METHOD FOR DETECTING URANIUM CONTAMINATION AND ESTIMATING 235U / 238U ENRICHMENT BY LOW-RESOLUTION GAMMA SPECTROMETRY

Abstract

The invention relates to a method for detecting artificial uranium contamination (8) in an object containing natural uranium, comprising: a) - positioning a scintillator (4) on the object to be examined, b) - obtaining at least one gamma radiation spectrum emitted by the surface to be examined, c) - selecting a first spectral region and a second spectral region of said gamma radiation spectrum, d) - measuring a first net count rate in the first spectral region and a second net count rate in the second spectral region, e) - estimating the artificial 235U enrichment based on the first and second net count rates. Figure 1A

Description

Title of the invention: METHOD FOR DETECTING URANIUM CONTAMINATION AND ESTIMATING THEREOF ENRICHMENT 23S U / 238 U BY LOW RESOLUTION GAMMA SPECTROMETRY

[0001] TECHNICAL FIELD AND STATE OF THE PRIOR ART

[0002] The invention lies in the field of rapid detection of artificial uranium contamination in an object, for example containing natural uranium in secular equilibrium; this may be the case, for example, of a stain of enriched uranium contamination on a concrete wall. By contamination stain, we mean a surface contamination that may range from a point contamination (hot spot of a few cm2) to a more extensive surface contamination (several m2). Such detection may, for example, be carried out during the dismantling and cleanup of a nuclear installation.

[0003] The invention is applicable, for example, to any uranium processing installation during dismantling, where residual uranium contamination could be present.

[0004] For example, uranium enrichment can achieve different levels, ranging from low enrichment (less than 1% by mass of 235U) to very high enrichment (more than 90% by mass of 235U), intended for different applications. However, the process of enriching natural uranium modifies the proportions of 238U, 235U and 234U, which implies variations in their activity percentages depending on the enrichment level.

[0005] These activities induce not only the emissions a specific to these uranium isotopes but also the emissions y and [3 of their descendants, from the filiation chains, as shown in the figure published in the article http s : / / fr. wikipedia.org / wiki / Uranium_238.

[0006] If we now consider a contamination by more or less enriched artificial uranium present on or in an object containing natural uranium, we understand that the different contributions (y, a and |3) of this contamination are superimposed on an ambient background noise (BDF) produced by the filiation chains of uranium, thorium and potassium, naturally present, for example, in the concrete of the walls and floors of the installation. We seek to characterize this background noise and to take it into account in the measurements of the net signal of the uranium contamination stain. Among the objects concerned, we can cite for example a conveyor belt, or samples or bags of rubble ("big bag") after stripping of a contaminated wall, or even packages of uranium-bearing radioactive waste (drums, boxes, etc.).

[0007] For example, in the context of the detection and characterization of uranium contamination spots, the invention makes it possible to meet three main objectives:

[0008] - verification of the decommissioning thresholds imposed by the safety authorities nuclear, which are for example of the order of Bq / cm2,

[0009] - rapid control of surfaces; for example, the total surface to be controlled can reach several hundreds of thousands of m2, which requires the implementation of rapid measurement techniques in order to control the entire surface area in a reasonable time;

[0010] - the estimation of the level of enrichment in 235U of the contamination spot, which provides additional useful information, for example, for the decontamination, characterization and management of the waste it generates, such as directing it to an appropriate storage channel.

[0011] Different measurement techniques are known for contamination control, ranging from alpha and beta counting to high and low resolution gamma spectrometry.

[0012] A first detection technique can be carried out by alpha counting, for example using ZnS(Ag) scintillator type contaminameters.

[0013] The advantages of this type of technique are as follows:

[0014] - the number of alpha particles emitted by uranium contamination is generated generally larger than the number of beta particles and even larger than the number of gamma particles. Therefore, the measurement times required for the detection of contamination by alpha counting are very short;

[0015] - the background noise associated with a count of alpha particles emerging from the concrete is low, because only surface emissions can reach the detector due to the short path of alpha particles in concrete, about 20 pm; this facilitates the identification of surface uranium contamination; according to an experimental estimate with a surface alpha source placed on uncontaminated soil, it contributes less than 1% of the alpha count of a contamination spot of a few Bq / cm2;

[0016] - it is possible to produce large surface detectors (for example 1000 cm2) at very low cost.

[0017] But this technique has drawbacks:

[0018] - alpha particles are easily stopped by matter. On average the path alpha particles are about 3 cm in air and 20 pm in concrete for an energy of 4.5 MeV (SRIM calculation). The detector must therefore be as close as possible to the contaminated surface to detect these particles;

[0019] - this method depends heavily on the surface condition of the object. For example, in the In the context of detecting contamination on the floor of an installation, the presence of possible dust and roughness in the floor can significantly reduce these radiation.

[0020] The technique of detection by beta counting is also known, for example using PVT plastic scintillators.

[0021] The advantages of this type of technique are as follows:

[0022] - the path of beta particles in matter is greater than that of alpha particles, for example about 1.5 mm in concrete for beta particles emitted by 234mPa, whose maximum energy is high: Epmax = 2.269 MeV. This method will therefore be more effective than alpha measurement if the contamination is deep, up to a few millimeters;

[0023] - depending on the 235U enrichment of the contamination and the number of beta particles emitted per second, beta measurement can be more effective than gamma measurement.

[0024] But this technique also has disadvantages: - in particular, the number of betas emitted for contamination with highly enriched uranium decreases sharply. Consequently, the measurement times required to differentiate the contamination signal from the background noise become very long, it can reach several hours to detect contamination levels of the order of Bq / cm2; - in addition, the lack of sensitivity beyond depths of the order of mm poses a problem for detecting deeper contamination or for analyzing large objects.

