science and technology to defend our national interests · 2019. 5. 29. · about the cover about...

May 1999

U.S. Department of Energy’s Lawrence LivermoreNational Laboratory

Science andTechnologyto DefendOur NationalInterests

Science andTechnologyto DefendOur NationalInterests

Also in this issue: • Catching the Meaning of

Acoustic Waves• Lighting the Way to Faster

Computers

Also in this issue: • Catching the Meaning of

Acoustic Waves• Lighting the Way to Faster

Computers

About the Cover

About the Review

• •

Lawrence Livermore National Laboratory is operated by the University of California for theDepartment of Energy. At Livermore, we focus science and technology on assuring our nation’s security.We also apply that expertise to solve other important national problems in energy, bioscience, and theenvironment. Science & Technology Review is published 10 times a year to communicate, to a broadaudience, the Laboratory’s scientific and technological accomplishments in fulfilling its primary missions.The publication’s goal is to help readers understand these accomplishments and appreciate their value tothe 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 (925) 422-8961. Our electronic mail address is [email protected].

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

© 1999. The Regents of the University of California. All rights reserved. This document has been authored by theRegents of the University of California under Contract No. W-7405-Eng-48 with the U.S. Government. To requestpermission to use any material contained in this document, please submit your request in writing to the TechnicalInformation Department, Document Approval and Report Services, Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, California 94551, or to our electronic mail address [email protected].

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 California nor any of their employees makes any warranty,expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness ofany information, apparatus, product, or process disclosed, or represents that its use would not infringe privately ownedrights. Reference herein to any specific commercial product, process, or service by trade name, trademark,manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring bythe United States Government or the University of California. The views and opinions of authors expressed herein donot necessarily state or reflect those of the United States Government or the University of California and shall not beused for advertising or product endorsement purposes.

Cov

er d

esig

n: G

eorg

e K

itrin

os

The DOE national laboratories have much tooffer the Department of Defense. At LawrenceLivermore, a plethora of science and technologyprojects is supporting DOD missions (see thearticle beginning on p. 4). The cover is acomposite illustration of several projects,including, from top to bottom, a fiber-compositesabot directing a projectile out of a weapon, asimulation of an electron beam heating a landmine, a shaped-charge warhead undergoingdesign evaluation by an optimization code, and a large explosive test used in developing modelsfor evaluating missile propellant accidents.

SCIENTIFIC EDITORDavid Eimerl

MANAGING EDITORSam Hunter

PUBLICATION EDITORGloria Wilt

WRITERSArnie Heller, Ann Parker, Sue Stull, DeanWheatcraft, and Gloria Wilt

ART DIRECTOR AND DESIGNERGeorge Kitrinos

INTERNET DESIGNERKitty Tinsley

COMPOSITORLouisa Cardoza

PROOFREADERCarolin Middleton

S&TR is a Director’s Office publication,produced by the Technical InformationDepartment, under the direction of the Office of Policy, Planning, and SpecialStudies.

2 The Laboratory in the News

3 Commentary by Milton FingerSupporting the DOD Is Part of Livermore’s National Security Role

Features4 Leveraging Science and Technology in the National Interest

Since its founding, Lawrence Livermore has worked well with the Department of Defense, accepting challenging assignments and providing services and products that support DOD missions.

12 The Revelations of Acoustic WavesLawrence Livermore scientists are finding ways to make sense of acoustic signals and extend their uses.

Research Highlight20 Pulses of Light Make Faster Computers

23 Patents and Awards

Abstracts

S&TR Staff May 1999

LawrenceLivermoreNationalLaboratory

Printed in the United States of America

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

UCRL-52000-99-5Distribution Category UC-700May 1999

Page 4 Page 20

Page 12

2 The Laboratory in the News

Lawrence Livermore National Laboratory

Lab performance rated excellent in ‘98The Department of Energy awarded Lawrence Livermore an

overall performance rating of excellent for fiscal year 1998.The annual assessment includes appraisal of the Laboratory’sperformance in science and technology as well asadministration and operations.

Laboratory science and technology programs received anoverall rating of outstanding, DOE’s highest rating, with anoverall score of 90.6 percent, up slightly from last year’s90.1 percent. The programmatic assessment is based onLaboratory self-assessment, peer review, and validation byprogram managers at DOE headquarters.

“For administration and operations activities, the appraisalprovides us with a valuable tool for measuring our progress inrelation to the performance goals set out in the contract,” saysJohn Gilpin, head of the Office of Contract Management atLivermore. While these are areas where the Laboratory cancontinue to improve, great strides have been made in cuttingadministration and operations costs and improving efficiencysince the results-oriented, performance-based contract withDOE went into effect in 1992, according to Gilpin. He alsonoted that improving productivity and performance ratingswhile reducing costs and organizational staffing “means agreater share of Lab dollars and resources are available for ourcore activities in science and technology.”Contact: John Gilpin (925) 423-1492 ([email protected]).

DOE-ASML team to develop new chip technologyThe Department of Energy recently announced an

agreement with ASML, an international supplier of lithographytools based in the Netherlands, to participate as a licensee in aproject led by the U.S. to develop a new technology forproducing computer chips.

The technology, extreme-ultraviolet lithography (EUVL),has the potential to make desktop computer chips that are ahundred times more powerful and have a thousand times thememory of today’s chips at feature sizes less than one-thousandth the width of a human hair. The pact with ASML is a major step toward realizing the commercial prospects andinternational acceptance of this new technology.

ASML joins U.S. lithography tool suppliers Silicon ValleyGroup and USAL in an ongoing cooperative research anddevelopment agreement (CRADA) between Extreme UltravioletLimited Liability Corporation—a consortium of U.S. computerchip makers Intel Corporation, Motorola, and Advanced MicroDevices—and DOE’s Lawrence Livermore, Sandia, andLawrence Berkeley national laboratories. The purpose of theCRADA is to demonstrate the feasibility of EUV lithographyand establish a pathway to commercialization.Contact: Gordon Yano (925) 423-3117 ([email protected]).

Senate holds first field hearings at LabMembers of the Senate’s Armed Services Strategic

Subcommittee traveled to Lawrence Livermore recently tohold budget hearings for the first time in the field rather thanin Washington. Their purpose was to take testimony fromDOE Assistant Secretary of Defense Programs Vic Reis, thedirectors of DOE’s three national security laboratories, andthe managers of the four DOE plants on the success of theStockpile Stewardship Program.

At the hearing, Senator Robert Smith, the subcommitteechair, and Senator Mary Landrieu, the ranking minoritymember, questioned Reis and the laboratory managers onwhether they believed the Stockpile Stewardship Programwould work well enough to merit ratification of theComprehensive Test Ban Treaty.

In written testimony, Livermore Director Bruce Tarterreported that stockpile stewardship in the absence of nucleartesting is working and that there is optimism that theprogram’s long-term goals are achievable, given sustainedsupport. Tarter, like Reis and the other DOE laboratory and plant managers, urged strong support of the fiscal year 2000 budget submission for DOE Defense Programs.Tarter detailed Lawrence Livermore’s own involvement inmaintaining a safe and reliable stockpile and providedupdates on the National Ignition Facility and Livermore’s role in the Accelerated Strategic Computing Initiative.

Tarter’s complete testimony is available on the WorldWide Web at http://www.llnl.gov/PAO/cbt7-testimony/.

Sustained spheromak dedicatedLawrence Livermore recently dedicated the Sustained

Spheromak Physics Experiment (SSPX), giving a new leaseon life to a magnetic fusion concept pioneered at Los AlamosNational Laboratory. A reinterpretation of old results fromLos Alamos fusion research by Livermore’s Kenneth Fowlerfound them more promising than originally believed. Therejuvenated experiment involves a collaboration of LawrenceLivermore, Los Alamos, and Sandia national laboratories;General Atomics; the California Institute of Technology; theUniversity of California at Berkeley and Davis; theUniversity of Wisconsin; the University of Washington; andSwarthmore College.

The spheromak generates magnetic fields by internaldynamo motion caused by turbulence in the plasma, the hotionized gas that serves as the reactor fuel. The objective ofthe SSPX is to better understand and, ultimately, control thephysics of magnetic fusion, a potential source of abundant,inexpensive energy.Contact: Keith Thomassen (925) 422-9815 ([email protected]).

S&TR May 1999

HE major mission of Lawrence Livermore NationalLaboratory is national security, with a principal emphasis on

providing a secure, safe, and reliable nuclear deterrent. We alsocontribute to national efforts to stem the proliferation of weaponsof mass destruction. Our core competencies, developed becauseof our focus on the nuclear weapons program and our currentresponsibilities in the Department of Energy’s StockpileStewardship Program, are acknowledged by the U.S. nationaldefense leadership to be an important science and technologyresource for the Department of Defense as well. Blue Ribbonpanels, the Congress, and key defense managers historically haverecognized the value in applying DOE’s weapon technologieswhere they can advance conventional warfare capabilities. This is a mutually beneficial strategy. When Livermore weaponstechnology “spins off” to address a DOD requirement, in manycases there is also a return benefit (“spin-back”) to DOE, whereinLivermore’s core competencies have been enhanced.

