(32d) Analysis of the Damage in Solidifying a Ceramic Waste Form

Summary Background A ceramic waste form (CWF) is being developed for long term storage of fission product and actinide salts obtained in the processing of spent nuclear fuel from the EBR-II reactor (Ref. 1). Because it is desirable to minimize the cracking of the product during the solidification and cooling process, a new theory which postulates that a permanent stress is set-in during solidification was developed which estimates the magnitude of the stresses produced when a CWF is cooled from a high temperature to the working temperature (Ref. 2). Recently, CWF1, first full-scale surrogate waste form, was produced at the INL. CWF1 did not contain any fission products, but all other properties were that of a standard CWF. It is approximately 1 m high, 0.5 m in diameter, weighs about 400 kg, and is formed in a stainless steel can 0.5 cm thick. The CWF furnace (containing an argon atmosphere) was ramped to 500 C at a rate of 1 C/min then held at that temperature for 40 hours. Then ramped to 925 C at the same rate and held that temperature for 75 hours. After the hold, the furnace control system turned off the heating elements. Forced cooldown was started about 15 hours later. The forced cooling system consists of an argon circulation pump and a heat exchanger. Hot argon from the furnace is cooled through the heat exchanger via the circulation pump, , and re-circulated back into the furnace. The flow is varied to maintain a circulation pump inlet temperature of 100 C. Post Processing Examination During processing the CWF material consolidated to half of its original volume due to melting and conversion to sodalite. It has the appearance and weight of a concrete cylinder but the stress strain characteristics of ceramic or glass. Three axial slices were cut out of CWF1: near the top, near the bottom, and in the mid plane. All three slices show that it was broken into many pieces. The midplane slice, shown in the figure, was fractured more than the upper and lower slice. The results of both circumferential and axial stress damage are evident in this cross-section. The circumferential strain that caused failure is indicated by the radial cracks. Since this was an axial slice, the machine cut almost parallel to breaks in the axial direction. Thus the pieces that have fallen out of the slice show that axial stress has pulled the CWF apart in that direction. Cracks tend to be located more toward the center, indicating that tension occurred there. A rapid furnace temperature decrease at the start of the forced cooldown was thought to be responsible for the damage. Process Modeling The only measurements available for CWF1 are three thermocouples (TCs) in the furnace insulation. Since no TCs were located on the CWF, these temperatures were inferred from the conditions measured in the furnace insulation. The furnace heat transfer coefficients in the cooling channel were estimated so that the calculated and measured temperatures in the insulation agreed. Since this model includes calculating CWF temperature, once this agreement was obtained, the temperatures were known in the CWF. This temperature distribution was used to estimate the stress in the CWF and determine when and where the damage occurred. Coolant flow was started 15 hours and rapidly cooled the furnace until the circulation pump inlet temperature reached 100 C (approximately 2.5 hours). The speed of the circulation pump continued at a rate slow enough not to exceed the circulation pump inlet temperature limit of 100 C and increased as the furnace cooled. The furnace thermocouple shows a sharp drop in the insulation temperature at 16 hours into the cooldown, due to the cooling gas flow started at 15 hours in the gas space between the CWF and the cylindrical furnace walls. In order to mimic the accelerated temperature decrease in the insulation at 16 hours, it was necessary to increase the cooling in the model at 15 hours in the cooling channel. The temperature of the gas was set to a constant 25 C and the heat transfer coefficient from the walls to the gas space was varied to produce the measured temperature profile in the insulation. Good agreement of the calculated and experimental temperature at the location of the two insulation thermocouples was obtained. Setting the gas space temperature to 25 C simulated a flow of gas that did not significantly increase in temperature as it flowed through the cooling channel. This method of estimating the heat transfer to the coolant was consistent with the input needed by the CWF STRESS code. To match the calculated and measured insulation temperature as a function of time, four different heat transfer coefficients were used in the gas channel at different times. The CWF surface temperature was calculated to decrease to the solidification temperature of 625 C after 44 hours of cooldown, and this was the time the solidification stress began to set in. The estimate from the stress code calculated the tension limit in the inner region was exceeded 82 hours after the start of cooldown. The maximum stress reached was 24,000 psi at room temperature. Data Comparison Comparing the calculation to the experimental data revealed that the rapid cooldown at 16 hours could not have caused the damage stress because CWF1 was still molten. Solidifcation started to occur at 44 hours which was when the solidification stress started to be set-in; therefore, the CWF1 was still molten and the cracking must have occurred well after this significant drop in the insulation temperature. At 40 hours after the start of cooling the data records the inlet temperature was low enough that flow no longer oscillated, causing the average flow to increase. This increased the heat transfer just as CWF1 was solidifying (625 C) and caused a significant set-in stress. The thermal stress during solidification is as large as the solidification stress so cancels it out in this temperature range. But continued cooling decreases the thermal stress leaving the net stress (solidification minus thermal) to increase more and more as room temperature is approached creating factures as the net stress exceeded the tensile stress limit at about 300 C. The STRESS code estimated the temperatures for the surface and center of CWF1 to be 100 C and 120 C respectively after 130 hours of cooldown when the CWF was removed from the furnace. Due to the significant solidification stress that had earlier set-in, the CWF continued to fracture as it approached room temperature where the total stress is calculated to be the maximum. This was verified by the cracking noises heard for the next 24 hours. Cooldown Recommendation To prevent fracturing during future formations of CWF, the cooldown cycle will need to be adjusted. Specifically, during the solidification phase the cooldown rate should be decreased a small amount to lessen the set-in solidification stress. This will keep the stress in the inner region below the tensile limit during cooling all the way to room temperature. The thermal stress component alone will exceed the tensile damage limit in the outer region but the solidification stress component of the thermal stress will counteract the thermal stress keeping the total stress under the damage limit. As the thermal stress decreases, the net stress will go into compression and stay there throughout the rest of cooling cycle, resulting a stronger product. The temperature drop with this scheme will always be much flatter than that which occurred in the CWF1 formation. This cooling method will result in lower stress levels. The CWF will cool down slower (175 hours) than the CWF1 run (130 hours). Conclusions Recently, the first full-scale surrogate waste form, CWF1, was produced at the INL. The temperature in the insulation indicated that the flow rate decreased exponentially between 14 and 17 hours. Surprisingly, the model indicates CWF1 did not sustain any damage during this temperature drop because it remained molten during this time. Instead, the fractures in the CWF occurred because at about 40 hours of cooling the inlet temperature was low enough that the flow no longer oscillated, causing the average flow to increase. This increased the heat transfer just as CWF1 was solidifying and caused a significant set-in stress. Fracturing continued until the CWF reached room temperature. To prevent fracturing during future formations of CWF the cooldown cycle will need to be adjusted. During the solidification phase the cooldown rate will be decreased a small amount to prevent the build-up of stresses. Reference 1. K. M. Goff, M. F. Simpson, S.G. Johnson, K. J. Bateman, T. J. Battisti, and S.M. Frank, ?Ceramic Waste Form Production and Testing at ANL-West,? American Nuclear Society Third Topical Meeting DOE Spent Nuclear Fuel and Fissile Materials Management, Charleston, SC, September 8-11, 1998. 2. C. W. Solbrig and K. J. Bateman, Modeling Solidification-Induced Stresses in Ceramic Waste Forms Containing Nuclear Wastes, to be published Acknowledgement Work supported by the U.S. Department of Energy, Office of Nuclear Energy (NE) under DOE Idaho Operations Office Contract DE-AC07-05ID14517.