Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools

Nuclear Engineering and Design xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes

Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools Martin Arlit a,⇑, Christine Partmann b, Eckhard Schleicher c, Christoph Schuster b, Antonio Hurtado b, Uwe Hampel a,c a

AREVA Endowed Chair for Imaging Techniques in Energy and Process Engineering, Technische Universität Dresden, 01062 Dresden, Germany Chair of Hydrogen and Nuclear Energy, Technische Universität Dresden, 01062 Dresden, Germany c Helmholtz-Zentrum Dresden-Rossendorf, Institute of Fluid Dynamics, Bautzner Landstraße 400, 01328 Dresden, Germany b

a r t i c l e

i n f o

Article history: Received 15 February 2017 Received in revised form 22 June 2017 Accepted 24 June 2017 Available online xxxx Keywords: Spent fuel pool Temperature measurement Thermal anemometry Grid sensor

a b s t r a c t Beside the nuclear reactor and its primary circuit the spent fuel pool is yet another safety-critical part in a nuclear power plant which has gained increasing focus after the Fukushima accident. Loss of coolant or enduring loss of cooling conditions would ultimately result in loss of cladding integrity at elevated temperatures with excessive release of fission products and hydrogen. To predict the available response time and to assess the efficacy of mitigating measures computer simulations can be employed. Their validity, however, needs to be proven by dedicated experiments at small scale but relevant thermal hydraulic conditions. For that purpose, the test facility ALADIN was designed, which enables conduction of experiments on a single BWR fuel element mock-up under loss of coolant and loss of cooling accident conditions. In this paper we introduce the facility and its instrumentation, with a focus on temperature sensors and a new thermal anemometry grid sensor for flow velocity measurement in one quadrant of the rod bundle with a resolution of one point per subchannel together with the affiliated calibration procedure for a potential application in superheated steam and air in a wide range of fluid temperatures. Ó 2017 Elsevier B.V. All rights reserved.

1. Introduction The spent fuel pool is an interim storage for spent fuel as well as fuel elements during revision periods in the reactor pressure vessel. Potential safety risks are the loss of cooling as a result of a longer persisting station black-out or the loss of coolant caused by a leakage. In such cases the integrity of fuel rods would be lost if a certain temperature limit is exceeded. This temperature limit depends on several factors, e.g. pool filling level and atmosphere, cladding material and individual properties of the cladding. The lowest temperature limit at which the integrity is just ensured and which should be used as a guide is at 565 °C. At this temperature time-dependent deformation (creeping) degrades cladding rigidity till failure after around 10 h (Smith, 1969; Nourbakhsh et al., 2002). A highly exothermic zirconium-air reaction occurs ⇑ Corresponding author. E-mail addresses: [email protected] (M. Arlit), [email protected] (C. Partmann), [email protected] (E. Schleicher), [email protected] (C. Schuster), [email protected] (A. Hurtado), Uwe. [email protected] (U. Hampel).

at cladding temperatures above 800–900 °C (Leistikow et al., 1975; Benjamin et al., 1979) and an exothermic zirconium-steam reaction above 1100 °C (Stuckert et al., 2011). The oxidation of zirconium by steam is accompanied by release of hydrogen gas presenting an additional hazard potential (Benjamin and McCloskey, 1980; Collins and Hubbard, 2001). This behavior has already been observed in the severe Three Mile Island accident at TMI-2, where a partial nuclear meltdown occurred as a result of a loss-of-coolant accident in the reactor pressure vessel (Kemeny, 1979). Against this background and the lessons learned there is an increased need for simulation codes that can correctly predict the three-dimensional and transient heat transfer in the spent fuel pool and the resulting time course of the dry-out accident (NEA, 2008; Kaliatka et al., 2010). For the validation of the simulation codes reliable experimental data are required. So far experimental studies have been conducted in the US for air-cooled conditions on full-sized Boiling Water Reactor (BWR) and Pressurized Water Reactor (PWR) assemblies to study fuel cladding ballooning and rapid zirconium oxidation (Lindgren and Durbin, 2013). Other experimental investigations were carried out on related aspects, but not originally with the focus on the investigation of spent fuel

http://dx.doi.org/10.1016/j.nucengdes.2017.06.034 0029-5493/Ó 2017 Elsevier B.V. All rights reserved.