Our DOD work is usually performed under Work for Others(WFO) programs, which are directly funded by the DOD, andsometimes under Memorandum of Understanding (MOU)programs where both DOE and DOD provide funds. In eithercase, DOD and DOE are mutually served. For example,Livermore develops sophisticated structural dynamics codes that can address the safety of nuclear weapons that are struck byfragments or projectiles. These are three-dimensional problemsthat require three-dimensional codes. These same codes weredeveloped to enable DOD to address the lethality issues of hit-to-kill missiles used to intercept reentry vehicles. They also aid thedesign of three-dimensional, explosively formed projectiles(EFPs) used in flyover shoot-down munitions. In turn, DOEbenefits from these applications. We couple the modeling andsimulation efforts with experiments, which helps to establishbenchmarks for the codes and often drives significantimprovements and new code developments.

We at Livermore take the view that our role in the nationalsecurity arena naturally includes working with DOD, inparticular helping to fill an important gap for DOD in the weapon research, development, demonstration, and productimplementation process. In general, universities, DODlaboratories, and not-for-profit research centers are dependedupon to insert new science and technology into a pipeline thatleads to production of warfare materials. DOD relies heavily on

T

3


Commentary by Milton Finger

universities to provide fundamental basic research (designatedas DOD 6.1 programs), and the DOD laboratories of themilitary services generally focus on early applied technology(DOD 6.2 programs, for example). Industry provides theoutput of the pipeline. DOD relies on defense industries forengineering development and production (DOD 6.3 andbeyond) such as munitions, armored vehicles, planes, ships,and electronic systems.

The national laboratories help to fill the gap between basicresearch and manufacturing in the research-to-implementationcycle (see figure). We integrate basic research developmentsinto focused, applied technology solutions, oftentimes byapplying innovative, “out-of-the-box” concepts. Livermore’sdemonstrated role is to keep the flow going by filling inresearch as needed, integrating and applying science, maturingtechnologies, demonstrating proofs of concept, and teaming,where appropriate, with DOD entities and handing off todefense industry.

The full scope of our DOD activities naturally involves awide spectrum of technologies and deliverables. We provideexpertise in conflict simulation; codes for hydrodynamics,electromagnetics, and explosive response and performance;new explosives; sensors; laser technology; new materials;manufacturing technology; missile defense; and spacetechnology and forensics, to name a partial list. The articlebeginning on p. 4 describes some of this work in detail,demonstrating some of the ways we leverage technologies fordual national security benefits. The projects represent a smallportion of our DOD work, which is spread over almost allLaboratory directorates and involves multidisciplinary efforts.

Supporting the DOD is Part ofLivermore’s National Security Role

n Milton Finger is Deputy Director, DOD Programs.

Leve

l of e

ffort

, dol

lars

6.1Basic

research

6.2Applied

technology

6.3Development

........ Production

UniversitiesDOD laboratories

and national laboratories

Defense industry

S&TR May 1999

pilots could plan safe flight paths andenvironmental air monitors couldestimate the plumes’ health effects.

During war and in less turbulenttimes, the Laboratory delivers servicesand products to DOD. They range froma new missile warhead to a new direct,in-line detonator that provides a safe,reliable electronic fuse used to initiateexplosives in munitions. LawrenceLivermore has also provided DODwith items such as the LX-14explosive, currently found insideDOD’s TOW, Hellfire, Javelin, andBAT antiarmor munitions.

A Two-Way Exchange DOD has long recognized that

Livermore—indeed, all the DOEnational security laboratories—hasunique capabilities that can be leveragedfor DOD purposes. Traditionally, DODhas provided additional funding to thenational laboratories to extend projectstoward conventional weaponsapplications. In one instance, tomaximize this leveraging, the fundingarrangement was formalized in 1985 in a Memorandum of Understanding(MOU), which established a jointmunitions program between the DOEnational laboratories and DOD.

The MOU program at Livermorewas managed for many years by

chemist Milton Finger, now theLaboratory’s Deputy Director for DODPrograms. He subsequently turned thatresponsibility to Al Holt, and currentlyDennis Baum manages the program.Finger says that “the program providesa window through which LawrenceLivermore can be aware of DOD needsand DOD can be knowledgeable of thetechnologies available at Livermore.DOD can challenge Livermore tocontribute innovative science andtechnology to attack pervasive problemsand grand challenges in the defensearena. In addition, Livermore can focusits efforts more efficiently andproductively to serve the dual interestsof DOE and DOD.”

Baum identifies the program’sprincipal technical areas as highexplosives, codes, nonnuclear weaponsdesign, fuses, demilitarization, sensors,and advanced materials. Both Fingerand Baum point to the program’sefficient integration with Laboratoryprojects and priorities. Consequently,Livermore resources are being usedmore fully and productively, and DODderives advantages from Livermore atthe same time that Livermore corecompetencies are enhanced.

The projects described here, whichare but a small portion of Livermore’sDOD work, demonstrate some of theways the Laboratory uses technologiesfor dual national security benefits.

Getting out of Tight SpotsOne day, U.S. soldiers under attack

in hostile, foreign terrain may findthemselves depending on a devicedeveloped with the help of LawrenceLivermore. To put an insurmountableobstacle between themselves and theenemy, they pull out a weapon called a

4


The Department of Energy and the Department of Defense have historically shared LawrenceLivermore’s wealth of national security resources.The results are more science and technology for theinvestment and better assurance that the nation’ssecurity and defense needs will be met.

HE three DOE national securitylaboratories—Lawrence Livermore,

Los Alamos, and Sandia—have atechnology base of interest to theDepartment of Defense. Their nuclearweapons technology can be leveraged toaddress the DOD nonnuclear securitymission. Therefore, it’s not surprising thatDOE and DOD have a long history ofcollaboration at the three laboratories. AtLawrence Livermore, that collaborationdates at least as far back as February 1956,when Edward Teller made a bold pledge todeliver to DOD a smaller, lighter warheadfor the Polaris missile and do so on anextremely short schedule. LawrenceLivermore scientists took up the challengeand made good on Teller’s promise. It wasone of many instances where scientistsfrom DOE national security laboratorieswere to fulfill DOD requests.

Later, during the Cold War, a NavyTrident test missile blew up andextensively damaged the testing range.Lawrence Livermore, working with Los Alamos and two Navy laboratories,unraveled the cause of the explosion,which led to the development of a safer,high-energy propellant to put the Tridentmissile back on track. More recently, inKuwait while the Persian Gulf War wasbeing waged, Livermore’s AtmosphericRelease Advisory Capability trackedsmoke plumes from torched oil wells so

T

Leveraging Science and TechnologyLeveraging Science and Technology

S&TR May 1999

munition designed to crater airfieldrunways. The portable four-stagemulticharge PAM—a demolitionmunition at once compact, light, and effective— was realized under the joint DOD/DOE MOU program.

During the fabrication and testing of the first PAM, the device would not work properly because the shockresulting from the rebar-destroying and hole-drilling charges caused thefuse in in the main penetrating chargeto malfunction. Livermore scientistsdeveloped a fuse that could survive the explosive shocks and detonate thelast charge at the appropriate time.

5


Livermore Work for DOD

bridge’s concrete rebar, the secondmakes a deep, narrow hole in the bridgepier, and the third penetrates to thebottom of that hole and detonates.Objective accomplished. The soldiershave hindered enemy mobility.

The genesis of the multistage PAMcan be traced to work during the 1980son a two-stage munition system for theAir Force. Livermore scientists evolveda two-stage munition based on the workof a defense industry contractor into awarhead for a 2,000-pound laser-guided bomb. The Defense AdvancedResearch Projects Agency sponsoredfurther development of a three-stage

PAM, or Penetration AugmentedMunition. Although compact andlightweight (approximately 35 pounds,33 inches long), it contains the power of four explosive charges and whendeployed, can effectively destroy bridges,runways, roads, and tunnels (Figure 1).

A typical demolition target for thesoldiers is a bridge. To destroy it, theymust detonate two PAM unitssimultaneously at the bridge pier. Theytrigger the PAMs’ propelling chargesand shoot the warheads directly into thestructure. The motion of the propellingcharge sets off each PAM’s other threecharges: one charge cuts through the

Figure 1. (a) The Penetration Augmented Munition(PAM) is a lightweight, compact weapon thatcarries a propelling charge to (b) set off three othercharges to effectively destroy bridges, runways,roads, and tunnels. (c) It is shown being deployedon a bridge pier.

(a)

(c)

(b)

Follow-through charge

Hole-drilling charge

Rebar-cutting charge

in the National Interestin the National Interest

S&TR May 1999

Michael J. Murphy, one of thedevelopers of the device, says that the PAM has been designated by theDepartment of Defense as a “TypeClassified Standard for Army SpecialOperational Forces Use,” meaning that DOD has made a firm decision toproduce and use it. It is now designatedas the XM150. Engineeringdevelopment, conducted at AlliantTechsystems and under U.S. Armysponsorship with Lawrence Livermoresupport, is complete.

