Today is the twentieth anniversary of a completely unnecessary accident with wide-ranging consequences: the botched test at the Chernobyl military reactor #4 in the former Soviet Union (now Ukraine).
Most nuclear power plants are a subset of thermal power plants (those that convert heat into electricity). Thermal power plants use a heat source to boil water which then turns a turbine connected to a generator. Nuclear power plants simply use a nuclear reactor as the heat source. Many reactor designs have been tried.
-Boiling Water Reactors (BWRs) are the simplest type of nuclear reactor. They consist of rods of low-enriched uranium fuel (between 3% and 4% uranium-235, the rest being uranium-238, as opposed to natural levels of 0.71%) carefully arranged in a pattern that expands on heating like a thermostat. This gives BWRs an equilibrium temperature; as a BWR heats up, it expands, and the power decreases, lowering the temperature. BWRs also take advantage of the fact that water slows down or moderates the neutrons produced and used in nuclear chain reactions. They only work on slow neutrons, due to their size and layout and the fuel they use. When a BWR is immersed in ordinary water (as it must be to function), the relatively high temperature it produces (~600 degrees Fahrenheit) boils the water. Without the water’s moderating effect, the reaction slows down as the water boils. These two passive effects–expansion (a negative temperature coefficient) and boiling (a negative void coefficient)–prevent any manual action from taking the reactor out of control. Usually the fuel also contains neutron-absorbing materials known as burnable poisons that are destroyed during the reaction, which maintain the same reactivity level over the life of the fuel (normally it would go down as more fuel is burned; these initially decrease the efficiency of the fuel to offset this effect). In fact, is very feasible in principle to design a BWR that cannot melt down. Of course, manual methods for control of the reactor are provided. These are silver or boron control rods, which absorb neutrons and slow down the chain reaction when inserted between the fuel rods. Adjusting the flow of water through the reactor can have the same effect. However, there are 16 known naturally-occurring BWRs in uranium deposits, which automatically regulated their power for millions of years without any intervention. Clearly, safety is an effect in nuclear reactors, not a feature (By the way, this is a major disagreement that we have with the way the nuclear industry did business until recently (and still does to some degree): they tried to design extra safety systems into reactors instead of designing the basic mechanism correctly. This method is known as defense-in-depth and has the singular effect of costing money as it cannot make a bad system work.). Safety problems have to be designed in, as we will see later.
-Pressurized Light-Water Reactors (PWRs) use the same basic reactor of a thermostat-like arrangement of low-enriched uranium fuel rods and movable boron or silver control rods. PWRs have a negative temperature coefficient as well, since they slow down on the expansion of the fuel, and a negative void coefficient, since they slow down when they lose the neutron-moderating effect of their cooling water if it boils. PWR fuels can also use burnable poisons. The difference is that instead of allowing the water to boil, the water is pressurized and simply transfers heat. PWRs are not as quick to shut down (i.e., not as idiot-proof; Three Mile Island happened in a PWR), but are more stable. Some PWRs, like the modern AP-1000, are so stable that they can be cooled without actual coolant flow (a BWR’s coolant flow is the boiling steam). PWRs also cannot be controlled by changing the flow of coolant, since they do not boil their coolant (although they would shut down if there was a loss of coolant and the remainder did boil; the Three Mile Island meltdown happened when a loss of coolant flow–and eventually lowered coolant levels–in an early reactor allowed residual radioactivity to heat up the fuel). PWRs do not easily follow a changing electrical load, either; it is not as easy to change the reactor’s power level because they are so stable. PWRs are more complex than BWRs, meaning more mechanical complexity, but can also minimize the amount of material exposed to radiation (isolating, for example, steam handling equipment and the generator). PWRs and BWRs are usually united under the heading of Light-Water Reactors (LWRs), which have some general shared characteristics:
LWRs cannot be refueled online. The reactor is inside a large steel pressure vessel, and the top must literally be unbolted to access the fuel.
LWRs produce fairly large amounts of plutonium.
