CN102999693B - A method for pre-discrimination of material characteristics of nuclear components based on 252Cf source drive - Google Patents

Abstract

The invention discloses a method for pre-discriminating material characteristics of nuclear components driven by ²⁵² Cf sources. It mainly includes the use of time-domain correlation models and signal processing methods for pre-discrimination work on the material thickness and material information of nuclear components. The time-domain correlation model and signal processing method includes three main aspects: Using the location of the neutron peak in the time domain correlation function between the source and the detector to interpret the material thickness of unknown nuclear components; Analyze the rising edge and falling edge of the neutron peak, put forward the neutron peak ratio method, and use it to eliminate the interference of non-nuclear materials and preliminarily judge the material of nuclear components; Using the peak count value of the time-domain correlation function between the source and the detector, the material of the unknown nuclear component is clarified and the judgment result is strengthened. Due to the unprecedented ability to remove non-nuclear materials and the ability to interpret the material and thickness characteristics of nuclear components, this method has achieved a good pre-discrimination effect on the material characteristics of nuclear components.

Description

A method for pre-discrimination of material characteristics of nuclear components based on 252Cf source drive

technical field

The invention belongs to the technical field of digital signal processing and verification, and relates to a pre-discrimination method for unknown nuclear component materials driven by a ²⁵² Cf source.

Background technique

The nuclear material/nuclear weapon identification system (Nuclear Material Identified System, NMIS) in the field of verification technology, that is, the nuclear material verification system, aims to measure the characteristic parameters of nuclear materials, so as to infer its field of use. Its important function is to detect the enrichment of nuclear materials (the enrichment of nuclear materials is an important symbol to distinguish between weapon grade and civilian grade, and it can be inferred from it that it has the level of the national nuclear industry). From the perspective of the way to obtain the radiation signal of nuclear materials, the verification technology route is usually divided into two types: passive and active. Limited to the low energy of uranium (U, Uranium) material spontaneous radiation, the verification technology for it requires active measurement, that is, injecting energy through an external excitation source to cause a chain reaction of nuclear materials, and obtaining necessary information by detecting its fission products , also known as the source-driven measurement method.

The verification principle of the source-driven noise analysis system is shown in Figure 1. First, the spontaneous fission source ²⁵² Cf in the #1 channel is the driving neutron source, and the spontaneous fission neutrons emitted by it are incident on the nuclear component to be tested and induce a chain reaction. The chain reaction produces several neutrons and gamma rays, which are captured by detectors in channels #2 and #3. The ²⁵² Cf source emits 4 spontaneous fission neutrons and 6 gamma photons per fission on average. Spontaneous fission neutrons enter the nuclear material to be tested to induce a chain reaction to produce a series of induced fission neutrons, that is, after the nuclear material to be tested is "stimulated", the neutrons carrying its relevant information are detected by the detector, and a multi-channel pulse is obtained neutron signal. The pulsed neutron signal obtained by the source-driven method can be analyzed and processed by NMIS time-domain signal, and the corresponding time-domain correlation signal label can be obtained. The traditional NMIS time-domain correlation signal particle counting composition is shown in Figure 2. The peak with a larger slope on the left side of the figure is the first detected direct γ photon, which directly passed through the nuclear component to be tested without any collision; It consists of photons, direct neutrons, scattered neutrons and induced fission neutrons, and they are roughly arranged sequentially on the time axis as marked in the figure. The production mechanisms and energies of direct neutrons, scattered neutrons and induced fission neutrons are different, so the time to reach the detector is also different, which causes the broadening in the time domain. Obviously, the count value of the source-detector signal time-domain correlation function in the time interval of induced fission neutrons (usually 20-100ns) is sensitive to the enrichment of the measured ²³⁵ U material. Therefore, this characteristic of the time-domain correlation signal of nuclear signals can be used to identify the enrichment of nuclear materials.

