ORNL-TM-1445 by Jon Morrow

HOME

a c.

HELP

L-L-

OAK RIDGE NATIONAL L A B O R A ~ Y d’t’5T-; o p e r a t e d by

U N I O N CARBIDE CORPORATION

f o r the

U.S. ATOMIC ENERGY COMMISSION

ORNL- TM- 14-43 COPY NO.

DATE

-i *

April -5, 1966

SIMJIITORS FO3 TRAINING MOLTETJ-SALT REACTOR EXPERIMENT OPEMTORS

S . J . Ball

A3STPACT

Two o n - s i t e r e a c t o r k i n e t i c s simulators were developed f o r t r a i n i n g ogerators of t h e Molten-Salt Reactor Experiment (MSm) i n nuclear s t a r t u p and power-level operating procedures. Both simulators were s e t up on general purpose, portable Electronic Associates, Inc., TR-10 analog comp u t e r s and were connected t o t h e r e a c t o r c o n t r o l and instrumentation system. The t r a i n i n g prograrLw a s successfully completed. Also, t h e r e a c t o r c o n t r o l and instrumentation system, t h e operating procedures, and t h e rod and radiator-door d r i v e s were checked o u t . Some minor modifications were made t o t h e system as a r e s u l t of the experience with t h e s e simulators.

-SED IN

NOTICE

FOR ANNCUTCEMENT

LyucItEAR SCIENCF ASS‘IIAZTS

T h i s document contains information of a preliminary nature and was prepared primarily for internal use a t the Oak Ridge National Laboratory. I t i s subject to revision or correction and therefore does not represent o f i n a l report.

LEGAL NOTICE T h i s report was prepored q s on account of Go*.rnment

sponsored work.

Neither the United States,

nor the Commission, nor ony person acting on beholf of the Commission:

A. Makes any worronty or representotion, expressed or implied, with respect to the accuracy, completeness, or usefulness of the informotion contoined In t h i s report, or thot the use of ony information,

apparatus, method,

privately owned rights;

or process disclosed In t h i s report may not infringe

6. Assumes ony l i a b i l i t i e s w i t h respect to the use of, or for damoges resulting from the use of any informotion. apparotus, method, or process disclosed in t h i s report.

As used in the obove, â&#x20AC;&#x153;person acting on behalf of the Commissionâ&#x20AC;&#x2122;â&#x20AC;&#x2122; includes any employee or controctor of the Commission, or employee o f such controctor. to the extent thot such employee

or contractor of the Commission, or .mpioyee

of such contractor prepares, disseminates,

provides access to, ony information pursuant t o h i s employment

or h i s employment w i t h such controctor.

or controct w i t h the Commission, 4

t; 4

* CONTENTS page Abstract 1.

3. 4.

5. 6.

................................................. Startup (Zero m e r ) Simulator ............................... Power Level Simulator ........................................ Time Required f o r Setup of Simulators ........................ Conclusions .................................................. Appendix ..................................................... 6.1 Details of Startup Simulation ........................... 6.2 Details of m e r Level Simulation ....................... 6.2.1 Neutron Kinetics Equations ....................... 6.2.2 Core Thermal Dynamic muations ................... 6.2.3 Radiator Effectiveness ........................... Introduction

6.2.4

1 j

]

Xenon Poisoning

..................................

5 5

8 13 13 15

15 19 19

21 22 23

5 1.

INTRODUCTION

Two reactor k i n e - - x simulators were developel f o r t r a i n i n g operators of t h e Molten-Salt Reactor Experiment (MSRE) i n nuclear s t a r t u p and powerl e v e l operation procedures. Both simulators were i n s t a l l e d a t the reactor s i t e , and were connected t o the reactor instrumentation and controls system. The operators were trained i n startup, or zero power, operation w i t h t h e simulator i n February 1965 and i n parer-level operation i n October 1965.

