Nuclear power plants produce power by creating large amounts of heat. In this respect, they're very similar to coal- and oil-fired power plants. They key difference is in how the heat is generated. Rather than burning fossil fuels, nuclear plants use the heat of a self-sustaining nuclear fission reaction.
A form, or isotope, of the element uranium (U) is the most common fuel for commercial nuclear power. U-235 is very rare. In natural uranium, 99.3% of the atoms are U-238, which is stable and not useful as fuel. U-235, as it is called, makes up only 0.711% of what we find in nature, and it's necessary to use very complicated means to enrich natural uranium to at least 3-5% U-235 before it can be used as reactor fuel. This is known as LEU (Low-Enriched Uranium), as opposed to HEU (Highly Enriched Uranium) which is the weapons-grade stuff ... 20-90% U-235.)
U-235 is useful because it has an atomic structure that isn't stable; atoms of U-235 are constantly decaying, or falling apart, releasing tiny, subatomic particles called neutrons. Neutrons are like bullets, moving very, very fast. When one of them strikes another atom of U-235, the struck atom splits and releases neutrons of its own. Meanwhile, the pieces of the former U-235 atom (usually isotopes of Iodine (I) and Strontium (Sr)) are sent spinning away at high velocities, and as they slow down by colliding with other atoms, their kinetic energy is converted to heat in the fuel. Sometimes the fission products are unstable and decay also, releasing even more neutrons. This process of splitting atoms is called nuclear fission.
The intensity of this reaction is proportional to fuel density. The closer together we pack our U-235 atoms, the greater the chance that a neutron from one will actually split another, rather than flying uselessly off into space. If we achieve a state where exactly one neutron from each and every fission causes another fission, we call the reaction self-sustaining. Except when starting up and shutting down, this is the region where we want the reactor to operate, because it will produce constant thermal power (heat).
In a nuclear reactor, then, our goal is to bring a great deal of fuel into close proximity, so that enough atoms will be split by flying neutrons to create a self-sustaining chain reaction. The reactor core at TMI contained 36,000 long, slim rods known as fuel pins. Each pin consisted of a thin zirconium alloy tube filled with uranium oxide fuel pellets, as shown in the drawing. The pins were bound together into a number of bundles or assemblies, and were held vertically in racks also made from a zirconium alloy. These racks hold the fuel pins spaced closely enough to sustain the reaction, but far enough apart that cooling water can pass among them to carry away heat.
With such awesome power at our disposal, we need a means of control. To reduce heat, we must reduce the number of atoms split, and therefore we must reduce the number of neutrons that are flying around. So, interspersed among the fuel pins are several rods made of hafnium, cadmium, or boron, elements which have the property of absorbing neutrons. These control rods are used like throttles for the reactor. As they're pulled out of the core, heat output increases. As they're pushed in, heat output decreases. In an emergency, or in the event of a loss of power, magnetic drive mechanisms on the rods release them, allowing them to fall quickly to their fully inserted positions in the core. This is called a reactor trip or SCRAM, and is a very fast way of shutting down the reactor.
(The word "SCRAM" originated in the first nuclear reactor, in which the main safety control rod was suspended above the core on a rope. To quickly shut down the reactor in an emergency, the rope was cut with an axe. The person responsible for doing this was called the Safety Control Rod Axe Man.)
The entire reactor core is encased in a large, thick-walled casing known as the pressure vessel. Cooling water is forced through this vessel, via side-mounted nozzles, to carry heat from the reactor core to the rest of the plant. The control rods enter the reactor vessel from the top; in-core instrument wiring enters the vessel from below. The diagram you see at left is an actual layout diagram for the core at TMI, looking down from the top, showing locations of control rods, inlet and outlet nozzles, and instrumentation.
The interior of the reactor pressure vessel is at once one of the most inhospitable places on earth, and also one of the most heavily instrumented. Thermocouples and flow instruments operate here in an environment where hard radiation can make ordinary metals turn brittle and crumble. The water here is hundreds of degrees above its normal boiling point and under a pressure of more than 2,000 PSI.
In order to ensure that power is produced most efficiently in all parts of the core, and also to ensure that the fuel "burns" evenly, the control rods are arranged in groups which are independently adjustable. The plant's design documents specify which rods are to be raised, and in what order, during each operating regime from cold shutdown to full power. Rod positions can also be altered as the core becomes depleted so that the fuel "burns" evenly among the fuel bundles.
