PRESSURE VESSEL AND PIPING SYSTEMS

Experiments were performed on 1/25th-scale models of a basic design of a nuclear power reactor to determine whether radioactive core material and chemically reactive coolant would remain contained and whether the heat withdrawal system would continue to function in the remote event of a core disassembly accident. Several reactor configurations were included in the tests so that it could be established which design among the alternatives provided the greatest safety. The sudden energy release of a nuclear excursion was represented by detonating an explosive charge at the core of each test assembly. Three series of tests were performed: the first with the core shields alone, the second with a simple vessel assembly in a tank of water and having cover plates to represent the shield deck, and the third with reactor models of more detailed construction. Water was used to simulate the sodium coolant. The effects were recorded by monitoring thrusts on the plug and deck plate, pressures at vital nozzles, strains of the vessel, vessel shroud, heat exchanger pipe, and tank wall, and by measuring final deformation of the vessel, vessel shroud, tank and other small components.

This experimental program was carried out to investigate the structural response of the primary containment (vessel, head, and primary loop) of the Fast Flux Test Facility (FFTF) when subjected to conditions of a hypothetical core disruptive accident (HCDA). A calibrated source was developed to simulate the HCDA and instrumented experiments were carried out with the source in several simple models and three complex models consisting of two 1/30-scale reactor structures and one 1/10-scale reactor structure with three primary loops. The simple model experiments provided basic response results that facilitated the design of the complex model experiments and that were then used in the development of computer codes suitable for HCDA descriptions.

The propagation of pressure pulses in water-filled thin-walled Ni 200 pipes was studied with small-scale experiments. Pressure and strain along the pipes were measured to help verify the modeling techniques used in the REXCO hydrodynamic structural code and to study the response of pipes used in the primary cooling loop of fast breeder reactors to pressure pulses generated during a hypothetical core disruptive accident (HCDA). Pressure pulses, produced by an explosively driven pulse gun, simulated pulses generated at the outlet nozzles of a liquid metal fast breeder reactor vessel during an HCDA. The pulses were of sufficient amplitude (2200 psi) to produce plastic deformation in thin-walled piping systems. The results of straight-pipe experiments showed that as the pressure pulse enters a thin-walled pipe, it causes plastic deformation and attenuates, the peak pressure dropping from the incident value of 2200 psi to the pipe yield pressure (350 to 550 psi) in about 6 diameters (18 inches). Elbow-pipe experiments showed that as the pressure pulse travels through a thick-walled elbow, there is no pressure gradient across the diameter of the elbow, and the pressure pulse is attenuated 15%.

A simulated HCDA loading was applied in simple, thick-walled and thin-walled, 1/30-scale models of a fast breeder reactor vessel. The loading and resulting vessel response were measured to verify the modeling techniques used in the REXCO computer code. The sodium coolant was simulated by water and the HCDA load was simulated by the expansion of the detonation products of a low-density explosive detonated in the vessel core. Piezoelectric quartz transducers measured the loading pressure in the core and on the core support platform, the vessel wall, and the cover. Foil strain gage measured circumferential and axial strains on the vessel wall. From the strain measurements and the posttest deformation profiles of the vessel, we calculated the strain energy absorbed by the vessel before and after slug impact by the coolant.

The energy loss of a sodium vapor bubble through heat transfer to the upper core structures in a sodium-cooled fast breeder reactor during a hypothetical core disruptive accident (HCDA) was simulated by using high-temperature, high-pressure steam (4046 ^oK, 200 bars) in a 1/30-scale model of the core of the Clinch River Breeder Reactor. The steam expands into a chamber either through an open passage or through an array of tubes inserted in the passage. The open passage simulated an upper core barrel with all the upper core structure removed during the HCDA, whereas the passage with the array of tubes simulated the empty subassembly ducts remaining in the upper core barrel during the HCDA. The peak pressure in the chamber was measured for both cases and the difference in peak pressures was used to compute the implied energy loss caused by heat transfer to the simulated upper-core structure. The decrease in work potential of the steam bubble was calculated from this energy loss.

