Thermodynamic investigations of passive decay heat removal from HTR cores and component behaviour

Nuclear Engineering and Design 121 (1990) 211-218 North-Holland

211

T H E R M O D Y N A M I C I N V E S T I G A T I O N S OF P A S S I V E DECAY H E A T R E M O V A L FROM ~ CORES AND COMPONENT BEHAVIOUR W. REHM, J. ALTES and H. B A R T H E L S Forschungszentrum Jiilich GmbH, Jiilich, Fed Rep. Germany Received December 1989

This paper describes core heatup transients for the THTR-300 and HTR-Module and the knowledge obtained with the experimental facility LUNA-HTR for the study of natural convection characteristics of different HTR primary circuits. In consideration of thermodynamic similarity principles, transients with heat removal via natural convection from a pebble-bed into the primary circuit were simulated, as well as calculated with the code THERMIX-CRAY-2D. On the basis of experiments, the code is furthermore used for the analysis of natural decay heat dissipation in the depressurization case of the primary circuit. Studies on HTR safety have indicated the great significance of processes resulting during temperature stressing of the concrete pressure vessel for the sequence and consequences of accidents, particularly those with unrestricted core heat-up. In the course of the accident the vessel is slowly heated up to very high temperatures. Over the last ten years the behaviour of PCRVs of different HTR (THTR, HTR-500) plants during core heat-up accidents has been analyzed. Besides the PCRV itself, also the top reflector suspension is experimentally investigated.

1. Thermodynamic characterictics of passive decay heat removal transients for the THTR-300 and HTR-Module power plant Gas-cooled high temperature reactors (HTR) with pebble-bed cores are being designed in Germany based / totalparticle 2500 ~ fCk,'e

|

.

.

.

._,I.

o

I

. . . .

.

/ 2000~

L

/(/ release

/

/

on experience from the AVR and THTR. In contrast to the medium-sized HTR, the small H T R does not need a separate afterheat removal system due to the inherent temperature stabilization below 1600°C which can be performed without active core cooling systems. However, the medium-sized HTR also shows a great safety potential for passive afterheat removal processes in core cooling accidents, particularly as long as the reactor

b) del~'essBiz~ T..

max

I

~soo--------/-

1400"

dewessurlzagon [ best estimte)

1200-

- - with

] vessel

~ 1000-

E

•

a

800-

a 000. ation o

~

0

--V///////////////Z,~ . . . .

0

1(30 Time [ h ]

~ 41111.

Liner cooling

200

21111.

Concrete Outw Surface

>

Fig. 1. Curves of the maximum core temperatures and average gas temperatures in the case of core heat-up accidents of the THTR with intact liner cooling system of the concrete reactor pressure vessel: (a) RuD: reactor pressurized (40 bar); (b) DES: reactor depressurized (from 12 h onwards from 50 to 1 bar after 75 h) (best-estimate for accident management study).

0

0

'

26o

460 Time [h)

eoo

e6o

lOOO

=

Fig. 2. Comparison of the temperature transients in the primary circuit for depressurization with and without an operating vessel cooling system (best-estimate for worst case study).

0 0 2 9 - 5 4 9 3 / 9 0 / $ 0 3 . 5 0 © 1990 - Elsevier Science P u b l i s h e r s B.V. ( N o r t h - H o l l a n d )

212

IV. Rehm et al. / Heat remoual for H T R cores

POS

!

:

: I ]Pos.

~

2

,i i i ,~ I L-jL-J , .'~-~

r-----r--q>~i

/

I

+7,00

,

oul COC I

o o°

c0oter

~ J ~IIo~

i~o°

elec_~,I' frical I heater I i_ _c,.o.._ L

:

,

.31s

,

~

!!

bed

.

.

~

,direction ~" of t~ flow

L (~[

L___~__,F__~_

....

turbo

•

*2

~ '

,,

50

L_.J

"

'

,

"

', I

i

' "

•

.

--- j p as 4.

~;:---~

,

k~pebble

.