Strong String and GlueEngineers in Livermore’s Mechanics

of Materials Group, led by SteveDeTeresa, were part of a LawrenceLivermore–Army Research Laboratoryteam that developed a fiber-compositesabot for DOD use. A sabot is alightweight carrier used both toposition a missile or subcaliberprojectile inside a gun tube and totransmit energy from the propellant tothe projectile (Figure 2). DeTeresa saysthat the sabot works much like a personthrowing a dart, where the thrower’sarm movement acts as both the

propellant-driving gas and the sabot’senergy-gathering pusher (Figure 3).

In general, guns operate with a fixedmass to be propelled out of the gun’stube. The sabot is necessary to transferpropellant energy but is a parasiticweight in terms of projectile targetperformance. Reducing the sabot’sweight allows greater projectilevelocity. The weapons thus penetratedeeper, with more lethal results. Butmaterials used to fabricate sabots canonly be as lightweight as they are strongenough to withstand great pressures andloads during gun-tube acceleration.Previously, the lightest weight sabotswere made of aluminum.

In the past, the search for lighterweight sabot materials focused onmetal composites. But researchers werecontinually frustrated by failure—metalcomposites simply were too brittle.Attention then shifted toward polymer-based composites, which were beingused extensively in thin structures foraerospace applications. Researchersbegan to consider fiber composites forcomplex shaped structures that neededto survive multidirectional stresses.

6



Livermore material scientists wereasked to help develop a new sabotbased on these materials.

DeTeresa relates that some engineersrefer whimsically to a fiber compositeas “string and glue.” It consists of high-strength carbon fibers, which must belaid down and oriented to yieldmaximal strength and handle maximalstress. Polymer is used to glue togetherlayers of these fibers in a processsimilar to that used to manufactureplywood. When layers are gluedtogether, the grains of adjacent layersare arranged either at right angles or atsome wide angle to each other. Once a piece of the material has beenfabricated, it can be machined into therequired form. Fairly thick pieces thatcan withstand high three-dimensionalstress are used for sabot material.

Although they have developed aneffective, extremely lightweight sabot,development team members continue toinvestigate which material combinationsand fiber architectures will provideever-greater material strength. They areeager to understand the material’s stressresponses and failure modes completely,particularly because thin sheets of thismaterial are used for safety-criticalcomponents in airplanes.

The team has developed models offiber-composite materials and issimulating their performance using theLaboratory’s DYNA and NIKEstructural response codes. One of themodels incorporates a misaligned fiber.By analyzing the effects of the imperfectfiber on material properties, theresearchers can address how to preventor minimize those effects. At the sametime, they are investigating cheaper waysof producing fiber-composite material.

The Army, the largest consumer ofadvanced carbon fiber composites inthe defense community, is using thefiber-composite sabot in the M829 A2kinetic energy projectile, the weapon of choice for antitank warfare.

SabotProjectile

Fin

Obturator

Burning propellant

Gun tube

(a)

(b) (c)

Figure 2. (a) Asabot transfersenergy from theweapon propellantto its projectile. (b) Livermore’sfiber-compositesabot is (c) part ofthe Army’s weaponof choice forantitank warfare.

As a result of the sabot work,Livermore holds a patent on the fiber-composite sabot’s structure andfabrication process. Livermore and theArmy Research Laboratory have wonan Army Service Award for developingthe sabot. The Livermore engineers arethe first non-DOD civilians to receivethis award.

Code Optimizes DesignComputational modeling and

simulation, already a key component ofLivermore problem-solving capability,will become even more dominant asDOE’s Accelerated StrategicComputing Initiative continues toincrease computational speed andpower. Not surprisingly, computer codedevelopment is flourishing atLivermore, and many scientists arewearing the dual hats of code developerand code user. Michael J. Murphy, whowas involved in the design of the PAM,is one of them. He and Ernest Baker ofthe U.S. Army Armament Research,Development and Engineering Centerhave developed a code useful foroptimizing warhead designs, includingshaped charges (warheads encased insteel or aluminum and consisting of ametal cone, or liner, backed by highexplosive). Murphy’s code is calledGLO (global local optimizer).

GLO directs physics codesimulations to optimize the warheaddesign. It is a powerful tool that savesmunitions designers time and producesrobust results.

Two key steps are involved inGLO’s work. First, it must incorporate adescription of an optimum design,based on the kind and degree of damagethat designers want the shaped charge toinflict. For example, the goal may be tocreate a hole of a specified size anddepth in a certain target. GLO runs thephysics codes and then compares thecalculated hole profile with the desiredhole profile. Figure 4 shows a

simulation of the shaped-chargedetonation, jet formation, andsubsequent penetration into a target.

The second step optimizes the designusing the results of the comparison fromthe first step. GLO is repeatedly linkedto the physics codes and adjusts theshaped-charge design until it obtains as close a match as possible to thespecified hole profile. Often, the codethat GLO directs is the two-dimensionalhydrodynamic code, CALE (C-languagearbitrary Lagrangian–Eulerian), inwhich is embedded a number ofparameters defining the overall size andgeometry of the shaped charge. For eachdesign considered, GLO specifies thevalues of the parameters that define thegeometries of the shaped-chargeexplosive and metal cone. CALEcalculates the mass and velocitydistribution of the jet for each shaped-charge design. GLO’s parameterschange over the series of calculations todescribe different configurations of theshaped charge.

The CALE calculations result in adefinition of the geometry of the jet of metal formed when the cone of aparticular shaped-charge configurationis compressed by the explosive charge.This definition is used by an analyticpenetration code to calculate the jetpenetration and the resulting targethole profile.

In a typical overnight optimizationrun, GLO can evaluate some 250 sets of parameters. The optimum designconfiguration is selected from thesesets. Murphy says that GLO is a “verydedicated assistant working unceasinglyto generate numerous iterations ofshaped-charge configurations.”

From TIGER to CHEETAHRon Atkins, head of Livermore’s

Energetic Materials Center, notes thatit’s usual for inventors to first try tomake their inventions work and thento try to understand how they work. That is certainly the case with high-explosive detonations. Scientists haveworked for over a century to understandthe physics of detonation properties ofexplosives long in use. Atkinscoordinates a group of projectsattempting to expand that understandingfurther in order to design safer and more powerful explosives as well as to formulate new explosives withproperties tailored to specificapplications.

One ongoing project is a code thatsimulates detonations and predicts theresults of detonating a specific mixtureof chemical reactants. The code isCHEETAH, a fast, scientificallyrigorous descendant of Livermore’sTIGER and RUBY thermochemicalcodes. Chemist Laurence Fried and

7


Livermore Work for DODS&TR May 1999

Figure 3. A sabot pushing the weapon projectile toward its target.

8


Livermore Work for DOD S&TR May 1999

Figure 4. (a) Physics codesimulation of the explosivedetonation, metal liner collapse,and stretching jet formation of ashaped charge. (b) Simulation ofthe jet impact with the target,penetration into the target, andfinal hole profile.

(a) (b)

colleagues designed CHEETAH toallow explosives formulators to predictdifferent starting molecules andformulation performance and, hence, to design optimized explosives withspecific characteristics.

The newest version of CHEETAH(described in S&TR, November 1997, pp. 21–23) is a particularly popular toolfor explosives formulators in that it ismore user friendly than earlier versionsand includes a database of 200 chemicalstarting reactants and 1,000 possibleproducts. This database saves a user theinconvenience of looking upthermodynamic constants for eachchemical. More significantly, the newCHEETAH tracks chemical reactionsdown to the molecular level to obtainvery accurate predictions of the velocityand energy of the detonation.

The earlier version of the codeassumed that all reactions occurinstantaneously, that all reactioningredients are consumed completely,and that thermal equilibrium is reached atthe same time. In reality, the chemistry ofa detonation is much more varied andcomplicated. Many different moleculesare involved, with some reacting moreslowly than others, and those slowchemical reactions require a long time toachieve thermochemical equilibrium.Moreover, a variety of chemical reactionstakes place during the explosivedecomposition of mostly large, energetic-material molecules into small, simpleproduct molecules. The explosivereaction products undergo materialchanges and occupy different states ofpressure, density, and velocity. All thesereactions must somehow be representedin the codes to obtain accurate predictionsof detonation pressure, velocity, andenergy of the detonation.

Fried implemented a kineticdetonation model, based on theWood–Kirkwood detonation theory,which provides equations of state forcomplicated mixtures of detonationproduct molecules. This model accounts

for the microscopic mechanical andthermal processes that occur in shockinitiation and detonation, and itcalculates chemical reaction rates at the molecular level. The calculationalresults showed CHEETAH effective for modeling many features of slowlyreacting explosives.