[0025] In view of the advantages and disadvantages of each of the techniques, it appears for example that alpha counting is the fastest method among those cited for the detection of surface contamination over a range of enrichments from less than 1% to more than 90%. However, the effectiveness of this method depends strongly on the surface condition of the object to be characterized, for example when ground contamination is covered by a layer of dust. As for beta counting, the constraint linked to the cleanliness of the ground is less of a concern but the considerable drop in beta emission rates in the case of significant 235U enrichments prohibitively lengthens the measurement times. Furthermore, these two methods do not make it possible to trace the 235U enrichment of the uranium present within the contamination.

[0026] Gamma spectrometry detection techniques can be implemented because they are much less sensitive to the depth of contamination or to the cleanliness and surface condition of the object being examined, since the gamma and X-rays of interest can pass through several cm of concrete;

[0027] High-resolution gamma spectrometry with a hyperpure germanium (HPGe) detector allows a fine analysis of the different lines of the gamma spectrum and to measure the 235U / 238U enrichment by intensity ratios (or surface or area, preferably net) of certain lines.

[0028] On the other hand:

[0029] - GeHP detectors are very expensive (about a factor of 10 on the price) in volume (and therefore efficiency) of similar detection compared to Nal(Tl) type scintillation detectors;

[0030] - the measurement times required to obtain a gamma spectrum with a sta uncertainty sufficiently low data on the lines of interest, allowing for example to detect contamination, to determine its isotopic composition, or even its depth profile, are too important to consider measuring very large potentially contaminated surfaces without multiplying the number of GeHP detectors, which would represent a prohibitive cost;

[0031] - this type of detector requires cooling at very low temperature with nitrogen liquid which complicates its exploitation in situ, outside the laboratory) or with an electrical system (cryogenerator), which increases the cost.

[0032] The technique of detection by low-resolution gamma spectrometry with Nal(Tl) type scintillators is also known.

[0033] The advantages of this type of technique are as follows:

[0034] - the required measurement times are relatively short (a few minutes) for detect contamination in the order of Bq / cm2. Large NaI(Tl) scintillators (e.g. 1000 cm2 surface area and 5 cm thickness) can be produced at a reasonable price, unlike higher resolution scintillators such as those based on LaBr3(Ce) or CeBr3 crystals, or the high resolution GeHP semiconductor detectors mentioned above.

[0035] The disadvantages of Nal(Tl) scintillators are their poor energy resolution, which does not allow the fine characterization of contamination (for example its depth profile or the enrichment in 235U) by the analysis of gamma or X lines classically carried out in gamma spectrometry (identification of the emitting radioelements thanks to the energies of the lines and their activities thanks to the net areas of the lines).

[0036] It is however possible to characterize uranium by energy bands with a Nal(Tl) detector as described in documents FR3104739 and WO2021122772. In these documents, the objective of the energy band analysis is the estimation of the natural uranium content in uranium ore samples or in a borehole, i.e. with a distribution of the natural uranium to be measured throughout the volume interrogated.

[0037] The problem posed by the present invention is, on the contrary, to detect and determine the enrichment of a low contamination in artificial uranium located on or in (in depth) a thick volume, for example a volume of concrete, containing natural uranium inducing a significant background noise to be subtracted.

[0038] Furthermore, the problem in documents FR3104739 and WO2021122772 was to distinguish the gamma signal of uranium from that of its descendants in its natural radioactive chains to take into account a possible uranium / radon imbalance in the chains of descent. The technique described in these documents does not allow the detection of artificial uranium contamination in an object containing natural uranium which generates background noise.

[0039] The problem therefore arises here of finding a new method and a new device for measuring or evaluating residual contamination by artificial uranium, generally of low activity (with for example a decommissioning threshold of the order of Bq.cm2), in an object containing natural uranium, for example in secular equilibrium, and other natural radioelements such as 232Th in secular equilibrium with its descendants and 40K, for example in a concrete wall of a uranium enrichment facility. This natural radioactivity creates a gamma background noise to be subtracted in order to detect and characterize the contamination by artificial uranium. For this type of measurement, it is sought to implement a rapid technique, to allow for example the analysis of all the surfaces of an installation in a reasonable period of time with a view to their decommissioning. Statement of the invention

[0040] A method according to the invention implements a measurement of the gamma emissions of uranium contamination in an object containing natural uranium, or natural uranium in secular equilibrium, by low-resolution gamma spectrometry, for example a detector of the NaI(Tl) type, preferably of large volume, for example between 100 cm3 (or 0.11) and 1 1, and / or with a large detection surface, for example adapted, or corresponding, to the surface of the contamination spot, for example between 100 cm2 and 1600 cm2 (the width and length of the detector may be between 10 cm and 40 cm, for example; its depth is between 5 cm and 15 cm, for example).

[0041] The invention makes it possible to estimate the presence, or not, of uranium contamination in an object as well as the enrichment in 235U present within it.

[0042] The invention firstly relates to a method for detecting artificial uranium contamination of an object containing natural uranium, this method comprising:

[0043] a) - the positioning of at least one scintillator on the object to be examined,

[0044] b) - the production of at least one gamma radiation spectrum emitted by the object to be examined,

[0045] c) - the selection of a spectral region of said spectrum, this spectral region preferably being at low energy, for example between 50 keV and 220 keV, of said gamma radiation spectrum,

[0046] d) - the measurement of the counting rate, preferably net, (without background noise, in said spectral region on the gamma spectrum emitted by the contaminated object.

