Determination of countermeasures
This section discusses how you decide which countermeasures are appropriate during a nuclear emergency. This is a difficult area, full of hard decisions, worried people, a lot of uncertainties and, probably, very little time to make the right decision.Countermeasures should do more good than harm (NRPB, IAEA references) however this is a field fraught with difficultly. It is hard to assess the harm that could be caused by exposure to radiation and not easy to assess the harm that could be caused by imposing countermeasures.
Harm caused by radiation - there is no doubt that large doses of radiation over a short period of time are damaging to health. What is less clear is the harm that might be done by small doses of radiation. Two types of radiation harm are known. These are the effects that happen quickly during and after large exposures and include burning and other tissue damage. These have a threshold dose below which no effects are seen and above which the effect worsens with increasing dose. It is unlikely that this class of effects will be noticed in any members of the public due to a nuclear emergency at a licensed site.
The second type of radiation effect is the longer term damage leading to cancer sometime after the exposure. This time delay can be as long as several decades after the exposure. The likelihood of these effects at low doses is very difficult to assess. The assumption currently used internationally, but not with universal approval, is that the chances of an ill-effect such as a cancer decreases with decreasing dose but there is no dose below which the chances are zero. This is why scientists and government advisors are reluctant to say what a “safe” dose is.
Accepting the linear no-threshold hypothesis (that is the idea that every little dose brings with it a little chance of getting cancer you would not otherwise have got later in life) and attempting to put a value to this and comparing it to the costs of countermeasures is not at all easy but, in the UK, we have the NRPB ERLs as guidance ( Katmal | Avertable dose page).
Another problem is that now we have an agreement on the amounts of dose averted that make the three key countermeasures worthwhile we need to be able to assess the avertable dose to individual members of the public during a nuclear emergency. This requires a somewhat tortuous chain of argument.
Source term elucidation
It may seem reasonably straightforward to predict the consequences of a nuclear emergency in terms of the total quantity of radioactivity released (Total Release in Bq) and the isotopic composition (a table showing the amount of key nuclides released and their relative quantities). It is simply a matter of working out how many becquerels are in the box at the time of the accident and what fraction would be released in the course of the accident. However, for operating gas-cooled reactors such as the Magnox and AGRs used in the UK it is not nearly so simple.
Radioactivity Inventory
When nuclear fuel is first placed in a reactor it is hardly radioactive at all. When it comes out it is highly radioactive and heat
generating as the fission process results in a wide range of radioactive fission products (see figure). The levels of radioactivity
can be predicted by mathematical modelling provided the isotopic composition of the fuel and the history of its exposure to neutrons
are both known to adequate accuracy. This requires a detailed 3D model of the reactor, its power history and a large library of parameters.
Short lived fission product concentration is determined by the recent neutron flux and hence power history whereas the long lived fission
product concentration is related to the integrated neutron flux or the “burn-up” of the fuel. Since individual fuel elements are in different
parts of the reactor they experience different neutron fields and therefore different power histories and burnup rates.
For continuously refuelled gas cooled reactors, in particular, the core contains
fuel elements that have a wide range of stay times. Hence each fuel element in a reactor has a different radionuclide composition. Faults on gas
cooled reactors do not affect the whole core just one or a few channels and the operators may be uncertain about which channels are affected.
We are not dealing with a simple well mixed box with a stable inventory.
The reactors are designed not to release radioactivity in an uncontrolled manner. This design incorporates a number of barriers through which
the fission products (we’ll set aside activation products for now) have to escape if they are to be released. The important barriers are the
fuel matrix itself, the fuel element clad and the primary circuit.
The rate at which radioisotopes can escape through these barriers depends upon the chemical and physical form of the radioactivity, the
extent of any damage to the barrier, and factors such as temperature of the fuel element, coolant flow and coolant composition.
Physical Damage of fuel. With usage Magnox fuel can become porous in particularly in certain regions of the element. If physical damage is caused to this region then there is both a pathway for escape and increased tendency for the fuel to oxidise which further enhances loss of radionuclides. Rate of oxidation, and therefore of release, depends on fuel porosity (which depends on operating history), fuel temperature (which depends on many factors, including reactor cooling and local damage to the cooling circuits) and coolant composition (which depends on reactor damage). There are mechanisms, other than oxidation, that can lead to release of radioactivity from the fuel but generally at a much lower level.
Coolant flow and Composition. Both of these factors can be sub-optimal in the event of reactor damage. Inadequate cooling in all (gas circulation failure for example) or part of the reactor (channel blockage for example) can lead to temperature rises which can cause physical damage, melting of clad or fuel and/or fuel or clad ignition. The introduction of air or water into the CO2 coolant can significantly enhance the rate of fuel and clad damage. The operator will be attempting to maintain the quality of the coolant composition and the adequacy of cooling.
