My notes (Nancy’s) on the Reactor Safety Training Course (R-800) from the NRC http://pbadupws.nrc.gov/docs/ML0210/ML021080117.pdf
The notes are an abbreviated gathering of parts that may be relateable to Fukushima from sections 3 and 4.
Steam can be sent from the RPV via a safety valve and piping into the suppression chamber (torus).
If the pressure in the suppression chamber (wetwell/torus) is higher than that in the drywell vacuum breakers equalize the pressure.
There is a cooling spray system for the drywell portion of containment and also cooling system in the suppression chamber that runs through a heat exchanger to cool the water in the suppression chamber.
Mark 1 reactors also have a direct vent system to the outside in the suppression chamber to relieve pressure. (There were attempts to open these at Fukushima)
Mark 1 containments were not specifically designed to handle these situations:
Reactor Vessel Rupture
Steam Generator Rupture
Large break (LOCA) Loss Of Coolant Accidents
These kinds of accidents produce higher containment pressure and more fission products. High heat caused by these kinds of accidents can also lead to failing containment by way of cracks or ruptures.
Areas of containment prone to failures are piping penetrations, electrical penetrations, hatches and airlocks. Isolation valves are provided on all pipes and ducts that penetrate the containment but can also fail.
Pressure suppression style containment designs are smaller and allowed a 1.0 wt.%/day leakage rate without having a significant safety impact.
Statistics for Peach Bottom are likely similar to units at Fukushima as they have similar designs.
Peach Bottom BWR Mark 1 Peak Design Basis Accident Pressure PSIG 49.1 (KPA 339)
Peach bottom design: 56psig, ultimate failure pressure 148 psig
Challenges to containments can occur during four time regimes:
1. at the start of the accident
2. prior to reactor vessel failure
3. at or soon after reactor vessel failure
4. long after reactor vessel failure.
These containment systems can build pressure over hours or days so they have cooling systems installed. Sprays, fan coolers, suppression pool cooling, or emergency core cooling recirculation cooling are installed and have to meet a “single failure” standard of durability.
These active systems do not act quickly enough to affect the initial blowdown during a large-break LOCA, but limit further pressure increases and are also beneficial during slower developing accidents.
These major accidents can cause the heat removal systems to fail. In some situations the design pressure may be exceeded early, but the ultimate failure pressure would not be reached for many hours or even days. The point where the RPV is breached, usually by melted fuel burning a hole through the lower portion of the RPV is of specific risk. Several things can happen at that point, even happening at the same time.
1. Steam spike
2. Steam explosion
3. Direct containment heating
4. Hydrogen combustion
5. Containment shell melt through
6. Downcomer failure (Mark II BWR)
Steam spikes or explosions can happen if there is water in the pedestal region below the RPV as molten fuel hits the pool of water. Water can build up in the pedestal due to a broken reactor cooling pipe or from containment cooling spray systems. Steam spikes are more of a problem if containment is already at a high pressure. A steam explosion can send out shockwaves that can damage the containment or the RPV supports. If the RPV supports fail the RPV can move significantly. RPV movement can cause more problems for pipe penetrations in the containment wall. A steam explosion can bypass the suppression chamber (torus) and cause the containment to over pressurize.
A containment that leaks slowly can leave the radioactive substances inside containment or the building, a large fast leak will dump large amounts of radioactive substances into the environment.
In a severe accident the reactor vessel “blowdown” (of steam) can be beyond the design basis of the containment structure. As heat and pressure increase before RPV failure containment can deform and create leaks around penetrations. This can also cause ruptures in the steel containment of a steel/concrete containment like at Fukushima. This is the likely outcome according to recent studies. The concrete structure of the containment is more likely to crack as the steel ruptures.
Three things can significantly add to the pressure in containment.
1. Initial reactor coolant system blowdown (into the torus)
2. Steam release when reflooding the reactor with water
3. Hydrogen release when reflooding the reactor with water.
If penetrations don’t leak steel containments have a larger margin between the design and ultimate failure pressures than concrete containments.
Direct Containment Heating (DCH) involves the ejection of the melt from the vessel at high pressure, thus spraying the molten material into containment. The small particles of fuel being sprayed into containment can super heat the atmosphere inside containment. This can rapid heat the containment area faster than a steam spike.
