We commissioned the Seawolf (SSN- 575) on 30 March 1957 after about three years of preparation at the Knolls Atomic Power Laboratory, Schenectady, New York, and the prototype plant in West Milton, New York, where the General Electric Company operated the laboratory for the Atomic Energy Commission.
Rear Admiral Hyman G. Rickover ran the program with a very capable group of naval officers and civilians. I dealt closely with engineering duty officers like Commanders Archie P. Kelley, Jack Hinchey, and Sherman Naymark; they were all from the top of their Naval Academy classes and had done postgraduate work at the Massachusetts Institute of Technology.
There was much activity in the nuclear-power field during the 1950s. Westinghouse Electric Corporation had developed the pressurized-water reactor (PWR) for the USS Nautilus (SSN-571) at the Bettis Laboratory near Pittsburgh, Pennsylvania, and at its prototype plant near Arco, Idaho. Research also had been intense in nuclear rocket propulsion, an aircraft propulsion plant involving a direct air cycle, another aircraft propulsion plant based on the liquid metal lithium contained in columbium to drive a gas- turbine engine, a very compact spacecraft power plant that used liquid metal as a heat-transfer agent, and a liquid- metal plant experiment aimed at breeding fuel.
Soon to be developed were a nuclear- powered merchant ship and a U.S. Army power plant for use in Antarctica. To illustrate the intensity of activity, I was visiting a prominent metallurgist at the Oak Ridge National Laboratory, Oak Ridge, Tennessee, when there was a loud “pop,” and he broke into purple prose as “the only two pounds of ruthenium in the world” floated out the window in a cloud of dense white smoke.
The contractors and the 3,000 people at the Knolls Atomic Power Laboratory had done a wonderful job developing the submarine intermediate reactor (SIR) power plant prototype, built into a steel sphere containment vessel 225-feet in diameter. The power plant destined for the Seawolf was initially designated the S1G (vessel-model-designer: submarine- first model-General Electric); later, when improved boilers were added, it went into the Seawolf as the S2G.
During a year-long, demanding course at Union College in Schenectady, the crew had been taught by the engineers who designed the plant; subsequently, each crew member was checked out on the plant in detail by then Lieutenant Commander Charles S. Carlisle, chief engineer.
As with all major hardware programs, the design of the new power plant began with the development of materials. Rick- over’s program was responsible for developing boron, beryllium, hafnium, 347 stainless steel, sodium, enriched uranium fuel, and zircaloy. Each of the new materials had to be tested for radiation resistance, strength at high temperature, resistance to thermal stress, weldability, corrosion resistance, and coolant containment—just to mention a few requirements.
The general approach in the advanced experiments was to try to raise the operating temperature—the key to efficiency in thermal engines. Fossil-burning plants had operated at higher temperatures using mercury and bismuth as coolants, but these had proved unsatisfactory for nuclear plants. One nuclear effort involved the very high temperatures achievable with fused salts of the fissionable material; an even higher temperature effort involved the use of vaporized fissionable material for a rocket engine.
Sodium—with a melting point of 97° Celsius and a boiling point of 887° Celsius, excellent heat-transfer conductivity, and high electrical conductivity—represented an intermediate high-temperature possibility. It had other useful characteristics, including the ability to wet stainless steel and then to be frozen to provide its own seal. Its high thermal coefficient of expansion gave it the added advantage of providing excellent reactor convection cooling—significantly better than that of pressurized water. But using it meant that operators had to be very careful in dealing with melting in the pipes to avoid pipe ruptures. In addition, sodium’s thermal neutron capture cross-section is low and does not slow the reactor.
Sodium’s major disadvantage as a coolant is that it becomes highly radioactive in the reactor’s neutron flux. With a half-life of 15 hours, it emits a strong beta ray of electrons with more than one million electron volts energy. In contrast, a clean pressurized-water plant shows little activity after a few minutes.
The use of electromagnetic pumps in the sodium plant meant that there were no moving parts, and all seals were welded or frozen. With all the stories about sodium’s chemical activity, water at loop temperatures and pressures is much more reactive than contained sodium. Whereas the PWR loops are frequently treated chemically, in a sodium plant the sodium is “cold trapped,” i.e., cooled to remove any pollutants. Sodium’s tendency to react with air required an inert atmosphere in the lower reactor compartment in case of an air leak.
