[en] Fuel-sodium interaction (FSI) produces a pressure load on the surrounding core structures. It is necessary to prove that the corresponding core deformation neither initiates a nuclear excursion nor renders the shut down system inoperable. This requires the knowledge of the initiating FSI pressure time history. The theoretical pressure time history here presented differs completely from all calculations known so far. In the new code MURTI (Multi Region Thermal Interaction) the mixing zone can be subdivided into subregions in each of which the interaction starts at some given time. In order to simulate the progressive mixing of fuel and sodium after sodium reentry into the subassembly, the reaction is initiated successively in several subregions with a small time delay. Each fuel mass is assumed to be initially molten and to be instantaneously fragmented into droplets with 117 μm radius and mixed with an equal volume of cold (approximately 900 K) liquid sodium. The heat transfer is calculated exactly assuming ideal thermal contact and the sodium obtains heat only from the fuel with which it is mixed. Therefore temperature balance is reached already after 2 msec and the mass ratio is that defined by the input. With these assumptions it is found that: there is no liquid pressure peak like with instantaneous mixing; the pressure reaches a saturation level after 2...3 msec; the pressure drops quickly after the acoustic period; the maximum pressure is in the range 350...400 bar. Detailed parametric studies on these results are presented

[en] A simple well-defined, shock-free pressure source was developed to simulate in hexcans circumferentially and axially uniform pressure pulses having peak pressures ranging from tens of bars to a kilobar, rise times from tens of microseconds to milliseconds, and pulse durations from milliseconds to seconds. The controlled expansion of a high-pressure gas in and out of a hexcan is obtained using a mixture of PETN (pentaerythritol tetranitrate) explosive powder and glass microballoons in a vented steel canister. The pressure source is calibrated in an enclosed rigid cylinder with axial vent ports and an internal volume equal to that of a hexcan. Pressure gages mounted flush with the cylinder wall along the length of the cylinder measure the loading pressure pulse. Upon detomation, the hot, high-pressure detonation products fill and expand out of canister (Chamber 1), fill and pressurize the space between the canister and the rigid cylinder (Chamber 2) and expand through the axial vent ports into the atmosphere. Thus the loading pressure rises initially from one atmosphere to a peak pressure and then decays back to one atmosphere, giving a triangular-shaped pulse in Chamber 2. For given chamber volumes, the venting area of Chamber 1 determines the pulse rise time and the venting area of Chamber 2 determines the pulse decay time. For fixed chamber volumes and venting areas the peak pressure varies directly with the charge mass. The theoretical analysis of the pressure source assumes quasi-steady flow of the detonation products. The analysis provides the venting port sizes and charge masses required to produce pulses of given shape and magnitude. Results show good agreement with theory

[en] Dynamic response of LMFBR primary containment to the hypothetical explosive accident is discussed. The amount of experimental and analytical effort given to the problem in Czechoslovakia is presented. The experimental program consists of a sequence of tests using models of increasing complexity in fast reactor primary structure. As the energy source, high explosives are used. The response of reactor materials subjected to dynamic load is investigated on an experimental facility based on Hopkinson-Davies bar which makes possible to study the influence of strainrate, temperature and irradiation on material properties. Theoretical effort of the Power Research Institute is based on the mechanistic approach. For the initial phase of the hypothetical accident, the programme SASIA was adopted. The course of nuclear excursion is solved by the TWEXCO code. Dynamic response of fast reactor primary containment to the core disruptive accident is treated by the CEFRA code. Both programmes are two-dimensional (r-z) Lagrangian hydrodynamic codes. They treat the media under consideration as compressible, inviscid and non-heat-conducting fluids. Hydrodynamic equations and equations of state of reactor materials are used for the calculation of the shock waves propagation, the damage produced by these loads and the motion of coolant. They used plate-shell equations to determine the dynamic response of the reactor containment. Containment materials may be elastic-plastic, strain hardening and strain-rate sensitive. The plug and cover are treated as rigid-body model. The validity of the codes performance has been verified through experiments. Some results of this comparison and application to a 600MWe fast reactor project are presented

[en] The analysis to be presented will restrict attention to the elastic part of the elastic-plastic constitutive equation used in several Fast Reactor Accident Analysis Codes and originally applied by M.L. Wilkins: Calculation of Elastic-Plastic Flow, UCRL-7322, Rev. 1, Jan. 1969. It is shown that the used elasticity concept is within the frame of hypo-elasticity. On the basis of a test found by Bernstein it is proven that the state of stress is generally depending on the path of deformation. Therefore this concept of elasticity is not compatible with finite elasticity. For several simple deformation processes this special hypo-elastic constitutive equation is integrated to give a stress-strain relation. The path-dependence of this relation is demonstrated. Further the phenomenon of hypo-elastic yield under shear deformation is pointed out. The relevance to modelling material behaviour in primary containment analysis is discussed