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034

2

M. Arlit et al. / Nuclear Engineering and Design xxx (2017) xxx–xxx

Nomenclature A A# , B# BWR C CVA cp d f1 Gr g K l Nu n Pel Pr Q_ c Q_ t R Ra Re TAGS U

v

area temperature coefficient of resistance boiling water reactor thermal conduction constant constant voltage anemometry specific heat capacity diameter function of Pr Grashof number gravitational constant curvature parameter length Nusselt number amount Joule heating power Prandtl number Pr ¼ lcp =k convective heat flux thermal conduction heat flux electrical resistance Rayleigh number Reynolds number Re ¼ qv lchar =l thermal anemometry grid sensor voltage velocity

pools under accident conditions (Aksan et al., 1992). A more detailed description of past and present studies, both numerical and experimental, can be found in the status report of the Nuclear Energy Agency (Adorni et al., 2015). Experimental investigations of the fuel load in an evaporating spent fuel pool or inside a reactor pressure vessel of a boiling water reactor has been conducted at the test facilities ADELA-I and ADELA-II at Technische Universität Dresden. In these projects preliminary simulations made with system thermal hydraulics codes were used to check their basic applicability. CFD codes have not been used so far. But this is considered as a crucial future need for a more accurate prediction of flow and heat transfer across multiple scales (Schuster, 2008; Schulz et al., 2014). The main objectives of the German joint project SINABEL (SIcherheit NAsslager BrennElement-Lagerbecken) are the experimental investigation of the thermal hydraulics in a boiling water reactor fuel element and an improved CFD modelling. For this purpose a replica of a 10
10 rod BWR fuel element was constructed and is operated under spent fuel pool accident conditions. Special care was given to create boundary conditions for the single assembly close to those in a fuel rack with many assemblies. This is achieved by having heated bordering channels representing adjacent fuel elements in a simplified way. With the planned integral experiments thermal hydraulics effects within the fuel assembly as well as directly above can be studied. In the following sections the investigated accident scenarios, the test facility design and its instrumentation will be presented. A stronger focus will be given to the special instrumentation, which had to be developed due to a lack of commercial products for this application.

Greek

a

b D#

l q

#

heat transfer coefficient thermal expansion coefficient overheating dynamic viscosity density temperature

Subscripts cal calibration char characteristic f fluid free free convection forced forced convection H heating level min minimum max maximum s sensor overheated T temperature level TC thermocouple w wire x fluid composition

have to be considered. In the first case it is assumed that the coolant is absent due to damage of the pool integrity. The experiment will start in the dry state (Fig. 1, left). In contrast the loss of cooling accident experiment will start in the flooded state. The temporal progress of the flooding state depends on the decay heat or the heat input to the experimental system respectively. If the heat cannot be dissipated sufficiently via thermal transport processes the coolant will reach boiling temperature. Then the water level will decrease due to the evaporation of the coolant (Fig. 1, right). Parameters of interest in both scenarios are the surface temperatures of the rods and their temporal development as well as temperatures and flow conditions in the subchannels. This requires the determination of the fluid temperature and flow velocity in the subchannels of the rod bundle. 2.2. The test facility ALADIN The ALADIN test facility is essentially a box-shaped enclosure containing a heavily instrumented BWR fuel assembly mock-up with some additional components to mimic conditions in a spent

2. Experiments and test facility 2.1. Investigated scenarios For the investigation of both accident scenarios, the loss of coolant and loss of cooling respectively, two different initial conditions

Fig. 1. Scheme of the states in a loss of coolant (left) and a loss of cooling (right) scenario for spent fuel pools.