Sodium has another important property—the neutron flux in the reactor is at an average energy higher than that of a water plant. This increases the possibility of converting Uranium-238 (U-238) to Plutonium-239 (Pu-239) fuel or Thorium-232 (Th-232) to U-233 fuel in quantities greater than the fuel used. This breeding, not important in an individual submarine, might yet turn out to be vital to mankind as the Greenhouse effect or a coming ice age demands more use of nuclear power until a fusion process is developed. Our great-grandchildren may look back on us as fools if we waste the U-238 that could have been converted. One of the powerful things Rickover did was to develop the seed and blanket cores that come close to breeding in the PWR plants and greatly extend core life.
Nevertheless, naval engineers have long been comfortable with high-pressure water and steam. Further, they have learned to keep variables separable, and to design systems that can be fixed one thing at a time. A typical naval propulsion system is made up of separate components: boilers, separators, superheaters, turbines, condensers, feed pumps, condensate pumps, reduction gears, clutches, lubricating systems, main circulation pumps, throttles, etc. These are spread around compartments to optimize operation and repair. Only in the more sophisticated engineering of gas turbines and rocket engines can these systems be unitized. Recent studies at Argonne National Laboratory and General Electric show the promise of a unitized, sodium- cooled plant with reduced welding, shielding, and other features. (See “Next Generation Nuclear Reactors,” Popular Science, April 1990, page 69.)
The PWR plant is not particularly handicapped if its components are separated, but the sodium plant is at a severe disadvantage in that shielding weight is greatly increased for the separate units if the plant is not unitized. The engineers considered various schemes to minimize shielding weight for hypothetical future plants. Perhaps it would have been possible to combine in one unit the reactor, evaporator, superheaters, steam separators, and use of convection cooling with some ingenious scheme for refueling. As it turned out, though, even with superheaters removed, the Seawolf s liquid- sodium plant was heavier than the PWR in the USS Nautilus (SSN-571), and we were also using saturated steam at slightly reduced power.
The PWR plant operated at a primary pressure of 3,000 pounds per square inch as opposed to 100-200 pounds per square inch for the sodium plant, and the PWR plant was inherently heavier, except for shielding. In the Seawolf, with non-unitized components, the shielding weight exceeded that of PWR plants. Even so, we found that after a day of high-power operations, anyone standing on the pier in the dark could perceive a blue glow of Cherenkov radiation under the lower reactor compartment.
With both Westinghouse and General Electric interested in commercial power plants, the advanced development went into the PWR at Westinghouse and the boiling water reactor (BWR) plant at General Electric. This happened before the Seawolf was commissioned, and Rick- over had already engaged the Duquesne Light Company in Pennsylvania to start commercial development as an avenue toward an aircraft-carrier reactor. Thus the die already had been cast: ships would use PWR plants.
The Air Force and Navy efforts at powering planes had long been handicapped by the circular argument of feasibility versus requirements. Each had to compete with the very rapidly changing technology of turbojet engines. A couple of billion dollars were spent on studies— none of which justified building anything. Just as the budgets were being slashed, the lithium-columbium loop had operated at 2,000° for 1,000 hours, which was a remarkable achievement. Thinking that it would be smarter to get something flying, arrangements were made to transfer the 60 megawatt Seawolf prototype reactor to a Navy program where the plan was to have it provide cruise power for the British Princess flying boat. Feasibility tests seemed favorable but the budget money went to turbojet engines.
There was a fundamental engineering problem with sodium coolant: The 347 stainless steel, which had to be used to contain the sodium, has a high coefficient of expansion but a low coefficient of thermal conductivity. Sodium is a marvelous heat conductor that can carry temperature waves around the loop. If such a wave causes a temperature gradient too intense, thermal cracking can result, especially where the metal varies abruptly in thickness, as at a tube header interface in a superheater. The engineers had anticipated this problem, and several designs had been built for both the evaporator and the superheater, but their designs were complicated by the need to keep water or steam out of the sodium. The water and steam were at higher pressures than the sodium, and a crack would allow water into the sodium where hydroxides would form a sludge tending to slow coolant flow. Worse, if hydrogen got into the reactor, it could over-moderate the neutrons and cause a reactivity surge.