LWR-produced plutonium is not weapons-grade. When used in a bomb, it tends to start its own mini-reactions before the bomb can go off. Such a bomb would use most of its fuel during the detonation command, leaving too little to actually explode. It’s been tried for 50 years in the USA, Britain, France, and Russia, and nobody has been able to make it work.
LWRs cannot produce more fuel than they consume, with one specialized exception. Uranium-233, uranium-235, and plutonium-239 are considered fuel; thorium-232 and uranium-238 can be turned into fuel. Only the net production of uranium-233 from thorium-232 is possible in a LWR.
LWRs require enriched uranium. Light water doesn’t moderate neutrons efficiently enough to use natural uranium; too many of the neutrons from the splitting atoms are wasted, and extra “splittable” or fissile atoms need to be added.
All American nuclear power plants use LWRs. They are the most common reactor type.
-Pressurized Heavy-Water Reactors (PHWRs) use heavy water (the heavy hydrogen in the H2O is twice as heavy as ordinary hydrogen) in place of light water. Heavy water is a better moderator than light water, which allows a PHWR to use natural uranium and produce more fuel than it consumes under almost all circumstances. PHWRs usually have subtle design differences from PWRs due to the high cost of heavy water and to allow exploitation of the better moderator. The most common PHWR uses heavy water as the moderator and coolant but in separate pressurized systems; this type is known as a CANDU (some more modern ones, using light water coolant and heavy water moderator, are known as Advanced CANDU Reactors or ACRs). CANDUs place their fuel rods in heavy-water-filled high-pressure tubes, which are embedded in a tank of heavy water. Heavy water is pumped through the tubes, which then boil light water to generate steam to turn a turbine. In LWRs, the coolant and moderator are the same thing: the water. In the CANDU, the coolant and moderator are separate, which can cause a problem: if the coolant absorbs neutrons (which moderators also all do to some degree), removing it will allow more neutrons to get through. Instead of slowing down the reaction, boiling the coolant can then speed it up. The design of each pressure-tube reactor is different, though, which is why the CANDU’s positive void coefficient is small enough for the other passive safety effects and active safety systems to compensate, but the other major design that uses separate water coolant and moderator–the Russian RBMK–can go out of control, and the ACR’s void coefficient is actually negative. A meltdown in CANDU reactors would probably involve only one pressure tube, and would immediately stop the reactor. CANDUs can also run directly on nuclear waste from LWRs and refuel online. They do not use burnable poisons but rather distribute old and new fuel with the online refueling system.
-Fast Breeder Reactors (FBRs) use no moderator (they use fast neutrons). This can remove the safety effect of the negative void coefficient, create a positive void coefficient, or a negative one, depending on the reactor design. Sometimes, the coolant (which usually absorbs neutrons) is between the reacto
r and a neutron reflector; if the coolant were removed from such a reactor, fewer neutrons would be absorbed and the reactor would speed up. If the coolant has nothing to do with the neutron flux (normally), there is no void coefficient. If the coolant tends to enable more reactions–for example if it reflects neutrons–there will be a negative void coefficient. FBRs usually operate at the edge of criticality and most FBRs have a strongly negative temperature coefficient. FBRs can only run on enriched uranium (higher enrichment than LWRs) or a type of fuel known as MOX (a mix of plutonium and uranium) because of their poor use of neutrons. FBR coolant cannot absorb or moderate neutrons, so FBRs use liquid metal as coolant (usually liquid sodium). The main value of FBRs, though, is that they produce large amounts of plutonium fuel from uranium-238. This allows the use of the other 99.3% of natural uranium. They can also consume nuclear waste much more thoroughly than CANDUs. There are as many types of fast reactors as there are slow reactors, but FBRs generally use a reactor with enriched uranium (at least 5.64% as opposed to 2.14% for LWRs and 0.71% for everything else) surrounded by a “blanket” of uranium-238.
