GB2236210A - Core catchers for nuclear reactors - Google Patents

Abstract

A core catcher (25) for containing nuclear core debris in the event of a breach in the reactor pressure vessel (10) caused by a core meltdown. The core catcher has a multilayer sandwich construction comprising a middle layer of interlocking tongue-and-groove jointed refractory (e.g. zirconia) tiles or bricks (40) sandwiched between inner and outer steel plates in the form of domes (30 and 32). The refractory bricks (40) are fixed against movement relative to each other and the inner and outer steel plates (30 and 32) by means of refractory cement. The inner steel plate (32) is sacrificial in the event that it comes into contact with molten nuclear material but gives the sandwich construction greater shock resistance during normal operational service. The outer steel plate provides the main structural support for the core catcher. <IMAGE>

Description

CORE CATCHERS FOR NUCLEAR REACTORS The present invention relates to core catchers for nuclear reactors.

Attention is rightly being focused on the safety of nuclear reactors in the event of accidents involving loss of coolants, coupled with failure of the emergency core cooling system to cool the reactor core. Although in properly designed reactor systems such an accident is a very remote possibility, nevertheless if it did occur, temperatures within the reactor vessel would rise rapidly to cause a meltdown of at least parts of the reactor core.

A subsequent melt-through of the reactor vessel would then threaten the containment boundary of the reactor system.

Hence, there is a perceived need for a means of preventing molten core debris and other material from coming into contact with the containment boundary, such a means being conveniently labelled a "core catcher".

Various types of core catcher have been proposed in the last two decades. For example, US patent number 3607630 basically proposes a water-cooled metal basin to catch, cool, solidify and retain molten core material dropped from the reactor vessel upon a melt-through of the vessel. The basin (stated to be of steel) provides a level floor upon which the molten core material may spread out in a thin layer and be cooled and be maintained solid against radioactive decay heat by transfer of heat to water passing through tubes in the basin.The water is supplied to the tubes from an elevated storage tank, and is discharged from the tubes through an outlet positioned at a height at least equal to the water-level in the storage tank. Waterflow through the tubes is started and maintained automatically by gravity-induced natural circulation caused by heat from the molten core materials.

An alternative approach is seen in US patent number 3702802, in which the core catcher comprises a ceramic oxide eutectic with a relatively low melting point, which can dissolve the reactor fuel when molten. The preferred material is interlocking basalt blocks, and these rest on a layer of depleted uranium dioxide or fire brick for insulation purposes, the outer layers being a steel containment vessel backed up by a concrete foundation containing water cooled pipes. If core meltdown occurs, the molten mass of fuel falls onto the basalt, melts a portion of it and dissolves in the resulting pool of molten basalt, the fuel thereby being dispersed over a larger volume so that its temperature is reduced to the melting point of the basalt and the available surface area for heat dissipation is increased.The insulating layer then ensures the integrity of the containment vessel whilst the molten pool gradually cools down and solidifies.

Unfortunately, both these prior proposals suffer from the disadvantage of being large, complex and very heavy, requiring high capital investment additional to the investment required for the reactor itself, and also imposing high maintenance costs. Furthermore, a core catcher for a nuclear reactor should ideally be as compact and shock-proof as possible to guard against damage from such contingencies as earthquakes, but neither of the above proposals seem to offer this to a very high degree.

According to the present invention a core catcher for receiving and containing molten nuclear core material from a nuclear reactor has a multilayer sandwich construction in which a core-catching layer is sandwiched between a sacrificial layer and a structural support layer, the core-catching layer including a stratum of bricks or the like having lap joints between the bricks, means sealing said joints against penetration by said molten nuclear core material and means securing said bricks against movement relative to each other and the sacrificial and structural layers, said sacrificial layer being sacrificial in the event that it comes into contact with molten nuclear core material, said bricks comprising insulating refractory material resistant to chemical attack by said molten nuclear core material, said structural layer providing the main structural strength of the core catcher and comprising metallic material capable of supporting said core-catching layer without the occurrence of substantial strain during reception, containment and eventual solidification of the nuclear core material.

For the purpose of the present specification and claims, the word "brick" should be taken to embrace blocks and tiles.

In the exemplary embodiment, the structural support layer comprises heavy gauge steel plate. The sacrificial layer also comprises steel plate, but of a lighter gauge.

The bricks may comprise one of the following ceramic substances: dense sintered alumina, beryllia, calcia, lithia, thoria or zirconia. The means sealing the joints between the bricks against penetration and the means securing the bricks against movement comprises refractory cement chemically and thermally compatible with the bricks.