Fried and his colleagues arecontinuing to improve CHEETAH byincluding the effects of high pressureand high temperature on chemicalkinetics. They will thus be able tomodel more complex, slow detonationbehavior such as shock initiation, hot-spot formations, and failure processes.They are also launching an effort to linkCHEETAH to hydrodynamic codes sothey can create even more completemodels of high-explosive detonation.This effort will serve not only DODexplosives formulation work but alsohelp Livermore fulfill itsresponsibilities to the DOE StockpileStewardship Program. In the case ofCHEETAH, DOE resources that wereleveraged to benefit DOD are in turnbeing leveraged to benefit DOEmissions at Livermore.

Codes to Assess SafetyIn addition to CHEETAH, other

Livermore codes are proving useful forevaluating explosive performance andeffects. For example, CALE is used inLaboratory projects to assess a varietyof explosive and nonexplosiveproblems. Livermore scientists areusing it for such applications assimulations to evaluate safety concernsat missile launch sites.

In April 1986, at 8.7 seconds into thelaunch of an Air Force Titan T34D-9space vehicle from Vandenberg AirForce Base, one of the vehicle’s solidrocket boosters failed. A portion of thebooster came loose and fell back downfrom an altitude of 18,000 feet at aspeed of 320 feet per second, hitting theground sideways. That piece weighedan estimated 130,000 pounds, including

110,000 pounds of solid rocketpropellant. At impact, it exploded and burned, releasing between 7 and30 percent of the propellant energy and causing significant damage atVandenberg.

This launch was representative of the one out of every 30 launches, onaverage, that ends in failure. Many ofthose failures result in explosions whenunburned motor segments fall back tothe ground. Launch safety officials needto know just how destructive and far-reaching such accidents can be. Butuntil recently, they have had onlyintuition and sparse data to rely on formaking their safety judgments.

The upgrade of the solid rocketmotor of Titan IVB, which uses a newpropellant and a new motorconfiguration with much longer andmore massive booster segments,prompted the Air Force to initiate aproject to better understand fallbackaccidents. One part of that program isbeing performed by chemical engineerJon Maienschein and his colleagues.They have developed a computermodel that describes in detail thepropellant response to fallbackaccidents and predicts the extent andeffects of their energy releases.Simulations using this model willenable launch safety personnel toassess and provide safeguards againstthe hazards of these accidents.

The model developed by Livermorescientists is called PERMS (propellantenergy response to mechanical stimuli).One part of the model describes how ashock front, generated by the impact of afalling booster rocket, causes ignitionand burning of explosive material. Dataabout shock initiation used in the modelwere obtained through field tests(conducted by Phillips Laboratory atEdwards Air Force Base) that used largeexplosive boosters to generate shocksfrom 30,000 to 40,000 times atmosphericpressures over a long duration (Figure 5).Experimenters shocked a sequence of

9


Livermore Work for DODS&TR May 1999

propellant samples up to 5 feet indiameter. Propellant mass in this test wasover 48,000 pounds. From these data,they developed a model for the initiationmechanisms and estimated the diameterof propellant necessary to support steadypropagation of a reactive shock wave.

The other component of the modelis a description of how the boostersegment fragments at impact.Fragmentation creates additionalburning surfaces as the propellantdeforms. The rate at which explosiveburning occurs is related to the sizeand hence surface area of thefragments. Input data for fragmentationand burning rate models were derivedfrom laboratory experiments. Theresulting models were validated usinglarge-scale (thousands of pounds ofpropellant) tests with either steelimpact plates or hollow implodingcylinders to simulate the propellants at pressures less than 15,000 timesatmospheric pressure.

These two descriptive componentsform the PERMS model, which isimplemented in the CALEhydrodynamic code. Once theconditions of the booster fallback arespecified, the model calculates thepropellant reactions, considersfragmentation effects, and tracks theprogress of reactions over time. Theforce of the propellant reaction istranslated into the equivalent TNTenergy release.

Adding Capabilities to the CodePERMS provided both significant

new information about propellanthazards and remarkably good estimatesof the explosive behavior that resultsfrom both the nose-on and side-onimpacts of booster motors falling fromthe sky. Now, its developers want tolook at historic data of incidents inwhich motor segments fell inorientations where three-dimensionaleffects are important. To study those

effects, they would use the three-dimensional Lagrangian–Eulerian code called ALE3D, which has itsorigins in Livermore’s DYNA3D code.

ALE3D has recently been improvedto better model the response ofenergetic materials to heat andexplosive processes. It is nowundergoing testing by Laboratorydevelopers Albert Nichols and RichardCouch. The upgraded ALE3D hasadditional thermal and chemicalcapabilities as well as calculationaloptions that allow it to accurately depictevents over time scales ranging frommicroseconds to days. It is designed tosimulate a typical fire scenario, forexample, by following the transport ofheat from the exterior of an explosivedevice to the explosive itself, followedby the thermal decomposition of theexplosive. The decomposition graduallychanges the material properties of theexplosive and induces motion.Depending on how the explosive is

10


Livermore Work for DOD S&TR May 1999

Figure 5. Data fordescribing the initiationand burning of explosivematerial are obtainedthrough large-scale testsperformed at Edwards AirForce Base, wherepropellants were shockedat 30,000 to 40,000 timesatmospheric pressure,which caused them toexplode. (Photographcourtesy of Dr. ClaudeMerrill, Phillips Laboratory,Edwards Air Force Base.)

S&TR May 1999 11



confined, the simulation will then depicta slow, relatively benign response or afast, catastrophic explosion, as happensin real life.

The code has successfully simulateda U.S. Navy “cookoff” safety test inwhich a slowly heated high explosive is deformed over a long time span (seeS&TR, June 1997, p. 11). It has alsobeen used in simulations to investigatethe use of electron beams for clearingland mines (Figure 6).

The developers are planning to domore testing, using different materialmodels for the chemical reactions andmixtures associated with the explosiveprocesses. They also look forward to

using ALE3D to solve other kinds ofproblems associated with the forging,casting, and extruding processes ofmanufacturing.

Continuing the CollaborationAs Lawrence Livermore scientists

and engineers fulfill their DOEmissions, they often find their worktying well to DOD needs andapplications. Thus, providing productsand services to DOD is both a naturalextension of their scientific andtechnical work as well as a fruitfulleveraging of research funding. Asidefrom accruing advantages to bothagencies and the Laboratory, this

CORY COLL received his A.B. in physics from Johns HopkinsUniversity and his Ph.D. in physics from the University ofPennsylvania. After working at Sandia National Laboratories,California (1974 to 1981), he joined Lawrence Livermore’sweapons program as a design physicist and participated in threeunderground tests at the Nevada Test Site. His career atLivermore was interrupted between 1984 and 1986 when he

became, first, staff to the Deputy Undersecretary of Defense for Strategic andNuclear Forces and, later, a program manager at the Defense Advanced ResearchProjects Agency (DARPA).

Coll returned to Livermore in 1988 as deputy program manager for AdvancedApplications in the Laser Programs Directorate and moved to the LaboratoryDirector’s Office in 1992. Currently, he is staff to the director of the Department of Defense Programs Office at the Laboratory.

About the Scientist

Figure 6. Recently, a representative of theU.S. Navy used Livermore’s ALE3D code toinvestigate the use of electron beams to clearland mines. The thermal-only model shownhere is a snapshot in time of a circular regionof a land mine being radiatively heated by anelectron beam.

leveraging ensures that science andtechnology at Lawrence Livermore arefully in step with national security anddefense requirements, whatever theymay be.

—Gloria Wilt

Key Words: ALE3D, CALE, Cheetah,explosives, Department of Defense (DOD),fiber-composite sabot, fuse, GLO (globallocal optimizer), Memorandum ofUnderstanding (MOU), PenetrationAugmented Munition (PAM), PERMS(propellant energy response to mechanicalstimuli), safety assessment, warhead.

For further information contact Cory Coll (925) 422-2103 ([email protected]).

S&TR May 1999

ROM analyzing speech torecording earthquakes, trackingsubmarines, or imaging a fetus,measuring and analyzing acousticsignals are increasingly important in modern society. Acoustic waves are simply disturbances involvingmechanical vibrations in solids, liquids,or gases. Lawrence Livermoreresearchers are developing advancedtechniques for extracting andinterpreting the information in thesewaves. In the course of extracting datafrom acoustic signals, the researchers

have developed complex andcreative algorithms

(mathematical relationshipsimplemented in computers)that at times mimic thereasoning processes of thehuman brain.

One leader of Livermore’sacoustic signal-processing

research is electronics engineer Greg Clark, who is involved in threedisparate acoustics projects: heart valveclassification, where acoustic signalprocessing is determining whether anartificial heart valve is intact or needsreplacing; oil exploration, whereLivermore experts are automating a keyprocedure used for locating undersea oildeposits; and large-structure analysis,where Livermore is preparing to useacoustic wave vibrations to assess theintegrity of several large mechanicalstructures in northern California.

Making sense of acoustic signalsrequires researchers to develop realisticcomputer models and developalgorithms for separating signals fromcontaminating noise.

Computer models may be based onprior knowledge about the source andunderlying physics of the signal, as theyare in the large-structure project thatstudies the San Francisco–Oakland BayBridge. Knowledge about the bridgeand a detailed numerical model guidethe development of signal-processingalgorithms for the project.