Note that depicted on the diagram are several neutron detectors, which are used to measure reactor power. The amount of neutron activity in the core is directly and reliably proportional to the thermal power the reactor is producing, so by monitoring neutron flux we can keep tabs on the reactor's heat output. The instruments are in three ranges because no one instrument could be accurate over the wide range of power levels the reactor covers during startup and shutdown. The source range instrument is used at extremely low power levels, primarily when the reactor is shut down or is being started up from a cold shutdown. The power range instrument covers levels found when the reactor is operating at high power. The intermediate range instrument covers levels in between the other two instruments.
The top of the reactor pressure vessel is a heavy plate known as the head. The head is removed during maintenance and refueling operations when the reactor is completely cold. It's secured to the top of the pressure vessel by dozens of bolts which are all carefully torqued. Since the pressure within the vessel is in excess of one ton per square inch, one can scarcely imagine the force that acts on the head when the reactor is at full pressure. These are strong bolts!
In Pressurized Water Reactors (PWRs), such as the Babcock & Wilcox system at TMI, the reactor core itself is housed within a massive steel bottle known as the Pressure Vessel. Three coolant circuits, or loops,are used in moving heat from the reactor core to the turbines which drive the generators. Each loop uses water, in one phase or another, as its coolant.
The primary or RC (Reactor Coolant) loop begins in the reactor vessel. Water passing among the fuel assemblies in the core is heated to several hundred degrees Fahrenheit, under a pressure of over two thousand pounds per square inch. This is the ultimate pressure cooker; the high pressure is necessary to prevent the water from boiling into steam, which wouldn't adequately cool the fuel elements.
Hot water leaves the pressure vessel through the outlet nozzle, along a pipe known as a hot leg, and enters the steam generator. This is a heat exchanger, basically. In the diagram at right, you can see that the primary water passes through tubes inside the steam generator, in the process giving up most of its heat to the cooler secondary water flowing around the tubes. The secondary water, under far lower pressure, flashes quickly into steam.
An important feature is that the primary and secondary water never mix, or even touch. Primary coolant stays inside the tubes, and secondary water stays outside. This keeps any radioactive contaminants and activation in the primary water, and out of the atmosphere.
In reality, the internals of a steam generator are much more complex than this. Babcock and Wilcox, designers of TMI's reactor system, use a peculiar type of steam generator called an OTSG, or Once-Through Steam Generator. Highly efficient and reliable, the OTSG has one drawback. It holds less secondary water than other, more conservative designs. This means that should feedwater (the flow of secondary water) fail for any reason, the OTSG will boil dry very, very quickly. This made TMI's emergency feedwater systems very important.
Once the primary water has expended its heat in the steam generator, it's sent back to the reactor vessel, by way of the cold leg, for another trip. Primary coolant is constantly recirculated. Mammoth main coolant pumps, or RC pumps keep the water moving through the core. Each pump at TMI was two stories high, and delivered nine thousand horsepower. Each pump was capable of circulating 360,000 gallons of coolant, at rated pressure, per minute!
Attached to the primary coolant loop is a tall, slim tank called the pressurizer. The pressurizer is normally about half full of water, and half full of steam. A tube at the bottom connects the tank to the reactor vessel, and at the top are relief valves which open to protect the system when too much pressure builds up. By using powerful electric heaters and water sprays, the operators can, to a limited extent, control pressure within the primary system by controlling the size of the bubble. The steam bubble, the only steam ever allowed in the primary system, serves also to cushion the system against shocks, and prevents pipe ruptures from sudden pressure spikes.
The relief valves at the top of the pressurizer bear some further description, given their pivotal role in the accident. The main, large-diameter relief valve was known as a PORV, or Power Operated Relief Valve. The one used at TMI was an Electromatic Relief Valve manufactured by Dresser Industries. It's shown in orange on the diagrams.
The Electromatic valve was known to be somewhat unreliable. The valve was carefully designed by Dresser to never, ever fail to open, under any conditions. Unfortunately, there were quite a few conditions that might cause it to fail to close. In fact, a lesser-known, far less disastrous loss of coolant accident occurred at Davis-Besse Nuclear Power Station some years before TMI, due to this same valve's failure to close. So, at TMI, a manually operated block valve was installed, so that in the event that the relief valve failed to close, the outlet could still be blocked, and a disastrous loss of coolant could be prevented. The block valve is shown in violet in the diagrams.
The nuclear power industry prides itself on allowing for every "credible failure" that a plant might endure. In fact, the loss of great quantities of primary coolant, through a stuck valve, pipe break, or structural failure had been, everyone thought, well predicted and allowed for.