[en] During the last four years experiments were carried out at JRC-Ispra in the frame of a collaboration contract between BELGONUCLEAIRE and EURATOM to stimulate in small scale vessels the occurrence of an Hypothetical Core Disruptive Accident (hcda) of the SNR-300 reactor. The 2D computer code SURBOUM-II has been developed in parallel with the experimental programme. That code computes the two-dimensional fluid flow within the system in case of a core explosion. The fluid is assumed uncompressible and the deformations of the concentric shell(s) and vessel(s) are calculated by means of the thin shell theory. Perforated dip plates which are included in the model are treated as a particular type of boundary conditions involving empirical pressure drop versus flow relationships. The first part of the interpretation work was the determination of the pressure-volume relationship of the slow burning charge used to stimulate the hcda. This has been achieved by an original trial and error method built in the code which fits bets the experimental impulse time records obtained in bare charge experiments fired in overstrong vessels. Other experiments carried out in overstrong vessels and involving perforated dip plate above the core to damp out the fluid impact on the roof were calculated and the comparison of the theoretical and experimental impulse time curves was satisfactory. Further experiments, including or not the perforated dip plate, carried out in yielding vessels were also interpreted. For those experiments the scope of the work was the comparison of calculated and measured deformation of the vessel. The agreement obtained is satisfactory, though the code seems to overestimate slightly the final deformation

[en] The present study concerns the effect of fuel meltdown in one subassembly on adjacent subassemblies containing undamaged fuel and flowing sodium. The molten fuel could result in a heat flux of up to 600 W/cm² on a portion of one face of each adjacent cut. This flux would result in a temperature drop of about 725 deg C through the duct wall. The effects of the accompanying thermal stresses are to be determined. Sodium boiling is not expected in the adjacent subassembly unless the duct and adjoining fuel elements are severely distorted. The current subassembly design has an array of 217-fuel elements (0.58 cm dia) in a hexagonal duct. The ducts are 0.30-cm thick and 6.45-cm across each face. In the experiments, heat fluxes up to 600 W/cm² will be imposed on up to a 30-cm portion of one wall of a full-scale duct. The sodium-cooled duct is mounted in an existing sodium loop facility. Heat is generated in the duct wall by high frequency (450 kHz) resistive heating (proximity heating). The resultant current flows in the outer skin of the duct (0.04-cm penetration), closely approximating the postulated surface heat flux. The proximity heating procedure is not well established and requires preliminary tests to achieve a uniform heat flux. Extensive thermal measurements will be made. Displacement transducers are positioned to observe bowing and rippling of the heated surface in the initial tests. Later tests may incorporate holographic mapping of the heated surface. An attempt has been made to analyze the behavior of a hexagonal duct using a cylindrical approximation. The analysis is a modification of earlier work. Experimental results are to be used in calibrating the STRAW structural dynamics code which is being expanded by Kennedy and Kulak (ANL) to include three-dimensional thermal loading effects

[en] In order to verify the finite-element codes, a series of experiments has been performed under quasi-static and dynamic conditions. Preliminary tests demonstrated that for the range of deformations expected in a single subassembly prior to failure, a shortened duct section of only 30.48cm in length was sufficient to provide a central test section over which axially uniform conditions prevailed. The deformation over the uniform range corresponds to two-dimensional plane-strain conditions, and a two-dimensional computer code can be applied for the analysis. Additional tests were performed using a series of Type 316 stainless steel hexcan specimens in which the ductility was varied from a fully annealed specimen to a brittle specimen with a hardness corresponding to 50% cold working. The corresponding strains and displacements were measured at critical points such as the midflat region of maximum deflection and the duct corners where failure was expected. Dynamic tests of hexcans were also performed using a pressure-time source designed to duplicate a postulated local event. The local event simulated was the failure, after some period of operation, of misloaded, overenriched fuel pins. The fuel pin failure was postulated to result in the release of molten fuel into several subassembly subchannels. The resulting pressure pulse had the general characteristics of a 1msec risetime to a peak of about 100bar, with a total pulse width of about 5msec. Dynamic pulse loading experiments were then performed, in which the pressure pulse plus midflat and corner strains and displacements were measured. The material properties were also carefully determined for each specimen because of the high sensitivity of the hexcan response to the elastic modulus, the yield point and the plastic flow stress-strain relationship

No abstract available

No abstract available

[en] For safety considerations of sodium cooled fast breeder reactors the mechanistic accident-initiating conditions must be studied. In previous investigations of such initiating accidents the models assumed axisymmetric configurations and in general neglected the coupling effects with the subassembly boundary. This paper presents a more precise treatment of the subassembly boundary and also provides feedback of the boundary response to the pressure source. This is accomplished by marking use of two computer codes: REXCO-HT and SADCAT. The internal hydrodynamics of the fuel subassembly is simulated by the REXCO-HT code which possesses certain models of fuel-coolant interactions (MFCI) to be used as a pressure source. The hexagonal boundary of the fuel subassembly is modeled by the SADCAT code. Since both codes involve explicit time integration, coupling between the two is effected at each time step. The pressure at the outside boundary of the REXCO-HT model provides the loading on the SADCAT model. Given the load, the SADCAT model yields the three-dimensional deformation of the hexagonal boundary. With the deformation known, the outside REXCO-HT model boundary is adjusted and the computation cycle of the coupling is completed. In effect, the coupling of the two codes substitutes a cylindrical vessel of the REXCO-HT code by a hexagonal duct. It is shown by the use of this procedure that the assumption of a cylindrical vessel of the same thickness as that of the hexcan is quite erroneous. The maximum deformation of the flat of the hexcan in the illustrative examples is larger by as much as one order of magnitude. The maximum strains at the inside CORNER of the hexcan are also underestimated by a similar amount