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034

3

M. Arlit et al. / Nuclear Engineering and Design xxx (2017) xxx–xxx

fuel pool storage rack (Fig. 2). It is connected to a water tank for controlled flooding and water removal. The 10
10 fuel element replica consists of electrically heated rods with a 2
2 water channel in the center. The rods have D ¼ 10 mm diameter and a total heated length of l ¼ 3600 mm. The pitch-to-diameter ratio in the bundle is P=D ¼ 1:24. As common in BWR assemblies the bundle is enclosed by a metal box. The assembly further comprises seven self-designed spacer grids. To reproduce a storage rack an additional quadratic channel with 10 mm wall thickness can be installed. To reproduce the heat source conditions in the storage rack bordering fuel elements are represented by columns of single electrically heated rods. The housing is thermally insulated and to avoid heat losses it was constructed without flanges. Consequently the accessibility for installations is limited to the top side. The axial heating power profile of the rods is in accordance with a characteristic power profile of a fuel rod at the end of a core cycle under conservative assumptions. This is roughly equivalent to a cosine curve with the maximum shifted towards the upper end (Kok, 2009). The profile is achieved by a variable gradient of wound heating wires in the core of the rods. The overall heating power can be varied to reproduce different decay heat patterns.

Table 1 Expected flow conditions during experiments in the ALADIN facility. Parameter

Value

Fluid

Air (Loss of coolant) Superheated steam (Loss of cooling) v min ¼ 0:01m=s; v max ¼ 0:5m=s #min ¼ 100 C; #max ¼ 500 C Laminar

Velocity range Temperature range Flow regime

2.3. Instrumentation needs The surface temperature of the rods is the most critical parameter for safety assessments. It has therefore to be measured at several positions. As during the dry-out phase there is water and steam in the fuel assembly, the water level has to be measured also. For the steam phase the steam temperature and the steam velocity in the subchannels is of importance. One selection criterion for the instrumentation is the expected parameter range during the experiments. This can be estimated from former experimental investigations (Schulz et al., 2014). Expected conditions are given in Table 1. The measurement of the steam velocity is an extraordinary challenge due to the expected high temperatures and the very low steam velocities. The bundle cross-section can be divided into four quadrants with assumed symmetric temperature and flow distributions (Fig. 3). For measurement of rod cladding temperatures quadrant I is densely instrumented. The neighboring quadrants II and III are instrumented at lower density only for the proof of the symmetry. Quadrant IV is not instrumented with thermocouples to have accessibility for other instrumentation installations. For flow measurement the measurement points should be in a region of a fully developed flow above the water level and with sufficient distance to a spacer. An adequate spatial resolution is one

Fig. 2. Sketch of the test facility ALADIN.

Fig. 3. 10
10 rod bundle cross-section divided into four quadrants with different instrumentation density and numbers of thermocouples on cladding nTC.

point per subchannel. The requirements for temporal resolution are low due to an expected laminar flow. Additionally, the flow will change only very slowly with time since the maximum decrease of water level is expected to be 0.40 m/h at highest heating power (200 W/rod) and 0.04 m/h at a lower heating power of 20 W/rod. One requirement for all sensors is to be minimally intrusive.

3. Instrumentation The selection of the instrumentation was made according to the requirements described in the section before. A comparison of the measurement parameters is listed in column 1 of Table 2. For some of the measurement quantities commercial sensors were available, others required the development of special instrumentation. Thermocouples type K are used for the measurement of the cladding and fluid temperature in quadrants I-III. To achieve minimal intrusiveness and low time response thermocouples of 0:5 mm diameter are used. They are fixed on the surface of the rods by point welded sheets of metal with a thickness of 0.05 mm. For the measurement of the cladding temperature the tips are also connected firmly to the surface with the help of a covering metal sheet (Fig. 4). Thereby, it can be assumed that the temperatures of the thermocouple tip and the heater wall are corresponding. The thermocouples for fluid measurement are mounted along a lance to measure the axial temperature profile in the gap between the central water channel and the bordering channel as well as below and above the bundle. Further commercial instrumentation is a pressure transducer for the determination of the filling level by measuring the hydrostatic pressure of the water on the bottom of the test facility. For the measurement of the filling level additional local probes are installed in several heights. These so called needle probes combine a phase discrimination by measurement of the electrical conductivity of the fluid and a temperature measurement (Schleicher et al., 2008). For flow measurement in scientific and engineering applications several techniques are available. Most of them are