But in the heart valve and oilexploration projects, knowledge of thesignals is lacking or cannot be linked toa strong physical model, at least at theoutset. For these cases, a “black box”

12


F

The Revelations of Acoustic WavesLivermore researchers are developing advancedtechniques to extract information from acoustic signals.

Figure 1. A heart with anartificial heart valve. The

valve consists of a disk heldin place by two struts that

let it flip open and shutduring pumping of blood

through the valve.

13


Acoustic Signal ProcessingS&TR May 1999

technique. Tests show it is moreeffective than x-ray procedures, whichcannot capture a clear image of movingheart valves.

An artificial heart valve isessentially a small ring with two smallstruts welded to it, opposite each other.One is an inlet strut, and the other, anoutlet strut. Each is two-legged. Thetwo struts hold a disk that flips up anddown to open and close the valve(Figure 1).

Over time, the struts can developcracks at or near their weld joints andbreak loose from the ring. Livermoresignal experts are trying to detect heartvalve failure indicated by one leg of astrut breaking loose from the ring. (Ifboth legs break loose, the valve losescontrol of blood flow to the heart,leading to death in two of three cases.)

To determine the condition of anartificial valve, the Livermoreresearchers use acoustic signals ofrecipients’ heart valve sounds that havebeen recorded at clinics using high-sensitivity microphones. They firstcollect a database of heart valve soundsof about 100 beats per patient,discarding heartbeats with too low asignal-to-noise ratio or withcharacteristics statistically different from the other beats. The team usesalgorithms to classify the recordings asindicative of a valve that is either intactor not. The algorithms scrutinize thefrequency spectra of each opening soundof a valve and select key features of thespectra, usually parts of certain peaks.A statistical pattern classifier—in thiscase, an artificial neural networktrained on the recorded acousticfingerprint of known faulty valves—decides whether the valve is indeeddamaged (Figure 2).

The team’s efforts are focused on thesounds made when the valve opens—sounds caused by the disk hitting theoutlet strut—because those sounds yielddirect information on the condition of


they can design a filter to monitor theenergy content in that part of thefrequency spectrum.

Hearing the Heart For medical diagnostic applications,

acoustic signals provide advantages inbeing noninvasive and harmless. Theseadvantages are being exploited by ateam of Livermore engineers who aredeveloping an acoustic processingtechnique for sifting through a seemingcacophony of heart and body sounds toisolate the few telltale signals of afaulty artificial heart valve. Theirtechnique would spare patients, manyof them elderly, from open-heartsurgery to determine if an artificialvalve needs replacement.

The four valves of the human heartcontinually open and close, allowingblood to be pumped through the heart’sfour chambers. When a valve becomesdiseased, pumping ability decreases.Prosthetic heart valves correct thisdeficiency and extend the life spans ofmany people with serious heartconditions. But Livermore engineer andproject co-principal investigator JimCandy points out that prosthetic valves,while extremely reliable, are eventuallysusceptible to long-term fatigue andstructural failure, as might be expectedfrom any mechanical device operatingover a long time.

The Livermore experts are workingto find ways to identify faulty heartvalves made by one medical devicemanufacturer whose heart valves were implanted from 1979 to 1986 in86,000 patients. To date, more than 600of these valves have failed, and morethan 300 people have died. A court-appointed panel is funding research tofind the best screening technique orcombination of techniques to determine,with a high degree of accuracy, if oneof the manufacturer’s implanted valvesis failing. Livermore’s acousticalprocessing method is a leading

model, derived only from the data (that is, input and output signals),without details of the underlyingphysics, is used to guide thedevelopment of algorithms.

For the three projects, Clark usesadvanced signal-processing techniques,including statistical neural networks,which are systems of computerprograms that approximate the operationof the human brain. Current uses forneural networks include predictingweather patterns, interpreting nucleotidesequences, and recognizing features inimages. In a supervised learning mode, a neural network is “trained” with largenumbers of examples and rules aboutdata relationships. This training endowsthe network with the ability to makereasonable yes–no decisions on whether,for example, a geologic data plotindicates a geologic layer that couldmark the presence of an oil deposit, orwhether energy in a frequency spectrumof a recording signifies a damagedartificial heart valve.

In every acoustic signal project, vitaldata must be separated from noise thatcontaminates and inevitably degradessignal quality. The noise is caused bothby the surrounding environment and thevery system recording the signals. Forexample, the remote system designed byLivermore engineers for monitoring largestructures can introduce noise into thesignal in the course of relaying it overcellular phone lines to the Laboratory foranalysis. In another example, the delicatesounds of a heart valve flipping up anddown can be buried by acoustic scatteringinside the body. Similarly, multiplericocheting reflections from underwaterexplosions can contaminate the precisedata needed to isolate geologic strata.

Often, Livermore engineers canreduce noise by using filters for certainfrequency spectra. For example, if theyknow that the structural failure of abridge will cause a vibrationalresponse at 5 hertz (cycles per second),

14


Acoustic Signal Processing S&TR May 1999

the strut. A closing valve, in contrast,causes the entire ring to vibrate, and that masks the strut’s vibrations. Clarkcompares the sound of a faulty valve to the thud of a cracked bell.

Unfortunately, the opening soundshave much lower intrinsic signal levelsthan the valve’s closing sounds. Candynotes that measuring heart soundsnoninvasively in this noisy environmentputs significant demands on the signal-processing techniques to extract thedesired signals, especially when thedata of interest last only 10 to20 milliseconds. “Finding the openingsound caused by a strut separating fromthe heart valve is more difficult thansearching for a single violin string thatis out of tune in an entire orchestra,”Candy says.

Clark points out that every valverecording is distorted by the bodycavity and the recording process itself.That is why Livermore researchers are conducting studies at a U.S. Navylaboratory in San Diego in whichacoustic sensors are submerged inwater while collecting the sounds ofvalves that have been surgicallyremoved from patients. Thissubmersion isolates the pure sounds ofboth intact and damaged valves. “Thetest will allow us to measure the heart

valve sounds without the acousticscattering effects caused by the body,”Clark says. “The pure data shouldallow us to mitigate the distortion inpatient recordings.”

Listening for OilSome of the same approaches taken

to analyze the subtle sounds of humanheart valves are used to locate theinterfaces of undersea geologic layers,particularly slate and sandstone layerswhere oil tends to accumulate aroundsalt domes, or “plugs” (Figure 3).Mapping these geologic layer interfaces(called event horizons or, simply,

events) helps geologists decide whereto site oil drilling rigs.

One such mapping project issponsored by the Department ofEnergy National Gas and OilTechnology Partnership. It is acooperative effort involving Shell Oilof Houston, Texas, as LawrenceLivermore’s industrial partner and aPh.D. student from the University ofCalifornia at Davis who is beingsupported by the Laboratory to workon the project.

The purpose of the Livermore–Shell–UC Davis project is to automatea technique used to analyze acoustic oil

Dataacquisition

Medical clinic

Signalextraction

Bad beat rejection

Signalprocessing

Spectra and beatfrequencyhistograms

Classifier

Statisticalneural nets

Featureselection

Measures ofclass separability

Feature extraction

Spectralfeatures

Beat spectrum

INTACTSLS

Salt dome, or "plug"

Oil and gas accumulation

Slate V1

V2

V3

V4

Sandstone

Oil and gas accumulation

Figure 3. Acoustic techniques can be used to locate the interfaces between undersea geologiclayers (usually slate and sandstone) and salt domes, where oil tends to accumulate.

Figure 2. The process forclassifying heart valves includesacquiring clinical data, rejectingheartbeats that have low signal-to-noise ratios or are statisticaloutliers, estimating signalspectra, extracting the spectralfeatures to be used todiscriminate between intact andnonintact valves, selecting onlythe most important features,and finally, classifying thevalves as intact or not.

15



exploration data. The current techniqueis a significant bottleneck because it isperformed manually, at great cost in timeand money. The project’s goal is to reduce manual effort to only about0.1 percent of the data processed.Current results show that this goal isachievable, says Clark.

Oil companies obtain acoustic signalsby having a ship set off underwaterexplosions that generate acoustic wavesthat reflect from geologic layers as deepas 15 to 20 miles under the sea. A 5-mile-long linear array of some 100 hydrophones, which is towed by the ship, measures the signals (Figure 4).

The signals are organized into two-dimensional images, called commonreflection point (CRP) panels, whichrepresent vertical sections of the earth(Figure 5a). The panels show multiplereflections of the acoustic waves as they bounce off various strata of earth(Figure 5b). “You get a horrendousnumber of reflections,” says Clark. “Itgets very messy to sort out.”

Sorting out the reflections dependson correcting for the velocity of thesound waves as they travel throughdifferent layers of the earth. If thevelocity computer model is correct, theimaged events appear as approximatelystraight horizontal (flat) lines in the CRPpanel, because the true depth of theevent horizon is approximately constantacross the panel.