, it_ / I

~[

"~

tO,O0

i

compressor

main loop warm-up loop

Fig. 3. LUNA-HTR, the experimental facility for simulating natural convection (Pmax= 40 bar, Tma~ = 300 °C).

remains under pressure. Computed results confirm that the natural convection in a pressurized reactor can be exploited as an effective heat transport mechanism from the core to the heat sinks [1]. The optimized THERMIX-CRAY-2D code simulates the heat removal in the primary circuit of an HTR for normal operations and incidents or accidents caused by core cooling failure, especially if natural convection, heat conduction and heat radiation conditions are dominant. Within the framework of safety review, new analyses were undertaken for the THTR-300 on temperature behaviour during assumed core heat-up accidents as a consequence of long-time failure of forced cooling under afterheat removal conditions. The analyses have indicated that with a reactor under pressure (RuD) the maximum core temperature is stabilized after about 200 h at 1600°C. Even with the assumed depressurization of the primary circuit (DES), this temperature level is only exceeded after two days and the maximum core temperature then increases after 125 h to about 2400 ° C (fig. 1).

The curve of the core temperatures appearing during the first few days results from the flow distribution of natural convection in the core and in the primary circuit. Afterheat removal by natural heat dissipation led, in comparison to previous analyses, to reduced temperature loads for the THTR. The temperature transients which are to be expected in the primary circuit during passive decay heat removal have been analysed using the THERMIX-CRAY-2D code with improved data and advanced models in an extensive manner also for a small modular power plant (Module). In addition to best-estimate calculations, conservative analyses have been performed with 10% additions for main influencing correlations (i.e. for the effective thermal conductivity of the pebble bed core). On this basis, risk-relevant temperature transients, for reactor under pressure (natural convection in the pressure vessel) and depressurization (heat conduction and thermal radiation in and out of the pressure vessel) with the vessel cooling system (surface cooler) or the primary

213

W. Rehm et al. / Heat removal for H T R cores

concrete cell as heat sinks, lead to the following facts, for which comprehensive sensitivity studies are underway: • The behaviour under loss of forced convection is characterized by core temperatures below 1100°C, due to the intensive natural convection in the reactor pressure vessel. The reactor design will be able to keep the pressure and temperature transients below the design values. However, a hypothetical failure of the vessel cooling system results in the safety valve of the primary circuit opening several times in the first four days at 69 bar. Accident-related natural convection, between the reactor core and the steam generator, is analysed in connection with the experimental test facility LUNA-HTR for the Module design [2]. • The behaviour of the depressurized reactor is characterized by a self-limitation of the maximum core temperature at about 1420 ° C (best estimate for slow depressurization), which will be reached in the first day, due to heat conduction and thermal radiation in the primary circuit. Conservative calculation (for fast depressurization) results in 1570 °C. A hypothetical failure of the vessel cooling system leads to a lower radial temperature distribution, as assumed in the past, and is also a result of the energy absorption in the concrete of the primary cell which acts as a natural heat sink in this worst case. Fig. 2 shows a comparison of the radial temperature transients with and without an operating vessel cooling system. As can be seen, in the first case the temperature in the reactor vessel will remain below 350 ° C and in the second case in the concrete below 500 o C.

2. Experimental facility for natmai convection and depressurization studies The experimental facility LUNA-HTR (_L~p Circulation of Natural Convection in the primary circuit of an HTR) simulates the natural convection, starting from a pebble-bed as the heat source, proceeds via a closed piping system to a heat exchanger and back again to the pebble-bed (fig. 3). Fig. 4 shows a selection of measured and calculated gas throughputs during an experimental period of 3 h and for the various conditions shown in fig. 4. Experiments have also been performed for a depressurization. During the depressurization phase the heat is mainly transported by heat conduction and thermal radiation which is described as an effective thermal conductivity of the pebble-bed core. The effective ther-

0.8 mL~

-- m - -

c=__~_t,]o~

V ~ ~ m l l m ~ q ~ l ~

0.4

~.

downward flow

po$.I/ail' (loophot)

0.2. z+~

_ _

.