Although theoretically, the mathematical analysis of system-related functions has been relatively complete, but it is only limited to relatively ideal situations. The working principle of NMIS determines that the characteristic function obtained on site must be compared with the calibration value to draw a conclusion - obviously, the traditional NMIS cannot know whether the component to be tested is a nuclear component or whether it is worth measuring, but only mechanically " As the signal generated by "nuclear materials", a fixed time interval is used to judge the enrichment degree. This not only has considerable blindness, but also may waste a lot of precious time and system resources, and may introduce very large errors, which will eventually affect the identification and interpretation results of NMIS and cause serious consequences. Figure 3 shows the simulation results of the source-detector cross-correlation function under several different situations under NMIS measurement, including: Different substances, including nuclear and non-nuclear materials; different material thicknesses; Enrichment of different nuclear materials. In order to highlight the problem to be explained, the gamma peak is truncated to a certain extent in the figure, focusing on the difference between the neutron peaks. It can be seen from the figure that, even in the case of no noise, no matter the neutron peak amplitude, downward trend or time of appearance, the difference between nuclear material and non-nuclear material, nuclear material with different enrichment degree and even material with different thickness They are all different, intertwined with each other. If the enrichment is judged by the integral method according to the traditional NMIS, some of the non-nuclear materials may be judged as nuclear materials, or nuclear materials with the same enrichment degree but different thicknesses are judged as nuclear materials with different enrichment degrees, etc. wait. It can be seen that there are inherent defects in judging the degree of enrichment simply by relying on the integral value of the time-domain correlation function.

Therefore, a material prediction method based on gamma and neutron time-domain correlation waveform resolution is proposed. Using it to predict the properties of unknown nuclear components can eliminate useless information and obtain nuclear component material characteristic parameters before NMIS operation. The pre-discrimination method of component material characteristics becomes the focus of the present invention.

Contents of the invention

The technical problem solved by the present invention is how to obtain information on the thickness of unknown nuclear components, remove the interference of non-nuclear materials, determine the characteristics of materials of unknown nuclear components, and judge the thickness and material of materials, so as to overcome the lack of pre-judgment mechanism in the existing system, which is easily disturbed by non-nuclear materials and The disadvantage of a single characteristic parameter.

Considering the characteristics and properties of NMIS and neutron pulse signals, the purpose of the present invention is to provide a method for pre-discriminating material characteristics of nuclear components based on ²⁵² Cf source drive, by making corresponding neutron Extraction and analysis of peak and γ peak signals, extraction of material and thickness of unknown nuclear components, and elimination of interference from non-nuclear materials. The method makes full use of information, has strong anti-interference ability of non-nuclear materials, and can obtain the characteristics of nuclear component materials, so that the source-driven noise analysis and signal processing methods based on pulsed neutron signals can achieve higher accuracy and expand its application range.

In order to achieve the above purpose, the present invention adopts the following technical scheme: a method for pre-discriminating material characteristics of nuclear components based on ²⁵² Cf source drive, firstly, adopting ²⁵² Cf source drive noise analysis method and a series of standard material establishment can reflect material thickness and material The neutron peak position, neutron peak ratio and gamma peak count value calibration database; then, use the ²⁵² Cf source driven noise analysis method to extract the neutron peak position, neutron peak ratio and gamma peak count value; finally, Comparing and calculating the extracted three numerical values with the calibration database to obtain the material characteristic parameters of the part to be tested;

The material characteristic parameters include thickness interpretation, non-nuclear material rejection, material interpretation and identification result enhancement; wherein:

(1) Thickness interpretation: Obtain the thickness information of nuclear components through the analysis of the neutron peak time point of the source-detector time domain correlation function;

Calculate the source-detector cross-covariance function of the pulsed neutron signals of nuclear materials acquired by the nuclear weapon inspection system based on the source-driven noise analysis method; divide the original data into block, while is a certain time point on the time axis in a certain piece of data, and the calculation formula for the unbiased estimation of a certain point in the pulsed neutron source-detector cross-covariance function is:

The unbiased estimate of the overall cross-covariance signal is thus:

in represent a signal complex conjugate of is another detector signal; is the length of the data block, expressed in discrete points, since the sampling interval is 1ns, the point spacing is 1ns; is the time delay; and is the total number of data blocks;

extract From the nature of the physical model, the neutron peak is regarded as a combination of the first half of scattered gamma photons, direct neutrons, scattered neutrons and the second half of induced fission neutrons. Experiments have proved that the first half of the neutron peak time is not sensitive to whether the material is a radioactive nuclear material, but only related to the travel of the particles in the material - the thicker the material, the greater the travel of the particles. The longer it takes for neutrons and scattered neutrons to reach the detector;

In fact, it is a time series composed of time points with different amplitudes, including two parts: neutron peak and gamma peak. γ peaks and neutron peaks always appear sequentially and do not interfere with each other. Therefore, it is only necessary to use the time interval method to The second peak is extracted, and the required neutron peak image and value are obtained;