If, as typically occurs, the initialvelocity computer model is incorrect,the event depths vary across the paneland do not appear flat. As part of aniterative velocity estimation process, anexpert must visually inspect the panelsto pick out event locations manually.The expert’s picks are then used asinput to refine the velocity model. Thisprocess is repeated several times, untilthe model produces events imaged asflat lines. The corrected panels arecombined to obtain a two-dimensionalimage of the subsurface strata to helpgeologists determine where to site anoffshore drilling platform.

When Clark inspected the CRPpanels at Shell facilities in Houston, he suggested a new approach foranalyzing reflection data. Instead ofanalyzing one signal or a few signaltraces at a time, as is conventionallydone, he proposed treating the set of 45 traces that forms a CRP panel as asingle image. “I pointed out that if youtreat the panel as an image, there’s awhole set of literature and a lot ofpowerful tools available to you,” hesaid (Figure 5c).

The Livermore team developed atechnique that breaks the panels intosmall pixels (picture elements) todetermine if the data represented withineach pixel are part of an event orsimply background noise. Thetechnique uses advanced algorithmsfrom the areas of automatic targetrecognition, computer vision, andsignal–image processing. For example,using algorithms similar to thoseemployed for computerized militarytarget recognition, the technique

Shot point

1 2 3 4 5 6 7

87654321Common reflection point

8

Figure 4. Underwaterexplosions are set off togenerate acoustic wavesthat reflect off geologiclayers and are received bya linear array of equallyspaced hydrophones.

16



scrutinizes the neighborhood of eachpixel at various orientations and scalesand looks for common features withneighboring pixels. Whenever possible,prior knowledge of known events isincorporated into the procedure. Theresults of the Livermore process comparefavorably with those attained by experts(Figure 6).

Clark has made many trips to Shell to collaborate with their scientists on the project. “It’s been a wonderfulmarriage of disciplines,” he says. “Iintroduced them to a lot of advancedsignal-processing techniques, and theyeducated me about their world ofexploration geophysics.” Clark’s Ph.D.research involved analysis ofseismograms for Livermore’sComprehensive Test Ban Program, so he already had an elementarygeophysical background.

Shell Oil estimates that theLivermore work will significantly affect the oil industry. Shell’s currentimplementation of the Livermorealgorithms already reduces data-pickingtime from about one work day to about90 minutes. Once the software iscompletely converted from researchcode to production code, Shell estimates

4.7

5Offset, feet in thousands

10 15

Tim

e, s

econ

ds

4.6

4.5

4.4

4.3

4.2

4.1

4.0

Automated picks

Human picks

Figure 6. Results of Livermore’s automated process for picking out geologic layer interfaces (calledevents) compare favorably with those attained by experts. The circles depict manual picks, while theblue line, barely visible behind the circles, depicts the automated picks. The manual picks andautomated picks overlap so closely they are hard to differentiate.

(x,y)

θ

θ

(x,y)

Shot

CRP panel

Hydrophone

(z)

(z)

(a) (c)(b)

Figure 5. (a) A series of common reflection point (CRP) panels that represent vertical sections of the earth. (b) A depiction of explosive shots causingthe wave reflections that are recorded by hydrophones and imaged into CRPs. (c) The CRP panels plotted side by side to form a mosaic.

17



the cost of performing a single velocityanalysis could be reduced from $75,000and 12 weeks to $6,000 and one week.An oil company performing 100velocity analyses per year could savenearly $7 million annually. Potentialannual savings for the U.S. oil industrycould amount to roughly $140 million.

Clark points out that it costs about$1 billion to erect an oil platform in theGulf of Mexico. Such costs make itcrucial to respond quickly to businessopportunities. The Livermore signal-processing techniques, providing timesavings of a factor of 12, couldsignificantly enhance the industry’sresponsiveness.

Vibrational FingerprintsAcoustic signal processing may also

make it possible to analyze vibrationsand thus assess large mechanicalstructures for damage after earthquakesor other destructive events. OneLivermore project is combining signalprocessing with advanced numericalmodels and new remote monitoringsystems to better understand largestructures and provide a unique way toquickly monitor them for damage.

“Our task is similar to that of theheart valve project: use sophisticatedsignal processing to enhance ourunderstanding of the way largestructures vibrate to find out if there isdamage,” says project leader andmechanical engineer David McCallen.

There is a critical need for a speedymethod to assess the integrity of astructure after an extreme event, addsMcCallen, who has worked withCaltrans (California’s Department ofTransportation) on earthquake-relatedprojects. Current procedures requirelengthy, largely visual inspections.

McCallen readily acknowledges thetechnical challenges of using vibrationmeasurements to determine the health ofa structure. “We’re asking a lot of oursignal-processing people,” he says.“We’re telling them, we’ll give you

enough data and insight into the structureso it won’t be a black box situation, andwe want you to tell us if there is damage,what it is, and where it is.”

The Livermore project involves threenorthern California case studies: theBixby Creek Bridge in Big Sur, the SanFrancisco–Oakland Bay Bridge, and theNational Ignition Facility (NIF), theworld’s largest laser under constructionat Lawrence Livermore. Each structure

previously has been studied atLivermore; as a result, a detailednumerical model exists for each(Figure 7).

The numerical models provideinformation useful for designingsensors (accelerometers) thatLivermore researchers will install onlarge structures for remotemonitoring. The models indicate whatfrequencies are of interest and also

Simulation model from Caltrans-sponsored research project

Simulation model from Lawrence Livermore–UCcollaborative research project

Large distributed structure with very broad- band frequency characteristics requires strong and weak motion measurements

Large distributed structurewith small amplitude, high frequency content motions

Remotely located structurewith severe accessibilitylimitations

Simulation models from programmatic work

SanFrancisco–Oakland Bay Bridge

National Ignition Facility

Bixby Creek Bridge

Figure 7. Livermore’s detailed numerical models of three large structures are the basis fordeveloping algorithms that analyze acoustic wave vibrations to determine whether thesestructures have sustained damage in an earthquake or other extreme event.

18



help determine the best locations for thesensors.

The monitoring system developed by Livermore engineers willcontinuously record, time stamp, andstore sensor data. Researchers cancontact the system at any time by cellularphone to download the data for analysis.

In developing signal-processingalgorithms, Clark and colleaguesintegrate data from the sensors, modelsof sensor noise and the structure’sunique environment, and a state–spacenumerical model. State–space models

(common in electronics engineering)transform standard finite-elementcomputer models (common inmechanical engineering) to render themmore suitable for signal processing.

The resulting algorithms comparenumerical model simulations withmeasurements of the real structure. Adiscrepancy between the two is a signthat the structure has suffered damage.Future algorithms will attempt todetermine the source of discrepanciesbetween the numerical model and thestructure. Statistical classifiers, including

artificial neural networks, could thencome into play to classify what kind of damage the structure has sustained.

Data were collected recently to test theability of signal-processing algorithms todetect differences between the numericalmodel and the actual structure. Anexperimental structure at the Nevada Test Site, a scale model of a five-storybuilding some 14 feet tall, was the testbedfor verifying these algorithms (Figure 8).

Engineers used the experimentalstructure both to evaluate the sensorsystems and to acquire data for

S&TR May 1999

Mode #1frequency = 5.8 hertz



Single-boardcomputer

Global positioning system

Analog-to-digitalconverter

Cellularphonemodem

Disk drive

Precision timing

Remote communication

Data storage

(a)

(b)

(c)

Figure 8. (a) A scale-modelbuilding at the Nevada Test Sitewas used to evaluate how wellsignal-processing algorithmscould detect damage fromearthquakes or other events.(b) Vibration data could beanalyzed at the experimental siteor downloaded in near real-timeto remote locations via a cellularphone included in the dataacquisition system.(c) Measurements of the scale-model building were comparedwith finite-element models todetermine whether the structurehad suffered damage as a resultof the simulated earthquake. Thefigure indicates the first threenatural modes computed with thecomputational model.

19



evaluating the signal-processingalgorithms. For the latter purpose, theysimulated an earthquake using a smallamount of explosives contained in arubber bladder. They also vibrated thestructure continuously to excite everyfrequency at which it might vibrate.

This summer, a dozen sensors willbe placed on the Bixby Creek Bridge torecord vibrational ground motion fromsmall earthquakes. Placement anddesign of the sensors were guided bythe existing numerical model of thebridge, made by Livermore engineersfor Caltrans to evaluate the retrofittedbridge in a large earthquake.

Also this summer, a remotelymonitored sensor and data acquisitionsystem will be installed at variouslocations on the San Francisco–Oakland Bay Bridge. The system willmonitor both ambient vibrations fromtraffic and wind and the structure’sresponse to ground motion from smallearthquakes. The guiding numericalmodel of the bridge is a product of aLivermore–University of California atBerkeley project (see S&TR, December1998, p. 18). Data from the system willallow researchers to identify thebridge’s “healthy fingerprint” andassess how well the signal-processingtools detect and identify discrepanciesbetween model simulations andmeasured structures.

Prototype instruments are alsobeing placed on NIF for long-termstructural monitoring. The monitoringwill help ensure that the giant laserfacility’s sensitive optical systems canperform under ambient vibration

conditions, which include traffic, air conditioners, and other “culturalnoise” effects, as well asmicroearthquakes. The NIF numericalmodel was made prior to beginning construction of the $1.2-billion laser facility.

