RU2785525C1 - Method for search and detection of sources of ionizing radiation - Google Patents

Abstract

FIELD: radiation.

SUBSTANCE: method for searching and detecting sources of ionizing radiation using stationary or moving scintillation spectrometric detectors over the surveyed area is proposed, which consists in the fact that the entire measurement time interval is divided into sections of exposure time t_exp, the duration of each of which is the same and is preferably (8-15)⋅t_e, then, from the measurement results for each of the energy windows, the values of the average background count rate b, the amplitude of the expected signal a_m and the position of the signal on the time scale tm are determined, after which the value of the parameter η is calculated by the formula

,

where t_e is a predetermined signal duration, then η is compared with the threshold q_o, determined by a given false alarm probability, and the decision to detect the desired radiation source is made if η > q_o, for any of the established energy windows simultaneously.

EFFECT: reducing the time of monitoring the object, simplifying the procedure for preparing for measurements, achieving a reduction in the thresholds for detecting radionuclides.

1 cl, 1 dwg, 1 tbl

Description

The claimed method relates to the field of radiation monitoring using scintillation spectrometric detectors and is designed to search for and detect sources of ionizing radiation (IRS), lost or deliberately hidden (in cases of illegal disposal of radioactive waste, etc.) or unauthorized carried through the portal monitor. The claimed method can be used in stationary portal radiation monitors (RM), as well as in wearable portable or moving RM (for example, automobile, helicopter), mobile radiation monitoring systems (RK) using scintillation detectors.

There are known methods for searching for IRS used in portable devices of operational RK [1-4] - hand-held radiometers-dosimeters of the MKS-A02 type, etc., in which a detection method is used to detect IRS, based on comparing the number of pulses N ₀ measured by the equipment for a strictly assigned control time t _n with a pre-calculated detection threshold q ₀ . The threshold q ₀ is calculated based on the background situation before arrival in the control zone, according to the formula

where b=B/t _b - average background count rate;

σ=(B) ^{one/ 2} - standard deviation (RMS) of the value AT , equal to the measured number of background pulses for a rigidly specified time t _b ;

m - coefficient ( m ≥ 4).

If N ₀ is greater than q ₀ , then a decision is made to detect the desired IRS, otherwise a decision is made to not detect it.

“Hard” detection thresholds similar to those described are used in the methods according to relatively recent patents No. 2655044 (2018) and No. 2317570 (2008). It should be emphasized that in the described methods (devices) the value of m is practically not substantiated and is selected during the operation of the device based on work experience or "intuitively"; accordingly, this leads to an overestimation of the detection thresholds in comparison with the theoretically justified ones and to an increase in the probability of missing a dangerous AI source.

Known methods for detecting sources of AI [5], which use some of the basic provisions, terms and conclusions from the theory of detection [6-8]. However, in [5], the mentioned provisions are used insufficiently consistently and correctly, and therefore also lead to an overestimation of the detection thresholds and an increase in the probability of missing. In the search and detection method described in [5], according to the authors, the Neumann-Pearson criterion is used, however, the detection threshold is actually calculated by (1). Obviously, the described method is almost the same as [1], except that the values of t _b and t _n can be set (adjusted) depending on the required accuracy and limitations on the measurement time. Moreover, the determination of the threshold according to (1) leads to a significant overestimation of the minimum detected signal in comparison with the detection theory, as we have already written above.

In the well-known "Method of monitoring moving objects for the detection of fissile nuclear materials" [9], ratios are mainly used in accordance with the detection theory [6-8], and corrections for Poisson statistics are applied to form the detection threshold calculated from a given probability of false alarms , which is especially important at low count rates. However, here, too, the main provisions of the detection theory are applied not quite correctly, which leads to an increase in the probability of missing AI.

There is a known method for detecting weak fluxes of ionizing radiation, described in patent No. 2140660 [10], which is completely and strictly based on the theory of detection [6-8].

In accordance with the detection theory [6-8], under conditions of complete a priori uncertainty (i.e., when there is no information about the object being detected, either about the presence or absence of an AI source on it, or about the magnitude of the signal from the AI source), the detector can only the Neyman-Pearson test , which is the most powerful test, should be used. The latter means that the application of the Neyman-Pearson criterion makes it possible to obtain the highest theoretically possible probability of detection for a given false alarm probability (and, accordingly, the lowest probability of missing).