Any coolant (water) still in the RPV will then be released into containment. This water can quickly produce hydrogen. If the oxygen and hydrogen mixture is of the right ratio and temperature, the molten fuel can act as an ignition source and cause a hydrogen explosion. If large enough it can threaten containment. Hydrogen explosions can happen in conjunction with other events to fail containment.
The likely melt through scenario in Mark 1 containment is for the melted fuel (corium) to reach the floor of the drywell, flow out of the concrete vertical pedestal tube and across to where the floor and steel liner on the wall meet. This type of failure can happen faster and is more likely than burning through the basemat concrete directly below the reactor.
Public risk is probably dominated by accidents in which substantial core damage occurs and the containment fails or is bypassed.
A disturbing note found in the document, that the initial mindset was locating a reactor in a low population area was considered a safety buffer and containment systems were created to protect the population nearby.
“containments began to evolve when designers realized that remote siting would not be practical in all cases. The first containments were provided for the Knolls Atomic Power Laboratory and Shippingport experimental reactors in order to allow them to be sited in more populated areas.”
The yellow containment cap bolt style used at Peach Bottom nuclear plant is the style used at Fukushima and is prone to bolt elongation (stretching) and flange separation (opening) under high temperatures. Accidents where high containment (drywell) temperatures occur are more likely to have the containment cap bolts stretch.
The silicon seal of the containment cap can also fail, it degrades significantly at temperatures over about 600 K (620 ‘F). These two flaws can create an outlet for containment in situations where the pressure and temperatures are both high.
**This is certainly a potential scenario as part of the unit 3 explosion. Considerable steam and debris was ejected high in the air but the containment cap or parts of it have not been found in the debris.
Corium (melted fuel):
**This quote talks about the failure in a Mark II containment but could likely happen in a Mark 1 containment.
A phenomenon of importance for Mark II BWRs is downcomer failure. While Mark II designs vary significantly, there is often the potential for molten material to flow across the floor and into the downcomers. This molten material may directly fail the downcomer or, possibly, lead to a steam explosion that fails the downcomer. Downcomer failure does not lead to immediate containment failure; however, the suppression pool is bypassed, thus negating its heat removal and fission product scrubbing capabilities.
** This could potentially happen at a Mark 1 if the corium didn’t melt through the drywell wall at the downcomer vent connection first. The location of the downcomer vent to the floor of the drywell varies at different Mark 1 reactor builds. Some are closer to the drywell floor than others. Those that sit at the junction of the drywell floor would be a candidate for this.
If the melted core is in containment the heat must be removed to prevent temperature and pressure buildup. The heat and pressure can weaken structures and make them more likely to leak or fail at lower temperatures. There are also non-condensable gasses created by the reaction as the corium destroys concrete it comes in contact with and add to the total pressure inside containment.
Base mat melt through happens when the melted fuel is in contact with the concrete for an extended period of time. It can cause hydrogen, non condensible gases, radioactive and nonradioactive aerosols all in large amounts. During this phase hydrogen explosions can occur if the mixture of hydrogen and oxygen are in the right ratios. Excess amounts of steam can suppress a hydrogen explosion, so adding water cooling at this point can remove the steam creating conditions favorable for a hydrogen explosion. The melted fuel can also burn through the base mat if it is not cooled sufficiently.
Venting can be done through the wetwell (torus) or through the drywell. Wetwell venting is preferred as it can scrub radionuclide releases through the wetwell. If the reactor has hardened venting pipes they are less likely to fail under high containment pressure.
When melted fuel comes into contact with water the following things could happen.
1. The water may act to cool and quench (refreeze) the molten core debris, and may limit the spread of molten core across the containment floor.
2. The debris may form a molten pool under the water, probably with an overlying crust layer, and remain molten.
3. An energetic fuel-coolant interaction may occur
Water in contact with corium does not always mean it will cool it. If the fuel forms a debris bed across the surface the water may not reach the area that needs to be cooled. Large amounts of forcefully escaping gasses can prevent water from reaching areas of the corium it needs to cool. This allows the fuel to burn through the concrete or metal where water can’t reach.
The biggest threat to containment is a steam explosion.