V6

pos.t/Helum

l (=,l, ot) 0

__.

•

-0.2

;

--I pos.4/Helll~

upward ~.~S~-"" flow j ~ :

l ( I o ~ II°t) t

I f ~ " ~~'11 V2

-°4-Vr

I

V9

-O.6f 0

l

÷4

pelwwe-I~D 1 Time [ h ]

2

heat

"

iTpos.4 3

Fig. 4. Natural convection circulation mass flows in the LUNA loop.

mal conductivity in packed beds of spheres had been measured for the high temperature range up to 2000 K in a special test facility [3]. In the temperature range above 1500 K, heat transport by thermal radiation becomes dominant. Fig. 5 shows the dependences of the effective thermal conductivity related to a characteristic radiation factor. The influencing physical properties, such as annealing effects of the matrix graphite, the surface emissivity, and the pebble-bed porocity, are updated for modern HTR fuel element. The far-reaching agreement of calculation and measurement demonstrates the basic suitability of the computer programs, and thus a further contribution has been made towards firming up the thermodynamic safety analysis of passive afterheat removal concepts for gas-cooled high temperature reactors.

3. PCRV investigations In the experiments the sections of the PCRV with an area of 1.0 × 1.5 m2 and a thickness of 0.5 m were heated by being suspended over an electric chamber furnace (fig. 6). There are 12 silit heating rods in the furnace chamber which make working temperatures of up to 1500 ° C possible. A preset accident temperature-

214

IV. R e h m et al. / Heat removal f o r H T R cores

~e A¢

-

4adT

3

1

----

equation of Breitbach/Barthels equation of Robold

0,9 -q-

q~

= 0,39 voidage of the pebble-bed emissivity

0,8 d

¢=1

balldiameter

Jks solid thermal conductivity 0,7 o

~: = 0,8

Stefan-Boltzman constant / radioactive unradiated graphite A3

0,6 ]~\

.- .~ radioactive radiated graphite A3

0.5 :~,6 0,4

02

03.

^7 (,p= 0,4z)

¢=0,1

OJ

A*~ (q~= 0,39)

J I 10.2

,,;

I'

"

I

I

:

A'~ (q~ = 0,24)

HTR--Core Temp. 1000°C 15000 20000 25000C

ll0"1

1,0 I

I

li 1101

1102

4 ~d T 3

2ks Fig. 5. The quasi-homogeneous model equations for determining the effective thermal conductivity in packed beds.

time curve can be simulated. There are several inspection holes with quartz glass in the sides of the furnace in order to observe the test specimen or to make a photographic record of the processes at the test specimen during the experiment. The test specimens are representative sections cut from HT R prestressed concrete pressure vessels on the original scale. In the case of the HTR-500 design the specimen consists of basaltic concrete and a 12-ram liner plate anchored with 5 / 8 in. bolts and cooling pipes (fig. 7). The model is equipped with thermocoupies and pressure gauges. The concrete block is jacket with steel plates so that the amount of released water and gas can be measured. The insulation consists of 10 cm Kaowool with 5-mm thick steel coverplates and a

two-train liner cooling system (LCS). A test was also carried out for the THTR-300 design with metal foil insulation and limestone concrete [4,5]. The test specimens were exposed to the temperature transient calculated for the hypothetical core heat-up accident, i.e. the maximum temperature at the cover plates was 1130 ° C after about 12.5 weeks in the case of HTR-500 (fig. 8). During the temperature increase both liner cooling systems were operating. The liner reached 55 o C. The insulation remained intact at these temperatures even when only one train of the liner cooling system was in operation. For this case the temperature of the liner increased to 85 o C. Only the edges of the cover plates were slightly lifted from the insulation, however, without dropping down.

W. Rehm et al. / Heat removal for H T R cores

Support rack

soQ

Fig. 6. Experimental setup.

One of the aims of the experiments is to verify the improvement of a code for calculating the water and gas release of the concrete coupled with the temperature distribution in a real PCRV of an HTR nuclear power plants, respectively. The code used is the USINT program taken over from the Sandia National Laboratoties, Albuquerque, USA [6]. With some improvements it is possible to calculate the release of water and gas as well as the temperatures in good agreement with the measured values. It was furthermore investigated up to which temperature of the cooling tubes and thus until which time a refeed of the liner cooling system is possible. For this the flow of cooling water in both liner cooling systems was stopped and the increase of temperature in the liner cooling tubes was measured. At a fixed temperature the flow was started again and the processes during refeeding were recorded. The temperature-time curve for a refeeding at 350 ° C tube temperature is given in fig. 9. It was possible to refeed the liner cooling system up to 450 ° C. Higher temperatures were not examined due to the limits given by the test facility. The time-span

