1.2.1. Single phase flow

In Section 2 several nuclear reactor designs as well as natural circulation applications have been described. The single phase natural circulation flow is driven by a gravity head induced by coolant density differences; the mass flow is established according to the balance between driving head and flow resistance losses. Because the (one component) density is a function of the temperature there is a functional interaction between heat exchange and natural circulation flow.

In general, the determination of the main parameters, e. g. average velocities, pressure drops, heat transfer and 1-D temperature fields for flow inside pipes or around structures, is based on established engineering practise. These parameters can be calculated with industrial codes as well as with thermal hydraulic codes used for nuclear reactor system behaviour.

Some uncertainties exist if two natural circulation flows with different densities caused by different temperatures are mixed. Then diffusion processes, turbulences and other mixing processes become important. Specific experiments would be capable to reduce the uncertainties; in some cases related experiments are necessary for a design.

If 2-D or 3-D flow fields can establish, the use of capable codes isnecessary; the codes listed in Section 3.2.2.2 can be used for calculations of this type.

Mixing processes and 3-D natural convection flow is important for deboration accidents for PWRs. For BWRs similar complicated flow fields exist for sequences with boron injection. For these processes experiments with detailed instrumentation are underway; the data can be used also for code validation.

The capability to calculate single phase natural convection flows with high confidence does allow the optimisation of systems and components, e. g. a decrease in flow resistances and an appropriate arrangement of heat sources and heat sinks.

Part 1

DESIGN CONSIDERATIONS

Natural convection and natural circulation flow and limits in advanced reactor concepts

R. B. Duffey

Atomic Energy of Canada Ltd, Canada

Abstract. Existing reactor designs and new concepts rely to varying degrees on heat removal processes driven by natural convection as a potentially important design feature or ultimate heat removal mechanism. This is independent of whether the nuclear core is cooled by water, gas or liquid metal, since in many shut down or emergency conditions forced cooling is assumed or predicted to be lost. However, using natural convection to advantage is possible, since it can provide significant cost-savings by the elimination of pumps and ancillary equipment and also can result in simplified and hence higher reliability safety systems. It is highly desirable to build on the inherent or existing heat removal processes than to graft design or add them on afterwards. The limits to the heat removal are set by the natural circulation flow and heat removal capability, so these need to be predicted with accuracy. The capability limit is determined by well-known physically linked parameters, including the flow rates, driving heads, heat sinks, fluid thermal expansion, and flow thermal and hydraulic stability. In natural convection plants, there are opportunities for the limits to be set by the absolute power output available from naturally convective flow, and the onset of instability in that flow. We are interested in the ultimate or maximum power output both in order to minimize power generation costs, and to determine how far the natural circulation designs can be developed. This paper reviews some of the fundamental equations and analytical solutions for natural convection flows, and examines their application to determine the limits of heat removal as a means of establishing simple criteria and fundamental design limits. This type of physical analysis can be used to investigate the flow and stability limits for a thermally expandable fluid, which encompasses the extremes of both low and supercritical pressure applications. To illustrate the approach, simple analytical expressions are derived for the ultimate or maximum heat removal. We can then relate the maximum thermal hydraulic limits to hypothetical reactor power output. The relationship between some of the various enhanced design features is then clear when seeking the ultimate or maximum safe power output at least cost. Hypothetical natural circulation designs are discussed as a basis.

1. INTRODUCTION

Existing reactor designs and new concepts rely to varying degrees on heat removal processes driven by natural convection as a potentially important design feature or ultimate heat removal mechanism. This is independent of whether the nuclear core is cooled by water, gas or liquid

metal, since in many shut down or emergency conditions forced cooling is assumed or predicted to be lost. However, using natural convection to advantage is possible, since it can provide significant cost-savings by the elimination of pumps and ancillary equipment and also can result in simplified and hence higher reliability safety systems. It is highly desirable to build on the inherent or existing heat removal processes than to graft design or add them on afterwards. The limits to the heat removal are set by the natural circulation flow and heat removal capability, so these need to be predicted with accuracy. The capability limit is determined by well-known physically linked parameters, including the flow rates, driving heads, heat sinks, fluid thermal expansion, and flow thermal and hydraulic stability

In natural convection plants, there are opportunities for the limits to be set by the absolute power output available from naturally convective flow, and the onset of instability in that flow. We are interested in the ultimate or maximum power output in order to both minimize power generation costs (both capital and operating), and to decide or determine how far the natural circulation designs can be developed. We call this a hypothetical design, to indicate the conceptual nature of the analysis.