[0047] The invention also relates to a method for detecting artificial uranium contamination of an object containing natural uranium, and / or for estimating the enrichment in artificial 235U of said object, comprising:

[0048] a) - the positioning of a scintillator, for example a Nal detector, on or against the object to be examined,

[0049] b) - the production of at least one gamma radiation spectrum emitted by the object to be examined,

[0050] c) - the selection of a 1st spectral region and a 2nd spectral region of said gamma radiation spectrum,

[0051] d) - the detection of a 1st net count rate, without background noise, in the 1st spectral region and of a 2nd net count rate, without background noise, in the 2nd spectral region,

[0052] e) - the estimation of the enrichment in artificial 235U on the basis of the 1st net counting rate and the 2nd net counting rate, for example on the basis of the ratio of these two counting rates.

[0053] The invention therefore proposes a gamma spectrometry method with a scintillator, for example of the NaI(Tl) type, preferably a large scintillator, to carry out rapid detection of contamination, less sensitive to the surface condition of the object and to the depth of contamination than alpha or beta detection, the radiation of which can be masked by small layers of dust, paint, cement, or other, which would cover the surfaces examined.

[0054] To identify contaminated areas, the invention uses low energy resolution gamma spectrometry with a scintillator (or scintillation detector), for example with a NaI(Tl) crystal, rather than high resolution spectrometry with a Ge semiconductor detector, which is much smaller and more expensive per unit volume. In addition, gamma spectrometry can provide information on uranium enrichment, information that alpha or beta detection cannot provide.

[0055] Low-resolution gamma spectrometry measurement solves environmental attenuation problems. Since gamma radiation is poorly attenuated by, for example, concrete, it can be detected even if the contamination is a few millimeters or even a few centimeters deep.

[0056] Preferably, the 1st spectral region contains at least one line, for example at 93 keV, and the 2nd spectral region contains at least one other line, for example at 185 keV.

[0057] The spectral regions implemented within the framework of the invention can be located at low energy (for example between 50 and 220 keV, for the detection and the estimation of 235U enrichment from uranium contamination. For enrichment, the estimation can be based on a ratio of counts in two energy sub-bands, included in the low energy band, here between 50 and 220 keV.

[0058] For example, the 1st spectral region is located between 50 keV and 130 keV and the 2nd spectral region is located between 165 keV and 215 keV.

[0059] Whatever the embodiment of the invention, the net count rate can be obtained by difference between the gross count rate and the count rate linked to the background noise, for example obtained by measurement or estimation of the latter, for example again by the production of at least one gamma radiation spectrum emitted by an equivalent uncontaminated object.

[0060] According to one embodiment of the invention, it is therefore possible to estimate the background noise of the measurement, in at least the 1st spectral region and the 2nd spectral region, the net counting rates being calculated by taking this background noise into account.

[0061] According to one aspect, a method according to the invention therefore makes it possible to distinguish between, on the one hand, the gamma signal of the object (for example a concrete object) or of its surface and, on the other hand, the natural background noise of the object or of its surface, on or into which the contamination is deposited or introduced.

[0062] According to a particular embodiment, the background noise can be estimated from several gamma spectra acquired on uncontaminated samples or objects, an estimate being for example obtained by average or by regression, for example with automatic learning methods such as multilinear regression by the least squares method.

[0063] According to yet another particular embodiment, the background noise, in at least the 1st spectral region and the 2nd spectral region, can be estimated from measurements in one or more regions at higher energy than the 1st spectral region and the 2nd spectral region, for example for which the contribution of a possible uranium contamination is zero or negligible (less than 1%).

[0064] For example, if the 1st spectral region contains at least one line at 93 keV and the 2nd spectral region contains at least one line at 185 keV, or if the spectral regions implemented within the framework of the invention are located at low energy, for example still between 50 and 220 keV, the background noise, in at least the 1st spectral region and the 2nd spectral region, can be estimated from measurements in a spectral region between 550 keV and 660 keV and in a spectral region between 1300 keV and 1600 keV.

[0065] An estimation of the mass percentage of 235U / 238U enrichment can be made from the enrichment indicator data, for example using a calibration curve.

[0066] A method according to the invention may comprise a prior step of detection by alpha and / or beta counting, and / or an additional step of high-resolution gamma detection.

[0067] Advantageously, a method according to the invention may comprise, for low or no contamination, a step of storing at least part of the spectrum data. This data may then be reused to produce background noise data.

[0068] The invention also relates to a device for detecting artificial uranium contamination of an object containing natural uranium, this device comprising:

[0069] a) - a scintillator;

[0070] b) - means for producing at least one gamma radiation spectrum emitted by the surface of the object or by the object to be examined,

[0071] c) - means for selecting a spectral region of said spectrum, this spectral region preferably being at low energy, for example between 50 keV and 220 keV, of said gamma radiation spectrum,

[0072] d) - means, adapted or specially programmed, for measuring the counting rate, preferably net (without background noise) in said spectral region on the gamma spectrum emitted by the contaminated object.

[0073] The invention also relates to a device for detecting uranium contamination of an object containing natural uranium, and / or for estimating the enrichment in artificial 235U of said object, comprising:

[0074] a) - a scintillator;

[0075] b) - means for producing at least one gamma radiation spectrum emitted by the surface of the object or by the object to be examined,

[0076] c) - means for selecting a 1st spectral region and a 2nd spectral region of said gamma radiation spectrum,

[0077] d) - means adapted or specially programmed to carry out the measurement of a 1st net counting rate, without background noise, in the 1st spectral region and of a 2nd net counting rate, without background noise, in the 2nd spectral region,

[0078] e) - means, adapted or specially programmed, for estimating a 235U enrichment indicator on the basis of the 1st counting rate and the 2nd counting rate.