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034

4

M. Arlit et al. / Nuclear Engineering and Design xxx (2017) xxx–xxx

Table 2 Quantities of interest and employed sensors (*functionality integrated in one sensor). Parameter

Commercial

Cladding temperature Fluid temperature

x x

Filling level

Special

Technique

Number of measurement points

x x

Thermocouple Thermocouple Temperature grid sensor* Needle probe Pressure transducer Flow velocity grid sensor*

123 10 16 12 1 16

x Steam flow velocity

x

Pel ¼ Q_ c þ Q_ t :

ð3Þ

The convective heat transfer

Q_ c ¼ aðv ÞAð#s  #f Þ

Fig. 4. Thermocouple for wall temperature measurement with its fixation on the rod.

not applicable in the ALADIN test facility due to the boundary conditions during the experiments. The accessibility for optical methods, like PIV or LDA, is constrained by the compact design of the test facility and the narrow gaps between the rods. Using local probes spatial resolution can be obtained in different ways. Traversing a probe sequentially in the area of interest has the disadvantage of a non-simultaneous sampling. Applying a multitude of local probes for simultaneous measurement entails a huge wiring effort and complex signal processing electronics. Many local probes for flow velocity measurement, such as Pitot tubes, are not applicable due to the low differential pressure between the dynamic and the static pressure induced by the flow. The use of commercial thermal anemometer probes is excluded because of their maximum operation temperature and given application range of velocity. Therefore, a new measurement system had to be developed that will be presented in the following section. 4. The thermal anemometry grid sensor (TAGS)

ð4Þ

depends on the temperature difference between the sensor surface #s and the fluid #f , the sensor surface area A and the fluid velocity dependent heat transfer coefficient aðv Þ. The heat conduction along a wire is given by

kw  Aw  ð#s  #f Þ ¼ C  ð#s  #f Þ Q_ t ¼ lw

ð5Þ

with the thermal conductivity kw , the cross-sectional area Aw and the effective length lw of the contacting wires. As the latter is only effective parameters, we lump all three parameters together into a constant factor C. Putting Eqs. (2)–(5) together yields

ð#s  #f Þ ¼ D# ¼

U 2H : R½aðv ÞA þ C

ð6Þ

Known quantities are U H , A and #f (at least it is required that #f is known), R can be measured directly, which gives #s via Eq. (1). As aðv Þ and C are commonly unknown it is common practice to use an experimental calibration curve for calculating v from R. While the interested reader is referred to thermal anemometry text books for further reading (Bruun, 1996), we will describe in the next chapter the thermal anemometry grid sensor design and then come back to Eq. (6) when describing our approach to measure velocity in superheated steam at any temperature.

4.1. Functional principle

4.2. TAGS design

A resistor may be used to sense the temperature of a medium. This method is called resistance thermometry. For a platinum metal resistive sensor, as used in this study, the resistance R is a function of sensor temperature # commonly given as a polynomial

The basis of the presented solution is a regular spatial arrangement of single temperature/thermal anemometry sensors operated in a matrix excitation-acquisition scheme. The principle is shown in Fig. 5a. Wires connecting the sensor elements (dotted lines) in one plane (transmitter wires) are successively or simultaneously excited with a dc voltage. Wires in the opposite plane (receiver wires) are used to measure the electrical current flowing through the sensor elements. Thereby the wires of the receiver plane have the electrical potential of a virtual mass. The measured currents are inversely proportional to the sensor elements’ electrical resistance. It should be noted here, that temperature grid sensors for spatially-resolved measurement of temperatures have been described in the literature already. Schaefer et al. presented a device for distributed temperature measurements on surfaces (Schäfer et al., 2013) and Ritterath et al. one for measurement in pipes (Ritterath et al., 2011). Operating a TAGS in superheated steam atmosphere and temperatures up to 500 °C is challenging and requires special care in the sensor design. Fig. 6a shows the developed TAGS. For mechanical stability the platinum sensors and connecting wires are mounted on a heat-resistant ceramic substrate (Fig. 6b). The electrical connection is realized with mineral-insulated cables. The TAGS is inserted into the ALADIN test facility from the top side of the rod (Fig. 6d).