In the exemplary embodiment, devised specifically for the pressurised water type of reactor, the bricks comprise dense sintered zirconia, preferably with a density exceeding 70% of the theoretical density of solid zirconia the cement being zirconia-based. Hence, besides the stratum of zirconia bricks, the core-catching layer preferably additionally comprises a bed of zirconia cement between the structural support layer and the bricks and a further bed of zirconia cement between the bricks and the sacrificial layer, the joints between the bricks being filled by further zirconia cement.

Preferably, the lap joints between the bricks comprise tongue-and-groove features on confronting portions of adjacent bricks.

To maximise integrity of the sandwich construction, it is substantially bowl-shaped. However, it is preferred that its radius of curvature - is greater near its axis of symmetry than near its edge to minimise the depth of any pool of molten core material held therein.

The surfaces of the core-catching layer, the sacrificial layer and the structural support layer may be substantially parallel to each other. Alternatively, the core-catching layer may increase in thickness towards the centre of the core-catcher such that the inner surface of the core catching layer and the sacrificial layer is less concave than the outer surface of the core-catching layer and the structural layer. To maximise cooling efficiency in cases where cooling is facilitated by contact of the outside of the core catcher with water, it is envisaged that the outer surface of the core-catching layer may be conical in form.

The stratum of refractory bricks conveniently comprises a plurality of concentric courses of bricks arranged around a centrally located brick.

The invention also embraces a method of constructing a core catcher for containing a molten nuclear core, comprising laying a first bed of refractory cement on a metallic structural support layer, laying a stratum of refractory bricks or the like on said first bed of refractory cement, the bricks having lap joints therebetween, filling said joints with refractory cement as said bricks are laid, laying a second bed of refractory cement on the stratum of refractory bricks, and seating a sacrificial layer on the second bed of refractory cement to form a multi-layer sandwich structure.

In an alternative method of construction, instead of laying the second bed of refractory cement on the stratum of refractory bricks before the sacrificial layer is seated thereon, the sacrificial layer is first positioned to leave a gap between itself and the stratum of refractory bricks and liquid refractory cement is thereafter pumped into the gap to substantially fill the same and provide the inner bed of refractory cement.

An embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which: Figure 1 is a schematic drawing showing a side elevation in axial cross-section of a reactor core pressure vessel and an associated core catcher in position within a containment vessel; Figure 2 is a view on section A-A in Figure 1 with the pressure vessel removed and part of the core catcher's inner skin broken away to show its interior construction; Figure 3 is a partly sectioned perspective view of a large scale model core catcher in course of assembly; Figure 4 is a schematic drawing showing a cross-sectional side elevation of the finished model core catcher of Figure 3 under test conditions; Figure 5 is a graph showing how temperatures in the model varied during the course of the test; and Figures 6 and 7 are partial views of schematic cross-sectional side elevations of further embodiments of the invention.

Referring now to Figures 1 and 2, a cylindrical pressure vessel 10 with a domed lower end 12 contains the core of a pressurised water-cooled nuclear reactor (PWR, not shown) and its associated apparatus. The nuclear reactor may be of any size up to 1300 MW electrical power.

The pressure vessel 10, supported and held at its top end (not shown) in any convenient known way, is located between side walls 15 and floor 20 which form part of the final containment structure 22 for the nuclear reactor. However, interposed between, and spaced from, the floor 20 and the domed end 12 of the pressure vessel 10 is a core-catcher 25, which is a circular bowl-or dome-shaped structure. The somewhat flattened or shallow shape of the dome helps to minimise the depth of any pool of molten material and core debris contained by core catcher 25, and maximise its surface area, thereby aiding more rapid cooling of the molten mass and minimising risk of criticality.

The core catcher 25 has a structurally strong outer dome 30 of heavy gauge structural steel (say, in excess of 50 mm according to the load bearing strength required) with an upturned rim 31 which is welded at arrows W to supporting structure comprising a cylinder 35 surrounding, but spaced away from, the reactor pressure vessel 10. The supporting cylinder 35 is also made of structural steel.

Its gauge is shown as approximately the same as the outer dome 30 of the core catcher, but will depend upon the overall size of the core-catcher and on whether any additional support structure is provided for the cylinder.

In this particular embodiment the cylinder 35 is partially supported and held at its top end (not shown) in any convenient known way, but part of the combined weight of cylinder 35 and core catcher 25 is also supported by structural steel A-frames 37, resting on the floor 20 of the containment structure. For clarity, only three of the frames 37 are indicated in the drawing, but there would of course be as many as necessary to adequately support the core-catcher, arranged in concentric circles about the vertical centreline V of the containment structure. As well as successive circles being cross-braced as shown, pairs of adjacent A-frames within each concentric circle are braced against each other by virtue of being mutually inclined and attached at their apices, thus forming a space-frame structure of a well-known type.The apices of the A frames 37 are also attached to circular rails R to avoid point loads on the underside of the outer dome 30.