[0079] The invention also relates to a method implementing such a device.

[0080] In a method or device according to the invention, the enrichment indicator can be obtained by calculating the ratio between the 1st net count rate and the 2nd net count rate. The net count rates are obtained after subtracting the background noise in each of the spectral regions concerned.

[0081] In a method or device according to the invention:

[0082] - the scintillator is preferably a Nal detector;

[0083] - and / or the 2 spectral regions (Ab A2) have different contributions from 235U and of the 238U;

[0084] - and / or the 1st spectral region contains at least one line at 93 keV and the 2nd region spectral line contains at least one line at 185 keV.

[0085] - and / or the 1st spectral region is located between 50 keV and 150 keV and the 2nd region spectral is located between 150 keV and 250 keV;

[0086] - and / or the enrichment indicator can be obtained by calculating the ratio between the 1st net count rate and the 2nd net count rate.

[0087] A device according to the invention may further comprise:

[0088] - means for carrying out an estimation of the background noise of the measurement, in at least minus the 1st spectral region and the 2nd spectral region, and means for correcting the count rates based on this estimate of the background noise;

[0089] - and / or means for storing data from gamma spectra of non-surfaces contaminated.

[0090] A device according to the invention may further comprise:

[0091] - an alpha and / or beta detector and means for implementing a step prior detection by alpha and / or beta counting;

[0092] - and / or a high-resolution gamma detector.

[0093] The technique implemented according to the invention does not primarily use high-resolution detectors, such as a germanium detector, which require a permanent cooling system; in addition, the detection surface of this type of device is limited by the size of the crystal, and in particular of the Ge crystal in the case of a germanium detector. The invention makes it possible to overcome this problem at a lower cost since scintillation detectors can be produced having a large surface area with a ratio approximately 10 times greater between efficiency and price, while being able to use them at room temperature (no cooling). Brief description of the drawings

[0094] [Fig. 1A] represents an example of the implementation of a detection of uranium contamination with a Nal detector;

[0095] [Fig. IB] represents an exemplary embodiment of a measuring chain which can be implemented within the framework of the present invention;

[0096] [Fig.2] represents gamma spectra obtained by MCNP simulation for quasi-surface uranium contamination with enrichments of 1 to 80% and for BDF due to natural gamma emissions in the volume of concrete, with a large volume Nal(Tl) detector;

[0097] [Fig.3] represents the identification, in a gamma spectrum, of the Al energy bands (here from 50 to 130 keV) and A2 (from 165 to 215 keV);

[0098] [Fig.4] represents the evolution of an aENR indicator after correction of the background noise and as a function of the mass percentage of enrichment 235u / 238pj •

[0099] [Fig.5] represents the evolution of an aENR indicator without subtraction of the background noise and as a function of the mass percentage of enrichment in 235U;

[0100] [Fig.6] represents spectra corresponding to background noise measurements carried out on an installation with a Nal detector;

[0101] [Fig.7] represents a comparison of the points calculated by the regression model and the experimental points;

[0102] [Fig.8] represents steps of a method according to the invention, for detecting uranium contamination by coupling alpha, beta and gamma measurement methods;

[0103] [Fig.9] represents steps of a method according to the invention, for detecting uranium contamination and estimating its 235U / 238U enrichment.

[0104] [Fig.lOA] and [Fig.lOB] schematically represent variants of a method according to the invention.

[0105] DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS

[0106] An example of implementation of the invention is given in [Fig. 1A].

[0107] This figure concerns a detection on a surface of a concrete block 2, but it could also be carried out on or in any other contaminated object, such as floors and / or walls of a building. Block 2 is for example of dimensions 100 x 100 x 30 cm3, for a density of 2.35. Reference 8 designates a spot of contamination by artificial uranium, for example of thickness 0.1 cm for a surface of 32 x 32 cm2. Block 2, more generally the object on or in which one wishes to make a detection, and preferably also an estimation of the enrichment, also contains natural uranium, or natural uranium in secular radioactive equilibrium.

[0108] The ASN methodological guide on the “Management of sites potentially polluted by radioactive substances”

[0109] https: / / www.asn.fr / l-asn-reglemente / guides-de-l-asn / guide-methodologique-gestion-d es-sites-potentially-polluted-by-radioactive-substances

[0110] indicates that, "when the period of the father is very clearly greater than that of the sons, the activities of the different sons come into equilibrium with that of the father. This equilibrium, called secular, is obtained after a time equal to approximately 10 times the period of the son whose period is the longest: 234U (T=2.46.105 years) for the 238U chain (T=4.47.109 years) and 228Ra (T=5.75 years) for the 232Th chain (T=1.40.1010 years). Considering that all the activities are then equal, the error made is less than 1%0".

[0111] The input face of the detector 4 is therefore deposited against this task. More gene Usually it is placed against the surface of the object on or in which it is desired make a detection or an estimate, as explained above.

[0112] This detector 4 is a scintillator, for example a NaI detector. In the context of the present invention, such a low-resolution gamma detector is used.

[0113] It can be of a relatively large volume, for example here it has a width L of 32 cm and a depth of 32 cm (both measured parallel to the surface against which it is placed) for a height h of 7.62 cm. The height of the detector, for example between 5 cm and 10 cm, is linked to the targeted efficiency for the gamma radiation to be detected, and to the limitation of the Compton scattering phenomenon responsible for the background noise continuum observed on the gamma spectra. More generally, the width and length of the detector can be for example between 10 cm and 40 cm, its depth for example between 5 cm and 15 cm, its overall volume for example between a few hundred cm3 and several liters, for example between on the one hand 100 cm3 (or 0.1 1) or 200 cm3 (or 0.2 1) and on the other hand 5 1 or 10 1.