R ¼ R0  ½1 þ A# ð#  #0 Þ þ B# ð#  #0 Þ2 

ð1Þ 

with the nominal resistance at reference temperature #0 ¼ 0 C and the temperature coefficients of the resistance A# ¼ þ3:9083  103 C1 and B# ¼ 5:775  107 C 2 (IEC 60751, 2008). The applied voltage U T has to be chosen so low that selfheating of the sensor has no significant influence and the sensor temperature # can be considered as equal to the temperature #f of the surrounding fluid. A resistor can also be used to measure fluid velocity according to the principle of thermal anemometry (Bruun, 1996; Lomas, 2011). For that the resistor has to be overheated by applying a sufficiently high voltage U H . In equilibrium the resistor attains a constant temperature #S and the dissipated electrical power

Pel ¼

U 2H R

ð2Þ

is balanced by the convective heat flux to the fluid (Q_ c ) and a conductive heat flux (Q_ t ) through the connecting wires according to

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034

5

M. Arlit et al. / Nuclear Engineering and Design xxx (2017) xxx–xxx

Fig. 5. a) Schematic connection diagram of the TAGS and b) employed platinum resistor.

Fig. 6. a)–c) Photography of the TAGS, d) TAGS in the test facility ALADIN.

The applied operation mode for the TAGS is constant voltage anemometry (CVA) with two cycles, one for measuring #s and one for measuring #f (Sarma, 1993). For measuring #f the transmitters are switched off until thermal equilibrium is reached. Subsequently resistance R is measured for each sensor and #f determined from the calibration curve. After that the sensor elements are heated by simultaneously applying the voltage U H to all sensors. After thermal equilibrium is reached again, resistance R is measured by switching each transmitter to U T for a very short interval while at that very moment all other transmitters are kept at ground potential. Requirements for this procedure are a short sampling time at one transmitter and a sufficient thermal inertia of the single sensor elements to avoid excessive cooling during the unheated phases. As the ceramic carrier is a flow obstruction attention has been given to the cross-point design. The platinum sensors are not directly mounted to the ceramic substrate but instead extend some way (approx. 3 mm) into the upstream direction via their contact wires (Fig. 6c). The flow constriction induced by the platinum sensor itself can be neglected since the coverage ratio defined by the sensor area compared to the free subchannel flow area is 2.5%.

ature fluctuations during the experiment compensation is necessary. Experimental calibration at several temperatures and extraction of the velocity from this set by interpolation is very complex. Another method is the usage of a temperatureindependent calibration curve as presented by Sarma and ComteBellot (2002), which corrects the Joule heating power by sequential measurement of hot and cold resistance. It is not directly applicable for the TAGS since the method neglects the heat flux over the supports and the used Nusselt numbers are for wires. Our concept for the measurement in a wide range of temperatures in superheated steam is inspired by Sarma and Comte-Bellot and consists in the transformation of the calibration curve. For this we reconsider Eq. (6) and note that the heat transfer coefficient is a function a ¼ f ðRe; Gr; PrÞ ¼ f ðv ; q; k; l; cp ; bÞ with the density q, thermal conductivity k, dynamic viscosity l, heat capacity cp and thermal expansion coefficient b of the liquid. Furthermore f depends on the resistor geometry. In the following we will derive a formula for f from the theory of mixed convection for a given resistor geometry. This model relationship can then be used to transfer a calibration curve f

l

ðcalÞ

Usually in thermal flow measurement calibration curves are determined for a specific fluid temperature. In the case of temper-

to

a

calibration

ðv Þ ¼ f ðv ; qðcalÞ ; kðcalÞ ;

curve

f

ðxÞ

ðv Þ ¼ f ðv ; qðxÞ ;

for another fluid (x) but the same sensor k ;l geometry. For small flow velocities the heat transfer coefficient can be derived from a mixed convection approach with ðxÞ