At least for reactors at the lower end of the above-quoted size range, it would be possible, if desired, to support core catcher 25 and cylinder 35 at its top from the walls 15 of the containment structure, thereby eliminating the need for support frames 37. However, in that case the thickness of the cylinder 35 relative to outer dome 30 would have to be increased to provide the requisite stiffness and load carrying capacity. An arrangement eliminating the need for support frames under the core catcher would of course be particularly desirable for isolating the core catcher from the effects of earth movements.

The core catcher 25 is of a multi layer sandwich construction and besides the outer dome 30 has a similarly shaped but lighter gauge structural steel inner dome 32 with upturned rim 33, the gap between the two domes being filled with dense refractory, preferably sintered ceramic tiles or bricks 40 secured in position with an appropriate refractory cement as described in more detail later. The upturned outer edge of the bowl-shaped sandwich construction is capped by a steel ring 38 which extends from the inside surface of the support cylinder 33, across the outermost of the bricks 40, and is welded to the upturned edge 33 of the inner steel dome 32 through a downwardly turned tapered flange 39.

The protective and insulative envelope formed by the core catcher 25 and its support 35 is of course pressure tight to a high degree. Either during normal operation of the reactor, or in the event of an emergency, the cavity 45 defined between this envelope, the walls 15 and the floor 20 may be filled with water. In the event of a melt-down in the reactor core, and a consequent melt-through of the bottom 12 of the reactor pressure vessel 10, the core catcher 25 retains the molten material and gradually releases heat to the water in the cavity 45, which boils off, taking heat with it. If the containment structure 22 is not open-topped, vents must be provided to allow steam to escape and so relieve pressure. However, the core catcher may be constructed sufficiently robustly to contain a molten core even without water in the cavity 45.Watercooling with only be absolutely necessary if "worst-case" analysis during design calculations shows that radiative, conductive and convective cooling with air in cavity 45 is inadequate. In either case, the supporting frames 37 allow satisfactory circulation of water or air under the core catcher 25.

The structure of the core catcher will now be described in more detail.

Whilst the exact specifications of the refractory bricks can be varied to match design criteria which may differ between reactor types and sizes, they should be of a consistently fine grain and a high density. We prefer sintered zirconia bricks for use with PWR's, the bricks having a minimum density of 4400 kg/m3, i.e. at least 72% of the theoretical density of solid zirconia. Furthermore, the top face of the bricks should be manufactured with a smooth finish for minimum interaction with flows of molten material and other debris thereover should a core meltdown occur. The inner steel dome 32 is of course sacrificial in such an event, its main purpose being to help protect and contain the bricks during normal operation of the reactor.

However, it would remain intact for a short time after impingement of the molten core material upon it, thus reducing the severity of the initial interaction between the core material and the cement and bricks.

Sintered zirconia was chosen as the preferred material for the bricks in the context of PWR's because of its high melting point, its resistance to chemical attack from the core materials of PWR's, its low coefficient of thermal expansion, its low thermal conductivity and its commercial availability in high-grade brick form. A molten nuclear core of the type mentioned, besides being at a very high temperature, contains materials which are chemically highly reactive. Our engineering judgement and research shows that zirconia is a suitable material under these conditions. Zirconia bricks are available from Refractory Mouldings and Castings, Limited, Kegworth, Leicestershire, England under the trade designation ZAL Zirconia Tile.

Clearly, it is necessary that the sintered zirconia bricks, or other refractory material used, give adequate thermal protection to the outer structural steel dome 30 of the core catcher. Their 'insulative effect must be sufficient to prevent any substantial high-temperature creep effects or other strain damage to the dome 30 over an extended period of time whilst the molten material gradually cools. "Substantial strain" in the dome.30 would of course be such as could lead to fissuring, breakage or movement of the bricks and their securing cement and consequent penetration by molten material. At the same time, the bricks must not insulate the core debris excessively from heat loss through the core catcher, because it is necessary to ensure cool-down and solidification of molten material within a reasonably short time-frame.We project, for instance, that for reactors in the size range 5 to 250 off thermal power, sufficient rigidity of the outer structural steel dome, combined with adequate cooling rates for the contained material, is maintainable with zirconia brick thicknesses in the range 50 to 200 mm. Regarding other dimensions of the bricks, practical requirements indicate that the bricks should be light enough to ease handling and laying and yet large enough to minimise the number required.

As can be seen in Figures 1 and 2, a tongue-and-groove lap joint configuration is adopted for the bricks 40 which comprise the middle layer of the sandwich construction, the tongues and grooves being provided at the mid-point of the bricks' thickness.