The Livermore team hopes thework will provide a structuralmonitoring capability for Caltrans thatcan also be applied to critical DOEsites such as hazardous materialfacilities. In this way, says McCallen,authorities could have a much betterhandle on assessing damage toimportant structures and determiningresponse and upgrade priorities.

Listening to the WorldAcoustic waves permeate the

natural and cultural world. The soundof a heart valve, the acoustic reflection

from a pool of oil, and the vibration of a building are only three examples.But techniques developed for theseresearch projects may well beapplicable to other research fields.“Everything we’re currentlydeveloping will help us to solveproblems in other areas,” says Clark.

—Arnie Heller

Key Words: accelerometers, acousticsignals, algorithms, artificial neuralnetworks, Bixby Creek Bridge, Caltrans,Department of Energy National Gas and Oil Technology Partnership, heart valves,National Ignition Facility, Nevada Test Site, numerical models, oil exploration, San Francisco–Oakland Bay Bridge, signal processing, structure analysis.

For further information contact Greg Clark (925) 423-9759([email protected]).

GREGORY A. CLARK received his B.S. and M.S. in electricalengineering from Purdue University in 1972 and 1974,respectively, and his Ph.D. in electrical and computer engineeringfrom the University of California at Santa Barbara in 1981. Hisresearch activities are in the theory and application of automatictarget recognition, computer vision, sensor fusion, patternrecognition–neural computing, estimation–detection, signal and

image processing, and automatic control. He joined Lawrence Livermore in 1974and is currently a principal investigator for several projects in the Defense SciencesEngineering Division. He has contributed to over a hundred technical publicationsand serves as a reviewer for several professional journals.

About the Scientist

20 Research Highlights S&TR May 1999


O inexorable is the demand for ever-greater computingpower that attempts are being made to exceed Moore’s

Law—that computer performance doubles roughly every18 months. This effort is particularly evident in ultrascalecomputing, also known as parallel multiprocessing. Parallelsupercomputers integrate as many as a thousand processorsinto a single system to achieve calculating speeds up toseveral trillion operations per second (teraops). However,further advances have been thwarted by bottlenecks inshunting data from processor to processor via traditionalelectronic interconnects.

A team of Lawrence Livermore researchers believes it hasfound a way to overcome communications limitations byreplacing the flow of electrons interconnecting the processorsof a supercomputer (or conceivably, the computers of anetwork) with pulses of light of different wavelengths. Bycombining Livermore advances in optoelectronics with off-the-shelf hardware, they are pointing the way tocommunication speed improvements up to 32-fold. Andbecause optical interconnects can be packaged very tightly,additional microprocessors can be added to a supercomputer(or more workstations can be added to a network) for muchgreater overall performance with no decrease incommunication speeds.

The technology development project is called lambda-connect (the Greek letter lambda represents wavelength inscientific notation). Funded by the Laboratory DirectedResearch and Development Program, the project has maderapid progress and resulted in the filing of four patents basedon different aspects of the project. Lambda-connect appears so promising that several supercomputer and computercomponent companies have begun discussions with theLivermore researchers on ways to incorporate the newtechnology into their products. It has also been well receivedby government agencies that need new technologies capableof processing unprecedented volumes of data in as short a time as possible.

According to electronics engineer and principal investigatorRobert Deri, ultrascale computers are essential for theDepartment of Energy’s Accelerated Strategic ComputingInitiative, which is developing capabilities to simulate nuclearweapon performance in lieu of nuclear testing. Ultrascalecomputers are also envisioned for climate and biomedicalsimulations as well as for specialized intelligence andDepartment of Defense missions.

The full potential of ultrascale computers has not beenrealized because standard approaches for sharing data amongtheir many processors have limited their speed. Becausetraditional wire cable connections can carry only one“message” at a time, data become backed up while waiting tobe processed or routed to another processor. These bottleneckssubstantially degrade computational performance, complicateprogramming, and cause inefficient use of memory. Simplyadding additional processors can compound the congestionwithout significantly improving performance.

Performance Gap Is WideningUnfortunately, the performance gap between

communications and other system components is widening.More demands have been placed on communicationscapabilities by more powerful computer boards (with morepowerful processors and more processors per board), fastermemory, increasing use of memory caches and sharedmemory, and new sensor systems that generate enormousamounts of data for processing.

Deri says that users need new tools to break the bottlenecksand meet the increasing communications demands to fully use computing power, memory, and sensor data. Severalsolutions have been offered, but they fail to relieve two keyproblems: inadequate throughput (the rate at which data flows)and high latency (initial time delay in transferring data).

The novel Livermore approach attacks both problems by building on the growing commercial practice of replacingelectrical connections with optical signals of a particular

Pulses of Light Make FasterComputers

Pulses of Light Make FasterComputers

S

Number of 1-gigaop nodes

Electronic

8-wavelength LambdaBus

Per

form

ance

, gig

aops

10 1001

10

100

Lambda-connect transmitter–receiver modules handling only eightwavelengths easily outperform electronic interconnects, particularlyas the number of nodes (processors or machines networkedtogether) increases.

wavelength. Optical signals transmitted along glass fibers arean attractive communication medium because they do notsuffer from electromagnetic interference and other drawbacksassociated with electrical signals.

The Livermore technology calls on wavelength divisionmultiplexing, or WDM, to vastly increase the utility of opticalconnections. Instead of a single wavelength, many differentwavelengths are carried by parallel, multimode glass fibers(MMFs) that are already in use in local area networks. In thatrespect, says Deri, “We don’t need to invent a whole newinfrastructure.” An MMF connection is about six to ten timeslarger than the ubiquitous glass fiber that carriestelecommunications signals. The larger cabling reduces costand improves reliability because it requires significantly looseralignment tolerances.

With lambda-connect, every parallel optical fiber within a cable carries data of different wavelengths, with eachwavelength assigned a destination. Thanks to filters developedby Livermore engineers, data are “source-routed,” with theirwavelength determining the ultimate destination.

Running like an Express TrainIn this way, says Deri, each processor can communicate

simultaneously with a large number of others withoutsignificant increases in cabling or processor complexity.Lambda-connect makes possible optical-fiber “expresschannels,” which like express trains, go to their designateddestinations directly, requiring no electronic routing. What’s

more, the number of processors can be increasedsignificantly for powerful performance boosts with nocommunications delay.

With standard electronics, Deri says, every communicationis like a local train making numerous intermediate stops.Electronic express channels are difficult because they strainthe processing capabilities of electronic interconnects and insome cases require such long cables that electrons cannottravel their lengths effectively.

The Livermore technology achieves unprecedented gains inbandwidth combined with significant decreases in latency.“The fact that all data travel simultaneously solves thebandwidth problem, and the fact that data all travel to differentdestinations solves the latency problem,” says Deri. He alsonotes that the data error rate has been measured at less than10–11, or 1 bit in 100 billion, meaning that the technologytransmits data essentially error-free.

The team is presently developing key components fortransmitter–receiver modules that can handle optical data ofup to 32 wavelengths. To date, their modules can route fourdifferent wavelengths on optical-fiber cabling and areintegrated with standard processor boards.

Achieving the project goal requires significant innovationin filter and microoptics technologies as well as advances invertical-cavity, surface-emitting laser (VCSEL) diodetechnology. Laser diodes and associated electronics inside thetransmitter–receiver modules turn electrical codes into opticalpulses at distinct wavelengths for data transmission to

21Optical CommunicationsS&TR May 1999


792.5–70

–60

–50

–40

–30

–20

–10

0

830

Opt

ical

pow

er o

utpu

t, ar

bitr

ary

units

Wavelength, nanometers858.12

Transmitters with four wavelengths of optical data have been developed; the project’s goal is a transmitter capable of handling 32 wavelengths. Thefigure shows a dual-wavelength transmitter and its optical output. Wavelengths are separated by about 30 nanometers (billionth of a meter).

22 S&TR May 1999


Raj Patel, Rick Ratowsky, Mark Emanuel, Henry Garrett,Holly Peterson, Bill Goward, Claire Gu (from the Universityof California at Santa Cruz), and Rhonda Drayton (from theUniversity of Minnesota).

Deri says that “leaders in the supercomputing field havetold us this approach is the most innovative and highlyleveraged use of optical interconnects they have seen.” Thefirst commercial products incorporating lambda-connecttechnology may appear as early as 2003. Deri notes, however,that adoption of the Livermore approach depends on continueddemonstration of its effectiveness to industry leaders. Theteam has also developed relationships with organizations thatare traditionally aggressive, early adopters of advancedcomputing technology.

Deri expects lambda-connect advances to be used by otherLivermore programs. The surface-mounted laser diodes, forexample, will be used in advanced diagnostics and sensors forphysics experiments. Furthermore, the project is generatinginterest in optical interconnects at the semiconductor chiplevel. However, the greatest impact on Livermore researchprograms will be the arrival of commercial machines usinglambda-connect technologies to boost computer performanceto record heights.