Different designs have differing heat removal characteristics. For example, single-phase gas and liquid metal systems rely on single-phase natural convection, whereas water — cooled designs can utilize boiling two-phase flow. The use of flashing-driven, natural-circulation systems are being considered in advanced boiling designs, and also the use of high pressure super critical water. This important development concept for innovative designs increases the thermal efficiency by using a higher coolant temperature. The large variations in fluid properties, primarily density and enthalpy, near the critical point also introduce the potential for flow instabilities similar to those in a boiling system.

Using natural convection to advantage is possible, since it can provide significant cost-savings by the elimination of pumps and ancillary equipment and also can result in simplified and hence higher reliability safety systems. It is highly desirable to build on the inherent or existing heat removal processes than to graft design or add them on afterwards. The limits to the heat removal are set by the natural circulation flow and heat removal capability, so these need to be predicted with accuracy. The capability limit is determined by well-known physically linked parameters, including the flow rates, driving heads, heat sinks, fluid thermal expansion, and flow thermal and hydraulic stability

To illustrate the approach, simple analytical expressions are derived to illustrate the ultimate or maximum heat removal. We can then relate the maximum thermal hydraulic limits to hypothetical reactor power output and cost. The relationship between some of the various enhanced design features is then clear when seeking the ultimate or maximum safe power output at least cost. Hypothetical natural circulation design are discussed as a basis.

TPNC regime occurs as a consequence of coolant loss from the primary system. Owing to this, both driving and resistant forces increase when decreasing mass inventory of primary system. Assigned the typical geometrical layout of PWR, the former effect, i. e. increase of driving forces, is prevalent at small decreases of mass inventories. The opposite occurs for larger decreases of mass inventories. The net result is a ‘peak’ in core mass flowrate versus primary system inventory (when primary mass flowrate decreases), as can be observed in Fig. 1. Forced convection, subcooled and saturated heat transfer regimes occur in the core. Condensation occurs inside the U-Tubes of SG. The average core void fraction is typically
less than 30%, whereas at the outlet values around 50% can be reached without occurrence of thermal crisis in the considered pressure range.

For testing the scaling laws, a two-phase natural circulation loop as shown in Fig. 11 was constructed. The loop was made of 50 mm NB (2” Sch 80) pipes except for the separator which is 150 mm NB (6” Sch 120) pipe. The separated steam is condensed and the condensate is returned to the separator. The vertical heater is direct electrically heated with a high current source. The loop was extensively instrumented to measure temperature, pressure, differential pressure, level and flow rate. Further details of the loop are available in Naveen et al. (2000). Prior to the actual experiments, pressure drop across one pipe segment was measured under forced flow conditions. This gave the following equation for the friction factor

f = 0.05042/Re003768 (38)

FIG. 11. Experimental two-phase FIG. 12. Steady state flow in a two-phase natural natural circulation loop. circulation loop.

1.1. History

In 1993 it was decided by the Ministry for Research and Technology, SIEMENS, the German Utilities and the Forschungszentrum Jiilich to build a test facility to study the effectiveness of the SWR 1000 emergency condensers; the facility was built in only 18 months.

From 1996-1998 the facility was used for a project within the 4th FP of the EU "European BWR R&D-Cluster for Innovative Passive Safety Systems"; seven partners participated. In 1997, in addition, seven partners of a EU-Concerted Action "BWRCA" took part in this project.

After the accidents in TMI-2 and Chernobyl, the study of phenomena, reactor system, and containment behaviour for sequences with core melts became a major research area. Compared to phenomena under DBA conditions, the phenomena that are present during accident sequences with core melt are very complex and, evidently, much less understood. The main phenomena influencing containment behaviour are

— the release, mixing and burning of hydrogen,

— the release of volatile fission products and its decrease due to condensation and sedimentation processes,

— the transport of solid or liquid fuel (melt) into the containment and its interaction with concrete and steel structures and with the atmosphere or water.

While loads on the containment wall or structures resulting from DBAs stay within design limits and are well below loads threatening the containment integrity, this is not the case for loads resulting from severe accidents. For these the containment integrity may be threatened or even lost.

A useful overview about severe accidents can be taken from [6]:

— Although severe accidents do not belong to design basis accidents, the knowledge about the loads resulting from these sequences are needed for probabilistic safety assessments and also for the design of advanced LWRs which might take core melt sequences into account in the design process;

— The modelling of phenomena related to severe accidents should be as realistic as possible to avoid unbalanced designs or unrealistic predictions of loads on the containment resulting in unrealistic predictions of doses outside the plant;

— The modelling of phenomena related to severe accidents can only be done with relatively large uncertainty bands due to the limited experimental basis but also due to the uncertainty in predicting reliably the accident progression;

— For the calculation of phenomena and system behaviour the major lumped parameter codes as CONTAIN, MELCOR and COCOSYS in addition to many codes for special purposes(e. g. melt/concrete interaction) are used. For the assessments which need a very detailed simulation of geometrical structures (e. g. hydrogen accumulation, deflagration to detonation-transition and burns) CFD-codes are used with increasing frequency. Nevertheless, validation and model improvements are still needed.