The second most important provision of the detection theory, which should be used in the practical implementation of the theory, is the need to calculate the likelihood ratio. For practical purposes, it is more convenient to use an expression equivalent to the mentioned likelihood ratio for sufficient statistics η [6], which in the particular case under discussion of detecting weak AI flows takes the form

where all the designations are the same as those introduced above (in the section describing formula (1) for the analogue - dosimeter drs-rm1401 [1]).

In physical terms, the value of η is a relative (or "normalized") value of the average count rate of the signal from the detected source a = nb (otherwise: the value of η is the relative value of the signal a , presented in such a form when the signal dispersion is equal to one).

The obtained value of η is compared with the threshold q _o , the value of which depends on the false alarm probability P _lt given by the operator and is determined from the tables of the normal distribution law. (The value q _o = 1 corresponds to the so-called "quantile of the normal distribution").

Thus, the detection method according to [7] is reduced to the following sequence of operations:

• before the beginning of control of objects, the threshold q _o is determined in accordance with the probability of false alarms P _lt specified by the operator according to the tables for normal distribution (in particular, when P _lt =0.05 q _o =1.64);

• measure the average count rate b due to background radiation, and the background measurement time t _b can be any, as a rule, relatively long, and is determined by the operational situation or is entered by the operator;

• the device is switched to the operational control mode, i.e. measure the number of radiation pulses from the controlled object N _o during the control interval t _n specified by the operator; calculate the average count rate of the additive mixture of signal and background n=N _o /t _n ;

• determine the value of sufficient statistics η according to (2);

• the obtained value of the parameter η is compared with the threshold q _o . If a

then they make a decision on detection (the “alarm” light or sound signal turns on); otherwise, a decision is made about non-detection (about the absence of a source), which is recorded in the report on the control procedure. Inequality (3) in the detection theory is called the decision rule.

The methods described above [1-5, 9, 10] are not suitable for search tasks; they can be used specifically and only for the tasks of detecting AI sources at controlled objects. Obviously, when monitoring objects for the detection of AI sources, the first key point is that the background is measured in advance, i.e. when it is known for sure that the object of control is absent in the detector's sensitivity zone. The second key point is as follows: the operator always knows whether the controlled object is present in the control zone (the so-called "presence sensors" are used for this, or the operator himself directly sees the control object and at the right time presses the "start" and "stop" buttons) ). Moreover, the time of the control measurement of radiation from the object corresponds to a value equal to the time the object is in the sensitivity zone of the detectors. And to make a decision about the presence or absence of AI on the control object in methods like [1-5, 9, 10], the result of radiation measurements from the object is directly or indirectly compared with the previously measured background level.

When conducting a search, a significantly different formulation of the problem arises: in the general case, when searching, it is always unknown whether or not a weak source of IR is present in the detector sensitivity zone. When conducting a search, it is required continuously, for a long time, sometimes reaching 3-10 hours or more, when moving equipment with detectors, to measure the radiation from the search area or continuously moving monitoring objects, and from the results of such continuous measurements to somehow extract information for making a decision about the presence or absence of an AI source in the search area. In this case, it should always be borne in mind that the background level in the search area can vary quite significantly (from tens to hundreds of percent of the original). In such a formulation of the problem, the methods [1-5, 9, 10] are in principle not suitable.

As the results of practical studies of the background over long time intervals [11–13] show, the background level, measured at averaging intervals of the order of 100–300 s or more, can change at rates of the order of 10⋅RMS in 20–30 min., and in 2– 3 hours (during the search) "float away" in absolute value by 50% or more even in cases of a fixed radiation monitor (RM). If for problems of radiation monitoring of objects (i.e., in detection problems), the problem of adaptation to fluctuations in the average background level can be solved quite simply (in the limit, under the most unfavorable conditions - by measuring the background level before monitoring each new object), then in search problems The problem of adapting to background changes is not easy to solve. Here, for example, such a technique is completely unacceptable: to leave the search area every 15-20 minutes to clarify the background level. Therefore, more often they resort to the only measure possible in these situations - to the method of overestimating (roughening) the detection thresholds.