Both simulators were set up on general purpose, portable Electronic Associates, Inc., TR-10 analog computers (borrowed from the Instrumentation and Controls Division analog computer pool). No special hardware (other than t h e computers) was required. Although most of the simulation techniques were straightforward, a few special techniques were devised. This report describes t h e two simulators. STARTUP (ZERO POWER) SIMULATOR

The s t a r t u p simulator, s e t up on one 9 - 1 0 analog computer (Fig. l), computed the reactor neutron l e v e l from 10- w t o 1.5 Mw as a function of control-rod-induced r e a c t i v i t y perturbations. The e f f e c t of nuclear power on s y s t e m temperatures was not included.

INPUTS

OUTPUTS

ORNL DWG. 66-4834

RE ACTOR INSTRUMENTATION

N E U T R O N LEVEL!

UL A

FISSION CHAMBER POSITION

R R - 8 100 R R- 8200 CONSOLE METER: CHANNEL 1 CONSOLE M E T E R : CHANNEL 2 SPECIAL: ON CONSOLE

LINEAR FLUX RANGE

Fig. 1. Diagram of Startup Simulator.

6 Tbe inputs t o the sirhulator were signals indicating t h e a c t u a l positions of the control rods, and t h e outputs (indicated on t h e reactor instrumentation) were log count r a t e , period, l o g power, and l i n e a r power. The l i n e a r flux-range input signal was taken f r o m the s e l e c t o r switch on the reactor console. The fission-chamber position readout was provided by a meter mounted on the console. The f i s ion chamber is t h e detector f o r the wide-range counting channel system.' The chamber position i s servo-controlled to give a constant output s i m a l , and the charriber posit i o n i s r e l a t e d t o the log of the nuclear power. The period interlocks and t h e flux control system were a l s o used.

The operators practiced the approach-to-critical experiment (in which p l o t s of inverse count rate vs rod position are used t o extrapolate t o the c r i t i c a l rod position) and rod-bump experiments f o r calculating d i f f e r e n t i a l rod-reactivity worth from measurements of s t a b l e reactor period. The simulator was a l s o used t o check out t h e flux servo controller. Rod position simals were obtained f r o m the three potentiometers normally used by the MSRE computer. The "S" curve r e l a t i n g rod worth and position was approximated f o r t h e regulating rod by a diode f'unction gene r a t o r (Mg. 2). The rod worth vs position relationship f o r t h e other two rods w a s l i n e a r .

k . E . B a l l e t al., MSRF: Design and Operations Report, pdrt V, Reactor Safety Analysis Report, ORNL-TM-732 (August 1964), pp. 96-98. 1

ORNL IWG. 66-4835

1.0

I-

0.8

0.6

0 0.4

n 0

a -I

e LL

I-

0.2

DISTANCE WITHDRAWN (IN.)

Fig. 2.

Simulator Approximation of Regulating Rod Worth vs B s i t i o n .

7 The analog c i r c u i t used t o compute r e a c t i v i t y from t h e three rod posit i o n s included the e f f e c t s of the position of one rod on the t o t a l worth of t h e others ("able 1).

Bible 1. Full-Scale â&#x201A;Źbd Worths

Pos it ion

Rod

Regulating Rod

Shims out Shims i n

Both Shims

Regulating rod out Regulating rod i n

Full Scale Rod Worth (Q W K )

2.6 1.3 5.8 4.5

Reactivity vs Position "S" curve I'

S" curve

linear linear

The neutron l e v e l computation was made by converting the k i n e t i c s equations t o logarithmic form? since the neutron l e v e l varied over eight decades. 'lho effective delayed-neutron precursor groups were used. The usual method of including the source term i n these equations was found t o be unsatisfactory, and a special c i r c u i t was used (see Sect. 6.1). The conversion of log power t o l i n e a r power wag approximated by using a squaring device t h a t gave adequate accuracy over each l i n e a r (1.5 decade) range (Fig. 3 ) . A voltage signal from the reactor instrumentation l i n e a r 2A.E. Rogers and T.W. Connolly, Analog Computation i n Ehgineering Design, pp. 334-7, I k G r a w - H i l l , New York, 1960.