[0114] Preferably, the detector is equipped with a shield 6, for example a lead shield, for example still 5 cm thick. This shielding makes it possible to avoid parasitic gamma radiation. It can be noted that the concrete block 2 emits background noise.

[0115] [Fig. 1B] therefore represents a detector 4 and the measuring chain 10 with which it is associated. This chain is composed of a HT / MCA base 12 (high voltage and multi-channel analyzer, or MCA for Multi-Channel Analyzer) whose role is to power the detector and reconstruct the spectrum from the electrical signals produced by the detected gamma radiation, as well as an acquisition computer or microcomputer 14 programmed to record and analyze the gamma spectra.

[0116] The measured data may be stored in the memory of the computer or microcomputer 14 for processing purposes according to the present invention. This computer or microcomputer is therefore preferably adapted or programmed to implement one or more of the steps described in the context of the present invention.

[0117] In the examples presented, the gamma emitters of interest for the detection of contamination are 238U, 234Th, 234mPa, 235U and 231Th with emission energies essentially included in a low energy region A, for example between 50 keV and 220 keV.

[0118] The table below shows the main gamma emissions of natural or enriched uranium isotopes (238U, 235U, 234U), and their descendants, after 40 years of evolution (data taken from the LARA database, see for example http: / / www.nucleide.org / Laraweb / index.php). The uncertainty on the last figure is given in parentheses.

[0119] [Tables 1] Isotope Gamma emitters Energies Y (keV) Intensities Y (%) 238£J 23 8^ 49.55 (6) 0.0697 (26) 234^ 63.30 (2) 92.38 (1) 92.80 (2) 3.75 (8) 2.18(19) 2.15(19) 234mp^ 766.361 (20) 1001.02 6(18) 0.323 (4) 0.847 (8) 235JJ 235U 185.720 (4) 143.767 (3) 163.356 (3) 205.316 (4) 57.0 (3) 10.94 (6) 5.08 (3) 5.02 (3) 23lTh 25.64 (2) 84.2140 (13) 13.9 (7) 6.70 (7) 234JJ 234^ 53.20 (2) 0.1253 (40)

[0120] As can be seen from this table, the counting rates in the low energy region A, for example between 50 and 220 keV (the counting rate is then noted CBrut L5o 22oj) are higher when there is a contamination of significant percentage of 238U activity; therefore, when the background noise signal (BDF) is taken into account, the difference between this BDF signal and the contamination signal will be clearer. In order to verify this difference, several MCNP simulations (Monte Carlo N-Particle Transport Code, Los Alamos National Laboratory, “Monte Carlo N-Particle (MCNP) Transport Code”) were carried out according to the geometry of [Fig.lA]. The proportions of 238U, 235U and 234U activity of the contamination spot correspond to different enrichments from 1 to 80% in 235U, and they vary respectively from 40 to 0.1% for 238U, from 2.6% to 3.2% for 235U and 57.4% to 96.7% for 234U.

[0121] The simulation results in the volume of detector 4 make it possible to reconstruct (see [Fig.2]) the different gamma spectra detected, due to the emissions of the BDF in the volume of concrete 2 and due to those of the surface contamination 8. In this [Fig.2], region A, in this example ranging from 50 to 220 keV, is the one which contains the most information relating to the uranium contamination, mainly by the lines at 63 and 92 keV of 234Th as well as the line at 185 keV of 235U.

[0122] The different curves in [Fig.2], in addition to the one corresponding to the background noise (BDF), correspond to various enrichments from 1 to 80%. Overall, these results highlight that the gamma spectra corresponding to the enrichments from 1 to 80% display, in the low-energy region A from 50 to 220 keV, higher counting rates than those of the background noise. The difference between the different contamination signals and the background noise signal of pure concrete therefore makes it possible to detect and quantify the contamination, by means of an estimation of the background noise and its subtraction as explained below.

[0123] These results show that contamination can be detected by establishing a spectrum of gamma radiation emitted by the surface, more generally by the object, to be examined.

[0124] The measurement times estimated by simulation to obtain a spectrum with a sufficiently low statistical uncertainty to detect the contamination and estimate its enrichment are, for example, of the order of a minute in low-resolution gamma spectrometry with the parallelepiped Nal(Tl) detector described above (volume of 32 x 32 x 7.62 cm3), for a low enrichment of 1% and a contamination level of the order of Bq / cm2.

[0125] Once the contamination has been detected, the spectral region A of interest can be separated into two sub-bands Ai and A2:

[0126] - the first (AJ) can for example be located around at least one of the 2 lines at 63 keV and 92 keV (preferably around 2 to have more signal), region associated with 238U with a contribution from 235U by the internal conversion X-ray emission of 231Th at 93 keV; here, this sub-band Ai is located between approximately 50 keV and 130 keV;

[0127] - the second (A2) is for example located around the 185 keV line, emitted by the 235 U; here, this A2 sub-band is located between approximately 165 keV and 215 keV;.

[0128] According to the invention, it is therefore possible to select 2 energy ranges in which the contributions of 235U and 238U are different, the numerical values ​​of the terminals of these ranges being able to be the subject of different choices.

[0129] In each of these 2 bands, we obtain a raw counting rate CBRUTAi and CBRUTA2. By subtracting the background noise associated with each of the sub-bands, CBDFAi and CBDFA2pn can establish an indicator, noted “ExR”, allowing the enrichment (mass ratio) of the contamination to be estimated:

[0130] [Math.l] ^RVT „BDF _ ^1565-2151 _ 165-2!5f ^,165-215] ^ENR “ ^NET ~ ^BRUT ^BDF *-

[50130] *-[50130r*-[50L30]

[0131] CnftAi and CNFTA2 being the respective net counting rates in each of the sub-bands Ai and A2, after subtraction of the background noise due to pure concrete.