1. generation of dynamic pressure loads (shock waves) that can fail the containment structure
2. generation of pressures and shock waves that can fail vessel support structures, leading to movement of the vessel and failure of containment piping penetrations
3. generation of energetic missiles that can vessel case will always be at low pressure, be thrown into the containment
4. generation of pressures and shock waves that can fail the drywell floor of a BWR
The magnitude of a steam explosion can be impacted by the following:
1. the amount of water available to participate
2. the composition of the melt, including the amount of unoxidized metals that may react during the explosion
3. cavity or pedestal region geometry as it may lead to confinement of the explosion or focusing of shock waves
4. transmission of shock waves through a water pool
5. pouring rate and contact mode, i.e., water on corium, corium on water, or jet ejection into water
6. fraction of the core participating.
3 ways the corium may create a steam explosion are falling into a pool of water, ejecting in a jet spray into a water pool or having water dumped on top.
Rapid quenching (watering) can also cause rapid oxidation of the metals in the fuel creating hydrogen.
Corium in the RPV pedestal can encounter a large amount of water if the containment sprays have operated or large quantities of water have flowed out through a break in the reactor coolant system. Then, if the sump and floor design allows, some of this water will overflow into the reactor cavity (pedestal). If the drywell spray system has not been used and water pipes have not spilled down into containment or the pedestal it is less likely to have water buildup below the fuel melt.
BWR containment types are susceptible to failure of the vessel supports, with relatively small amounts of water present.
There is also the possibility of corium flowing across the drywell floor and down the downcomers into the suppression pool, failing the downcomers with a steam explosion or as a result of meltthrough. For the three BWR types, drywell failures from steam explosions contribute noticeably to the overall containment failure probabilities.
Core Concrete Interactions (CCI)
In addition to melting through the basemat corium can destroy vessel supports and other local structures that can indirectly lead to containment failure.
Corium can penetrate concrete at the rate of several inches (tens of cm) per hour.
The intense heat in corium is sufficient to melt concrete, release gasses and melt the remaining oxides into the corium mass. This can include the metal reinforcing structures like rebar that are in the concrete. All of this addition to the corium makes it harder to cool.
If the corium is stopped before it melts though the basemat is based on these factors.
1. type of concrete and aggregate used in the structure
2. basemat thickness
3. cavity size and geometry
4. melt mass in the cavity
5. melt composition
6. Presence of overlying water
The presence of an overlying water pool does not guarantee that the corium will be cooled. A crust may form over the melt, and heat transfer may be insufficient to remove all heat from the melt. A number of experiments have shown minimal effect of water on concrete ablation rates. However, overlying water can reduce fission product releases even if it does not cool the debris.
So all the water at Fukushima may help keep the radiation releases down but may not be doing much to stop the corium from burning through structures. It was also noted that the big steam shows seen at Fukushima after earthquakes could be due to cracks in the crust over the corium due to the shaking.
Concrete loses its structural integrity before it actually melts. The process of damage to melting takes place between 30-1000 celsius. The point where the concrete loses structural integrity varies based on the make up of the concrete and is of particular risk at failing vessel supports in the BRW reactor.
The large amounts of steel in the reactor cavity (pedestal) can add to the melted mass as does the portion of the RPV that may melt with the fuel. The gasses produced due to the core concrete interaction can exceed those produced by the zirconium cladding on the fuel. A few thousand pounds of combustible gas can be created by the core concrete interaction. These gasses can spontaneously ignite if there is not a water pool over the melted fuel. These gases are noncondensible, they can lead to significant pressure buildup that can not be removed using sprays or suppression pool cooling.
In the absence of an overlying water layer, Core-Concrete Interactions produce dense clouds of aerosols. Aerosols can be created in minutes based on the VANESA chart. Aerosols released from CCIs can significantly impact the concentrations of radionuclides in the containment atmosphere. In a situation where there are aerosols being created and the containment fails, large amounts of radionuclides can be released into the environment. These aerosols can also plug air filtration systems. If containment failure is delayed long enough the aerosols can settle, releasing less to the environment.
Direct Containment Heating:
When vessel failure occurs at a pressure of a few hundred psi (several hundred kPa) or more, the melt will be ejected as a jet into the reactor cavity. The fuel can be fragmented by jet breakup and swept out of the cavity into the containment where it will heat the atmosphere (direct containment heating [DCH]). This thermal energy can be transferred to the containment atmosphere through radiative and convective heat transfer. This heat transfer will be very rapid, with much of it occurring in a matter of seconds if particles remain airborne and continue to encounter cool atmospheric gases.