POWER SIGNAL FROM LOGARITHMIC CALCULATION

Fig. 3.

Approximate bg-to-Linear Conversion.

8 range selector c i r c u i t was subtracted from the l o g power signal, and t h i s difference was then converted t o the l i n e a r signal. The equations and analog computer c i r c u i t used f o r t h e s t a r t u p simul a t o r are given i n Sect. 6.1

POWER U:vEL SIMULATOR

The power l e v e l simulator, set up on two TR-10 analog computers (Fig. 4), simulated the k i n e t i c behavior of t h e MSRE f o r power levels

ROD PosiTioNs~

RADIATOR DOOR POSITIONS RADIATOR AIR AP REACTOR CONTROL MODE

1 -e

u -

“ L A ? - T -

0 -

NEUTRON LEVEL: LINEAR POWER LOG POWER LOG COUNT RATE PERIOD FISSION CHAMBER POSITION REACTOR INLET TEMPERATURE

REACTOR OUTLET TEMPERATURE RADIATOR SALT OUTLET TEMP.

RR-8100 RR-8200 CONSOLE METER: CHANNEL t CONSOLE METER: CHANNEL 2 SPEC’AL :ON

T R -202 -A 5

T I -202-A2

RADIATOR SALT A T

Td 1 2 01-A

RADIATOR HEAT POWER FLOW (CONST) X AT

X p R-201-A

Fig. 4.

Diagram of Parer Level Simulator.

between 0.5 and 12 Mw. The inputs were signals indicating the a c t u a l positions of t h e rods and the radiator doors and t h e a c t u a l pressure drop of t h e cooling a i r across the radiator. The outputs were neutron l e v e l s and temperatures. The usual nuclear information and key system temperat u r e outputs were indicated on the reactor instrumentation. The reactor power-level servo controller and radiator load control systems were a l s o used. m e r e a c t i v i t y inputs from control-rod position signals were computed as i n the s t a r t u p simulation. The neutron l e v e l computation (using two delayed-neutron precursor groups ) solved t h e linear, rather than logaranges rithmic, k i n e t i c s equations. Only the 0 t o 1.5 and the 0 t o 1 5 on t h e reactor l i n e a r power channels were operational. Conversion from l i n e a r t o l o g power was approximated using a square-root device (Fig. 5 ) .

.I675

=O-lOMW

Q.40 t I O V :

= SJUULATOR APPROXJMATION

DESJRED OUTPUT

COMPUTED LINEAR POWER -0

Fig. 5 .

(Mw)

Approximate Linear -to -Log Conversion.

10 Other r e a c t i v i t y inputs t o the power l e v e l simulator were f r o m computed xenon poisoning, noise, and f u e l and graphite temperature changes. The xenon-poisoning computation (Fig. 6 ) was included as an option. In consideration of t h e long time-constants of xenon buildup and decay, the equations were time scaled t o run a t t e n times real time. ORNL DWG. 66-4839

P IODINE

- b X e (FUEL)

DECAY 1 >*=2.9XlOJSEC-'

) .X=(CRAPHITE)

STRIPPING (DOMINANT) DECAY BURNUP

xF(ToTAL) >>,,I

DECAY Ax = 2.1 X lo-' SEC-' BURNUP

1.66X10-6 PSEC-'

Steady- State Xe Poisoning When P = 10 Mw: 6 K Fuel = -0.7% 6 K Graphite = - 0.79 5'0 Fig. 6. 'Ihe r e a c t i v i t y of usual simulators t h e noisy output of resistance feedback

Diagram of Xenon Poisoning Computation. noise input w a s included t o o f f s e t complaints t y p i c a l about how "smooth" t h e flux output i s compared w i t h a c t u a l reactors. An operational amplifier w i t h high (40 megohms) w a s used a s the noise source.