4.3. Transformation of the calibration curve

ðcalÞ ; cp ; bðcalÞ Þ ðxÞ ðxÞ ; cp ; bðxÞ Þ

ðcalÞ

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034

6

M. Arlit et al. / Nuclear Engineering and Design xxx (2017) xxx–xxx

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  3  3 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3 k k 3 a ¼ a3free þ a3forced ¼ þ Nuforced Nufree : lfree lforced

ð7Þ

and the characteristic lengths lfree and lforced as well as the Nusselt numbers Nufree and Nuforced for free and forced convection. The critical point now is to find an appropriate empirical Nusselt correlation. For the given sensor geometry the free convection is well described by the correlation for a vertical wall (Churchill and Chu, 1975) as

h i 1 2 Nufree ¼ 0:825 þ 0:387ðRa  f 1 Þ6   9 169 0:492 16  f 1 ¼ ½1 þ Pr

ð8Þ

ð9Þ

correct shape of the calibration curve. Instead the correlation for a cylinder with a coextensive diameter d

Nuforced ¼ 2 

0:55 K 0:5

þ

10 0:95  9 K 0:1

ð11Þ

and curvature parameter

K¼

l lchar 1 4 lchar   ¼  q ðd=2Þ2 v d Re

ð12Þ

proved to be appropriate (Gampert, 1973) with d the coextensive diameter of the cylinder. Note that in the correlation of Gampert lforced ¼ d. 4.4. Experimental validation For experimental validation we first obtained a calibration

q2 gbD#l3free  Pr Ra ¼ Gr  Pr ¼ g2

ð10Þ

with the Prandtl number Pr, the Rayleigh number Ra and the Grashof number Gr. The expansion coefficient is b ¼ 1=ð273:15 C þ #f Þ for gases. The characteristic length lfree ¼ lchar is the quotient of the surface area A of the sensor and the circumference length of the area Aproj of the projected sensor in flow direction. For forced convection the approximation of a plate that is orientated lengthwise to the flow failed in reproducing the

ðv Þ for air at room temperature (#f ¼ 20 C). For this purcurve f pose the TAGS was positioned downstream of a multi-orifice plate with one orifice directly upstream to each TAGS sensor element. The uniform distribution of the air flow downstream the orifice as well as the correlation between the adjusted flow rate V_ and ðcalÞ

the measured gas velocity was validated with a traversing single platinum sensor operated in CVA mode. In step 2 the model curve according to Eq. (6) was fitted to the experimental calibration curve with the fluid parameters of air (#f ¼ 20 C). The fitting variable of the resulting model f ðv ; CÞ is the factor C. In step 3 the calibration curve is transformed with the new fluid properties ðxÞ qðxÞ ; kðxÞ ; lðxÞ ; cðxÞ and #f and the determined factor C. The flow p ;b

Fig. 7. Scheme of the calibration curve transformation procedure.

velocity is then determined from the new calibration curve with the measured D# ¼ #s  #f . The procedure is illustrated in Fig. 7. For validation of the approach the same type of platinum sensor like in the TAGS with the same orientation was mounted concentrically in a heatable pipe of diameter 0:016m at an entrance length of 0:8m. The latter one is sufficiently high to ensure the formation of a parabolic velocity profile in the pipe. The fluid properties can be varied via the fluid temperature #f and the use of the air and CO2. The latter one was chosen since the handling and the adjustment of flow conditions is much simpler than for superheated steam. The experimental procedure is equal to the presented calibration procedure. First an experimental calibration curve with air at room temperature conditions was generated. For validity of the converted theoretical calibration curves corresponding experimental calibration curves were exemplary acquired and compared (see next section).

