GB2522256A

GB2522256A - Location of criticality

Info

Publication number: GB2522256A
Application number: GB1400932.8A
Authority: GB
Inventors: Malcolm John Joyce
Original assignee: Lancaster University Business Enterprises Ltd
Current assignee: Lancaster University Business Enterprises Ltd
Priority date: 2014-01-20
Filing date: 2014-01-20
Publication date: 2015-07-22
Also published as: GB201400932D0

Abstract

The present invention is a method of spatially locating a nuclear criticality 101 using a plurality of detection units 201, each comprising a gamma ray photon detector 203 and a neutron detector 205. The method uses data from each detecting unit regarding detected gamma photons and neutrons to estimate the distance from each detection unit to the location of the criticality, said estimated distance measurements being combined to determine the most likely spatial position of the criticality event. Preferably, a rapid rise in gamma photon activity starts a timing process at each detection unit 201, with the subsequent arrival time of a temporal peak of detected neutrons being used to estimate the distance from that unit to the criticality. Each neutron detector may additionally comprise a neutron spectrometer so that the energy spectrum of detected neutrons may be used in the distance measurements.

Description

LOCATION OF CRITICALITY

Criticality is perhaps the worst accident scenario associated with environments where fissile nuclear material is stored, processed and/or in use. Although such events are very rare, no facility dealing with such material can afford to be complacent about this possibility and all are required to monitor continuously.

Current safety philosophy is concerned principally with the reliable detection of criticality. It is one object of the present invention to obtain more information during the process of detection so that the consequences of a criticality event may be better handled.

Current criticality detection devices are often based on a measurement philosophy of two from three' responses in a triad of Geiger counters. This is simple and effective and incorporates an automatic level of redundancy. However, such a device yields only the information that it has detected the signature of a criticality event. It gives no information about the location of the identified criticality.

Criticality events are transient occurrences which are influenced strongly by the physical environment in which they occur. In general criticality events are more persistent in solid media. In liquid media the released heat may generate bubbles which temporarily cause a local dilution and halt the criticality. As a bubble moves away a new criticality may occur. Thus in a liquid medium several related criticality excursions may occur in succession (as is believed to be the case at Tokaimura in Japan on 30 September 1999).

After a criticality event (because of the associated high residual radiation levels) it can be very difficult to gain access to a facility to determine exactly what has happened, and to locate the origin of the event. This can result in significant delay before remedial action can be taken. During this time more radiation may be emitted, and accrued by personnel (even to the extent of fatalities) and may cause further damage to facilities.

Locating the source of a criticality is a significant challenge. The gamma-photon burst from a criticality is used as the trigger for existing alarms. Comparing arrival times for a single burst across a set of detectors in different locations might offer a potential means for triangulation on the basis of time-of-flight. However gamma photons travel at the speed of light (ie traversing one metre in about 3.3 nanoseconds) making this an extremely difficult timing challenge at present.

One objective of the present invention is to use detection of neutrons at a plurality of detection units as a means to estimate the spatial location of a criticality. Neutrons travel more slowly than photons, making the timing problem tractable.

A further objective of the present invention is to estimate the spatial location of a nuclear criticality sufficiently quickly to be of use in emergency situations.

The present invention comprises a method of locating spatially a nuclear criticality, using a plurality of detection units in known locations, where each detection unit comprises a first detector for detecting gamma photons and a closely located second detector for detecting neutrons. The detection units are each operably connected to a shared means of computation which in turn is operably connected to ancillary systems.

The method uses data from each detection unit regarding detected gamma photons and detected neutrons, and estimates the distance from each detection unit to the location of the criticality. The method combines the estimated distance from each of a plurality of detection units to estimate a spatial location for the criticality.

Each detection unit may additionally comprise pre-processing electronics.

The method may use a rapid rise in gamma photon activity to start a timing process at each detection unit.

The arrival time of the temporal peak of detected neutrons at each detection unit may be used to estimate the distance from that detection unit to the criticality.

The neutron detector may additionally comprise a neutron spectrometer so that the energy spectrum of detected neutrons at each detection unit may be used to estimate the distance from that detection unit to the criticality.