The bricks 40 are laid into a substantially uniform layer of compatible refractory bedding cement 42 (zirconia cement for zirconia bricks) on outer dome 30 and are arranged in circular courses 60-67 around a central circular key brick 50. Each course of bricks 60-67 comprises a short thick-walled cylinder, individual bricks comprising sectors of the cylinder, being quadrilateral in plan view. Location features 55 are provided for the key brick 50 to initiate accurate assembly, comprising a square steel boss fixed to the interior of the outer dome 30 at its centre, and a correspondingly shaped recess in the key brick 50, this brick being installed first and the others being installed around it.

It will be apparent that the bricks in course 66 must conform substantially to the knuckle-shape where the outer and inner steel domes 30,32 transition to their cylindrical upturned rims 31 and 33. Also, the bricks in the outermost course 67 must conform to the cylindrical rims 31,33 of the domes and to the transition 37 between those rims and the thicker supporting cylinder 35.

It will further be apparent that in order to conform accurately to the shape of the core-catcher's outer and inner domes 30,32 the bricks 40 in courses 60-65 must have a slight two-dimensional curvature on their upper and lower faces, i.e. they must be slightly"dished" in form.

An examination of Figure 2 also shows that the dimensions of the bricks, and their relative dispositions, are such that the radial joint lines between abutting bricks in each of the courses 60 to 67 are not coincident with the radial joint lines in the adjacent course. This helps to prevent the formation of extended "lines of least resistance" with respect to penetration by molten material.

Provision of the tongue-and-groove lap joint features aids accurate assembly of the courses of bricks and helps prevent penetration of molten debris between bricks. The radially outer cylindrical face of the key brick 50 has an annular groove which receives corresponding tongues projecting from the confronting radially inner partcylindrical faces of the bricks comprising course 60. The bricks in their successive courses 60-65 are all similarly provided with grooves on their radially outer partcylindrical faces and tongues on their radially inner part-cylindrical faces.Although the bricks in course 66 have radially inner part-cylindrical faces with tongues which mate with grooves on the radially outer partcylindrical faces of the bricks in course 65, the consequence of the change between the dome shape and the upturned cylindrical rim of the core catcher is that the confronting faces of the courses 66 and 67 are planar except for the provision of the tongues in the bricks of course 67 and the grooves in the bricks of course 66.

Besides the above-described interlocking between the courses 60-67, the bricks within each course are circumferentially interlocked with each other in similar fashion, one side-face of each brick being provided with a tongue and its other side-face being provided with a groove.

For brick thicknesses of the order of 70 mm, we prefer the size of the tongue on each brick to be 10 mm radius, the mating groove being 12 mm radius. For thicker bricks, the tongue and groove sizes may be scaled up.

As already mentioned, the bricks 40, starting with key brick 50, are laid into a substantially uniform layer 42 of refractory bedding cement which is trowelled onto the inner surface of the outer dome as the courses of bricks are built up. For bricks of the above thickness, we ensure that there is a 5mm minimum thickness of bedding cement 42 under each brick using feeler gauges. All joints between adjacent bricks are limited to a maximum gap width of 2mm using feeler gauges and are filled with a jointing cement.

A uniform layer 43 of refractrory infill cement is also applied to the top surfaces of the bricks by trowel as they are laid, and we ensure that a 3mm minimum thickness is maintained for this cement layer in order to fill up substantially all the space between the bricks and the inner dome.

After the bricklaying and cementing operations have been completed and after setting of the cement, the inner steel dome 32 of the core catcher is placed in position on top of the infill cement layer to complete the sandwich construction. Assuming water-based cements are used, the cement is cured by heating the inner dome 32 and venting moisture in the cement through small outlet holes in the inner dome 32 which are blanked off and permanently sealed, e.g. by welding, after curing is complete. It is most important to ensure that substantially all water has been driven off from the interior of the sandwich structure because core debris reacts violently with water. Finally the upturned edge of the core catcher is capped by the steel ring 38.

Depending upon the accuracy with which the bricks 40 are laid, the infill cement layer applied, and the inner dome 32 manufactured, this procedure could result in an air gap in certain areas between the inner dome and the cement, but this is not thought to be a problem.

As an alternative to hand-trowelling of the infill cement onto the top surfaces of the bricks, a pump-filling option can be chosen. In this, the bricks are left without a top layer of infill cement after laying and instead are provided with a few spaced-apart appropriately dimensioned lands of hardened cement onto which the inner dome 32 is lowered and positioned so as to obtain the required spacing between the inner and outer domes, the edge of the core catcher then being capped by ring 38. Infill cement is then pumped as a viscous liquid into the interspace left between the top surfaces of the bricks and the inner surface of the dome 32, the pumping being done through a hole at the bottommost point of the inner dome and venting occurring through outlet holes in the top-most part of the inner dome. Once the cement appears at the outlet holes, they are temporarily plugged and the cement pressurised to increase its density: The cement is then left to set, the plugs are removed, and moisture in the cement driven off through the holes as described previously. Finally, after drying, all holes in the inner dome are blanked off as before.