This can lead to very rapid and efficient heat transfer to the atmosphere, possibly accompanied by oxidation reactions and hydrogen burning. It is clear that in extreme cases high pressure melt ejection and direct containment heating can produce pressures that threaten structural integrity of the containments.
From the RPV molten material will be ejected as a liquid stream, as the liquid corium level in the vessel drops, gas blowthrough will begin to occur. High velocity expanding gas flow provides the motive force for ejecting corium from the reactor cavity (pedestal). Some experiments indicate that the gases exiting the reactor cavity can contain as much as 50% hydrogen during some phases of the blowdown. The fuel melt is fully dispersed from the vessel and cavity long before (gas) blowdown is complete. Gases exiting the reactor cavity may have velocities of several hundred feet per second. Any trapped corium may result in subsequent core debris-concrete interactions within the reactor cavity.
Vessel failure may occur at a small opening, such as an instrument tube, or as a result of a larger rupture. This can be important for ex-vessel steam explosions or for the gas blowdown phase.
The degree to which the debris can be transported to the top of the containment affects the resulting pressure rise. Some containments have a fairly open path around the reactor vessel to the upper containment. Melt can be dispersed upwards from the cavity through the annulus around the RPV into the refueling canal and upper dome.
Melted fuel rich in metal will tend to result in higher DCH (Direct Containment Heating). As metals oxidize the combustion of any hydrogen produced by the oxidization adds to the DCH.
The presence of water in the reactor cavity could result in some quenched debris, thus partially mitigating the DCH threat. But a jet of molten material entering a pool of water will often lead to a steam explosion.
With small levels of water, the experiments show that the initial contact with molten debris produces a steam explosion that blows the remaining water out of the cavity. High temperature melt injection into model reactor cavities filled completely with water have produced dramatic steam explosions of sufficient magnitude to threaten structural damage. Water also provides an additional source of hydrogen that could lead to rapid hydrogen production.
Though they have automatic depressurization systems, boiling water reactors could be susceptible to high pressure melt ejection in some accidents. Hydrogen combustion in the drywells of inerted boiling water reactor containments will not be especially important, but pressurization as a result of heat transfer to the drywell atmosphere might be significant in these containments that are typically much smaller than pressurized water reactor containments.
In reactor accidents the conditions inside containment prior to hydrogen combustion may include elevated temperature, elevated pressure, and the presence of steam. With 75% additional nitrogen, the atmosphere is inert. This corresponds to 5% oxygen at the limit of the flammable region.
Dry hydrogen/air explosions:
Common sources of ignition are sparks from electrical equipment and from the discharge of small static electric charges.The addition of a diluent, such as steam, will increase the required ignition energy substantially.
Deflagrations are flames that generally travel at subsonic speeds relative to the unburned gas.
Deflagrations propagate mainly by thermal conduction from the hot burned gas into the unburned gas, raising its temperature high enough for a rapid exothermic chemical reaction to take place.
For hydrogen concentrations between 4.1 and 6.0%, there will be upward propagation from the ignition source. Hydrogen concentrations between 6.0 and 9.0% will produce both upward and horizontal propagation, and hydrogen concentrations above 9.0% will produce propagation in all directions, although the upward propagation may be faster than the downward propagation.
In laboratory experiments that when hydrogen:air mixtures with hydrogen concentrations in the range 4-8% were ignited with a spark, some of the hydrogen was not burned. The range of incomplete combustion corresponds to the range in which the mixture is above the flammability limit for upward propagation, but below the flammability limit for downward propagation.
A detonation is a combustion wave that travels at supersonic speeds relative to the unburned gas in front of it. For near stoichiometric hydrogen air mixtures this speed is about 6600 fps (2000 m/s).
The transition from deflagration to detonation and is still not completely understood after more than 50 years of investigation. We can say that, in most postulated reactor accident scenarios, deflagrations are much more likely than detonations.
Hydrogen air mixtures near stoichiometric (about 29% hydrogen, two parts H2 to one part 02) are known to be detonable. Mixtures departing from stoichiometric, either in the hydrogen-lean or hydrogen-rich direction are increasingly more difficult to detonate. It has been observed that “detonation limits” are functions of geometry and scale, and not universal values at given mixture concentrations, temperatures and pressures. A detonable mixture requires adequate hydrogen and oxygen, but not too much steam.