—Arnie Heller

Key Words: embedded systems, lambda-connect, Moore’s Law,multimode glass fiber (MMF), optoelectronics, photonics, vertical-cavity, surface-emitting laser (VCSEL) diode, ultrascale computing,wavelength division multiplexing (WDM).

For further information contact Mark Lowry (925) 423-2924 ([email protected]).

wavelength-encoded destinations. The optical output of theVCSEL lasers is emitted perpendicular to the semiconductorwafer surface (which is the dominant plane shown in thefigure below). This surface-normal emission requires packagemounting innovations over the more conventional edge-emitting laser diodes.

Deri says that the team is well suited to develop therequired components because microoptics, microassembly,and photonics are all Livermore strengths. Team membershave four R&D 100 awards in the area of photonics and over75 years of accumulated photonics experience. The teamincludes co-principal investigator Mark Lowry andinvestigators Mike Larson, Steven Bond, Mike Pocha,

Optical Communications

A cluster of vertical-cavity, surface-emitting lasers (VCSELs) that turnelectrical codes into optical pulses for data transmission. The cluster isa closely spaced, two-by-two configuration that allows coupling into asingle multimode fiber. Twelve of these clusters are coupled into fiber-ribbon arrays that are 12 fibers wide.

23


Each 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 thework done at the Laboratory.

Patents and Awards

Patent issued to

Andrew M. Hawryluk

Gary D. Power

Stephen E. SampayanGeorge J. CaporasoHugh C. Kirbie

George J. CaporasoStephen E. SampayanHugh C. Kirbie

Paul G. CareyPatrick M. SmithJohn HavensPhil Jones

Patent title, number, and date of issue

Forming Aspheric Optics byControlled Deposition

U.S. Patent 5,745,286April 28, 1998

Ground Plane Insulating Coating forProximity Focused Devices

U.S. Patent 5,780,961July 14, 1998

Enhanced Dielectric-Wall LinearAccelerator

U.S. Patent 5,811,944September 22, 1998

Dielectric-Wall Linear Acceleratorwith a High Voltage Fast Rise TimeSwitch That Includes a Pair ofElectrodes between which AreLaminated Alternating Layers ofIsolated Conductors and Insulators

U.S. Patent 5,821,705October 13, 1998

Plastic Substrates for Active MatrixLiquid Crystal Display Incapable ofWithstanding ProcessingTemperature of over 200°C andMethod of Fabrication

U.S. Patent 5,856,858January 5, 1999

Summary of disclosure

Controlled deposition of a material onto a spherical surface of anoptical element to form an aspheric surface of desired shape. Areflecting surface can then be formed on the aspheric surface byevaporative or sputtering techniques. Aspheric optical elements aresuitable for deep ultraviolet and x-ray wavelengths. The reflectingsurface may, for example, be a thin (about 100-nanometer) layer ofaluminum, or in some cases the deposited modifying layer mayfunction as the reflecting surface.

The ground plane of a microchannel plate is coated with a thin layerof aluminum oxide that does not cover its pores, so its performanceis not affected. The thin dielectric coating greatly improves thespatial resolution of proximity-focused image intensifiers. Thephosphor screen can be run at 9,000 volts, compared with 3 kilovoltswithout the coating.

A dielectric-wall linear accelerator comprising a stack of paired fast and slow Blumlein modules. The stack is shaped as a hollowedround cylinder through whose core charged particles are accelerated.To withstand acceleration gradients that can reach 20 megavolts per meter, a novel insulator structure is used to construct a dielectricsleeve that fits tightly into the core. The insulator comprises flatannular rings of fused silica, with thicknesses on the order of1 millimeter, arranged with their planes perpendicular to the coreaxis. At least one metal is deposited and diffused into each of twosides of the fused-silica, flat-annular rings. The rings are fusedtogether into one hollow cylinder by applying enough heat andpressure to weld, braze, or solder the metal-to-metal interfaces.Exothermic multilayer foils can also be sandwiched in the stackunder pressure and then ignited to flash bond the fused-silica, flat-annular rings together.

A high-voltage, fast rise-time switch that includes a pair ofelectrodes in between which are laminated alternating layers of isolated conductors and insulators. A high voltage is placedbetween the electrodes sufficient to stress the voltage breakdown of the insulator on command. A light trigger, such as a laser, isfocused along at least one line along the edge surface of thelaminated alternating layers of isolated conductors and insulatorsextending between the electrodes. The laser is energized to initiate a surface breakdown by a fluence of photons, thus causing theelectrical switch to close very promptly.

Bright-polarizer-free, active-matrix liquid-crystal displays areformed on plastic substrates. The primary components of the display are a pixel circuit fabricated on one plastic substrate, anintervening liquid-crystal material, and a counter electrode on asecond plastic substrate. The pixel circuit contains one or more thin-film transistors (TFTs) and either a transparent or reflectivepixel electrode manufactured at sufficiently low temperatures toavoid damage to the plastic substrate. Fabrication of the TFTs can be carried out at temperatures less than 100°C.

Patents

24


Ronald Natali recently traveled to DOE Headquarters inWashington, D.C., to represent the Laboratory’s HazardousMaterials Packaging and Transportation Safety(HMPTS) Assurance Office at the Hammer Awardceremonies. These awards are given annually by VicePresident Al Gore’s National Partnership for ReinventingGovernment to federal and contractor employees who havecontributed significantly to making government moreefficient and cost-effective. The HMPTS Assurance Officewon as part of a group of DOE contractors, the SuppliersQuality Information Group (SQIG), that shares supplierassessment information to save money by eliminating theneed for each contractor to evaluate the same suppliers.

HMPTS is responsible for making sure all packaging andcontainers purchased by the Laboratory for transportinghazardous materials and waste meet applicable regulatoryrequirements. The office, like all SQIG contractors,contributes to and shares in SQIG’s database of supplierevaluation information gathered from assessment visits. In fiscal year 1998, 36 percent of the HMPTS supplierassessments were done through the SQIG database, resulting

in a significant cost saving for the Livermore program andultimately DOE. SQIG participants are also working tostandardize the assessment process, work beneficial to both DOE and vendors.

Laboratory scientist Grant Logan has received the FusionPower Associates Leadership Award for his nearly 25 years of contributions in both magnetic and inertial fusion energy. The award is presented each year to individuals who have shown“outstanding leadership qualities in accelerating the developmentof fusion.”

Logan, who has been a Livermore employee since 1975, isdeputy director of DOE’s Heavy-Ion Fusion Virtual NationalLaboratory and coordinator of inertial fusion energy technologyfor DOE’s Fusion Energy Virtual Laboratory for Technology—both based at Lawrence Livermore. The citation on Logan’saward states, “Your outstanding leadership qualities, yourinnovative contributions to both magnetic and inertial fusionenergy programs, as well as to fusion power and fusionapplications in general, have provided researchers a rich array of options to explore.”

Awards

Patent issued to

Robert S. Glass

Alex V. HamzaMehdi BaloochMehran Moalem

Richard F. Post

Patent title, number, and date of issue

Urea Biosensor for HemodialysisMonitoring


Process for Forming Silicon CarbideFilms and Microcomponents


Fail Safe Controllable OutputImproved Version of theElectromechanical Battery


Summary of disclosure

An electrochemical sensor capable of detecting and quantifying ureain fluids resulting from hemodialysis procedures. The sensor is basedupon measurement of the pH change produced in an aqueousenvironment by the products of the enzyme-catalyzed hydrolysis ofurea. The sensor may be fabricated using methods amenable to massfabrication, resulting in low-cost sensors and thus providing thepotential for disposable use. The sensor could be used in treatmentcenters for determining the hemodialysis endpoint and in home teststo determine whether dialysis is necessary.

Silicon carbide films and microcomponents grown on siliconsubstrates at surface temperatures between 900 and 1,700 kelvins via carbon-60 precursors in a hydrogen-freeenvironment. Selective crystalline silicon growth can be achieved on patterned silicon–silicon oxide samples. Patterned silicon carbidefilms are produced by making use of the high-reaction probability of carbon-60 with silicon at surface temperatures greater than 900kelvins and the negligible-reaction probability for carbon-60 onsilicon dioxide at surface temperatures less than 1,250 kelvins.

Mechanical means are provided to control the voltages induced inthe windings of a generator/motor. In one embodiment, a lever isused to withdraw or insert the entire stator windings from the cavitywhere the rotating field exists. In another, voltage control and/orswitching off the output is achievable with a variable-couplinggenerator/motor. A stator is made up of two concentric layers ofwindings, with a larger number of turns on the inner layer. Thewindings connect in series. One or both windings can be rotated with respect to the other. The design for the stator assembly ofelectromechanical batteries provides knife-switch contacts that are in electrical contact with the stator windings.

Patents and Awards S&TR May 1999


Leveraging Science and Technology in the NationalInterest

A sampling of current projects at Lawrence Livermor

science and technology to defend our national interests · 2019. 5. 29. · about the cover about...

Documents