It is prudent to ensure that the minimum gap between the bricks and the inner dome is at least 5mm to allow for easy pumping of the viscous cement.

Assuming that sintered zirconia bricks are used, the the three types of water-based cement mentioned above are preferably of the silica/alumina/zirconia mortar type as known in the furnace lining industry, such as set out below: (a) Bedding cement Type: Zirconia with Sodium Silicote binder Zalsil (TM) bedding cement.

(b) Infill Cement Type: Zirconia with Sodium Silicote binder Zalsil (TM) bedding cement.

(c) Jointing Cement Type: Zirconia Cement - ZAL (TM) jointing cement with a particle size of 150 pin.

With the inner dome 32 welded in position and all holes blanked off, the integrity of the finished core catcher can be checked with full radiography, dye penetrant and leak tests.

From the above description of manufacture it will be understood that since the refractory bricks are close fit interlocked and cement jointed together, surrounded by cement and encased between two steel domes, the lower one being very rigid, it is unlikely that the bricks 40 will move or be damaged even under shock conditions. The structure will also be adequate to resist radiation induced material degradation throughout the life of the reactor.

With a core catcher installed as in Figure 1, the accident sequence presenting the greatest threat to the containment structure is as follows (only a general indication of times from the start of the accident sequence for each phase are given in the parentheses as such times vary according to the reactor size and the operating power history of the reactor immediately prior to the accident):- For some reason, such as an unforseen accident involving loss of coolant and failure of emergency cooling and shutdown systems, core materials melt and collapse to the bottom of the pressure vessel 10 (minutes to hours).

A complex process involving solidification and re-melting of the core debris eventually results in failure of the pressure vessel 10 either by a combination of thermal shock and wide-spread melting, or by thermally induced mechanical effects.

Molten material and other debris pours or jets from the vessel failure site onto the core catcher inner steel dome 32 with some risk of erosion of the infill cement as the inner dome melts (minutes to hours).

The core catcher temperature then rises and heat is transferred from the lower region into the cavity 45.

If cavity 45 is only an airspace, cooling is mainly by means of radiative transfer to the walls and floor of the containment structure and by convection of air in cavity 45 and the core catcher temperature rises rapidly to a high but not destructive value (several hours).

Alternatively, if the cavity is water-filled, the temperature of the water in the cavity rises and boiling begins, thus enhancing the cooling effect until the core catcher becomes uncovered whereupon, unless more water is added, heat transfer is reduced and the core catcher temperature rises again (several to many tens of hours).

Eventually, the radioactive decay heating rate reduces significantly and temperatures start to decline (days).

The accident design criteria for the core catcher are related to the accident sequence described and are as follows: a target failure probability derived from a fission product release risk assessment chosen to satisfy public safety criteria, etc.

to withstand the pressure impulse or mechanical impact during a pressurised melt-through of the reactor vessel to maintain integrity during: (a) the initial thermal transient arising from the impact of hot molten material and other debris with the upper surface of the core catcher; (b) the medium term with adequate heat transfer from the lower surface of the core catcher; (c) the long-term cooldown of the captured debris.

To provide sufficient heat transfer from the molten debris to the water, if present, the containment structure around the core catcher and to the environment so as to cool down in a controlled manner and minimise the vapourisation of gaseous and volatile fission products.

To minimise at all times the degradation of the containment boundary.

In the design of the core catcher, a number of important features need close attention if the design is to meet these criteria: (i) The dimensional tolerances for many of the components are preferably controlled within a few millimetres. Particularly critical are brick thickness and profile, the forged steel domes' radii and the cement thickness beneath and between the bricks.

(ii) As mentioned previously, care should be taken to ensure that the rate of heat transfer through the refractory layer, and the rate of heat loss from the outer structural steel dome, are such as to ensure reasonably rapid cooldown of the captured core, maintain rigidity and avoid the production of any substantial strain in the dome due to heat-softening of the steel.

(iii) Careful control of brick manufacture is crucial to obtain the characteristics already detailed.

(iv) The correct brick profile must be maintained throughout laying to make sure that the inner dome will fit.

(v) During storage and laying of the bricks humidity should be controlled to ensure uniform drying after laying, since corrective action is extremely difficult at a later stage.

Our preferred method of manufacture of zirconia bricks comprises the fabrication of plaster and wood formers using profiles generated by computer aided draughting techniques.

Onto these formers are laid the moulds for the bricks, and the moulds, which are slightly bigger then required to accommodate shrinkage, are filled and then fired. A separate former shape is required for each course of -bricks. Acceptance of the bricks for quality follows visual, dimensional, ultrasonic, weight and destructive testing. In our opinion bricks having surface or internal cracks wider than 1.5 mm are not acceptable.