CAPCN resembles CAREM in the primary loop and steam generators, while the secondary loop is designed just to produce adequate boundary conditions for the heat exchanger. Water enters the heated section from the lower plenum. The nuclear core is reproduced by electric heaters. The heated water flows up through the riser to the upper plenum where a vapor interphase exists. The water exits this plenum through an outer volume in contact with the steam generator. The steam generator has two coils, once through, secondary inside. The subcooled water flows down through a downcomer or cold leg to the lower plenum. Natural circulation flow may be regulated by a valve in the cold leg and a by pass to the bottom of the riser.

The secondary loop pressures and cold leg temperatures are controlled through feedback loops operating valves. The pump allows the regulation of the flow. The condenser is an air cooled type with flow control.

Both loops allow automatic control and can be pressurized by nitrogen injection.

This rig was constructed according to ASME for the following primary parameters: 150 bar and 340°C. The primary loop may operate in saturated or subcooled regimes, with a heating

power up to 300kW and different hydraulic resistance. The circuit configuration allows the study of stationary states similar to CAREM conditions of pressure, specific flow and enthalpy. Height was kept in a 1:1 scale.

FIG. 4. CAPCN simplified process and instrumentation diagram.

The test facility named as Middle East Technical University Condensation Test Facility (METU-CTF) was installed at the Mechanical Engineering Department of METU. The experimental set up consisting of an open steam or steam/gas system and open cooling water system is depicted in the flow diagram of Figure 1 [4].

Steam is generated in a boiler (1.6 m high, 0.45 m ID) by using four immersion type sheathed electrical heaters. Three of these heaters have a nominal power of 10 kW each and the fourth one has a power of 7.5 kW at 380 V. All the heaters can be individually controlled by switching on or off. The boiler tank was designed to withstand an internal pressure of 15 bar, at T = 20 °С, and was tested at this pressure. The maximum operating pressure of the tank is 10 bar. To ensure dry steam at the exit of the boiler, a mechanical separator directly connected to the exit nozzle was installed. The boiler tank was thermally insulated to reduce the environmental heat loss.

Compressed air can be supplied either to the boiler tank or to the steam line via nozzle after the orifice meter on the horizontal part of the pipe, which connects the boiler and the test section.

The pipe connecting the boiler tank and the test section has a length of approximately 2 m and an ID of 38.1 mm. The pipe was connected to the boiler tank via an isolation valve. This isolation valve is used to isolate the boiler until inside pressure of the tank is increased to a predetermined level.

The test section is a heat exchanger of countercurrent type, that is steam or steam/gas mixture flows downward inside the condenser tube and cooling water flows upward inside the jacket PiPe-

The condenser tube consists of a 2.15 m long seamless stainless steel tube with 33/39 mm ID/OD. The jacket pipe surrounding the condenser tube is made of sheet iron and has a length of 2.133 m and 81.2/89 mm ID/OD.

A total of 13 holes were drilled with an angle of 300 at different elevations along the condenser tube length to fix the thermocouples for inner wall temperature measurements. Similarly, 15 holes were drilled radially at different elevations for installation of the thermocouples to be used for cooling water temperature measurements. The jacket pipe was thermally insulated to reduce environmental heat loss. Ten thermocouples were fixed to a 2 mm diameter Inconel guide wire and installed at the central temperature measurements.

The experimental text matrix has been constituted by pure steam and steam/air mixture runs and the effect of NC gas has been analyzed by comparing the pure steam runs with mixture runs with respect the temperature, heat flux, air mass fraction, and film Reynolds number. The range of the measured parameters; Pn = 2-6 bars, Rev = 45,000-94,000, and Xi = 0 %-52%.

Relief Valve

Cooling Water Outlet

Isolating Valve

Jacket Pipe

(HI/89 ID/OD, L: 2133 mm)

the experimental heat transfer coefficient to a reference, pure steam, heat transfer coefficient. The reference heat transfer coefficient is calculated from Nusselt theory. Moreover, the enhancement of heat transfer coefficient due to the shear stress of the gas on liquid film is considered, fshear, and conveyed to the correlation. The other effects enhancing the condensation heat transfer coefficient are also taking into account, fothers, and are correlated in terms of liquid side Reynolds number, ReL. The suppression of the condensation heat transfer coefficient by the accumulation of the NC gas at the interface is clarified and denoted as f2. In this present study, both the enhancement and the suppression factors given in UCB formulation are modified by considering mixture side Reynolds number and the Sherwood number defining the radial concentration gradient of NC gas respectively.

f Type Correlation Modified by Sherwood Number

f = ff (1)

where

f is the degradation factor, f1 is the enhancement factor, and f2 is the suppression factor.

f1 = fshear, f others (2)

f shear =TL where; (3)

d 2

51: Film thickness without interfacial shear stress 5 2: Film thickness with interfacial shear stress.