There are many methods for searching and detecting radioactive objects used in devices of the type [13] and related to the field of radiation monitoring. These methods can be used to search for and detect radioactive objects on the ground, as well as the spatial distribution of radioactive sources during nuclear logging of wells. However, these methods and their corresponding devices [14] are close to the claimed only in name; in fact, they necessarily involve the use of many stationary posts with complex radiation monitoring devices equipped with collimators. An overview description of this group of methods and devices designed for continuous radiation monitoring of the area around the NPP and for operational monitoring of the area in the event of an accident is given in [13]. Obviously, this group of devices [13, 14] is not related to the discussed problems of searching for AI sources.

It is worth mentioning the review [15], which describes in detail several different ways of processing data from radiation monitors. The review [15] actually describes all the methods for detecting AI sources that we have listed above (including the method for detecting AI by the authors of this application under patent No. 2140660 [10]). However, in [15], only cases of using stationary radiation monitoring devices are considered (it is the latter that are designated by specialists by the term “radiation monitors” - RM, or “portal” RM), and the tasks of searching for AI by means of mobile (mobile) radiation monitoring devices are not at all affected. . In addition, the review [15] does not at all address issues related to solving the problems of categorization (identification) of detected radionuclides (RN).

There is also a method for searching and detecting AI, described in patent No. 2242024 [16].

Its main distinguishing feature is that it is not required to measure the background before starting the search and detection procedure.

The essence of the AI search and detection method described in [16] is as follows:

The entire measurement time interval is divided into sections of exposure time t _exp , the duration of each of which is the same and is preferably (8 - 15 )⋅ t _e .

Further, from the measurement results, the values of the average background count rate b , the amplitude of the expected signal a _m and the position of the signal on the time scale t _m are determined, after which the value of sufficient statistics - (parameter η ) is calculated using a formula similar to (2), but differing in that instead of the control time of the object t _n substitute the value of "effective signal duration" t _e .

Next, η is compared with a threshold q _o determined by a given probability of false alarms, and the decision to detect the desired object is made if η > q _{o .}

The formulas used in patent No. 2242024 [16] are described and justified in [17] in more detail.

The method [16] has a number of significant advantages compared to those described above [1-5, 9, 10], the main of which are:

• there is no need for preliminary measurement of the background before the start of the search and (or) detection procedure, and thus the total time of control is reduced;

• it is possible to search for AI sources under conditions of a significantly non-stationary background;

• it is ensured that the greatest detection probability is achieved with the given detection parameters, in particular, with a given false alarm probability, and, accordingly, the achievement of low detection thresholds close to theoretically limiting ones [6-8].

However, it is impossible not to emphasize one important drawback characteristic of the described method [16], which is that the method according to [16] does not solve the problem of categorization (identification) of detected radionuclides (RN).

Closest to the claimed is the detection method described in the abstract of the report [18], in which, simultaneously with the task of detecting AI, the problem of categorization (identification) of detectable radionuclides (RN) is solved. Such a task, of course, is a valuable option in many specific cases of radiation monitoring, in particular, in the following:

• in addition to the most important function of detection (detection) of a radionuclide (RN), the task of detecting a specific technogenic (technogenic) RN (for example, ¹³⁷ Cs) is solved in the presence of radiation from radiopharmaceuticals that have not yet decayed, which are in the body of a controlled person as a result of medical procedures;

• when it is necessary to identify technogenic radionuclides illegally transported (smuggled) into the control object against the background of declared ROPs.

However, the authors of [18] do not even mention the detection criteria that they use in their methods (algorithms). Judging by the text of the theses, as well as by the long-standing application [19] of the authors of [18] and by the texts and notes to the tables in the ATOMTECH catalog [20], the authors of [18] use the so-called “hard” or “simplified” method (algorithm) detection, similar to that described above in the formula (1). Those. the authors of the prototype [18], judging by their texts, are not concerned at all with the use of known approaches of the detection theory.

Note:

Here we refer to the catalog [20] because the authors of the prototype [18] represent the enterprise "ATOMTEH" JSC "MNIPI".