-An important subset of FBRs are Integral Fast Reactors (IFRs). An IFR is a fast reactor with the entire fuel cycle onsite. Most probably, an IFR would be fueled by nuclear weapons material mixed with depleted uranium left over from enrichment (enrichment increases the amount of uranium-235 in uranium by taking out some uranium-238; the U-238 is stored in barrels at the enrichment plant but is still useful). After that, however, an IFR could extract 99% of the energy in its fuel, as opposed to current reactors’ 3%–meaning that 60 years of operation could be fueled by one truckload of fuel. This efficient use of fuel means that IFRs could generate all of the electricity for the United States for the next 500 years using only the uranium left over from enrichment. IFRs are also passively safe, meaning that they rely on the physics of the reactor, not active controls, to prevent accidents. IFRs are geometrically arranged to have an equilibrium temperature: excessive heat makes the fuel expand, which disrupts the reaction and slows it down. Because IFRs cannot melt down, they do not need high-temperature fuel, and can use high-efficiency metal fuel instead of ceramic as in other reactors. IFRs can be configured to either consume or breed fuel, but would probably be used to extend the current uranium supply while consuming nuclear waste and making fissile material useless for bombs–three supposedly daunting problems.
-Some nuclear power plants are not thermal plants. These reactors are cooled by gas, usually carbon dioxide or helium, and moderated by graphite (although one, Lucens in Switzerland, used heavy water). Since there is no liquid coolant, there is no void coefficient, but the geometry of the reactor can still be used for passive safety. The Pebble-Bed Modular Reactor (PBMR) is one advanced example. Older gas-cooled, graphite-moderated reactors include Britain’s AGR and MAGNOX reactors. Graphite is an excellent moderator; these reactors can use natural uranium. Graphite is also flammable, but these reactors can prevent fires by not reaching ignition temperatures and using an inert gas coolant, and those without passive safety can use a water flood system to extinguish a fire. There are as many ways to build a gas-cooled/graphite-moderated reactor as there are to build a LWR or fast reactor. Done right, they can work just as well.
-An important type of nuclear reactor which is not used in nuclear power plants is a weapons-production reactor. Only the isotopes uranium-233, uranium-235, and plutonium-239 at greater than 90% purity and in the absence of certain others (plutonium-242, erbium) work in bombs. Reactors normally produce plutonium that is a combination of isotopes 239, 240, 241, and 242, and use uranium that is at most 4% uranium-235. Plutonium-239 is produced from uranium-238, and plutonium-240 and up are produced in order from plutonium-239. Weapons-production reactors then must remove fuel rods before more than 10% of the plutonium-239 becomes plutonium-240. This condition can be reached in a matter of days, depending on geometry. To remove the fuel, the 150,000-pound reactor head must be unbolted and laid to the side by a crane while the oldest fuel is removed, new fuel put in, and the rest rearranged to control power (burnable poisons cannot be used, since the presence of any in a bomb would render it completely useless). The world record for doing this is 15 days, and the average in the United States is 38. Weapons-production reactors also cannot achieve high burnups and consequently waste a lot of fuel. Clearly, weapons production is incompatible with a power program, which must be on as much of the time as possible and must conserve fuel. A major design influence for the reactors at Chernobyl was to try to fuse these two incompatible aims.
-The type used at Chernobyl was the Water-Cooled/Graphite-Moderated Reactor (RBMK in Russian). The general class of water-cooled and graphite-moderated reactors (LWGRs) is not bad in and of itself, but have to be designed correctly. The RBMK is a particular kind of LWGR that is a textbook example of how not to do pretty much everything in reactor design.
A little history: the Soviets were not very good at industrial-scale uranium enrichment or heavy water production, and needed electricity and nuclear weapons. They lacked real heavy industry, but needed to build large plants. They did about as much as they could with the BN-600 breeder, but that was a civilian program, and not a very large-scale one at that. They decided to kill four birds with one stone: build a weapons-production reactor that would generate large amounts of electricity without any enriched uranium or heavy water. The RBMK was the result.