[0132] Based on the numerical simulation, it is possible to establish the relationship between this aENR indicator and the actual enrichment as represented in [Fig.4] (for enrichments ranging from 1% to 85%).

[0133] It can be seen from this figure that, when the background noise is subtracted from the spectrum before calculating the aENR indicator, the latter has good sensitivity over the entire enrichment range and without ambiguity in the estimation of the enrichment for a given aENR value (bijective relationship).

[0134] When the background noise is not subtracted from the energy regions for the calculation of the aENR indicator, the graph in [Fig.5] is obtained, which represents the evolution of a ENR as a function of the enrichment. This figure shows that, without subtraction of the background noise, the indicator is not suitable for estimating the enrichment because its sensitivity is very low, with a maximum difference of approximately 0.035 between the minimum and maximum values ​​of aENR which varies between approximately 0.45 and 0.49 (i.e. less than 10% relative variation while the enrichment varies over a range of 1% to 85%). In addition, certain values ​​of the aENR indicator can give several enrichment values ​​(non-bijective relationship beyond approximately 40% enrichment); the background noise is therefore subtracted before calculating the indicator.

[0135] For example, the estimation of the 235u / 238pj enrichment by reading the Al and A2 energy sub-bands makes it possible to differentiate between a low enriched contamination (0 to 20%) and a highly enriched contamination (more than 20%).

[0136] As explained above, a characterization of the background noise emitted by the material, for example concrete, of the inspected installation allows a better estimation of the net counting rates due to contamination in the energy bands. Since the background noise can vary depending on the U, K and Th contents in various materials, for example the constituent materials of concrete (cement, sand, aggregates), and the radon concentration in the ambient air (the latter is dependent on weather conditions), two estimation methods are possible, described below.

[0137] According to a first method, it is possible to estimate an average of the low-energy background noise count rates from several noise gamma spectra background acquired on an uncontaminated object, for example one or more certified uncontaminated areas, but representative of the object or area to be controlled (i.e. on the same installation), because there may be fluctuations in the uranium, thorium and potassium levels, particularly in concrete.

[0138] As an illustration, the acquisition of 47 background noise spectra of the concrete of an installation made it possible to estimate an average value of the counting rates of the BDF at low energy as presented in [Fig.6]. At the top of this figure are indicated the number of each measurement and the corresponding counting rate. From these different spectra, an average of 12.9 c / s can be deduced with an uncertainty (standard deviation) of 1.0 c / s, i.e. a relative uncertainty of 8%. In general, this type of preliminary study can therefore be carried out on any site to be controlled.

[0139] According to a second method, the different BDF spectra are used, for example those of [Fig.6], to establish, by multilinear regression, a model which makes it possible to determine the BDF of the gamma spectrum of a controlled zone from energy bands, for example Bb B2 etc, where the contribution of possible uranium contamination is negligible, for example above 220 keV.Using raw counts in areas likely to characterize the three natural contributions to the gamma BDF (for example around the 609 keV line for uranium, the 1460 keV line for 40K; the contribution of 232Th around the 2614 keV line is not taken into account explicitly in the model but indirectly through the two previous areas, because then the model gives better accuracy and a lower standard deviation on the estimated background noise, 232Th having a contribution in the 609 keV area through the 583 keV line of 2O8T1, a descendant in the 232Th chain), we obtain by regression, with the least squares method, an expression of the form: .

[0140] [Math.2] CbDF [50-150] ~ ^609 X Q550-660] + #1460 X Qi300-1600]+ #0

[0141] where C|55o-66oj and Qijoo-ieooj are, respectively, the raw count rates in the spectral regions around 609 keV (natural 238U region) and around 1460 keV (40 K region). In this preliminary study of the BDF we were interested in an energy band from 50 to 150 keV rather than those considered above, but the principle of BDF estimation is the same, especially for the "enrichment" bands between 50-130 keV and 165-215 keV. The calculation by multilinear regression allows, from experimental data (here the 47 BDF spectra of [Fig.6]), to find the constants a609, a i460 and a0 in order to reduce the gap between the CBDF lso-isoj values ​​calculated by the previous formula and the CBdf [so-isoj values ​​from the experiment; in other words, the counts for the BDF in the low energy regions can be estimated from counts in higher energy regions: these are the high energy lines energy which induces, by Compton diffusion effect, the continuum of the spectrum, particularly at low energy.

[0142] To test the method, the sample of 47 spectra was split into a subsample of 42 spectra for model training and 5 spectra for testing, drawn randomly. As output, we obtain the graph in [Fig.7], which allows a comparison of the experimental points (CBDF isoisojobservé) and the points calculated by the model (CBDf lso-isoj modèle).

[0143] [Fig.7] shows that, with this model, the calculated values ​​are close to the experimental values ​​(distribution of points close to the line Y=X). The coefficients are a609 = 5.4, ai460 = 5.0 and a0 = - 8.8 c / s and the uncertainty associated with CBDFl5o-i5oj is 0.4 c / s. It can be noted that it is smaller than with the first approach using the average (uncertainty of 1 c / s). The regression approach (for example by machine learning) gives better results than the first (simple average).