The interfacial shear stress is influenced by both the interface velocity and the mixture side velocity. Moreover, the entrainment from a liquid film is associated with the onset of disturbance waves at the interface and, in general, depends on both the vapor and the liquid flow rates. In fully turbulent flow, above a film Reynolds number of 3000 the condition for the onset of entrainment depend mainly upon the vapor velocity [5]. For this reason, the fothers in Equation (2) is correlated as,

The build up of NC gases at the interface and its back diffusion into the core constitute a primary problem. The accumulation of NC gas at the interface is the principle reason for the mass diffusion resistance in radial direction, which causes lower condensation rates. This effect is encompassed into the correlations with the aid of air mass fraction in UCB correlation. However, air mass fraction is not defining the ongoing process, which is originally governed by concentration gradient formed between the interface and the core. Therefore, the Sherwood number is used instead of air mass fraction in the suppression factor, f2. When the variation of f2 is investigated with the Sherwood number, the segregation of individual runs from each other is observed and this situation is attributed to inlet pressure which is, therefore, superimposed into the correlation as air mole fraction. Under the light of these arrangements, the suppression factor is formulated as follows.

f2 = 1 — C3.Shrrz3 (5)

where

Shrr = yg. Sh (6)

yg: Mole Fraction of air Sh: Sherwood Number

C1-C3 and zi-z3 are the constants to be determined.

Results and the comparison of the correlations are given in section 4.2.

Boundary conditions were set as follows: room temperature of 293 К (20°C) in each part of the facility and atmospheric pressure of 1 bar. The level in the pools is at the elevation of 10 m, which is the equivalent of 2.85 m measured from the bottom of the pool. The initial velocity of the fluid is 0 m/s everywhere in the facility. The input heating power was 99 kW in equal radial distribution in the 3 heated channels of the core. Upper plenum was filled with water totally. Table II shows the performed scenarios and status of the isolation valves. The initial condition of the experiments did not fully correspond the condition of the reference reactor in a LOCA situation after the depressurization period. The initial temperature is close to room temperature when density differences, which create the natural circulation, doesn’t exist yet.

TABLE I. THE GEOMETRICAL CHARACTERISTICS OF THE MODIFIED PACTEL FACILITY

TOPFLOW will be equipped with advanced two-phase instrumentation mainly and adapted and developed in Rossendorf, such as wire-mesh sensors, needle-shaped conductivity probes with integrated thermocouple, gamma and X-ray tomography and passive ultrasonic droplet probes. Additionally, laser-doppler anemometry and a phase-doppler particle analyser are available. Two of these devices, the needle-shaped conductivity probes with integrated thermocouple and the wire-mesh sensors will be described in detail.

Advanced needle probes are equipped with a micro thermocouple substituting the traditional electrode. These probes combine a local phase indication with a fast temperature measurement, so that the temperature can be correctly related to the instantaneous phase state. This allows to distinguish between steam and non-condensable gas in the gas phase and can be used to measure the sub-cooling in the liquid phase and temperature gradients at the inter­phase boundary [5].

FZR has developed wire-mesh sensors for gas-liquid flows, which allow a fast visualisation of transient gas fraction distributions in a tube [6]. The function is based on the measurement of the local instantaneous conductivity of the two-phase mixture. The sensor consists of two electrode grids with 16 electrodes each, placed at a small axial distance behind each other. The conductivity is measured at the crossing points of the wires of the two grids. One plane of electrodes is used as transmitter, the other as receiver plane. During the measuring cycle, the transmitter electrodes are activated by supplying with voltage pulses in a successive order. The currents arriving at the receivers are recorded. This procedure is repeated for all transmitter electrodes.

Sensors of this kind were applied at FZR to an air-water flow test loop in a vertical pipeline (inner diameter D = 51.2mm) as well as to a cavitating flow behind a fast acting valve. The high resolution of the sensor allowed to obtain bubble size distributions and to study the evolution of the flow structure along the pipe [7]. The maximum time resolution available to perform these experiments was 1200 measurements per second. Recently, the measuring rate was increased to 10 000 frames per second with sensors of 16 x 16 measuring points. In the result it is possible to visualise and quantify individual bubbles or droplets at a much higher flow velocity, than before.

It is planned to apply this type of fast flow visualisation to different test sections of the TOPFLOW facility. In the first experiments, the flow pattern in a vertical pipeline of 200 mm diameter will be studied. A special developed sensor will allow to achieve a spatial resolution of 3 mm at a measuring rate of 2 500 frames per second. For this purpose, the sensor must consist of 64 transmitter and 64 receiver wires (64 x 64 measuring matrix).