A simplified simulation of t h e thermal k i n e t i c s of the MSRE w a s used which was based on previous studies of reactor dynamics.3 The core was represented by two f u e l "lumps," or nodes, and the graphite by one. Six more lumps were used t o represent t h e r e s t of t h e system. The thermal c h a r a c t e r i s t i c s a r e summarized i n Table 2.

The heat removal rate f r o m the r a d i a t o r i s controlled by varying t h e a i r flow through t h e radiator; hence, the radiator s a l t o u t l e t temperature i s affected by s a l t i n l e t temperature, a i r i n l e t temperature, and a i r f l o w r a t e changes. A simple but f a i r l y accurate way of simulating the heat removal i s t o make use of t h e relationship of radiator cooling "effectiveness" a s a function of a i r flow r a t e . Cooling effectiveness i s defined as t h e r a t i o of the a c t u a l temperature decrease of t h e hot f l u i d t o t h e temperat u r e decrease i n an i d e a l ( i . e . , i n f i n i t e heat-transfer surface) heat exchanger:

3S.J. B a l l and T.W. Kerlin, S t a b i l i t y Analysis of the Molten-Salt Reactor Experiment, ORNL-TM-1070 (Dec. 1965).

Table 2.

MSm Thermal Characteristics Used i n the Power k v e l Simulator

7.6 (two

Core t r a n s i t time, sec

lumps ) Graphite time-constant, sec

200.0

a Heat exchanger t o core t r a n s i t time, sec

10 .o

Core t o heat exchanger t r a n s i t time, sec

6.67

Radiator t r a n s i t time, sec

6.67

a Radiator t o heat exchanger t r a n s i t time, sec

10.0 (two

lumps )

a Heat exchanger t o radiator t r a n s i t time, sec

5 -0

Heat exchanger "effectiveness" factors a t steady stateb: rn I

T P=

T = 0.7029

rn I

T Pl

= 0.4478

Tsi Tso

Tsi

= 0.2971

= 0.5522

Holdup time i n heat exchanger i s included i n t h e other t r a n s i t times.

P, primary; S, secondary; i, i n l e t ; and

outlet.

Tsalt i n

- 'salt

out

EC 4

The s a l t o u t l e t temperature i s computed from

'salt

out

Tsalt i n

Ec(Tsslt i n

- Tair i n 1 '

The calculated cooling effectiveness as a f'unction of a i r flow r a t e and the l i n e a r approximation used i n the simulator are shown i n Fig. 7.

ORNL DWG. 66-4840

COOL ING EFFECT IVENESS = TSALTIN- SALT OUT

TSALT IN - TAIRIN

0.05 (3

Z -

A PPROXfMATlON

0 0 U

COOLING AIR FLOW RATE Fig. 7.

(Yo)

MSRE Radiation Cooling Effectiveness vs A i r Flow Rate.

The a i r flow rate W

through t h e r a d i a t o r w a s computed from

a' where

K = constant adjusted t o give 10 Elw cooling a t f u l l a i r flow, APa

%,%

= measured a i r pressure-drop s i g n a l across t h e radiator, = measured r a d i a t o r door positions (inches raised).

Conversion of t h e analog computer voltages representing temperatures t o signals compatible w i t h t h e Fbxboro ECI instruments was done w i t h straightforward resistance divider networks. The equations and analog computer c i r c u i t used f o r t h e power l e v e l simulator are given i n Sect. 6.2.

