ML20052H874
| ML20052H874 | |
| Person / Time | |
|---|---|
| Site: | Big Rock Point File:Consumers Energy icon.png |
| Issue date: | 05/17/1982 |
| From: | Schwartz A BIER, MILLS, CHRISTA-MARIA, ET AL, MICHIGAN, UNIV. OF, ANN ARBOR, MI |
| To: | |
| Shared Package | |
| ML20052H869 | List: |
| References | |
| NUDOCS 8205240109 | |
| Download: ML20052H874 (7) | |
Text
ff fff f f
/ f ff ff/
d91 YM DEPARTMENT OF MATHEMATICS ANN ARBOR, MICHIGAN 48109 i
347 WEST ENGINEERING BUILDING
.n
..,,. (313t 764-6485 G :.
- l <C U l d llj TESTIMONY OF DR. ARTHUR SCIIWARTZ 1 lu y 17,1982 4
liv l i e nie in Ar tliur J.
dc hi nertz.
I con at Ausociate Pr otensor o t' ilull evnial t c d et lhe Ltiversity oi l li tl ei a es i. Tlie bull. ot in y Leuclii ng t unui L L t ut e eni or-e,id graduale level tuurneu in upplied nia t h ensa t i c s.
Ilr rutent consultinq < rid r eseer cli acti vi ty lius oeen in the area of randoen pr ocesu theory (a brunch at probability theor// cpplied to iiidustr i ul pr obl enis.
During tiie per i od 197 ce-1930 I did r esearch on risk asueebbenent und on uc c u s t. un Leetified bef or e the U.
5.
Seoule and lluliursel Instituus of Heu1 Lb concer ning the.epplication of r i si:
u n bes sinen t to recombinunL Dlm research. In 1979 I testified before a Joint Seoule/Huuse conuni L lev of the lii chigen State Louislulur e concerning the ri di: of eccidents ire nuclear power plants Oli llu y Olh I uas est ed Lu review the r ist unulyseu of circraft cr osi tu ut Big i;oc k Nut l t er Foucr Plant. ~Ilie necusery documents arr i ved ute l lu v ]>llli. Sa tice Llien I lie v" devoted nio s t.
oi iny t i nie Lo e evi ewilig Li n um. I have had not had surficient L i ine to de<elop a r i sl. ussesmut>n t ut in y oun. I uin nol certain Lhut e reliable Lt clinical u s e u s en.un t in ieunible withoul greal effort und conside'uble L i ine. Therefore 1 have 1 i nni Led an y ei f or ts lo r evi ew.
I uhall e:: p l ei n in Lhe following pages why I believe that the unul vses perf arnied by
- t. h e Air Force Officers and consult.snts are Luthr I cel l y inadequale.
820S240109 820518 PDR ADOCK 05000155 C
Page 2 Documents Reviewed.
1.
Risk Analysi s of a Catastrophic Event at Boyshore Strategic Training Range. Date stamped 2 Jan 1980. Reference to 11aj. Betourne. Signed by Col. Heinrich.
2.
Testimony of K.
M.
Campe. (Docket IJo. 50-155) 3.
Analysis of Richard J.
Camp (tiemor andum to Col. Clark, det.ed 26 April 1971) 4.
Tuulimony of Captain William P.
Hickey, Jr.
(Docket No. 30-155-OLA) 5.
Deposition of Maj. Gary Betourne and Mr. Clayton Thomas (Docket no.
30-155-OLA)
Duckground Canisderations Assessing the prababi1ity of a rare event ie always difficult for al least two reasons.
l (u) Intuition is a poor guide. We f r equentl y ignore probabilities I
as high as 1/10,000/ year since we face risks of death of that order over which we have li ttle con tr ol.
(b) An event with probabi1ity 1/10,000 may very wel1 not occur at all in 10,000 repeti tions of the trial or even in 20,000 such repetitions (the probability of the latter is over 10%). Such events may occur with even greuler f r equency Lhon expected. Conversely, estimating the probability of a rare event by observing the relative frequency of ils occurence is highly unreliable.
In order to assert thol a probability is less than 1/1,000,000/ year il iu therefore imperative that a rigorous analysis be made.
Conwiderable effort must be exerted to consider all factors and their relaLionships to the fullest possible degree. Simpl y overestimating some terms does not guarantee that the result will overestimate the final probability unless we are certain that no other terms or relationships are grossly underestimated (or overlooked).
Findings The documents I reviewed do not meet the standard of rigor I believe is necessary to establish that the probability of an aircraft crashing into the Big Rock Plant is less than 1/1,000,000/yeur for the following reasons:
(a) The r ange of accident scenarios is limited. It is not at all clear lhet every effort has been made to envisege all possible accident sequences that might lead to the catautr ophe under consideration. It must be remembered that a sequence whose probability is on the order of 1/1.000,000 is very remote but still significant to this calculation.
It thus becomes necessar y to consider poorly trained, inebriated or mesiLal1y unstable personnel. Likewise iL is necessary to consider very unusual end sudden changes in the weather. It is also necessary to consider bizarr e equipment mal f unctions.
(b) Ther e is in oli the analyses the enplicit assumption that the Uiq Rock Plant in neither more nor less likely to be hit by aircraft I
d Page 3
)
\\
than any other piece of real estate of the same size in the Bayshore Range. This is not obviously true since the plant is located ver y near the larget area and the point where the planes reverse direction.
Pr esuaiab l y the probebility of mishap may be influenced by these iactors.