Five experiments were performed to help evaluate the structural integrity of the reactor vessel and head design and to verify code predictions. In the first experiment (SM 1), a detailed model of the head was loaded statically to determine its stiffness. In the remaining four experiments (SM 2 to SM 5), models of the vessel and head were loaded dynamically under a simulated hypothetical core disruptive accident (HCDA). Models SM 2 to SM 4, each of increasing complexity, systematically showed the effects of upper internal structures, a thermal liner, core support platform, and torospherical bottom on vessel response. Model SM 5, identical to SM 4 but more heavily instrumented, demonstrated experimental reproducibility and provided more comprehensive data. The models consisted of a Ni 200 vessel and core barrel, a head with shielding and simulated component masses, an upper internals structure (UIS), and, in the more complex models SM 4 and SM 5, a Ni 200 thermal liner and core support structure. Water simulated the liquid sodium coolant and a low-density explosive simulated the HCDA loads. In the static loading experiment, head deflection and strain were measured as a function of applied pressure. In the dynamic loading experiments, pressures were measured in the core, along the vessel wall, and on the cover. Strains were measured on the vessel wall, on selected UIS columns, and on studs that hold the models to the support stand. Accelerations were measured on the head and the core support platform.

The effects of upper core and upper internal structures on the work potential of hypothetical core disruptive accidents (HCDA) were studied in a transparent 1/30-scale model of a typical demonstration size loop-type Liquid Metal Fast Breeder Reactor (LMFBR). Water at room temperature simulated the sodium coolant. High-pressure nitrogen gas (1450 psia) or flashing water (1160 psia) simulated the qualitative features of the expansion of sodium vapor and molten fuel. The upper core structure (UCS) simulated an array of empty hexcans, and the upper internal structure (UIS) simulated the control rod and flow guides. The expanding bubbles and the motion of the coolant simulant were studied using pressure transducers, water surface displacement gages, and high-speed photography. The results showed that internal vessel structures, through their hydrodynamic influence alone, markedly attenuate the potential of an HCDA transient for damaging the reactor vessel.

Six experiments were performed to evaluate the structural response of an above core structure (ACS) to increasingly energetic hypothetical core disruptive accidents (HCDA). The tests were performed in a thick-wall 1/20-scale model of a liquid metal fast breeder reactor (LMFBR) demonstration plant to provide information on ACS displacement as a function of HCDA energy so that computer codes might include ACS displacement when calculating core release energetics. The 1/20-scale ACS models include a 5-kg aluminum block with dimensions that simulate the overall geometry of the ACS and the four columns that support the ACS over the core of the reactor. Each column included an 8-in.-long, 0.7-in.-diameter 0.050-in.-wall Ni 200 tubular section that received all of the deformation upon HCDA loading. The columns were supported by a rigid structure that simulated the geometrical constraints provided by the cover of the reactor. Careful attention was paid to the dimensions and masses of the ACS to assure prototypic response. Loads were provided by the detonation products of a low density explosive in a rigid core structure. Instrumentation included pressure transducers to measure the loads on the ACS and in the core, strain gages to measure column forces, and displacement gages to measure column deformation. The structural response of the ACS was characterized by axial compression and two modes of buckling of the tubular portions of the ACS columns. For the base loading levels, axial compression exceeded the elastic limit at 200 ms and slug impact occurred at 2.3 ms. At higher energy levels, these times decreased to about 150 ms and 1.7 ms, respectively. The two buckling modes (overall column buckling and local buckling of the tubes as cylindrical shells) began at almost the same time early in ACS response and continued until the column collapsed.

Tests were performed in which a simplified 1/8-scale model of the intermediate heat transfer system and relief system of a liquid model fast breeder reactor was subjected to a simulated sodium-water reaction in a steam generator. Pressures in the intermediate heat exchanger (IHX) and in the pipe were measured. The flow of water through the relief system was photographed and its velocity was measured. The forces on the relief system elbows resulting from the fluid flow were also measured. The tests were performed primarily to validate pulse propagation codes used for design and for direct use as design data.