In addition, the statement that vehicles shield the background up to 30% is valid only for neutron radiation (and, accordingly, for the neutron detection channel), but not for gamma radiation. This is evidenced by our data obtained as a result of long-term operation of the Soratnik complexes [21]. The presence of a vehicle in the "field of view" of the detector, according to our data, has practically no effect on the level of the gamma background. (Note that this statement of the authors of this application is based on the rich experience of actual measurements by means of the Soratnik complexes, which have both neutron and gamma channels. The Soratnik complexes have been in operation for more than 17 years [21]). Therefore, based on the above experience, it seems to us that the methods (algorithms) for taking into account background radiation, described in [18], are unreasonably complicated. It should also be added that in the practical application of the methods described in [18], it is completely unclear to the operator of the RK complex how to use such algorithms. Neither the sequence (algorithms) of applying one or another “background spectrum with a given limiting contribution” out of the five listed in [18], nor the criteria for their application remain unclear. Possibly, the authors of [18] have in mind a stationary RM located in a carefully studied territory. But then such an approach contradicts even the title of the work [18] "Fast deployable radiation monitor". It is logical to assume that "rapid deployment" implies every time a new, hitherto unexplored area. And therefore, starting the search and detection procedure, the operator, as a rule, does not have a priori information about the nature of the background spectrum.

The proposed invention solves the problem of eliminating the need for a preliminary measurement of the background level, and thereby reducing the total time of monitoring, categorization (identification) of detected radionuclides (RN) and lowering the detection thresholds of RN.

To solve these problems in the claimed method of searching and detecting AI sources, the sequence of operations is proposed as follows:

1) Immediately upon arrival in the search area and turning on the detecting system, energy windows are set corresponding to the detected sought or illegally (unauthorized) transported radionuclides.

2) Next, they begin to measure the average count rate of the additive mixture of signal and background n(t) in the exposure time sections (time intervals) t _exp , the duration of each of which is the same and is preferably (8-15)⋅ t _e , where t _e is effective signal duration. Moreover, such measurements are carried out in all installed energy windows simultaneously.

3) At the end of exposure interval No. 1 with duration t _exp , from the values of the vector n ₁ (t) measured in interval No. 1, determine the values of the average background count rate b, signal amplitude a _m and the position of the signal maximum on the time scale t _m for all established energy windows at the same time.

4) Calculate the value of sufficient statistics - parameter η - according to the formula

5) Received parameter value η compared with the detection threshold q _o determined by the given probability of false alarms, i.e. apply the decision rule (3). If a η > q _o , then a decision is made to detect (about a positive result of the search for the AI source in interval No. 1); otherwise, they decide on the absence of a signal. Moreover, the calculation of the value η according to (4), and the application of the decision rule (3) is performed on all the established energy windows simultaneously.

6) In parallel with the procedure for processing the realization n ₁ (t), further measurements of the values of n(t) in the established energy windows are continued.

7) At the end of interval No. 2, the vector n ₂ (t) is subjected to processing, exactly the same as described in p.p. (2)-(5).

8) Next, sequentially repeat the steps according to p.p. (1)-(7) for intervals ## 3, 4, 5, etc. up to the completion of the search in a given area or until the completion of the radiation monitoring procedure in the case of a stationary RM.

The effective duration of the signal is determined from the relation t _e = k _k ⋅r _o /v _o , where r _o is the shortest distance between the detector and the expected location of the desired AI source; v _o - the average speed of movement of the detecting system (or vice versa: the movement of the control object past the stationary RM); k _k - coefficient, the value of which is determined by the design of the detector (for a flat detector k _k =2.0; for an omnidirectional detector k _k =2.8).

Details related to the substantiation of the above formula (4), as well as the calculation of t _e and the threshold q _o are described by the authors of patent No. 240024 [16] in the article [17].

It should be noted that the claimed method applies only to the case of using scintillation detectors in RM, which allow measuring the spectra of detected radionuclides. These may be detectors based on inorganic crystals (preferably such as NaI-Tl, CsI-Tl, etc.) or detectors based on scintillating plastic.

The application of the proposed method, as well as the prototype, due to the spectral selectivity allows you to detect man-made radionuclides illegally smuggled (smuggled) into the control object against the background of declared RH (categorization or identification of detected RH).