In order to test whether a design like that described above had the ability satisfactorily to withstand a core meltdown and meet the above criteria, an engineering test model 70 was constructed as shown in Figure 3 and was tested as shown in Figure 4. In our opinion the model was sufficiently large, at 1.6 metres outside diameter, to give a high degree of confidence that the test is a realistic valid representation of actual circumstances for full-size reactors.

The construction procedure for the test model 70 will be described with reference to Figure 3.

Sintered zirconia bricks as previously described but with a thickness of 72 mm were utilised, with an outer structural steel dome thickness of 52 mm and an inner sacrificial dome thickness of 20 mm. Zirconia bedding, jointing and infill cements are previously described were also used, with a bedding cement thickness of 5 mm and an infill cement thickness of 3 mm.

First of all, the bricks 72 to be used were checked for size, weight, consistency and finish, and dry assembly of the bricks was carried out in the outer dome, the bricks being marked to represent their position. A reference plan was produced by this method, so that the result could be repeated. Shims 75 were used to ensure that a constant 5 mm gap was maintained between the undersides of the bricks and the inside of the outer dome 80 during dry assembly.

Wet assembly was begun by installing the central round brick 90 on a central steel boss (not shown) welded to the inner surface of the outer dome 80, as already mentioned in connection with the embodiment of Figure 1, ensuring by means of feeler gauges that there was a 5 mm minimum thickness of bedding cement under the brick 90.

Provided on the brick 90 and integral with it was a boss (not shown) acting as a central location feature for the foot 95 of a post 97 supporting a sail-type gauge plate 100. The foot 95 acted as a journal bearing for the bottom of the post 97, so that the gauge plate 100, which was fixed to the post 97 by welded straps 105, could be rotated around the interior of the test model. The other end of the post 97 was journalled in a bearing 110, which was supported centrally of the dome 80 by means of arms 115 extending horizontally between the bearing 110 and the rim of the dome 80.

Gauge plate 100 aided accurate installation of the bricks 72, its bottom profile 120 being at a height 20 mm above the desired line of the top surface of the bricks to allow insertion of a gauge to check dimensions.

After installation of the gauge plate 100, the remaining bricks were installed course by course. As it was laid, each brick was independently checked with a special gap gauge inserted between the gauge plate profile 120 and the top surface of the brick to ensure that the 20 mm gap was being maintained. This ensured that there was always a 5 mm minimum thickness of bedding cement under each brick, assuming of course that the shape and dimensions of the dome 80 and the bricks were within allowed tolerances.

In fill cement was then applied to the top surface of the bricks by trowel. A 3mm minimum thickness of infill cement was ensured by using a gap gauge reading of 17mm between the gauge plate profile 120 and the top surface of the cement. After the infill cement had hardened, the gauge plate assembly was removed and the locating boss on central brick 90 was machined off level with the top of the infill cement layer. The sacrificial inner steel dome (not shown) could then be lowered onto the infill cement layer and secured.

Having constructed the model core catcher 70, it was tested as follows.

The model 70 was installed in an electric arc furnace 130 as shown in Figure 4, being supported at the bottom on a circular support wall 135. The top removable portion of the electric arc furnace is not shown, being of a standard industrial type. Such a furnace was used to allow the simulation of radioactive decay heat power by controlling the temperature of the molten pool within the core catcher.

The test was initiated by filling the core catcher as shown with several tonnes of molten maganese/molybdenum steel 140 at about 16400C. On completion of the pour the temperature of the molten steel was estimated to be 15500C.

The roof of the furnace was then lowered into position, the carbon arcs were inserted and the temperature was raised to follow, as nearly as could be managed, a control profile with respect to time which was ' determined from a theoretical model representing the radioactive decay heat characteristic of a melted core for a nuclear reactor of the PWR type.

During the test, temperatures were monitored by means of thermocouples A to E which had been built into the model during its assembly at the points shown in Figure 4. The temperatures as determined by these thermocouples are shown in Figure 5 on a linear scale of temperature against a contracted scale of time in hours, the curves being labelled with the thermocouple designations. Also shown as a dashed line T is the temperature profile representing the theoretical radioactive decay heat characteristic mentioned above, to which the temperature of the molten metal was approximately matched by switching the power to the electrodes on and off. The irregular solid line M is the actual temperature of the melt as measured by dip thermocouples (not shown).

It will be noted that thermocouple B was destroyed approximately 45 minutes after initiation of the test.

This probably occurred when the inner sacrificial steel dome 145 of the model core catcher melted.

In order to maintain the level of the melt 140 during continued oxidation and vapourisation at the high temperatures involved, the melt was recharged after 3 hours 45 minutes and 8 hours respectively. After 15 hours the electrodes were removed and lid 150 placed over the melt.