[0144] From similar regression models, one can estimate the background noise in each of the regions of interest for the estimation of 235U / 238U enrichment (see above, in connection with [Fig.3]) and deduce the quantities CBDF[5o-i3ojet CBDF[i65-2i5j-The equation for calculating the aENR indicator then becomes:

[0145] [Math.3] ^AND _ c[ 165-2151 aENR~ r^Ër c [50-130] ^■GROSS ^BDF _ H165-2r5rMt6Ml5| _ —■ ^RUT ^BÜF ~ H5IH 30]~ MSO-I30| C[i65-215n®609 XC[550-660]+ «1460 XC[1300-1600]+ iBRVT { i \ C5O-i3On,^6OT X ^1550-660]+ ^1460 x Q1300-1600]+ ^o)

[0146] Steps of an example of a method according to the invention are given in [Fig.8]. Such an example of contamination detection, for example on the floors of an installation, combines the alpha, beta and gamma measurement methods.

[0147] Previously, one may for example have scanned all or part of a building with alpha or beta detector(s) to find where the suspect areas are. The case of a floor given here is therefore only one example among others.

[0148] In a first step, a rapid alpha (resp. beta) detection of a suspect surface is carried out, or an alpha (SI) or beta (S' 1, if the alpha measurement is not suitable) measurement is carried out to identify the areas, for example the surfaces, contaminated; this step can be carried out systematically across an entire installation.

[0149] Then (step S2), it is possible to carry out a control step by spectrometry gamma.

[0150] In a third step (S3), it is possible to carry out a low-resolution gamma spectrum, in accordance with the present invention, which gives rapid confirmation of a first estimate of the 235U enrichment of the contamination spot.

[0151] As a variant (step S'3), or in addition, it is possible to carry out, by high-resolution gamma spectrometry (with a germanium detector) a fine characterization, long-term, as well as an accurate assessment of 235U enrichment and an estimation of the depth of contamination.

[0152] In other words, the present invention mainly implements a low-resolution gamma spectrometry technique, but can be coupled:

[0153] - to alpha and beta counters for initial detection of contamination uranium;

[0154] - and / or possibly to a measurement by high-resolution gamma spectrometry for a finer characterization.

[0155] [Fig. 10A] represents a first variant of a method according to the invention, for which a counter 4a, alpha or beta, is used, to make a first measurement of a contamination 8 on the surface of a volume 2, then a scintillator 4 as described above (for example a NaI detector) is used to carry out a second measurement, according to the present invention.

[0156] [Fig. 10B] represents a 2nd variant of a method according to the invention, for which a scintillator 4 as described above (for example a NaI detector) is first used to carry out, according to the present invention, a first measurement of a contamination 8 on the surface of a volume 2, then a high-resolution 4g gamma counter is used to make a second measurement.

[0157] It is possible to combine the diagram of [Fig.lOA] with that of [Fig.lOB], a measurement by a method according to the invention being preceded by a measurement by a 4a alpha or beta counter ([Fig.lOA]), and followed by a measurement by a 4g gamma counter ([Fig.lOB]).

[0158] Since alpha and beta measurements are more sensitive than gamma measurement (except in the case of an unfavorable surface condition of the object or contamination profile as in the case of deep contamination, or when the 235U enrichment of the contamination is high for beta measurement), alpha and beta measurements can be used for a rapid preliminary search (of the order of a second) for contamination.

[0159] Once a contamination is identified, a gamma measurement Nal(Tl), of slightly longer duration (of the order of a minute), can be carried out. From this gamma spectrum and a subtraction of the background noise as explained previously, the presence of a contamination can be confirmed with the net count CNETA in a spectral band A (for example 50 keV - 220 keV) and an enrichment indicator aENR can be calculated from the net counts CNETai and CNETa2 in 2 sub-bands Al (for example 50 keV - 130 keV) and A2 (for example 165 keV -215 keV) of the band A.

[0160] Detailed steps of an example of such a method according to the invention are given in [Fig.9]. Here again, all or part of a building may have been previously scanned with alpha or beta detector(s) to find where the suspect areas are. The case of a floor given here is therefore only one example among others.

[0161] The steps of [Fig.9] are described below:

[0162] Steps S10, S' 10 and S1 1 correspond to steps S1, S' 1 and S2 described above.

[0163] It is then possible (step S12) to read the counting rate in the low-frequency region. CBruta energy then (step S13) subtract the background noise CBDEA to obtain CNETA • CBDFAcorresponds to the raw background noise count rate estimated for example by a linear regression model from experimental BDF spectra which may have been acquired beforehand.

[0164] If CNETA is positive, contamination is present since the value obtained after subtracting the background noise is the surplus of detected counts, due to the presence of contamination.

[0165] In this case (step S14), a reading of the counting rates CBRUTAi and CBRUTA2 is carried out in the sub-bands A1 and A2 around the characteristic contamination lines.

[0166] Then (step S15), the counting rates CNETAi and CNETA2 can be obtained after under traction of the BDF counts CBDEAi and CBDEA2 which may have been estimated as explained previously, for example by linear regression models from previously measured experimental BDF spectra.

[0167] Then (step S16) from these net counting rates, the indicator aENR is calculated and the mass percentage of enrichment 235pj / 238pj csj deduced (step S17) from a calibration curve 235u / 23spj _ ^a^) pre-established by numerical simulation, for example such as the curve in [Fig.4].

[0168] If CNETA (as calculated in step S13) is less than or equal to zero (it should be noted that with the statistical fluctuations of the counts, and to the extent that the BDF is subtracted, we can have a negative net count if there is no contamination (or very low contamination), no contamination is detected and the spectrum obtained corresponds to the natural background noise. This spectrum can therefore be recorded (step S18) in a library dedicated to the background noise spectra acquired for example throughout the inspection of the floors of the installation.This library can be exploited as it grows to update, for example by multilinear regression, the models allowing to estimate, for each spectrum of the future areas to be monitored, the BDF count rates in the low energy regions of interest (according to the example taken here: from 50 to 220 keV, for example again from 50 to 130 keV and from 165 to 215 keV) from the raw counts in the higher energy regions (for example from 550 keV to 660 keV and from 1300 keV to 1600 keV). This makes it possible to reduce the uncertainty linked to the estimation of the contribution of background noise and to improve the sensitivity of detection of possible contamination, as well as the precision on . uranium enrichment in the event of detected contamination. Alternatively, the background noise can be obtained by one or more measurements on an uncontaminated object but containing natural uranium in secular radioactive equilibrium.