The means of computation may generate a human-readable representation of the location of the criticality.

The means of computation may provide the location of the criticality in machine-readable form to ancillary systems.

The means of computation may control automatic signage and/or audio.

The means of computation may send the location of the criticality to one or more remote locations.

The method may establish the distance to the location of the criticality from a respective detection unit within one second, one hundred milliseconds, ten milliseconds or one millisecond of the onset of the criticality.

Figure 1 shows a schematic representation of the present invention.

Figure 2 shows the Watt probability density function (prior art) for neutron energies.

Figure 3 shows transit time probability density for neutrons.

Figure 4 shows a means of estimating the location of a criticality according to one embodiment of the present invention.

The present invention comprises a plurality of detection units (201) each comprising a detector (203) for gamma photons, a closely located detector (205) for neutrons, and an optional computational pre-processor (not shown). It further comprises an operable data connection to a shared central computational processor (301) to which it sends radiation data, summarised radiation data or other data derived from radiation data. The processor (301) may send signals and/or data to other equipment (303).

The gamma photon detector (203) may comprise a very fast silicon diode. A suitable such diode is a silicon avalanche photodiode (for example Si APD Hamamatsu).

The neutron detector (205) may comprise a low-hazard, organic scintillation detector. A suitable such detector is a Scionix EJ3OG detector of type 127 A 76/3 M -EJ309 -X -Neg +VD20 -El -X-Neg.

Conventional 3He neutron detection systems cannot be used in the present invention because they need first to thermalise the neutron flux. This requires an additional physical process during which critical timing information would be lost.

A suitable analyser to process the outputs of the detectors is a Hybrid Instruments MFA1.2.

At the point of a criticality (101) there is an emission of radiation (121). After an initial gamma photon burst there follows neutron radiation. The neutrons have mass and so typically travel at a fraction of the speed of light. For example, a neutron travelling at 4% of the speed of light would travel across of a cell of dimension 10 metres in about 800 nanoseconds, and thus arrive about 770 nanoseconds after the respective gamma photons. Thus the gamma-photon event may be used as a temporal reference for the time taken for the neutrons to traverse the distance from the criticality (101) to the detection unit (201).

Embodiments of the present invention contain two or more detection units (101) as appropriate for local circumstances.

In a certain embodiment eight detection units (201) are placed, one in each corner of a cuboid storage cell for nuclear materials. Each detection unit (201) has a known three-dimensional position. Its role is to detect incoming radiation (121), to detect a criticality event (101), and to provide radiation information to allow a processing unit (301) to estimate the distance to the criticality (101).

The data processing system (301) may use data from a plurality of detection units (201) to estimate the most likely spatial position of the criticality event (101), and to communicate this information onwards to other systems (303).

Consider the time of flight of radiation (121) from the criticality event (101)to the detection unit (201). The gamma photons travel at the speed of light, so: s=ct (equationi) where s is the distance between the location of the criticality event (101) and the detection unit (201), t is the transit time for the gamma radiation over the said distance s and c is the speed of light.

The neutrons have mass and travel more slowly: s = v t (equation 2) where s is as above, and t is the transit time for a neutron with speed v over the said distance s.

The speed of a neutron is related to its energy: E q = 1⁄2 mv2 (equation 3) where E is the energy of a neutron (in electron volts, as is conventional in nuclear physics), q is the charge of an electron, m is the mass of a neutron and v is as before.

The time of criticality (101) is unknown, so it is not possible directly to discover t,,, ortrl, but it is possible directly to measure the time difference (delta): O = t -t (equation 4) Combining equations (1) to (4) with simple algebra yields equation 5 s = O ci ( (m c2 / 2 E q0)1/2 -1) (equation 5) Hence measuring O and E allows an estimate of the distance between the location of the criticality event (101) and the detection unit (201). All the other elements in equation are well known physical constants.