Two emergency core cooling systems served TMI in addition to the primary loop's heat removal capacity. The HPI (High Pressure Injection) system consisted of three pumps, two electric and one steam-driven. These pumps were designed to shove water into the core by brute force in huge quantities during a loss of coolant accident.
A depressurized system could also be cooled by the LPI (Low Pressure Injection) or makeup system, which is similarly equipped. Finally, if all else were to fail, a huge tank holding nearly a million gallons of borated water could be dumped, via gravity, into the core by the core flood system.
For safety's sake, all components of the primary cooling loop are located a huge concrete structure known as the containment dome. The walls of the containment are 12 feet thick, and reinforced with high- strength steel. It is designed to withstand the direct impact of a jet airliner, and can withstand incredible pressures from within as well.
The secondary loop is the power generating loop, and begins at the outlet of the steam generator. Steam from these outlets is sent through a series of dryers, which remove excess water vapor from the steam. From there, it's through a set of safety and throttle valves and directly into the turbines.
Turbines are like massive, enclosed windmills. Steam is directed against the blades of these windmills, spinning the shaft on which they're mounted. To improve efficiency, there are usually at least two stages to each turbine. A high pressure stage is first, and the steam gives up most of its energy there. Then, after being dried again, it is passed to the low pressure stages to expend the rest of its energy.
The shaft of the turbine is connected to a generator, which turns the mechanical power of the rotating shaft into electrical power. Power then leaves the plant through the switchyard.
Turbines, despite their immense size and imposing power, are quite delicate devices and must be treated with the utmost care. Water droplets, if allowed to strike the blades, will erode them -- wear them away! For this reason, a vacuum is developed in the main condenser and the low pressure turbine. This prevents the steam from condensing into water until it reaches the tubes of the condenser, and it also greatly reduces drag on the low-pressure turbine blades.
Should the turbine ever have to be tripped (shut down), we see another aspect of how delicate it is. The turbine must never be allowed to come to a complete stop when it's hot. If this happens, the turbine shaft will sag between its support bearings, throwing the blades out of alignment and incurring a seven-figure repair bill. So, machines called turning gear are used to slowly rotate the shaft, like a rotisserie, until it cools.
After leaving the turbine, the steam is piped into a condenser, which is actually quite similar to a steam generator -- in reverse. Cold water from the tertiary or "circ" water system runs through tubes in the condenser, and when the steam strikes these tubes, it condenses back into water, in the same way that your breath fogs a cold window. The condensed water is now called condensate, and is sent through a set of heaters to bring it back to a usable temperature.
Next, the water passes through a set of ion-exchange tanks, or polishers as they're often called. These remove excess mineral content and impurities from the water and ensure its purity. The water going into the steam generators needs to be free of minerals to avoid the buildup of "scale" (calcium and other deposits) inside the steam generators which would require a long "outage" to clean out.
After this stage, the water is known as feedwater, and is sent into a group of very powerful feedwater pumps, which force it back into the steam generators to be boiled for another cycle.
There are two types of cooling towers in use today, forced draft and natural draft. The towers at TMI, pictured here, are of the natural draft type, characterized by their imposing height and distinctive hourglass shape. Natural draft cooling towers have come to be a symbol of nuclear power, and certainly they're the most memorable sight to be seen at Three Mile Island. The hourglass shape is designed to cause a "natural draft" (i.e. a natural convective flow of air upward through the tower), as the name suggests.
Inside the towers, hot water from the condensers is pumped partway up the tower. It is then sprayed out onto a series of baffles, channels, and vanes known as the fill. These structures slow and spread the falling water throughout the tower's inner cross-section, dividing it into fine droplets and exposing maximum surface area to the air.
As the water falls, the fine droplets give up some of their heat by evaporation. Meanwhile, the warmed air rushes upward, carrying away the water vapor and drawing more cool air into the openings at the bottom of the tower. The remaining liquid water, which has now been greatly cooled, falls to the bottom of the tower and is collected.
The evaporative nature of the cooling towers' operation explains the large clouds of water vapor (similar to the foggy breath you exhale on a cold winter day) that we see above the cooling towers. This is one of the only outward signs that a nuclear power plant is in operation. The photo at left was taken after the accident, so there are no vapor clouds.
This process, by its very nature, evaporates a lot of water. To replace that volume, fresh water is drawn in from the Susquehanna River and added to the reirculating volume of tertiary, or "circ" water. It is seldom necessary for circ water to be released back to the river, although it would be entirely safe to do so.
The cooling towers exist primarily as an environmental protection measure. Dumping hot water into the river would be substantially cheaper, but would cause irreparable damage to the river's ecology.