[0169] In a method according to the invention, the object examined, which contains natural uranium, may further contain at least potassium and / or thorium.

[0170] The industrial applications of the invention are broad, including, for example, the cleanup and dismantling of uranium processing facilities (such as an enrichment plant, a fuel fabrication workshop or a research and development laboratory) as well as the analysis of radioactive waste containing uranium (such as that resulting from cleanup and dismantling operations).

Claims

Claims

1. - Method for detecting contamination (8) in artificial uranium of an object containing natural uranium, comprising: a) - positioning at least one scintillator (4) on the object to be examined, b) - producing at least one gamma radiation spectrum emitted by the object to be examined, c) - selecting a 1st spectral region (Ai) and a 2nd spectral region (A2) of said gamma radiation spectrum, d) - measuring a 1st net count rate, without background noise, in the 1st spectral region and a 2nd net count rate, without background noise, in the 2nd spectral region, e) - estimating the enrichment in artificial 235U on the basis of the 1st net count rate and the 2nd net count rate.

2. - A method according to claim 1 wherein the scintillator has a volume of between 0.1 and 10 1.

3. - Method according to one of claims 1 or 2, in which the scintillator is a NaI detector.

4. - Method according to one of claims 1 to 3, in which the 2 spectral regions (Ai, A2) have different contributions of 235U and 238U.

5. - Method according to one of claims 1 to 4, in which the 1st spectral region contains at least one line at 93 keV and the 2nd spectral region contains at least one line at 185 keV.

6. - Method according to one of claims 1 to 5, in which the 1st spectral region is located between 50 keV and 150 keV and the 2nd spectral region is located between 150 keV and 250 keV

7. - Method according to one of claims 1 to 6, in which an estimation of the background noise of the measurement is also carried out, in at least the 1st spectral region and the 2nd spectral region, the net counting rates being calculated by taking this background noise into account.

8. - Method according to claim 7, the background noise being estimated from one or more gamma spectra acquired on at least one uncontaminated object, or stored.

9. - A method according to claim 7 or 8, wherein the background noise is estimated from measurements in higher energy regions than the 1st spectral region and the 2nd spectral region.

10. - A method according to claim 9, wherein the background noise is estimated from measurements in higher energy regions than the 1st spectral region and the 2nd spectral region, for which the contribution of possible uranium contamination is zero or less than 1%.

11. - A method according to claim 9 or claim 10, the background noise, in at least the 1st spectral region and the 2nd spectral region, being estimated from measurements in a spectral region between 550 keV and 660 keV and in a spectral region between 1300 keV and 1600 keV.

12. - Method according to one of claims 1 to 11, further comprising an estimation of the mass percentage of enrichment 235u / 238pj | using a calibration curve.

13. - Method according to one of claims 1 to 12, the object further containing at least potassium and / or thorium.

14. - Method according to one of claims 1 to 13, comprising a prior step of detection by alpha and / or beta counting.

15. - Method according to one of claims 1 to 14, further comprising a high-resolution gamma detection step.

16. - Method according to one of claims 1 to 15, further comprising, for low or no contamination, a step of storing at least part of the spectrum data.

17. - Method according to one of claims 1 to 16, the enrichment indicator being obtained by calculating the ratio between the 1st net counting rate and the 2nd net counting rate.

18. - Device for detecting artificial uranium contamination of an object containing natural uranium, comprising: a) - a scintillator (4); b) - means (2) for producing at least one gamma radiation spectrum emitted by the object to be examined, c) - means (14) for selecting a 1st spectral region and a 2nd spectral region of said gamma radiation spectrum, d) - means (14) adapted to estimate a 1st net count rate, without background noise, in the 1st spectral region and a 2nd net count rate, without background noise, in the 2nd spectral region, e) - means (14) adapted to estimate a 235U enrichment indicator on the basis of the 1st net count rate and the 2nd net count rate.

19. - Device according to claim 18, in which the scintillator has a volume between 0.1 1 and 10 1

20. - Device according to one of claims 18 or 19, in which the scintillator is a NaI detector.

21. - Device according to one of claims 18 to 20, in which the 1st spectral region contains at least one line at 93 keV and the 2nd spectral region contains at least one line at 185 keV.

22. - Device according to one of claims 18 to 21, in which the 1st spectral region is located between 50 keV and 150 keV and the 2nd spectral region is located between 150 keV and 250 keV.

23. - Device according to one of claims 18 to 22, further comprising means for carrying out an estimation of the background noise of the measurement, in at least the 1st spectral region and the 2nd spectral region, and means for correcting the counting rates on the basis of this estimation of the background noise.

24. - Device according to one of claims 18 to 23, further comprising means for storing data of gamma spectra of uncontaminated surfaces.

25. - Device according to one of claims 18 to 24, further comprising an alpha and / or beta detector and means for implementing a prior detection step by alpha and / or beta counting.

26. - Device according to one of claims 18 to 25, further comprising a high-resolution gamma detector.

27. ​​- Device according to one of claims 18 to 26, the enrichment indicator being obtained by calculating the ratio between the 1st net counting rate and the 2nd net counting rate.

Images