215

between failure of the liner cooling system and reaching the tube temperature of 4 5 0 ° C was 17.3 h, i.e. sufficient time is available for restoring or emergency feed of the LCS (e.g. by fire brigade). In a further test only one train was refed at 350 o C. As in all the other tests, it was again possible to cool down the model to a steadystate temperature condition. For the specimen with THTR-300 design the maximum temperature in the test was 650 o C. After these tests a long-term failed liner cooling system was simulated. The liner and concrete reached a maximum temperature of 7 7 0 ° C after 8 days in the quasi-steady-state condition. The only phenomenon observed was a slightly increased deflection at the ends of the cover plates. No failure of the insulation occurred. The same behaviour was observed in the THTR-300 specimen. Therefore the temperature at the cover plates was increased, i.e. the theoretical possible highest temperatures were exceeded (see fig. 8). This further increase in temperature caused the cover plates to fold downwards at 1400 ° C without the bolts and insulation as a whole dropping. In this case the liner and concrete were heated up to 1100 o C. After the test the insulation was removed. The bolts showed no cracks or other defects. Also, the liner was completely intact. Then the block was sawn in half and the inside of the tubes inspected. No cracks were observed. Cylinders were drilled out of the concrete from different places and the strength was determined.

4. Top reflector suspension of the THTR-300 In the case of a core heat-up accident with failure of liner cooling, for the THTR-300 temperatures of approximately 980 ° C will occur after about 5.5 h at the metallic top reflector suspension with a pressurized reactor. A 40% failure probability was calculated for material strengths extrapolated to the high temperatures involved. Due to this high probability and the uncertain material data, an experiment was carried out to check the calculations and the mode of failure [7]. In the experiment only the crossbolt and the lower tie rod were examined (fig. 10) where the highest accident temperatures occur. The crossbolt and tie rod were designed true to the original. The lower graphite block was replaced by a metal specimen with the same stiffness relations. The total load present in the original was simulated in the experiment by additional weights. The metal specimen was heated from below. The temperature rise in the metal specimen was simulated in accordance with the calculated temperature curve for the

216

IV. Rehm et aL / Heat removal for H T R cores

Suspension

'[1

l

i

!,

I ' 0

I

,

~

Basaltic Concrete

o~ g

Gridof bolts

!oo ,oo !i I ' r

¢

--I I f Liner

o u~

i

\

Steam outlet

~..~

c~ i

io?

- t ~ Steel Jacket

~8' '!l

il [] ,i f!!!i}?:!7)!)~!il:'":":"" ........ ~ .... I~7<:~:',~-~:.':::' ,/, "~--"

~

/

/,

1" Foil lmm

//

Z, Foils0,3mm

/

li',~I [ ]

n

" / '/

III

\

--

~

\

Coverplate

/

-I

COO

/

M 2 4 ~ 150

qelson =TypHP

Fig. 7. Modelblockwith HTR-500design.

14001200"

j

~ lOOOo

== 800600 E t- 400

Q~

o 0

i,.

J

e. 200

\ 0

0

,

14.;".St

'

,

,

.

.

.

.

.

,

r

,

,

,

,

10

14.8.a7

I

[

/'

6 • 16,6 mm

l

i!i

Liner Cooling Tubes

1 SO0

Koowoot

m

14.g'.e7

,

,

i

,

,

,

,

,

,

20

14.10.e7

14.1'1.07 14.1z.a7

Time

-~

[weeks]

- - - ~ [month]

Fig. 8. Temperature-dine cm've for the cover plates.

F--

JI ;I

217

W. Rehm et aL / Heat removal for H T R cores

Prestressed concrete pressure vessel

400' Refeedingof ColdWater

Tube of the Lim~ CoolingSystem ( LCS)

i

41HI

~--

-~

,

3. Tie r 0 d j

300-

T

50 mm from LCS V////////////A

.

]

o200oflcrete

Q

$

200 mm

Q,

--[ 2. tierod

1

!

I

. . . .