TIME R E Q U m D FOR SETUP OF SIMULATORS

IIhe engineering and c r a f t time required t o develop, i n s t a l l , and check out t h e simulators and t o t r a i n t h e operators i n t h e i r use w a s as follows ( a l l values i n man-weeks):

Startup Simulator

Fower Level Simulator

Development

1.6

1.4

Set up and check out

0.7

1.2

Lecturing on use

0.3

1 .o

0.3

0.4

Engineering Labor

Craft Labor

Installation

4.0

Total

5. CONCLUSIONS The two on-site t r a i n i n g simulators were developed and operated satisf a c t o r i l y as p a r t of t h e MSRE operator t r a i n i n g program. Besides t h e obvious function of t r a i n i n g the operators, t h e simulators served as a

14 means of checking out the reactor instrumentation and control system, the operating procedures, and the rod and radiator-door drives. Some minor modifications w e r e made t o the system as a r e s u l t of t h i s experience w i t h t h e simulators.

All manipulations required t o operate t h e simulated reactor were done from the reactor console, and t h e readout devices were p a r t of the standard reactor instrumentation.

6 . APPENDIX

6.1 Details of Startup Simulation The neutron k i n e t i c s equations are , b

5 dt

i* [ k ( l - 4 ) -

hiCi

+ S

where

= neutron population,

= t i m e , sec,

1* = prompt neutron lifetime, sec, k

& pi 'i

= reactor multiplication, = t o t a l delayed neutron fraction,

= e f f e c t i v e delayed neutron fraction for i t h precursor group with fuel s a l t circulating,

= decay constant f o r i t h precursor group,

ci S

it' precursor population,

= rate of neutron production by source.

R e w r i t e Eqs. (1)and ( 2 ) , assuming kpT M

-- "Pi - hiCi dt 1*

dCi

Divide EQs. ( 3 ) and (4) by n:

and

--1* - nBi I* -

bf3-i

- B,

- 6k

1dn

Gat

+I6

'iCi

(3' )

i=1 1 dCi-

--n dt

AiCi n

(4'

Define new variables: dn

M = - -

reciprocal period,

w = -S n

kSubstitute i n t o a s . ( 3 ' ) and (4'):

6 (5) dVi

Pi - - - p dt

1.V.

- mi

%e usual method of computing the source term i s as follows: noting t h a t Wn = S and

therefore

or The analog computer can usually solve a f i r s t - o r d e r d i f f e r e n t i a l equation such as EQ. (7) f o r W; however i n t h i s case, W becomes so small when n >> S t h a t the voltage representing W i s within t h e noise l e v e l of t h e amplifier, s o f u r t h e r computation w i t h it i s meaningless. To avoid t h i s problem, t h e relationship between W and log n was approximated as shown i n Fig. 8.

j i

:: k 5

V k al

1.4

rl I

f A0

4 )

I; I? I,

..-

-r cl

' H E

xi L

P 03

19 6.2 6.2.1

Details of Power Level Simulation

Neutron Kinetics Equations

Equations (1)and ( 2 ) of Sect. 6.1,with two delayed-neutron precursor groups, were used. An analog c i r c u i t (Fig. 1 0 ) developed many years ago5 w a s used t o solve these equations. This c i r c u i t i s superior 30 most of those published i n t h e l i t e r a t u r e , mainly because of t h e way i n which the amplitude scaling i s accomplished. A key point i n the scheme f o r simulating the equations i s the use of a small feedback capacitor f o r t h e integration of the neutron l e v e l equation, r a t h e r than solving d i r e c t l y f o r dn/dt and t h n integrating with a conventional large-feedback-capacitance integrator In Fig. 10, amplif i e r 1 (which solves f o r n ) has a feedback capacitor of 10 I* wf. The amplifier gain i s 1/10 1* Rin(sec-l), where Rin i s i n megohms. With the assumption t h a t a l l input r e s i s t o r s a r e 0.1 megohm, Eq. 1 can be rearranged t o show t h e desired form of the inputs t o amplifier 1, as follows:

dn 1

dtF

(kn

- knp, - n

l*AIC1

1*X2C2).