Through ^{the 252} Cf source-driven noise analysis system, the pulsed neutron signals generated by the standard nuclear material at different thicknesses are obtained and time-domain correlation analysis is performed to generate a series of neutron peak calibration data as a reference standard and calibration for obtaining material thickness part of the database;

Accurately locate the time when the neutron peak point appears, compare it with the material thickness calibration database, and obtain the thickness of the unknown nuclear component material;

(2) Elimination of non-nuclear materials and preliminary judgment on the material of nuclear components: use the neutron peak ratio method to eliminate non-nuclear materials and preliminarily judge the material of nuclear components;

Based on the experimental method of nuclear physics, starting from the source fission event, through probabilistic events such as fission event conditional probability and fission chain related conditional probability, the following conclusions are inferred:

a. The count value of gamma rays is affected by both the thickness of the material and the atomic number of the material, and is related to the atomic number when the thickness is similar;

b. The count value of the neutron peak varies with the change of material and thickness, but due to the presence of induced fission neutrons in nuclear materials, its falling edge is significantly slower than that of non-nuclear materials, that is, the neutron count on the falling edge accounts for the total count Correspondingly, due to the absence of induced fission neutrons, the falling edge of non-nuclear materials does not have the above characteristics;

The neutron peaks extracted from the source-detector cross-covariance function are used for rising/falling edge analysis, and a new pattern recognition parameter r is introduced here, which is the proportion of pulse rising/falling edges:

In the formula is the rising edge integral value, is the falling edge integral value, and Indicates the time interval of the rising edge of the neutron peak, Indicates the time interval of the falling edge of the neutron peak, is the cross-covariance function of the source channel and the detector channel;

According to the relevant principles of nuclear physics and particle transport methods, the integral intervals of the neutron peak ratio method are determined. Considering that the induced fission neutrons are dominant within a certain time interval after the neutron peak, the time starting point of the falling edge is taken as 10 ns after the peak point, and the integration interval is 80 ns. Similarly, the interval of the rising edge is set as 4ns, and the end point is set as 2ns before the neutron peak point. If it is assumed that the peak time point is t ₀ , the above formula can be expressed as:

Carry out time-domain correlation analysis on the pulsed neutron signals of standard nuclear components of different materials, and obtain the obvious difference in the neutron peak ratio r , classify according to the r value, and use it as a part of the calibration database;

Compare the calculated r value with the calibration database, eliminate the interference of non-nuclear materials, and preliminarily judge the material of nuclear components;

(3) Clarify the material of nuclear components and strengthen the identification results: use the gamma peak counting method to further identify the subtle differences in materials that cannot be accurately distinguished in the first two parts, and strengthen the identification effect;

According to the material thickness identified in step (1), the cross-covariance functions of materials with different thicknesses are classified, and the count value of the γ peak point is extracted from it;

Carry out time-domain correlation analysis on the pulsed neutron signals of standard nuclear components with the same thickness and different materials, and obtain the obvious difference in the count value of the gamma peak point, classify according to the count value, and use it as a part of the calibration database;

Compare the count value of the measured γ peak point with the calibration database to obtain the material information of the nuclear components;

Verify the material information with the information obtained in step (2), and further distinguish different types and thicknesses of core components.

The signal acquisition frequency may be 1 GHz, and of course it may also be 500 MHz or other frequencies.

The length of the data block is 1024 bytes, and of course other lengths such as 256, 512, 2048, etc. can also be used.

Compared with the prior art, the present invention has the following beneficial effects:

1. The traditional NMIS cannot know whether the unknown component is a nuclear component, whether it is worth measuring, and which calibration database should be compared with it, but only mechanically "takes" it as a signal generated by nuclear materials, and applies a fixed time interval for subsequent identification and judgment. Therefore, the pre-discrimination mechanism and method first adopted in the present invention eliminates the interference of non-nuclear materials without increasing the burden on the NMIS system and without changing the system structure.

2. Due to the identification and interpretation of the position of the neutron peak in the time-domain correlation function between the source and the detector, NMIS can extract a new feature of the geometric thickness of the unknown nuclear component material from the traditional feature label, which is different from the traditional feature value. Value, so as to correspond the geometric thickness with the calibration database, clarify the database comparison object, and reduce the possibility of misjudgment caused by inaccurate database correspondence.