Fig. 8. Experimental (x) and modeled (-) calibration curves for air (left) and CO2 (right).

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034

M. Arlit et al. / Nuclear Engineering and Design xxx (2017) xxx–xxx

The results for fluid temperatures #f ¼ 23 C, #f ¼ 200 C and #f ¼ 400 C are presented in Fig. 8. The respective theoretical and experimental calibration fit very well. The coefficients of determination are varying between 0.999 and 0.843. This proves that the transformation procedure can be used with the model for the specific platinum sensor given above. The uncertainty of v mainly results from the measurement uncertainty of #s and #f , that increases with increasing #f . Considering the fact, that the sensitivity S ¼ DðD#Þ=Dðv Þ decreases with increasing #f (Fig. 8), the maximum uncertainty of the velocity dv ¼ 1=S  dðD#Þ is 39%. To counteract the sensitivity can be increased by adapting the heating voltage U H . The time response of the given platinum resistor is T 95 ¼ 5s for an air flow under the conditions #f ¼ 20 C and v ¼ 0:5m=s. Concerning the scenarios presented in section 2.1 for the experiments in the dry state the fluid properties of air and for the boil-off experiments of steam have to be inserted into the model. In spite of the presence of a mixed phase air/steam cannot be excluded pure phase properties are used, since the portion of the unexpected phase may be small and influences the measurement result very slight. Note, that this method may be readily applied to other sensor types, but has to be checked carefully according to the above described procedure. Particularly it might be necessary to adapt the Nusselt number correlations. This experimental validation proves the capability of the TAGS sensor to consistently measure gas velocities for different fluid and thermodynamic conditions (temperature, density and viscosity). 5. Summary and conclusions In this paper the test facility ALADIN and the associated instrumentation for the investigation of the thermal hydraulics in a fuel element during dry-out after a loss of cooling accident was presented. For this purpose a mock-up of a BWR fuel element in original dimensions was constructed. The fuel rods are represented by electrically heated rods. For the reproduction of the conditions in a storage rack peripheral components are reproduced by single rods of the neighboring fuel elements. The temperatures of the rod surface and the fluid temperature in the subchannels are measured with high spatial resolution using 148 thermocouples. A new and specially developed instrument is the thermal anemometry grid sensor (TAGS). It enables the determination of flow velocities in the center of the subchannels by a grid-like arrangement and matrix-multiplexed operation of single platinum resistor sensors. Potential applications are the measurement of rising steam during boil-off experiments and circulating air during heat-up experiments with a dried rod bundle. Acknowledgements