It used a graphite moderator in an inert-gas-filled container–but the reactor could reach ignition temperature, so if oxygen were to somehow leak in while the reactor was at full power, it could very easily catch fire.
Light-water-filled pressure tubes, which contain the fuel rods and control rods, go vertically through the graphite. Consequently, the refueling machine must be mounted on top of the reactor instead of at its side, making it too tall for a containment building.
The control rods use a graphite tip, followed by a section of water, and then the actual control rod, supposedly to make emergency shutdowns faster. The tiny problem with this is that graphite and water are moderators, so inserting a control rod briefly raises the reactor power. If enough control rods are withdrawn (more than 181 out of 211), an emergency shutdown is actually dangerous. So did the Soviets put an interlock system–not even a notoriously unreliable Soviet computer–in charge of the control rods? Of course not. It’s completely manual.
The main problem, though, is that the RBMK has a large positive void coefficient, so large that it overshadows any other engineered safety features. The water in the pressure tubes absorbs more neutrons than the graphite does, so if it boils, the moderator is made more efficient and the reaction speeds up. Obviously, this is a safety problem: if you lose coolant, the reaction should slow down to prevent a meltdown. Furthermore, the RBMK was actually designed to boil the water in the tubes! If the geometry were done better, if heavy water were used as the moderator instead of graphite, if the pressure in the tubes were higher, or if the enrichment levels were higher, the positive void coefficient could be engineered into a negative one, just like the CANDU’s (much smaller) positive void coefficient became a negative one in the ACR. These measures, of course, would defeat the RBMK’s original purpose, so they were not used.
The net effect of the positive void coefficient and the badly-designed geometry was a positive temperature coefficient (i.e., if the temperature goes up, the rea
ctor speeds up, raising the temperature, making it speed up more, until power is manually reduced or it melts down). Interestingly enough, the RBMK is fairly stable at high power, when there is so much steam in the tubes that the water isn’t a significant factor. Only at low power can the amount of water in the tubes change significantly with fairly minor boiling.
The positive temperature coefficient is such that neutron poisons–isotopes that absorb neutrons and inhibit the reaction–can drastically affect power. The fission product xenon-135 in particular can cause power to spike if it is present in large quantities and suddenly disappears. A perfect cause for such an event would be high-power operation, followed by a reduction in power, then a power surge, and a loss of coolant pressure. Sound improbable? Read on.
The RBMK’s last major fatal flaw lies in near-laziness on the part of the designers. In order to increase the capacity of the RBMK for newer models, the layout of the reactor was not redesigned. They simply made it longer. This works fine until you get to the point where you have two separate critical masses. Normally, the control rods only have to be inserted a certain distance into the reactor for it to start to lose power. If there are two critical masses, the first must be completely stopped and the control rods must reach far enough into the second to stop a power surge. This takes 18 seconds in an RBMK. A power surge can easily destroy the reactor in under six. To try to mitigate this obvious safety problem, floor-mounted control rods were added to help shut down the second critical mass. Unfortunately, floor-mounted control rods cannot drop into the reactor; they have to be driven, and a loss of electricity renders them completely useless.
Other, more minor flaws exist in the control system, backup power, and miscellaneous systems. They contributed to the accident, although in a more minor way.
Still, the RBMK can be operated properly. If it is, it can provide reliable, clean electricity. A list of things–which have a low probability of happening during normal operation and some of which are incompatible with normal operation–need to be in place in order for an accident to happen:
1. The reactor must be at low power to be unstable.
2. More than 181 control rods must be withdrawn.
3. The reactor must have been operating for a long time.
4. There must be a breach of the pressure tubes and/or an oxygen leak into the moderator.
This is with the reactor designed as it was. Almost every design feature of the reactor allowed the accident to happen. Had even one major flaw in the reactor not been there, the accident would not have happened. For instance, if the control rods were fast enough to shut down the reactor before a power surge could destroy it, the power surge that happened during the accident could have been controlled. If the geometry of the reactor had given it a negative temperature coefficient, the positive void coefficient wouldn’t have mattered. Had the pressure tubes been horizontal, not vertical, they might have built a solid, American-style reinforced concrete containment dome instead of a warehouse with a concrete lid. The design was truly a perfect storm of bad engineering. But how did the accident happen?