The present invention in comparison with the prototype has the following advantages:

• the need for preliminary measurement of the background level is eliminated, and thus the total time of object control is reduced both in cases of stationary RM and when a mobile RM passes by the controlled object;

• the need to take into account the background spectrum with a given limiting contribution from various natural radionuclides [18] is eliminated (based on “long-term simulations”, as formulated in [18]), which greatly simplifies the operator's work;

• the solution of the problems of searching for lost RM or survey of the contaminated territory is provided;

• a decrease in the detection thresholds of PH is achieved, since proposed inventionis based on the main provisions of the theory of detection (as opposed to the prototype).

To confirm the last advantage - the reduction of the detection thresholds of RH in comparison with the prototype [18] - we performed comparative calculations.

It should be noted that it was difficult to quantify the gain from the application of the proposed method in terms of the “minimum detectable activity” Q _min parameter, because the prototype [18] does not clearly indicate the design features of “pilars” (plastic scintillators), the background count rate at the output of the pilars in the corresponding energy windows, as well as the detection parameters (false alarm probability, detection probability, the formula by which the detection threshold is calculated, or at least a verbal description of the detection procedure).

Therefore, the following approach was used for comparative calculations.

Since the authors of the prototype [18] represent the ATOMTEKH company, we used the data given in the catalog of this company [20], where the table on page 40 shows the detection thresholds for the case of using BDKG-19 detection units based on NaI(Tl) 63x160 mm and at the same time (unlike [18]), there are sufficiently detailed data on the detection conditions. Figure 1 shows a screenshot from the specified catalog [20], where the values of the minimum detectable activity Q* _min (detection thresholds) are indicated.

For comparison, we calculated the values of the detection thresholds (equal to Q _min ) for cases where the measurement results would be processed by the method proposed in this application for the invention (of course, using the same detection units BDKG-19). And the necessary data on the design features of detection units based on NaI(Tl) and, most importantly, on the values of the average background count rate b in the corresponding energy windows, the authors of this application are known based on the operating experience of the Soratnik complexes, which include similar detection units with NaI(Tl) crystals.

The results of comparative calculations are shown in Table 1.

Designations adopted in table 1:

b- average background count rate, 1/s;

Q* _min - minimum detectable activity for the prototype (threshold

detection, is borrowed from Fig. 1, column 4);

Q _min - minimum detectable activity for the claimed

way;

K \u003d Q _min / Q * _min - gain in terms of detection thresholds of the claimed

way in relation to the prototype.

Calculations were made for the following detection conditions:

• probability _of false alarms Рlt = 0.05;

• probability of detection Р _det = 0.80;

• object speed v _o = 5 km/h;

• distance to the object r _o = 1 m.

The results obtained show that the values of the expected gain K for the proposed method in terms of detection thresholds range from 1.5 to 4 - 5 relative units, depending on the specific radionuclide.

Sources of information

1. Search microprocessor dosimeter DRS-RM1401. Technical description and operating instructions. - Minsk: SP Polimaster, 1997.

2. Dosimeter-radiometer DRBP-03. Passport SSCC 14.00.00.000 PS. - M.: VNIIFTRI, 1996.

3. Radiometer-dosimeter MKS-06N INSPECTOR. The passport. - M.: LLP "Green Star", 1996.

4. MKS-A02. Manual. DKCI.411168.002 RE. /G. Dubna, Moscow region: SPC Aspect, 2000, 19 p.

5. Kirillov V.M., Suprunov V.I. Detection of moving sources of ionizing radiation. Measuring technology. - 1994. N 8, p. 63.

6. Van Tries G. Theory of detection, estimates and modulation. T. 1. - M.: Sov. radio, 1972, 744 p.

7. Tikhonov V. I. Statistical radio engineering. - M.: Radio and communication, 1982, 624 p.

8. Zaks Sh. Theory of statistical inference. - M.: Mir, 1975, 776 p.

9. Method for monitoring moving objects to detect fissile nuclear materials. // Gorev A.V., Zaitsev E.I., Ivanov A.I. Patent No. 2150127 dated May 27, 2000.

10. Method for detecting weak fluxes of ionizing radiation. // Viktorov L.V., Kruzhalov A.V., Shein A.S., Shulgin B.V., Shulgin D.B. Patent of the Russian Federation No. 2140660. dated 10/27/1999.

11. Statistical characteristics of gamma and neutron background fluctuations. Andreev V.S., Viktorov L.V., Petrov V.L., Shein A.S. VI International meeting "Problems of Applied Spectrometry and Radiometry". Abstracts of reports. M.: 2002. p.32.