7

(BMBF) under the grant number 02NUK027. Responsibility for the content of this publication lies with the authors. References Adorni, M., Esmaili, H., et al., 2015. Status Report on Spent Fuel Pools under Loss-ofCooling and Loss-of-Coolant Accident Conditions – Final Report (# NEA-CSNI-R2015-2). Nuclear Energy Agency. Aksan, S.N., Stierli, F., et al., 1992. Boil-Off Experiments With the EIR-NEPTUN Facility: Analysis and Code Assessment Overview Report (# NUREG/IA-0040). NRC, PSI. Benjamin, A.S., McCloskey, D.J., 1980. Spent fuel heatup following loss of water during storage. Nucl. Technol. 49, 274–294. Benjamin, A.S., McCloskey, D.J., et al., 1979. Spent Fuel Heatup Following Loss of Water During Storage (# NUREG/CR-0649). Sandia National Laboratories. Bruun, H.H., 1996. Hot Wire Anemometry: Principles and Signal Analysis. Oxford Univ. Pr, Oxford. Churchill, S.W., Chu, H.H.S., 1975. Correlating equations for laminar and turbulent free convection from a vertical plate. Int. J. Heat Mass Transf. 18, 1323–1329. Collins, T., Hubbard, G., 2001. Technical Study of Spent Fuel Pool Accident Risk at Decommissioning Nuclear Power Plants (# NUREG-1738). U.S. Nuclear Regulatory Commission. Gampert, B., 1973. Grenzschichttheoretische Probleme des Aerodynamischen Schmelzspinnprozesses (Ph.D. thesis). TU Berlin, Germany. IEC 60751, 2008. Industrial Platinum Resistance Thermometers and Platinum Temperature Sensors. Kaliatka, Ognerubov, V., et al., 2010. Analysis of the processes in spent fuel pools of Ignalina NPP in case of loss of heat removal. Nucl. Eng. Des. 240, 1073–1082. Kemeny, J., 1979. Final Report of the President’s Commission on the Accident at Three Mile Island (Washington, DC, USA). Kok, K.D., 2009. Nuclear Engineering Handbook. CRC Press. Leistikow, S., Berg, H.V., et al., 1975. Studies on High Temperature Steam Oxidation of Zircaloy-4 Cladding Tubes (# KFK-2195). Lindgren, E.R., Durbin, S.G., 2013. Characterization of Thermal-Hydraulic and Ignition Phenomena in Prototypic, Full-Length Boiling Water Reactor Spent Fuel Pool Assemblies After a Postulated Complete Loss-of-Coolant Accident (# NUREG/CR-7143 and SAND-2007-2270). Sandia National Laboratory. Lomas, C.G., 2011. Fundamentals of Hot Wire Anemometry, 1. Cambridge Univ, Press. NEA, 2008. OECD-IAEA Paks Fuel Project Final Report (#NEA/CSNI/R(2008) 2). Nuclear Energy Agency. Nourbakhsh, H., Miao, G., 2002. Analysis of Spent Fuel Heatup Following Loss of Water in a Spent Fuel Pool. Users’ Manual for the Computer Code SHARP (# NUREG/CR-6441). US Nuclear Regulatory Commission. Ritterath, M., Öztürk, O.C., et al., 2011. Thermo-resistive mesh sensors (TMS) for temperature field measurements. Flow Meas. Instrum. 22, 343–349. Sarma, G.R., 1993. Analysis of a constant voltage anemometer circuit. In: Instrumentation Measurement Technology Conference. IEEE, pp. 731–736 (IMTC/93, Conference Record). Sarma, G.R., Comte-Bellot, G., 2002. Automated constant voltage anemometer for measurements with fluid temperature drifts. Rev. Sci. Instrum. 73, 1313–1317. Schäfer, T., Schubert, M., et al., 2013. Temperature grid sensor for the measurement of spatial temperature distributions at object surfaces. Sensors 13, 1593–1602. Schleicher, E., Da Silva, M.J., et al., 2008. Enhanced local void and temperature measurements for highly transient multiphase flows. Instrum. Measur. IEEE Trans. 57, 401–405. Schulz, S., Schuster, C., et al., 2014. Convective heat transfer in a semi-closed BWRfuel assembly in absence of water. Nucl. Eng. Des. 272, 36–44. Schuster, C., 2008. Experimental investigation of the rod load in an evaporating spent fuel pool. In: Proceedings Annual Meeting Nuclear Technology. Hamburg. Smith, C., 1969. Calculated Fuel Perforation Temperatures, Commercial Power Reactor Fuels. NEDO-10093, September. Stuckert, J., Grosse, M., et al., 2011. Results of Severe Fuel Damage Experiment QUENCH-15 With ZIRLO Cladding Tubes (# KIT-SR 7576), KIT Scientific Reports. KIT Scientific Publishing.

The work is part of the research project ‘‘SINABEL” and is funded by the German Federal Ministry of Education and Research

Please cite this article in press as: Arlit, M., et al. Instrumentation for experiments on a fuel element mock-up for the study of thermal hydraulics for loss of cooling or coolant scenarios in spent fuel pools. Nucl. Eng. Des. (2017), http://dx.doi.org/10.1016/j.nucengdes.2017.06.034