b'P ROVING THE P RINCIPLE1970. As it turned out, operators did not On March 28, 1979, at the Three Mile of the water in the pressure vesselremove the reactor after the tests Island 2 (TMI) nuclear power plant near flashed to steam. One thing after anoth-because of the complexity of auxiliary Ha r risbu rg, Pennsylvania, the main er went wrong with instruments, equip- piping and other systems around the pumps circulating the secondary coolant ment, computers, and human judgment.reactor. But they did open the door stopped running. This prevented heat During the next sixteen hours, a third ofbetween tests to facilitate preparations removal from the primary cooling sys- the core melted, although no one knewfor the next test. 3 tem. The turbine shut down, and the what this fraction was until much later.reactor likewise. Decay heat continued to The hot core material did not meltLOFT engineersthe ones at the heat the water near the core. This caused through the reactor vessel, let aloneRogers Hotelhad created computer a pressure surge and forced open a pres- down to China. About twenty curies ofmodels predicting how different types sure relief valve. Emergency pumps radioiodine were released to the of emergency core cool- environment. 6ing systems would sup-posedly perform if a In the immediate effortreactor lost its coolant. to understand the condi-Some systems pumped tion of the reactor,new cooling water into inspectors from thethe core; others injected NRC arrived fromcooling water from pres- Washington, D.C., andsurized tanks. The pur- concluded that the zir-pose of the LOFT conium-clad metal wasexperiments was to pro- interacting chemicallyvide empirical valida- with the hot steam totionor notfor the create hydrogen gas.theory behind the com- They feared that a bub-puter models. 4 ble of the gas couldinterfere with the flowThe first nuclear test on of cooling waterDecember 10, 1978, imi- INEEL 79-9362 through the core.tated a double-ended guillotine break, Three Mile Island Nuclear Station in 1979. Analysts speculated that a large gaswhere the coolant flooded from both bubble could explode and blow openends of a broken pipe. This was pre- the containment shell. 7sumed to be among the worst kinds of began to restore circulation. As pressureaccidents. The computer model had pre- subsided, the pressure relief valve should Urgent questions about the hydrogendicted that the fuel temperature could have closed, but it stayed open. came to INEL scientists. Within twen-n c yrise to 1,350F and that the emerge ty-four hours, they had modified thesystem would restore cooling water The operators didnt know, couldnt piping at the Semiscale facility to rep-within 90 seconds. The computer proved see, and hadnt been trained to imagine resent the situation at TMI and deliv-to be conservative. The coolant was that the valve was open. Because of ered reassuring information about therestored within 44 seconds, and the other events and faulty indicators, they hydrogen bubble. Fear of an explosionmaximum temperature of the fuel rose to believed too much water was entering lifted. President Carter visited the TMIabout 1,000F. More tests were planned the vessel and shut down the emer- plant for a briefing on the condition ofto imitate other accidents, but the assess- gency pumps. They started, then the facility, a gesture that soothed thement of what kinds of breaks to test was stopped the primary coolant circulating country and calmed Harrisburg citizens.about to change. 5 pumps. Water pressure fell, and some A support team flew from INEL to224'