3. Determine the material of unknown nuclear components through the neutron peak proportion method and the gamma peak counting method, thereby reducing the difficulty of subsequent feature extraction and signal processing. In addition to eliminating the interference of non-nuclear materials and obtaining parameters such as the thickness of unknown nuclear components, the invention additionally subdivides the types of nuclear component materials, thereby providing beneficial information for obtaining more accurate and comprehensive inspection results.

Description of drawings

Figure 1 is a schematic diagram of the ²⁵² Cf source driven noise analysis and signal processing system;

Figure 2 is a schematic diagram of the composition of source-detector time-domain correlated signal particles;

Fig. 3 is a comparison diagram of neutron peaks of time-domain correlation functions of different substances and different thicknesses;

Fig. 4 is the flow chart of material pre-discrimination of the present invention;

Fig. 5 is that the embodiment of the present invention thickness is 2cm different material neutron peak appearance time comparison;

Fig. 6 is the comparison of the neutron peak appearance time of different materials with a thickness of 0.95cm according to the embodiment of the present invention;

Fig. 7 is the comparison of the neutron peak appearance time of different materials with a thickness of 0.45cm according to the embodiment of the present invention;

Fig. 8 is a histogram of comparison of final recognition results according to an embodiment of the present invention.

Detailed ways

The present invention will be described in further detail through the following examples, but the present invention is not limited to the scope of the examples.

See Fig. 4, a pre-discrimination method for material characteristics of nuclear components based on ²⁵² Cf source drive: first, use ²⁵² Cf source drive noise analysis method and a series of standard materials to establish the neutron peak position, neutron peak ratio and gamma peak count value calibration database; then, use the ²⁵² Cf source driven noise analysis method to extract the neutron peak position, neutron peak ratio and gamma peak count value; finally, the extracted three values are compared with the The above calibration database is compared and calculated to obtain the material characteristic parameters of the parts to be tested;

The material characteristic parameters include thickness interpretation, non-nuclear material rejection, material interpretation and nuclear component material identification, and the specific steps are:

(1) Acquisition of thickness information: The thickness information of nuclear components is obtained by analyzing the neutron peak time point of the source-detector time domain correlation function. (See Figure 1) The spontaneous fission source ²⁵² Cf is used as the driving neutron source (#1 channel), and the spontaneous fission neutrons emitted by it are incident on the nuclear component to be tested and induce a chain reaction; the chain reaction will generate several neutrons and Gamma rays are captured by the detectors of channels #2 and #3; ²⁵² Cf sources emit 4 spontaneous fission neutrons and 6 gamma photons per fission on average. Spontaneous fission neutrons enter the nuclear material to be tested to induce a chain reaction to produce a series of induced fission neutrons, that is, after the nuclear material to be tested is "stimulated", the neutrons carrying its relevant information are detected by the detector, and a multi-channel pulse is obtained A neutron signal; a pulsed neutron signal obtained by a source-driven method; collecting neutron pulse data to obtain a time distribution of neutron detection counts produced by a neutron source and an unknown nuclear material excited by the neutron source, the time The form of the distribution is a neutron pulse sequence composed of "0" and "1"; then the collected neutron pulse sequence is subjected to a covariance operation and a cross-covariance function is obtained; the neutron peak point of the cross-covariance function is extracted position to obtain thickness information.

(2) Elimination of non-nuclear materials and preliminary judgment on the material of nuclear components: use the neutron peak ratio method to eliminate non-nuclear materials and preliminarily judge the material of nuclear components. According to the following formula:

In the formula is the rising edge integral value, is the falling edge integral value, and Indicates the time interval of the rising edge of the neutron peak, Indicates the time interval of the falling edge of the neutron peak, is the cross-covariance function of the source and detector channels.

Carry out time-domain correlation analysis on the pulsed neutron signals of standard nuclear components of different materials, and obtain the obvious difference in the neutron peak ratio r , classify according to the r value, and use it as a part of the calibration database;

Compare the calculated r value with the calibration database, eliminate the interference of non-nuclear materials, and preliminarily judge the material of nuclear components;

(3) Clarification of the material of nuclear components and enhancement of identification results: further identify the nuances of materials that were not accurately distinguished in the first two parts through the gamma peak counting method, and strengthen the identification effect.

According to the material thickness identified in step (1), the cross-covariance functions of materials with different thicknesses are classified, and the count value of the γ peak point is extracted from it;

Carry out time-domain correlation analysis on the pulsed neutron signals of standard nuclear components with the same thickness and different materials, and obtain the obvious difference in the count value of the gamma peak point, classify according to the count value, and use it as a part of the calibration database;

Compare the count value of the measured γ peak point with the calibration database to obtain the material information of the nuclear components;

Verify the material information with the information obtained in step (2), and further distinguish different types and thicknesses of core components.