(c) The problem of dependency (common mode f ailur e) is siot adequately tr ealed. Ib j or Belourne's statement (page 83 of deposition) that mul tipl ying hi s es t i ana t i on by 100 would correct any underestimation due to ignorisig dependency is not accepteble. He has inul tipl ied four numburu all of wh i cl e are small:
P(RF) x P(GE)
P(C) a 3/430 Tlie offact of dependency could be (hypotlietically) to replace up to three probebili ties by condi tional probabilities close to unity. Even if the two largest probabili ties were replaced in thi s enanner the final probability would be increased by a factor of 7,000. In general wi nen the probability is calculated by multiplying several small terms toyelher under the assumption of independence when in fact dependence does e :ist, il is possible lo have an error of sever al orders of magnitude. I cannot estimale how much dependency there is between the variables. I stronyly suspecL there is same. There is no a priori reacon to assume Llial it would not inflate the result by more than a f actor of 100.
(d) The analyses all seem to model the catastrophe as consisting of a normal fliyht with some rendom perturbation as it passed through a cri Lical.one such l.h a t this random per turbation was large enough La result in the crash. That is to say: the flight is normal in every respect u,: c up t that the etircraf t crashes into a ncclear power plant.
Seriouu, uccidents frequently involve greater and more complex departures f r om normality. The unelyses under r evi ew er e far too siinple La be convince me that they have considered the depth, variety and compl ex i t y of conditions that might lead to the type of catastrophe under con cui der ali os s.
Respectfully submiLted, l
M Arthur J.
Sunwertz Associate Professor l
Department of Mathematics l
Ti t e Uni ver si ty of fli c h i g an l
Ann Arbor, Michigan l
{
l l
i l
Arthur J.
Schwartz CURRICULUM VITAE Professional Experience:
Instructor of Mathematics, Wayne State University, 1960-61 Instructor of Mathematics, Princeton University, 1961-63 Research Associate and Assistant Professor, Columbia University, 1963-65 Assistant and Associate Professor, University of Michigan 1965-Invited Participant in Special Year in Topological Dynamics at the University of Maryland, 1970-71 Education:
B.S.
Eng.
Wayne University 1954 M.A.
Wayne State University 1958 Ph.D.
Wayne State University 1960 Publications:
1962 - Certain Conditions for the Existence of Invariant Linear Functionals, Proc.
A.M.S.
1963 - A Generalization of a Poincare-Bendixson Theorem to Closed Two Dimensional Manifolds, Am. J.
Math.
1965 - Common Periodic Points of Commuting Functions, Math. Math.
J.
1965 - Limit Sets of Foliations (with R.
Sacksteder), Ann. de L' Institute Fourier.
1968 - Flows on the Solid Torus Asymptotic to the Boundary, J.
Diff.
Eqns.
1968 - Poisson Stable Orbits in the Interior of the Solid Torus, in Topological Dynamics, an International Symposium (Auslander, Gottschalk, Eds.)
1970 - The Depth of the Center of 2-manifolds, (with E.
Thomas),
Proceedings of Symposia in Pure Mathematics, XIV.
1972 - Systems of ordinary Differential Eqns., (with Jack Goldberg),
liarper and Row.
Curriculum Vitae Arthur J.
Schwartz Page 2 Courses taught ordinary differential equations.
Stability and oscil-lation theory.
Operational methods.
Transforms and-distribution theory.
Complex variables.
Linear algebra and matrix methods.
Applications to statistics and geometry.
Linear programming, optimization and network theory.
Computer programming and computer science.
Numerical analysis.
Probability theory and stocahstic processes.
Applications to signal image processing.
Dif ferential Geometry.
Administrative experience I am acting coordinator and one of the organizers of the Energy Studies Program at the University of Michigan.
I am co-organizer with Prof.
B.A.
Taylor of the lecture series Mathematics in Industry.
Consulting experience.
F.
Joseph Lamb Co. (K. Michael Day)
]
Air flow Sciences Corp. (Robert L. Gielow)
Background
I have worked as a draftsman and a designer for industrial l
oven manufacturers.
I was chief engineer of an army radio station.
I keep abreast of developments in microcomputer and maintain my l
own Apple system.
h I
1
b CERTIFICATE OF SERVICE I hereby certify that_the attached Motion for Leave to File Testimony and Testimony of Dr. Schwartz was served on the attached list by United States Mail, first class-postage prepaid, on the CN
/6 day of May, 1982.
<~ a i
1
?
Ate.ic Safety and Licensing JerspF Colle, F. squire E:ard Panel Ishe, Lintn3n and Beale 3
l'. S. Nuclear Regulatory 1120 Conne :icutt' Ave, N.W.
Corrissien Suisc 325 Washingten, D.C.
20555 Washington, D.C.
20036 Peter B..Bloch,
. Esq., Chairman Atomic Safety and Licensing
),-
Board Panel
.G*
U.S. Nuclear Regulatory l
.],
Commission
- R.
Washington D.C.
20555
- K'
\\
'l '
Dr. Oscar H. Paris i-Atomic Safety and Licensing Board Panel U.S. Nuclear Regulatory Docketine and Service Section Commission office of the Secretary Washington D.C.
20555 U.S.
Nuclear Regulatorj connission Fir. Fredrick J. Shon Washingten, D.C.
20555 Atomic Safety and Licensing John O'Neill, II Bodrd Panel Reute 2, Bcx 44 U.S. Nuclear Regulatory Haple City, MI 49664 Commission Washington D.C.
20555 Janice E.
Moore, Esc.
Counsel for NR; Staf f U.S. Nuclear Re gulatory Corrission Washington, D.C.
20555
-ss..> :
,E Y ~ Y 0' l'
i-
~
l 4 : L_.: e _,. +.
e f
i-
.e O
e e