The solution was to develop a design with a double barrier between the water and the sodium in the heat exchangers— tubes within tubes—and a complex double header. There had to be provision for a third fluid between the tubes for heat transfer. The third fluid, at an intermediate pressure, could be monitored for pressure; increased pressure would mean a steam-to-interspace leak, while a pressure drop would indicate that the third fluid had leaked into the sodium. In an early test run at the prototype, pressure went up in the superheater third-fluid system that contained the sodium-potassium (NaK) metal alloy, which was liquid at room temperatures. The solids formed tended to block the leak, but it was found that the high alkalinity produced caustic corrosion cracking, and the superheaters had to be replaced.
One of the most bizarre efforts of the program ensued as mercury replaced NaK in the third-fluid system. We had taken over about half the annual production in the United States, and it arrived in hundreds of 70-pound ceramic bottles. Handling it in an enclosed space without killing everyone by mercury poisoning was difficult.
The next superheater leak that occurred was from mercury to steam and it showed as mercury vapor out of the air separators in the engine room. It was a pinhole leak; we vented the vapor and operated until one of the mercury pressures indicated that mercury was leaking into the sodium, although with little effect on anything like reactivity.
Rickover arrived at the plant about 0100 one morning for an inspection and asked Archie Kelley and me for the good news first. We told him that the miracle long-sought by the alchemists was here— we were making gold out of mercury at about $.73 per day using about 15% power. The kindly old gentleman (KOG, as we called him) had a sense of humor, and it didn’t fail him.
Most of the important prototype tests had been completed: pile oscillator tests showed the remarkably accurate predictions of the physicists, the power-level flow tests controlled by the magnetic amplifiers showed how well a World War II-era, exelectronics technician who had gone on to earn a doctorate on the GI Bill, had designed the system. This was of great importance because the way to avoid temperature waves in the sodium was to match the sodium flow with the reactor power level—the ratio remained remarkably constant. All kinds of scrams (reactor shutdowns) and recoveries had been conducted, and a 100-hour full-power run had been completed successfully.
Finally, it was time for more of us to spend time at the Electric Boat Division to get the ship ready to run. I had been commissioning executive officer of the USS Stickleback (SS-415) at Mare Island, California; commissioning commanding officer of the USS Trutta (SS-421) when she came out of the reserve fleet; and commissioning CO once again of the USS Harder (SS-568) at Electric Boat. It was not a new drill for me, and we had a wonderful crew—85% of whom were later commissioned.
I dealt with many fine people. I remember well Captain Ralph Kissinger, an EDO who ran the program for the Supervisor of Shipbuilding; Mr. Carlton Shugg, president of Electric Boat; Commander Williston Shor, Rickover’s representative there; and Mr. J. William Jones, program manager and later president of Electric Boat—he had also been construction manager for both prototypes. We were joined by our executive officer. Commander (later Vice Admiral) Robert Y. “Yogi” Kaufman. What followed was a remarkable team effort.
I don’t think the power-plant crew ever made an operational mistake. After we removed the superheaters and conducted very successful trials, the USS Seawolf was commissioned on 30 March 1957.
We operated almost two years before the first core approached the end of its life. I knew that another year’s supply of fuel remained in West Milton, that the boat could be refueled in a couple of weeks in Groton, Connecticut, and that we could be off and running to provide some of the many services required. But, realizing that the state of the art favored the PWR plant, I had, at Rickover’s request, written a letter recommending that at the expenditure of fuel the Seawolf be converted to a PWR plant. The Navy could not afford the cost of keeping two complete systems of power generation going indefinitely with the enormous pool of talent and laboratory assets required. I’ve never more regretted having to do anything.
I’d long watched yards put up to 600 workers down through 26-inch holes into a submarine while building—like “repairing a watch through the stem-hole.” So in the final months I sent a message to the Chief of Naval Operations, copies elsewhere, recommending that the Seawolf burn up the remaining 12-month supply while Electric Boat separately construct a reactor compartment through the open ends to be substituted for the old one in drydock with only weeks out of action. The reaction was strong. Evidently the inflexibilities of budgeting prevailed. Electric Boat, perhaps under instruction, expressed concern about “out-of-round problems.” Perhaps, more to the point, they had just heard of the massive challenge of cutting a Skipjack (SSN-585)-class boat in two to add the Polaris hull segment. They needed a manpower sump. Such is the path of progress.
Some 14 months later, the Seawolf was recommissioned as a PWR boat. [This is the first of two articles on the Seawolf the second, covering initial fleet operations with her liquid-sodium reactor, will appear in the Summer 1992 issue of Naval History.]