After 24 hours temperature recording was stopped. External examination of the model core catcher 70 when cool showed no visible signs of damage. The model was dismantled by flame cutting the top ring of solidified melt and removing the solid melt debris. This revealed that the bricks had survived with little or no damage, close examination showing a maximum penetration into the joints between the bricks of about 20mm.

The results of this test verified the soundness of our inventive concept, particularly because many of the pressurised water reactor accident transients considered in arriving at the design would involve lower decay heating than that assumed by the dotted decay curve shown in Figure 5, and also because the model core catcher was tested without benefit of the effective water cooling of its underside which we in practice prefer, as described in relation to Figure 1. However, the test also gives reason for confidence that even if the water in the reactor cavity 45 was absent, integrity of the core catcher outer dome 30 would be maintained.Furthermore, cognizance is taken of the fact that designers and operators of reactor systems might prefer not to use water in the containment vessel for cooling, relying instead on either pumped flow or natural convection of other coolant liquids or gases, such as air.

As will now be apparent to those skilled in the art, appropriate scaling of the thickness of the bricks and the inner and outer steel layers will be necessary to ensure adequate structural strength and rigidity whilst the core catcher is actually catching and holding a large molten core. One design parameter to which particular attention should be paid if it is believed that the core catcher will have to contain large amounts of molten radioactive material, is the radius of curvature of the core catcher dome structure: the greater the amount of molten radioactive material, the more important it is to ensure that its depth is minimised by allowing it to spread out over the largest available area.This maximises cooling rates and avoids criticality, and can be achieved (as in the exemplary embodiment) by appropriately minimising the radius of curvature of the core catcher's dome shape, at least near its axis of symmetry.

Another factor of relevance to the design of large core catchers is the need to assemble their inner and outer domes from a number of sections, these being welded together on-site because of difficulties with transport.

This may necessitate construction of the core catcher with the aid of scaffolding or other temporary support.

Whilst the above exemplary embodiment of the invention has contemplated shapes of core catcher in which the surfaces of the sacrificial inner steel layer, the core catching middle layer and the structural outer steel layer run parallel to each other, it would be possible to depart from such a configuration without going outside the scope of the invention.For example, as shown in Figure 6, it would be possible to have a shallow dome-shaped outer structural layer 600 of constant spherical radius r1 (apart from its peripheral upturned rim, not shown) with matching outer surface of the middle refractory layer 602, combined with a central region of the inner surface of the middle layer which has a larger radius of curvature R, thus forming a very shallow, or even flat, surface for the retention and spreading of molten debris, but surrounded by a circumferential lip region of increased curvature r2 for retaining the core debris, with a matching inner sacrificial layer 604, the differences between the outer and inner layers in rate of change of curvature with distance from the centreline 606 being taken up by increasing the thickness of the bricks and/or one or both of the cement beds towards the centreline 606 of the middle refractory layer 602.

As a further example, Figure 7 shows that it would be possible to have a bowl-shaped inner surface of the middle layer 702, with radius of curvature (say) r2 and with matching inner sacrificial layer 704, but at least the central portion 708 of the outer surface of the middle layer being a shallow conical shape, with a matching outer structural layer 700, the thickness of the middle layer again increasing towards the centreline 706 of the core catcher. As shown in Figure 7, the cone-shaped surface 708 and the matching structural layer 700 can advantageously transition to spherical forms (with radius, say, rl) towards the outer region of the core catcher.An advantage of having a cone-shaped outer surface for at least the central region of the core catcher is that in the case of contact with water in the cavity 45, a cone-shape is more efficient overall than a dome shape for facilitating escape of vapour from the region next to the surface in the middle of the underside of the dome, when water in contact with the surface is boiling. Consequently the rate of cooling of the core catcher is increased by the provision of the cone-shaped outer surface.

Although only structural steel has so far been specified as a constructional material for the sacrificial and structural layers in the above exemplary embodiments of the invention, it would of course be possible to use other materials provided that their properties were superior to, or not greatly inferior to, the properties of structural steel. For instance, it would be possible to use a high-grade alloy steel, or even a superalloy, for the outer dome of the core catcher. This would reduce problems with heat softening of the structural layer and/or would allow for a greater rate of heat transfer through the bricks, which could then be made of a material with a higher thermal conductivity than zirconia, not substantially inferior refractoriness and appropriate resistance to chemical attack from nuclear core debris.

Although zirconia has been specified above as the most suitable refractory material for the capture and retention of core materials following the meltdown of a core in a PWR, and is also thought to be suitable for boiling water reactors, zirconia may not be so suitable for other types of nuclear reactor having different core materials and this must be taken into account during design. Plainly, the cement chosen for joint sealing and securing of the bricks would have to be chemically and thermally compatible with the chosen brick material. Thus if alumina bricks were utilised instead of zirconia, alumina-based cements would also be utilised.