Detonations can also start from deflagrations that accelerate to high speeds pushing shock waves ahead of the burn front until at some point shock heating is sufficient to initiate the detonation. Sources of such highly accelerated flames are high speed jets coming from semi-confined regions and flames passing through fields of obstacles. Many obstacles that might potentially cause flame acceleration, such as pipes and pressure vessels, are present in the lower sections of most containments. Very fast burns may also occur due to the presence of a very intense ignition source, such as a jet of hot combustion products formed subsequent to ignition in some adjoining semi-confined volume. The detonability of a mixture is increased (cell size is decreased) with increasing temperature
Because of the containment geometry, the shock waves may be focused in local regions, such as the top center of the containment dome, giving rise to large local peak pressures and impulses.
Hydrogen concentrations can happen in these places
1. near the hydrogen release point
2. under ceilings or in the dome due to the rise and stratification of a low density plume, or 3. near steam removal locations such as ice condensers, suppression pools, and fan coolers.
Missiles may be generated when combustion (deflagration or detonation) occurs in a confined region or when a propagating combustion front produces dynamic pressure loads on equipment. Such missiles may pose a threat to the containment structure itself, as well as representing a potential threat to safety and control equipment.
For the hydrogen to burn, it is necessary that at some location the hydrogen:air:steam mixture be within flammability limits. Hydrogen that enters the containment may start to burn as a turbulent diffusion flame. A diffusion flame is one in which the burning rate is controlled by the rate of mixing of oxygen and fuel.
Combustion can begin either because of an outside ignition source, or because the mixture temperature is above the spontaneous ignition temperature. The spontaneous ignition temperature is in the range of 9591076 °F (515-580 -C). Above this temperature, combustion can occur without external ignition sources such as electrical sparks.
The same quantity of hydrogen at TMI could easily cause a detonable mixture in a BWR containment.
In general, there are very few regulations and guidelines dealing with beyond-design basis accident phenomena in reactor containments. For example, there are no specific rules dealing with CCIs, ex-vessel steam explosions, or direct containment heating. Hydrogen control has been an exception to this approach, significant regulations were passed following the TMI-2 accident.
After TMI the NRC ordered BWR containments to be inerted with nitrogen. These containments are small enough that relatively low levels of zirconium oxidation could produce detonable mixtures in containment. This doesn’t prevent hydrogen issues outside of containment or due to leaks.
The hydrogen rule was set up to address only degraded core accidents and not full scale melting and vessel breach. NRC rules assume watering will be resumed and won’t have a large burst of hydrogen so most design basis accidents are not addressed by NRC rules.
Because of its relatively small enclosed volume and drywell floor area, the BWR Mark I containment structural boundary is particularly vulnerable to failure by overpressure or by direct contact attack should molten core and structural debris leave the reactor vessel.
Numerous analyses of the potential for early failure of the containment pressure boundary due to direct interaction with corium have been published. At Peach Bottom, drywell meltthrough is the most important mode of containment failure.
As indicated, there is a single doorway to direct any flow from the inpedestal region toward the opposite portion of the drywell wall. The distance from the point directly underneath the reactor vessel centerline to the closest point of the drywall wall is only about 7 m (22.85 ft.).
***Unit 1 could be smaller due to it being a very early BWR.
The carbon steel sump cover, which occupies a central rectangular section of the pedestal region floor. The existence of the sumps is a factor with respect to the potential for debris spreading since erosion of the thin sump cover would permit a significant fraction of the emergent debris to be retained in the sumps. The sump volume is about 5.7 m3 (200 ft.3) at both Browns Ferry and Peach Bottom.
Another pertinent plant-specific difference involves the entrances to the vent pipes, which lead to the pressure suppression pool the lower lip of the vent pipe shielded opening is located on the sloping drywell wall a short distance (about 0.61 m [2 ft.]) above the floor. At Peach Bottom, one of these vent pipes lies directly opposite the pedestal doorway.
Quick melt scenario:
The liquid zirconium oxide mixture then drains into the containment over a period of about five minutes. The same debris composition to be released into the containment over the next 150 minutes.