(8 1

The quantity kn i s generated from n and 6k as shown i n Fig. 10. Typically k w i l l vary between 1.005 and 0.98 f o r control studies. Owing t o the inherent inaccuracy of the multiplier, it i s advantageous t o l e t the f u l l scale output of t h e (6k x n ) multiplier be only a few percent of kn. In the simulator, t h e voltage representing zero 6k w a s o f f s e t , i . e . , -1.5% < 6k I; +0.5$, because of the apparent deadband i n the quarter-square multiplier when one input oprates around zero v o l t s . The quantity kn& i s generated from 0.1 kn

100 1

pT J

pot 2 s e t t i n g

0.1 U

1 megohm input

t o amplifier 1 the gain reductions

thus allowing a reasonably large gain s e t t i n g on pot 2.

The 1* AiCiterms a r e obtained by first taking the Iaplace transform of EQ. 2;

which rearranged i s

5By E.R. Wnn (deceased), Instrumentation and Controls Division.

.ORNL DWG. 66-4843

-6K

.I kn

I I L

1 2

1 i

-10x2 OTHER D E L A Y GROUPS 4 AS REQUIRED

NOTE! RUPL/F/€R GAIN OF ? IMPLIES ?-MEGOHM INPUT RESISTOR; W I N OF ?O= O.?M

Fig. 10. Analog Circuit Designed by E.R. Mann for Neutron Kinetics Equations.

where S i s t h e Laplacian argument. Solving f o r the output of integrator 4 [e(4)]: de (4 )

= 9 4 )

Multiplication of e

(4 1

- -5”(4)

0.1 hi kn

by 100 pi gives -10

which i s seen from Eq. ( 9 ) t o equal -10 l*A.C. as required f o r generating dn/dt i n Eq. (8). Again, because the ampltfier gains were reduced, the gains on t h e pi pots could be increased.

Tis c i r c u i t

-4

c l e a r l y shows t h a t f o r small values of l* (e.g., 10 4 1 0 ’ s e c ) t h e feedback capacitor f o r amplifier 1 w i l l be very small and thus w i l l have a negligible e f f e c t on t h e response of n f o r t h e slow variations normally encountered i n control studies. Under these condit i o n s t h e negligible e f f e c t of t h i s capacitor implies that t h e neutron k i n e t i c s are independent of 1*, and f o r m precursor groups, the neutron k i n e t i c s can be described by m d i f f e r e n t i a l equations, r a t h e r than (m + 1) equations. This simplification i s useful when the k i n e t i c s equations are solved on a d i g i t a l computer, because the maximum computation time inverval is usually governed by the l*/& time constant and must be made q u i t e small t o give stable (and accurate) answers. 6.2.2

Core Thermal Dynamic Equations

The f u e l f l o w i n the core is approximated by two f i r s t - o r d e r lags i n series, and heat t r a n s f e r takes place between the f i r s t f u e l lump and the graphite. The nuclear importancesof t h e two fuel lumps are equal. Fortyseven percent of t h e nuclear heat is generated i n each f u e l lump. ’Ihe remaining 6$ i s generated i n the graphite. The heat balance equations used f o r t h e core are as follows:

F i r s t f u e l lump dFc

-d t-

- 0.263

0.017

FG+

0.246 Tci + 0.0329 n;

Second f u e l lump

- co- - - 0.263 at

U A

Graphite dTG - - -

Tco

0.263 Fc

- 0.005 FG+ 0.005 Fc

0.0329 n;

0.00084 n.

Temperatures are i n OF, time i s i n seconds, and neutron l e v e l n i s i n megawatts. As discussed previously, the lags due t o holdup and heat t r a n s f e r i n the loop external t o t h e core were represented by six f i r s t - o r d e r lags. Each l a g i s described by t h e equation

a x -- 1- (xin at T

- Z) ,

where T i s the time constant of the lag.