E k- 1 0 0 -

> Graphite layers

10 h 10 rain I

0

11

10

Time

[hi

:

Fig. 9. Temperature-time curves during refeeding. Core Fig. 10. Suspension of the top shield and the top reflector.

experiments with the material strengths calculated from the experiment and is now established as << 10-lo for 9 8 0 ° C (fig. 11). accident in question. The first test specimen was heated three times. The first time 1000 ° C was achieved without failure and the second time 1130 ° C, despite a 1.6 times higher dead load. No permanent changes in length were found in the cooled state. Failure only occurred during the third heating at 1162°C after a 1.9 times higher load had been applied by additional weights. Dismantling revealed that creep rupture at the undercut thread (screw thread of the tie rod in the bolt) had led to failure. Up until the time of failure, the location of rupture had been exposed to a temperature of over 1000 ° C for approx. 10 h, of which 6.5 h were with a higher load than was present in the top reflector of the THTR-300. The second specimen gave the same failure temperature. A third model specimenincludes the metal-graphite reaction. Failure probability was redetermined after the

.?-~

t o. ot

- ~ . - - TIE ROD XIOCrNiT 18 9

f

/

R

2 TIE ROD 15M03 - - 3 TEE ROD 15Mo3

060.d ~ 0.50-

J

0.400.300.20 /

0/00.00

,

900

|

950

1000

1050

1100

1150

1200

1250

TEMPERATURE (DEGREE C) Fig. 11. Failure probabilities as a function of temperature.

218

IV. Rehm et al. / Heat removal for H T R cores

5. Summary Performed theoretical a n d experimental research a n d developing work indicates that the n a t u r a l heat dissipation processes, such as n a t u r a l convection, heat conduction, a n d t h e r m a l radiation, can b e exploited as a n effective heat t r a n s p o r t m e c h a n i s m from the core to the o u t e r heat sink u n d e r a f t e r h e a t conditions. A c c o r d i n g to this, the core t e m p e r a t u r e s are self-limited, especially in c o n n e c t i o n with special design features, below values which will b e u n i m p o r t a n t for fission p r o d u c t release out of the spherical fuel elements of small, m o d u l a r H T R s , particularly. In consequence of a core heat-up accident in medium-sized H T R s with a long-term failure of the liner cooling system, n o failure of the insulation, of the liner or of the prestressed concrete pressure vessel will occur. Sufficient time is available for the refeeding of the liner cooling system. A t e m p e r a t u r e increase in the whole pressure vessel c a n b e avoided even if only o n e train of the liner cooling system is in operation. T h e failure of the top reflector suspension of the T H T R - 3 0 0 will occur at a b o u t 1 2 0 0 ° C , i.e. in the longer term.

References [1] W. Rehm, W. Jatm and K. Verfondern, Present results and further developments on safety analysis of small and medium-sized HTRs for core heatup accidents, SMIRT '87, Post-Conference, Lausanne, Switzerland, 24-26 August 1987. [2] H. Barthels, W. Jahn and W. Rehm, Experimental and theoretical studies on natural convection for the passive afterheat removal from an HTR, Presented Paper at the International Conference on Thermal Reactor Safety, Avignon, France, 2-7 October 1988. [3] H. Barthels and M. Schiirenkr~imer, The effective conductivity in packed beds of spheres especially in the core of a high temperature reactor, KFA report, Jii1-1893 (Feb. 1984). [4] J. Altes, G. Breitbach, K.H: Escherich, T. Hahn, M. Nickel and J. Wolters, Experimental study of the behaviour of prestressed concrete pressure vessels of high temperature reactors at accident temperatures, Trans. 9th Int. Conf. on SMiRT, Lausanne, Volume H (1987) 189-194. [5] J. Altes, K.H. Escherich, T. Hahn, M. Nickel and J. Wolters, Behaviour of the prestressed concrete pressure vessel of the HTR-500 at severe accident temperatures, Trans. 10th Int. Conf. on SMiRT, Anaheim (1989) Paper H 05/1. [6] J.V. Beck and R.L. Knight, User's Manual for USINT, NUREG/CR-1375, SAND 79-1694 (1980). [7] J. Altes and W. ViSlzer, Zuverl~issigkeit yon Komponenten yon Hochtemperaturreaktoren unter St/Srfallbelastungen, VDI Symposium 'Zuverl~issigkeit yon Komponenten technischer Systeme', Miinchen, 20-21 April 1989, pp. 20]-213.