Slow melt scenario:
Debrisrelocating into the lower plenum is quenched by the water there. The remaining water is boiled away under the impetus of decay heating and the lower plenum steel structures are subsumed into the surrounding debris. Penetration failures occur at a time when only liquid metals are present as the debris temperature increases after lower plenum dryout. Oxide melting (and release) follows the initial release of metals.
For the Mark I liner failure study, the slow melt scenario probability density functions are such that the initial pour can be approximated as 14 m3 (494 ft.3) of a mixture of stainless steel with 30% zirconium metal at 100 K (180°F) superheat. This initial pour, which occurs over a period of 20 minutes, is followed by a relatively slow release of 15 m3 (530 ft.3) of oxides (mixed with 15% zirconium metal) over a period of 100 minutes.
The material super heat is critical to the containment wall (Debris liquids (metals) have a higher effective conductivity for heat transfer into the wall surface.). Upward radiation from the molten pool held in the core region (of the RPV) might cause melting of the upper reactor vessel internal structures so that the molten liquids, when released into the lower plenum, would include a large quantity of stainless steel.
The melted debris (and fuel) falling from the reactor vessel would first fill the drywell sumps, then would spill over the pedestal region floor. As sufficient height is accumulated over the floor, flow would begin through the pedestal doorway. Initial spreading as the flow enters the drywell region would be slight, but after contact with the drywell wall, the flow would separate into two branches, each flowing along the wall in a nearly one-dimensional fashion. These two branches then meet at a position diametrically opposite to the doorway. During the spreading process, the flowing debris radiates to the overlying atmosphere (or water) and transfers heat to and ablates the underlying concrete. Gases released from the concrete promote oxidation of the metals carried with the debris, and the associated energy release serves to increase the debris temperature.
MELTSPREAD and CORCON analyses support the idea that the depth of debris at the wall, the initial superheat, and the duration of superheat employed for the shell failure analysis are appropriately conservative. In other words, the values used in considerations of heat transfer to the shell are higher than those that would be produced in a best-estimate analysis.
Structural analysis of the Mark I shell in localized contact with debris, which is of composition (oxidic) and depth (20 cm [8 in.]) was assumed to be covered with water. Containment pressure was represented as remaining constant at 0.2 MPa (29 psia). Creep rupture was predicted to fail the shell at 1533 K (2300 ‘F), which is about 240 K (430 ‘F) lower than the carbon steel melting temperature. This temperature of 1533 K (2300 ‘F) was then adopted by the Mark I failure study for use as the best-estimate failure temperature for the drywell shell.
If a significant quantity of water overlies the drywell floor at the time of initial debris release, then all three of these parameters would be affected favorably, from the standpoint of promoting the survival of the shell. The depth of debris adjacent to the wall would be reduced, because more of the debris would freeze within the pedestal region, and less would reach the shell.
The effectiveness of overlying water in cooling crusted debris is not well understood, insulating crusts would develop only after the superheat is lost. Since heat loss mechanisms to water from a superheated corium melt are straightforward and the associated heat transfer is large the effectiveness of water in reducing superheat is not a matter of controversy.
The portion of the shell that is in contact with water above the debris acts as an efficient cooling fin. A cooling pathway is established from the debris into the shell and up through the fin to the overlying water. This accelerates the elimination of the debris superheat.
It is not yet possible to make definitive resolution as to whether or not it is worth risking the destructive potential of steam explosions (in the pedestal) in order to reap the beneficial aspects of water on the drywell floor.
The study concludes that there is a “virtual certainty” of shell failure if the containment floor is dry at the time of initial release, but that early shell failure is “physically unreasonable” if the drywell is flooded with water to the lower lip of the vent pipe openings, a depth of about 61 cm (2 ft.)
This article would not be possible without the extensive efforts of the SimplyInfo research team
Join the conversation at chat.simplyinfo.org
© 2011-2016 SimplyInfo.org, Fukuleaks.org All Rights Reserved Content cited, quoted etc. from other sources is under the respective rights of that content owner. If you are viewing this page on any website other than www.simplyinfo.org (or www.fukuleaks.org) it may be plagiarized, please let us know. If you wish to reproduce any of our content in full or in more than a phrase or quote, please contact us first to obtain permission.