The Chernobyl accident was doubly unnecessary because it occurred during a test. The Soviets wanted to learn whether the primary electricity supply (the main generator) would last long enough for the backup (a diesel generator) to start (the fact that it didn’t start immediately is another minor RBMK design flaw). Secondarily, the test could prove to the world that their RBMK was as advanced as the safe American reactors, so they chose their most advanced RBMK–Chernobyl Unit 4.
In the afternoon of April 25th, the test was supposed to begin with the reactor being taken off the grid. However, unexpected load prevented the reactor from being disconnected until that evening. When it was disconnected, there was not enough time to bring the reactor safely down to low power. The test should have been aborted at this point, but the operator in charge of the test, who had never been trained on the RBMK (he was trained on nuclear submarines), ordered that the reactor be lowered to about 30% power. The operators conducting the test went too far, though, and lowered the power to about 1%. Instead of aborting the test (their second chance), they raised power to a little over 6%. During this time, the burnable poison xenon-135 started to build up, as power was too low to consume all of it. At this point, the number of control rods in the core was reduced manually to try to offset the xenon. The equipment which was supposed to draw the current from the generator and diesel backups was then turned on, which unfortunately was the coolant pump system. The positive void coefficient then performed its only good function and lowered the power of the reactor, since higher coolant flow would result in fewer voids, and more control rods were removed to make up for this. At this point, only six to eight control rods of a required 30 were in the reactor. The primary electricity supply was then switched off. As the coolant flow decreased, the water in the pressure tubes started to boil, and the positive void coefficient kicked in. The reactor accelerated out of control–over 1,000% of rated power–and an operator, unaware of the number of control rods in the reactor, panicked and scrammed it. The graphite tips of the control rods entered the reactor, the power spiked, and fuel pellets started to melt and pop out of their rods into the coolant. The remaining water flashed to steam, rupturing some of the pressure tubes, blowing the concrete lid off of the reactor, and setting it down nearby at a 75 degree angle. The lid pulled the rest of the tubes out with it.
Meanwhile, in the second critical mass in the lower half of the reactor, the xenon escaped through the broken tubes. Since the xenon was keeping the lower half of the reactor under control, and the control rods were by now melted or blown out, the lower half of the reactor experienced a second power surge, which melted fuel and caused the remaining fuel rod cladding to burst. The operators detected the second explosion and tried to scram the reactor again. It didn’t work.
Simultaneously, air rushed into the reactor from the outside. The graphite was obviously far past its ignition temperature and burst into flames. Radioactive material went with it: approximately 5% of the fuel was ejected during the two explosions and subsequent fire.
The operators reacted to this by trying to determine the radiation level in the building. The plant’s two detectors didn’t work. A third was brought in, and registered what they thought were ridiculously high levels. Apparently not noticing the fire, they tried to pump water into the reactor to prevent a meltdown. By that time, of course, the reactor no longer existed in any recognizable form, so what happened wasn’t strictly speaking a meltdown. It was certainly not a nuclear explosion, which would have required a dedicated device much different from any reactor in existence (much less the RBMK) with materials not present at the site and precise timing.
When they saw that pumping water had no effect, they notified Soviet authorities, who prepared a typical Soviet response: they tried to put out the fire by dumping clay, limestone, sand, and the neutron poisons boron and lead on what remained of the reactor building. This only acted as insulation and made the overheating worse, and was stopped. Thousands of people, known as “liquidators,” were brought in to put out the fire–without radiological protection. Forty-seven of them would die from radiation poisoning. It took over a week to put out the last pieces of graphite.
The day after, April 27, the nearby city of Pripyat was finally evacuated. It took the Soviets almost a month to evacuate a 30-kilometer-radius exclusion zone.