6.2.3

Radiator Effectiveness The p l o t of radiator cooling effectiveness vs a i r flow w a s calculated

Tsalt i n Tsalt i n

- Tsalt out --

Tair i n

- exp

[-(1

Nl)N2]

N1 exP‘ [-(1

- N1)N2]

where “‘p

N1 =

)salt

(WCp)air

= mass flow rate, lb/sec, = specific heat, Btu/lb-OF,

U = overall heat t r a n s f e r coefficient, Btu/sec-ft

A = heat t r a n s f e r area, ft2.

2 0

d’

Since a i r flow i s perpendicular t o t h e tubes, the heat t r a n s f e r coefficient on t h e a i r side was assumed t o vary as t h e 0.6 power of flow r a t e . 6.2.4

Xenon Poisonina

Even when time scaled by a f a c t o r of 10, the xenon t r a n s i e n t s a r e very very slow, and care had t o be taken t o avoid large e r r o r s due t o integrator d r i f t . Manual d r i f t - c o n t r o l pots were added t o both integrators i n the circuit. Fuel xenon was assumed t o build up a t a r a t e equal t o iodine production, since the xenon stripping time-constant i s small compared w i t h those for decay, burnup, and diffusion t o the graphite.

Since simulation pressed t h e limitations of t h e accuracy of the computer, some of the coefficients had t o be f i e l d s e t t o give proper steadys t a t e output values. Although there would have been a number of advantages i n speeding up the computation even more, it was not done because we d i d n ' t want t o have t h e xenon dynamics confused with t h e reactor thermal dynamics. Fig. 11 i s the analog computer c i r c u i t used f o r the power l e v e l simulator.

c!'

Fig. 11. Analog Computer Circuit for t h e F d e r Level Simulator.

ORNL-TM-1445 INTERNAL DISTRIBUTION

1. R.G. Affel S.J. m i l S.E. Beall E.S. Bettib 10. R. Blumberg 11. E.G. Bohlmann 12. C . J . Borkowski 13. G.A. Cristy 14. J.L. Crowley 15. F.L. Culler 16. R.A. b n d l 17. J . R . Engel 18. E.P. Epler 19. D.E. Rrguson 20. C.H. Gabbard 21. R.B. Gallaher 22. A.G. Grindell 23. R.H. Guymon 24. P.H. Harley 25. C.S. Harrill 26. P.N. Haubenreich 27. V.D. Holt 28. P.P. Holz 29. T.L. Hudson 30. P.R. Kasten 31. R . J . Ked1 32. T.W. Kerlin 33. S.S. Kirslis 34. A. I. Krakoviak 35. M . I . Lundin 36. R.N. won 37. C.E. Mathews 38. H.G. WcPnerson 39. W.B. McDonald

2-7. 8. 9.

40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

53. 54. 55. 56. 57. 58. 59. 60. 61-62. 63. 64-93. 94. 95. 96-110. 111.

112-113. 114. 115.

116. 117.

H.F. McEuffie C.K. McGlothlan H.R. Byne A.M. Perry H.B. Piper B.E. Wince M. Richardson H.C. Roller H.C. Savage D. Scott H.E. Seagren M . J . Skinner A.N. Smith P.G. Smith R.C. Steffy R.E. Thoma G.M. Tolson W.C. Ulrich B.H. Webster A.M. Weinberg MSRP Director's Office, Rm. 219, 9204-1 Central Research Library Document Reference Section Laboratory Records Department Laboratory Records, ORNL R.C. ORNL Patent Office Division of Technical Information Extension Research and Development Division, OR0 D.F. Cope, OR0 R.G. Garrison, AEC, Washington H.M. Roth, OR0 W.L. Smalley, OR0 M . J . Whitman, AEC, Washington

EX"G DISTRIBUTION

118. S.H. Hanauer, Univ. Tenn. 119. R.F. Saxe, North Carolina S t a t e Univ. 120. S.E. Stephenson, Univ. Ark.