As indicated above, for any specific type of reactor core, the choice of core-catcher materials will depend upon an analysis of such factors as the chemistry of the molten core, its temperature, and the time taken for radioactive heating to subside. Both the core-catching refractory layer and the structural support layer of the core catcher must of course be materials with appropriate expansion coefficients, thermal conductivities, mechanical strengths and refractoriness. In the- case of the core-catching layer, the bricks must have a low degree of reactivity with the molten core materials.

Using such factors to select the brick material, we would advise selection from the group of ceramics consisting of zirconia, alumina, beryllia, lithia, thoria or calcia.

Claims (19)

Claims:

1. A core catcher for receiving and containing molten nuclear core material from a nuclear reactor, the core catcher having a multilayer sandwich construction in which a core-catching layer is sandwiched between a sacrificial layer and a structural support layer, the core-catching layer including a stratum of bricks or the like having lap joints between the bricks, means sealing said joints against penetration by said molten nuclear core material and means securing said bricks against movement relative to each other and the sacrificial and structural layers, said sacrificial layer being sacrificial in the event that it comes into contact with molten nuclear core material, said bricks comprising insulating refractory material resistant to chemical attack by said molten nuclear core material, said structural layer providing the main structural strength of the core catcher and comprising metallic material capable of supporting said core-catching layer without the occurrence of substantial strain during reception, containment and eventual solidification of the nuclear core material.

2. A core catcher according to claim 1 in which the structural and sacrificial support layers comprise steel plate of heavy and relatively lighter gauges respectively.

3. A core catcher according to claim 1 or claim 2 in which the bricks comprise one of the following substances: dense sintered alumina, beryllia, calcia, lithia, thoria or zirconia.

4. A core catcher according to claim 1 or claim 2 in which the bricks comprise dense sintered zirconia.

5. A core catcher according to claim 4 in which the bricks have a density in excess of 70% of the theoretical density of solid zirconia.

6. A core catcher according to any one of claims 1 to 5 in which means sealing said joints against penetration and means securing said bricks - against movement comprises refractory cement chemically and thermally compatible with the bricks.

7. A core catcher according to claim 4 or claim 5 in which means sealing said joints against penetration and means securing said bricks against movement comprises zirconia-based cement.

8. A core catcher according to claim 1 or claim 2 in which the stratum of bricks comprises zirconia bricks, the core-catching layer comprising in addition a bed of zirconia cement between the structural support layer and the bricks and a further bed of zirconia cement between the bricks and the sacrificial layer, the joints between the bricks being filled by further zirconia cement.

9. A core catcher according to any one of claims 1 to 8 in which the lap joints between the bricks comprise tongue-and-groove features on confronting portions of adjacent bricks.

10. A core catcher according to any one of claims 1 to 9 in which the sandwich construction is substantially bowl-shaped.

11. A core catcher according to claim 10 in which the radius of curvature of the bowl shape is greater near its centre than near its edge.

12. A core catcher according to any one of claims 1 to 11 in which the surfaces of the core-catching layer, the sacrificial layer and the structural support layer are substantially parallel to each other.

13. A core catcher according to claim 10 or claim 11 in which the core-catching layer increases in thickness towards the centre of the core-catcher such that the inner surface of the core-catching layer and the sacrificial layer is less concave than the outer surface of the core-catching layer and the structural layer.

14. A core catcher according to claim 13 in which the outer surface of the core-catching layer and the structural layer are conical in form.

15. A core catcher according to any one of claims 1 to 14 in which the stratum of refractory bricks comprises a plurality of concentric courses of bricks arranged around a centrally located brick.

16. A method of constructing a core catcher for containing a molten nuclear core, comprising laying a first bed of refractory cement on a metallic structural support layer, laying a stratum of refractory bricks or the like on said first bed of refractory cement, the bricks having lap joints therebetween, filling said joints with refractory cement as said bricks are laid, laying a second bed of refractory cement on the stratum of refractory bricks, and seating a sacrificial layer on the second bed of refractory cement to form a multi layer sandwich structure.

17. A method of constructing a core catcher according to claim 16, except that instead of laying the second bed of refractory cement on the stratum of refractory bricks before the sacrificial layer is seated thereon, the sacrificial layer is first positioned to leave a gap between itself and the stratum of refractory bricks and liquid refractory cement is thereafter pumped into the gap to substantially fill the same and provide said second bed of refractory cement.

18. A core catcher for containing molten nuclear core materials substantially as described in this specification with reference to and as illustrated by Figures 1 to 3, 6 and 7 of the accompanying drawings.

19. A method of constructing a core catcher for containing molten nuclear core materials substantially as described in this specification with reference to and as illustrated by Figures 1 to 3 of the accompanying drawings.