explosivematerials. The concepts presented in the document were usedto evaluate the lightning protection systems of the DeviceAssembly Facility at the Nevada Test Site.■ Contact:Richard Hasbrouck (510) 422-1256 ([email protected]).

Abstracts

This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor the University of Californianor any of their employees makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information,apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service bytrade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or theUniversity of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or the University of California andshall not be used for advertising or product endorsement purposes.

U.S. Government Printing Office: 1996/784-071-30013

Dual-Band Infrared Computed Tomography:Searching for Hidden Defects

Federal highway officials estimate that 20% of the country’shalf-million two-lane bridges are structurally deficient. InApril 1996, the Federal Highway Administration began roadtesting a dual-band infrared computed tomography (DBIR-CT) system developed at Lawrence Livermore that images inthree dimensions the defects in the subsurface of bridge decksby using two thermal infrared bands to sense time-dependenttemperature differences. Mounted on a converted motorhome, the system can scan a 3-meter-wide (10-ft) lane on abridge and locate defects while traveling at 40 km/h (25 mph). This system enables inspectors to peer beneath aroadbed’s surface to find and evaluate hidden trouble spotsbefore they become hazardous. In addition to its applicationsfor the nondestructive inspection and evaluation of bridgedecks, DBIR-CT has a number of other applications bothinside and outside the Laboratory. Recently, for example, ithas been used to detect corrosion within the skins of aircraftand differentiate it from other benign subsurface anomaliessuch as excess sealant. LLNL has a Cooperative Research andDevelopment Agreement with Bales Scientific Incorporatedto commercialize DBIR-CT technology.■ Contact:Nancy Del Grande (510) 422-1010 ([email protected]).

lawrence livermore national laboratory livermore ... · our site. we also have identified numerous...

Documents