The neutrons emitted from a criticality event (101) are well known to follow an energy probability density function. This is the well known empirical Watt spectrum and is shown in Figure 2. In Figure 2, the x-axis represents the energy of a neutron in millions of electron volts. The y-axis represents a probability density P(E) such that the probability of the energy of a neutron lying in a small energy range dE is given by the product P(E) dE. The probability density is given by equation 6 where E is given in millions of electron volts (MeV): P(E) = 0.4865 sinh ((2 E)112) e" (equation 6) The most probable energy of a neutron given by equation 6 is 0.720 MeV. This means that if neutrons from a criticality are classified into equal width ranges of energy, then the range around 0.720 MeV is expected to show the highest count of neutrons.

Neutrons with higher energies travel faster (as per equation 3) than less energetic neutrons, and so the higher energy neutrons reach the detection unit (201) more quickly than less energetic neutrons.

It is possible to calculate the transit times of neutrons. This is done from equation 7 (derived from equations 2 and 3): = s ( m / ( 2 E q))1/2 (equation 7) The transit time in nanoseconds is given approximately from the neutron energy in millions of electron volts by equation 7A, which rearranges to give equation 7B: t = k s / E112 (equation 7A) = k2s2 / t2 (equation 7B) Where k = ( m I ( 2 q0))112 = 72.3 [when t is in ns, s is in metres and E in MeV] For the specific case where s = 1 Om we have: = 723 I [1I2 (equation 7C) E = 7232 / t2 (equation 7D) Neutrons with energy of 0.720 MeV take 852 nanoseconds to cover 10 metres.

Neutrons with energy of 10 MeV take 228 nanoseconds to cover 10 metres.

Gamma photons moving at the speed of light take 33 nanoseconds to cover 10 metres.

It will be clear that for longer distances the transit time is proportionally longer and vice versa.

It is possible to transform the Watt distribution (given in terms of neutron energy) into a derived temporal probability density (given in terms of transit time). Note that (as is well known in mathematical statistics) because of the change of variable, an additional differential term enters the equation (because dE and dt are not equal).

Pt(t) = d/dt ( E(t) ) . P(E(t)) (equation 8) The first term is given by differentiating equation 7B, and the second term by substituting equation 7B into equation 6, giving: Pt(t) = { 2s2k2/t2}. 0.4865 sinh (2ksIt) . exp -(s2k2It2) (equation 9) a This temporal probability density of equation 9 is shown in Figure 3. In Figure 3, the x-axis represents the transit time from criticality (101) to detection unit (201) of a neutron from the Watt distribution in nanoseconds. The y-axis represents a probability density P(t) such that the probability of the transit time of a neutron lying in the small time range dt is given by the product P(t).dt.

Each of the four plots shows the probability density of transit time over a different distance. From left to right these represent distances of 5, 10, 15 and 20 metres from criticality (101)to detection unit (201).

Looking at the plot for 10 metres separation, the most probable transit time across 10 metres is close to 440 nanoseconds. This means that for a momentary criticality event (101) which is 10 metres from a detection unit (201), when the neutrons are classified into equal width ranges of transit time, then the range around 440 nanoseconds is expected to show the highest count of neutrons.

Further calculation (differentiating equation 9 and setting the result equal to zero to find the maximum) shows a linear relation between peak arrival time and separation distance. The temporal peak travels at about 44 nanoseconds per metre. Subtracting the transit time of gamma-photons across the same distance, gives a figure close to 41 nanoseconds per metre of separation for the temporal neutron peak following the gamma trigger.

The analysis above yields two techniques for estimating the distance between criticality (101) and detection unit (201).

A first method requires detection of neutrons arriving in the period of about 2,000 nanoseconds immediately following a gamma trigger. Their arrival times (after the gamma trigger) are collected and classified into equal width time slots so that the time-slot of peak arrival may be found. Then dividing the peak arrival time by 41 nanoseconds per metre gives an estimate of the distance between the detection unit (201) and the criticality (101).

If there are sufficient neutrons detected to determine the peak time-slot, this first method can provide a distance estimate within a few micro-seconds of the criticality event.

At Tokaimura the maximum neutron emission rate from the criticality has been estimated in the literature to be about 1.4x1016 neutrons per second, and the average over the period of the accident to be about 7x1013 neutrons per second.

It is reasonable to assume an intrinsic efficiency of 1% for the neutron detector (205).

It is reasonable to assume that at a distance of 5 metres from the criticality event (101) a neutron detector (205) subtends 4x1 o4 of the flux (and lxi O of the flux at ten metres).