In-circuit behaviour. Radionuclides released from the fuel enter the coolant circuit. Their subsequent behaviour depends on their chemical and physical form, the composition and temperature of the coolant, the rate of circulation of the coolant and the nature and temperature of the surfaces it encounters as it circulates.
Release from circuit. If the reactor’s cooling circuit is damaged (breached) and there is a leakage of coolant then any radioactivity in that coolant will also be released. The operator will be attempting to prevent release by controlling the reactor pressure and by attempting to repair or isolate the breach. They may also be able to “blow down” the reactor via a controlled route which includes particulate and iodine filters to reduce the quantity of radionuclides released.
The rate of release from a reactor is seen to depend on a wide range of parameters associated with the reactor’s operating history, the nature and
extent of the damage and the operator’s actions to return the reactor to a safe state. All of these factors change both the quantity of radionuclides
released and the isotopic composition. Both of these can important impact on the potential for public dose. Therefore public dose cannot be predicted
to within orders of magnitude for any particular fault initiator due to the large number of variable involved.
Note. Different reactor designs such as AGR, PWR and BWR use different fuels, different fuel clads and a range of moderators and coolants at different temperatures and pressures so have different failure modes and pathways to atmosphere but the same principle of a number of barriers applies and the extent of penetration of any one barrier can depend on a number of factors making the actual release very hard to predict (Safety cases tend to be based on pessimistic assumptions at every stage to give a "conservative" estimate of the release - an overestimate).
Atmospheric Dispersion
It would generally be expected that any activity (radioactive dust or gas) that was released from a reactor building would move downwind with the same speed as the wind while being spread out by atmospheric dispersion.
There are many mathematical models of atmospheric dispersion. The most commonly used computer codes in the UK include:
R91 - a relatively simple gausssian dispersion model with correction factors for building effects.
ADMS A more modern dispersion code widely used in a number of industries
NAME A sophisticated long-range dispersion code
The UK Atmospheric Dispersion Modelling Liaison Committee maintains a watching brief on atmospheric dispersion and commissions key research in the area.
In the emergency response phase there is an effort to model where the radioactive material being discharged from the damaged facility is going. The utilities tend to use R91 style computer codes to model the dispersion relatively near to the site (out to a few kilometres) and the Met. Office use NAME to model the dispersion further afield. This is playing to the strengths of the different models. R91 is simple to understand and needs very little data to drive it as it is a straight line gaussian model. NAME is a more complex code using wind field data from the Met. Office's weather forecasting capability and is better suited to long range dispersion.
A comparison of these to models is given in a HPA report. This identifies that "there is a disparity (of up to a factor of approximately three) between the time integrated air concentrations in air derived using NAME and those derived using R91, most notably in the near field". It identifies R91 as being the more conservative model (it gives the higher values) and identifies the different dispersion parameters as the root cause of the differences.
Off-site Monitoring
The organisations providing countermeasure advice to the protect the public do not rely solely on estimates of the potential source terms from saftey cases and estimates of the dispersion of activity. Instead off-site monitoring is used to identify exactly where the radioactivty has got to and how much is there. There are a number of systems in use. These include:
Perimeter Monitoring Systems which involve a series of detectors placed around the site to detect and quantify any radioactivity crossing the site boundary. These may just be gieger counters, or similar, measuring dose rate in air or may have more sopisticated systems measuring particulate activity or identifying the isotropic composition of the release.
Mobile monitoring (off-site monitoring) These involve vehicles that drive out into the plume to preagreed points or as directed by a controller and take a series of measurements that might include: dose rate in air, particulate activity in air or spectroscopic information.
RIMNET This is a system of fixed monitors distributed across the UK orginally set up after the Chernobyl accident to monitor radiation plumes from abroad but now being expanded to have more detectors around UK nuclear facilities.
The value of off-site monitoring is it provides a more reliable picture of where the radioactvity is getting to than modelling could achieve on its own.
More details about off-site monitoring can be found elsewhere on this web-site
Public Dose
The mathematical modellers supporting the decision makers attempt to combine all of the information coming from the site about the current release rate, how it will change, and how long it will continue for with real data from off-site monitoring and models of the dispersion based on the day's weather to predict the distribution of dose to the public. To do this they use well established models to estimate the dose rate to a person standing near or in the plume of radioactive material and include cloud dose (the dose you get being near a radioactive cloud), inhalation dose (the dose you get from breathing in radioactive dust) and ground gamma dose (the dose you get from radioactivity that has settled on the ground and other nearby surfaces).
The decision process
Countermeasure advice to protect the public would be discussed and agreed in the Strategic Coordination Centre (SCC)