Work was begun on a shack to surround the destroyed reactor until a perm
anent structure could be built. It was assembled mostly by remote control, and is not at all structurally sound. A new shelter for the reactor is planned–hopefully before the “sarcophagus” currently in place caves in.
Long-term effects are not as well-known. The most credible report–the September 2005 report by the UN–predicts approximately 4,000 cancer deaths. Of course, anti-nuclear groups point to the 100,000 people who have died in Ukraine since the accident and ask whether only 4% of them could have been killed by Chernobyl. They point to tragic birth defects and cancer, especially pediatric cancer. However, birth defects and cancer happen in other places as well, and for most cancers there is not a statistically detectable increase in the area around Chernobyl. That does not mean that we do not value or care about these people’s lives. But determining the specific Soviet pollution that caused their cancer is very difficult. Blaming it on our personal bogeymen, whatever they may be, is fundamentally disrespectful. These thousands of personal tragedies deserve more than a knee-jerk response.
The policy implications are huge. Future RBMK construction has been rightfully stopped. Unfortunately, Chernobyl politically affected construction of reactors that have about as much to do with each other as a coal burner and a gas turbine. Why should a harebrained stunt at a uniquely terrible Soviet reactor condemn nuclear power that’s done right? After all, one doesn’t associate chemical processes with each other; coal mining disasters don’t bring calls to stop using natural gas.
Chernobyl showed us how much one has to really, really try to cause a nuclear reactor accident. The RBMK is a very bad design, as we have seen. However, even with every design problem in the RBMK, human error on a test that should not have been run in the first place was required to cause an accident. Had the reactor been operated correctly in a culture of safety, Chernobyl never would have happened. Had the reactor been designed correctly in a culture of safety, it could have been abused even more excessively than was done and Chernobyl never would have happened. Does it really say anything about the nuclear industry, nuclear technology, or the general concept of nuclear energy? No. It says a lot about the Soviet system, and others like it, where irresponsibility was rampant. They didn’t do much of anything right, and nuclear power is no exception. We can be just as irresponsible, possibly even more than the Soviets, by pushing energy policy off on the current teenage generation’s grandchildren. We could get away with it. With widespread enough extraction and infrastructure investments, natural gas could replace nuclear power and coal for perhaps 40 years. Once we run out of that, coal could work for electricity and as a feedstock for motor fuels for perhaps another 75. Then the grandkids get to figure out where to get electricity and hydrocarbons. They get to start over, reconstituting nuclear technology that could sit abandoned for 115 years. I wouldn’t be surprised if some of them tried to move to the Moon if it got bad enough. The economic conditions are perfectly set for such a disaster. This frustrates me, because I know it could be better: objective analysis shows good nuclear power to be orders of magnitude better than good chemical power. Even bad nuclear power is significantly better than good chemical power–Chernobyl’s 4,000 deaths are dwarfed by the human toll of coal fumes, which is approaching one million since the last American reactor order. We could, I suppose, waste money and time on renewables (which by definition have a production cap–oil could be renewable so long as it’s not depleted faster than it’s produced), or “soft energy,” or magic wiffle dust, but XYZ power source, no matter how vast it may be, is useless unless you can collect it. We could get into the numbers–1,400 watts per square meter, 20% absorbed by the atmosphere, 35% reflected off clouds, then spread out a factor of two or three because of the angle of inclination to the Sun, 5%-15% conversion efficiency, 14,767.75 terawatt-hours global consumption, but I leave that exercise to you. I don’t “believe” that nuclear power is the answer. I think, and analyze, and struggle, and come to the conclusion that I honestly didn’t want to draw:
Nuclear is Our Future
(and if it isn’t, it should be)
In closing, I ask you: Do you, personally, have the courage to tackle the energy crisis? Will you do what you can, in your corner of the world?
Filed under Alternatives, Chernobyl, Health, Industry Performance, International, Physics, Radiation, Safety