Assuming the average rate of neutron production above, at a distance of five metres each neutron detector experiences a neutron flux of about 2.8 x 1 010 neutrons per second, and about 1 8 neutrons per second detected. However the resolution of the Hybrid Instruments MFA1.2 analyser is currently limited to 106 events per second, so this is the limiting step in the analysis process.

About 100 events (or more) are needed for reliable detection according to this first method, so in principle the location of a criticality event may be detected in about 200 microseconds.

It is realistic to foresee improvements in the near future of electronics applicable to this analysis, which would reduce the processing time and allow a faster rate of neutron detection and faster detection of location of a criticality event.

A second method requires measurements of neutrons arriving in the period of a few seconds following a gamma trigger. The neutron energies and arrival times are collected. The energies are classified into equal width energy slots so that most probable energy may be found. This is assumed to correspond to the 0.720 MeV peak of the Watt distribution. The arrival time of the earliest arriving such neutron is then used with equation 5 to give an estimate of the distance between the detection unit (201) and the criticality event (101).

The second method requires a neutron spectrometry analyzer such as the Hybrid Instruments MFA1.2.

The following paragraphs apply to both detection methods above.

The principal source of experimental uncertainty in these methods is the resolution of the time of arrival. This depends on the resolution of the apparatus used. Using a Hybrid Instruments MEAl.2 analyser, the timing jitter is ±6 nanoseconds. This can be used to give upper and lower estimates of the distance. For example if the peak arrival time is 300 nanoseconds, the distance may be estimated as 7.31 ± 0.15 metres. This translates spatially as a spherical shell. There may be further uncertainty if the neutron distribution departs from the Watt model.

By using data from a plurality of detection units (201), software running in a processing unit (301) may co-ordinate multiple estimates of location and may generate an indication of the spatial location of the criticality (101). As is well known this indication may take a range of formats, for example a probabilistic heat map, determination of the co-ordinates of a most likely single location, or other representations. It will be evident to those skilled in the art how to develop a range of algorithms for this task.

As an example, Figure 4 shows how data from three detection units (201) may be used to propose three spherical shells, and then the region of overlap (251) indicates the most probable location of the criticality (101).

Having determined a most likely region and/or probable regions for the criticality (101), the computer (301) may send signals to ancillary equipment (303). Such signals may comprise pre-determined evacuation instructions (for example via intelligent signage, pie-recorded audio announcements, Sc) to aid escape of personnel. Probably the most significant benefit to arise from the present invention is the ability to assist emergency controllers to determine the most appropriate action to take, including automatic and/or semi-automatic selection and/or operation of appropriate escape routes. In criticality scenarios, radiation levels may be exceptionally high and significant injury and death may result very quickly. Therefore the efficacy and speed of emergency procedures is crucial.

Additionally, having been informed of a most likely location and/or probable locations of a criticality (101), personnel (for example in a control room) may be better informed as to measures to take to prevent repeat criticality and to restrict the spread of radioactive materials.

Equally, in due course, knowledge of the location of the (then past) criticality (101) may make permanent recovery of the facility easier and safer.

The processing unit (301) may automatically send data from the detection unit (201) and/or derived data (such as location estimations) to at least one remote location (303) to serve as backup in case it (301) eventually ceases to function for any reason following the criticality event (101).

All electronic devices and computing devices forming part of the present invention should be appropriately radiation-shielded as is well known.

Clearly it is not easily possible in practice to test an embodiment of the present invention by generating a nuclear criticality.

Testing of an embodiment may be accomplished by positioning a plurality of detection units (201) within a suitably shielded room (for example at the UK National Physical Laboratory), and then placing a suitable radioactive source (for example 241AmBe and/or 252Cf) in the room. Data from the detection units (201) may be processed as described above to estimate the spatial location of the source. The test may be repeated with the source at a number of different positions in order to assess the spatial resolution of the embodiment under test.

While the present invention has been described in generic terms, those skilled in the art will recognise that the present invention is not limited to the cases described, but can be practised with modification and alteration within the scope of the appended claims. The Description and Figures are thus to be regarded as illustrative instead of limiting.