12. Viktorov L. ., Mogilnikova Yu.A. Variations of the neutron field of the Earth. // In: "Problems of spectroscopy and spectrometry". Yekaterinburg. USTU. 2000. S.95-104.

13. Khazanov D. B. Construction of systems for assessing the radiation situation in the area of the nuclear power plant. (Review). // Nuclear Instrumentation. (Issues of atomic science and technology). Issue 2. 1985. P.3.

14. Device for searching and detecting radioactive objects. // Mukhin V.I., Muslimov R.Kh., Samosadny V.T. Patent of the Russian Federation No. 2160909 dated 12.20.2000.

15. Review of domestic radiometric and spectrometric systems that can be used for accounting and control of nuclear materials. // Federal State Unitary Enterprise VNIIA im. N.L. Dukhov.

http://www.vniia.ru/rgamo/literat/obzor/doc/obzorrus.pdf.

16. L.V. Viktorov, K.V. Ivanovskikh, Yu.G. Lazarev, V.L. Petrov, A.S. Shein, B.V. // Patent of the Russian Federation No. 2242024 dated 10.12.2004.

17. Viktorov L.V., Shein A.S. Algorithms for searching and detecting sources. In: Mobile complexes for radiation monitoring. Collection of scientific developments. / Under the editorship of professors B.V. Shulgin and A.V. Kruzhalova. Yekaterinburg: UrFU, 2011, pp. 51-73.

18. Rapidly deployable radiation portal monitor with the function of categorization of natural and technogenic radionuclides // Antonov A.V., Alexeychuk I.A., Bystrov E.V., Konovalov E.A. //Abstracts of the XV International Meeting "Problems of Applied Spectrometry and Radiometry", October 7 - 11, 2019 Kazan PEI DPO "UC" MEDTECHATOM ". pp. 34-36.

19. Method for detecting fissile materials. / Isakov A.I., Antonov A.V., Benetsky B.A. and others // Application: 94020227/25. 06/01/1994

20. Devices and technologies for nuclear measurements and radiation monitoring. Product Catalog. Research and Production Unitary Enterprise "ATOMTEH" JSC "MNIPI".

http://ftp.atomtex.com/catalogues/catalogue_ru.pdf.

21. Mobile complexes of radiation control. Collection of scientific developments. / Under the editorship of professors B.V. Shulgin and A.V. Kruzhalova. Ekaterinburg: UrFU, 2011. 137 p.

Table 1 The results of comparative calculations of the detection thresholds of the proposed method and prototype Qty
Detection Blocks
(DB) Nuclide Background, b ,
1/s Q _min (kBq)
when using the proposed method Q* _min (kBq)
for the prototype [18] K
win
o.u. 1 DB ²⁴¹ am 10.4 99.2 430 4.3 ^137Cs 4.8 96.6 220 2.3 ^60Co 3.2 65.5 100 1.5 2 DB ²⁴¹ am 20.8 60.4 320 5.3 ^137Cs 9.6 54.6 160 2.9 ^60Co 6.4 35.2 70 2.0

Claims (4)

A method for searching and detecting sources of ionizing radiation using stationary or moving scintillation spectrometric detectors across the surveyed area, which consists in the fact that prior to the start of measurements in the spectrometric path, energy windows are set corresponding to the desired radionuclides, then in the process of monitoring the object or the movement of the detecting system across the surveyed area perform continuous measurements of the average count rate of the additive mixture of signal and background n(t), process the received measurement information, compare the obtained measurement results with the threshold for all energy windows simultaneously and make a decision on the presence or absence of illegally carried or transported radiation sources in the object of control and at the same time obtaining information on the categories of detected radionuclides, characterized in that the entire measurement time interval is divided into sections of exposure time t _exp , the duration of each of which is the same and is preferably (8-15)⋅t _e , then from the measurement results for each of the energy windows, the values of the average background count rate b are simultaneously determined , the amplitude of the expected signal a _m and the position of the signal on the time scale t _m , after which the value of the parameter η is calculated by the formula

, where t _e is a predetermined signal duration, then η is compared with a threshold q _o determined by a given probability of false alarms, and the decision to detect the desired radiation source is made if η > q _o , for any of the established energy windows simultaneously.

Images