The signal acquisition frequency may be 1 GHz, and of course it may also be 500 MHz or other frequencies.

The length of the data block is 1024 bytes, and of course other lengths such as 256, 512, 2048, etc. can also be used.

In terms of the enrichment degree example in this embodiment, a circular casting is used as the unknown core part, and its materials are Fe, Au and U with enrichment degrees of 90.15% and 93.15%, respectively, and the thicknesses are 0.45cm, 0.95cm and 2cm respectively. The process mentioned in step (1) is shown in Figure 4. The obtained source-detector time-domain correlation function is identified for neutron peak position, and then the distinction of material correlation function images with different thicknesses is obtained, as shown in Fig. 5, Fig. 6 and Fig. 7. The recognition results are shown in Table 1.

1 Comparison of neutron peak time points in different materials:

The step (2) is carried out as follows: a. Perform neutron peak extraction and analysis on the source-detector time-domain correlation functions of 12 experimental samples of different materials and thicknesses; b. Calculate neutron peaks between different experimental samples The proportion of neutron peaks, and compare the differences between them, so as to eliminate non-nuclear materials; the proportion of neutron peaks in different materials and thicknesses is shown in Table 2.

Table 2 Comparison of neutron peak ratio r in different materials and thicknesses:

Afterwards, according to the principle shown in step (3), the discriminant analysis of γ peak counts for materials with the same thickness was carried out, and the results are shown in Table 3. Then compare it with the calibration data to obtain the material of unknown nuclear components and strengthen the results of the first two steps. Shown in Fig. 8 is the comparison histogram of the final recognition results.

Table 3 Comparison of γ peak counts of different materials:

After the above calculation is completed, the thickness feature information obtained in step 1 is fused with the material feature information obtained in steps 2 and 3 to obtain a pre-discrimination result that fully reflects the properties of unknown nuclear components.

Finally, it is noted that the above embodiments are only used to illustrate the technical solutions of the present invention without limitation. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be carried out Modifications or equivalent replacements without departing from the spirit and scope of the technical solution of the present invention shall be covered by the claims of the present invention.

Claims (1)

1. one kind based on ²⁵²the pre-method of discrimination of nuclear component material characteristics that Cf source drives, is characterized in that, first, adopts ²⁵²cf source drives noise analysis method and series of standards material to set up neutron peak position, the neutron peak when count value nominal data storehouse, γ peak that can embody material thickness and material; Then, adopt ²⁵²cf source drives noise analysis method to carry out the when γ peak count value extraction of neutron peak position, neutron peak to UUT; Finally, three numerical value extracted and described nominal data storehouse are compared and calculate part materials characteristic parameter;

Described material characteristic parameters comprises thickness interpretation, the rejecting of non-nuclear material, material interpretation and recognition result and strengthens; Wherein:

(1) material thickness interpretation: obtain nuclear component thickness information by source-detector time domain correlation function neutron peak time point analysis;

1. its source-detector cross covariance function is calculated by the pulsed neutron signal of the nuclear material obtained the nuclear weapon checking system based on source drive-type noise analysis approach; The time-domain pulse train signal collected by detector is divided into k block, and i is point sometime on time shaft in a certain blocks of data, any computing formula of unbiased esti-mator of the pulsed neutron source-detector cross covariance function of employing is:

${\mathrm{Cov}}_k\left(\mathrm{\τ}\right)=\frac{1}{M}\underset{i}{\mathrm{\Σ}}{x}_k^{*}\left(i\right)y_k(i+\mathrm{\τ})$

Thus the unbiased esti-mator of whole cross covariance signal is:

$\mathrm{Cov}\left(\mathrm{\τ}\right)=\frac{1}{K}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}{\mathrm{Cov}}_k\left(\mathrm{\τ}\right)$

Wherein represent a certain road signal x _kcomplex conjugate, y _kit is then another road detector signal; M is data block length, represents with discrete counting, and because sampling interval is 1ns, then dot spacing is 1ns; τ is time delay; And K is total number of data block;

2. the neutron peak of Cov (τ) is extracted: Cov (τ) comprises neutron peak and two, γ peak part; Utilize time interval method to be extracted at second peak in Cov (τ), namely obtain required neutron peak image and value;

3. pass through ²⁵²cf source drive-type noise analysis system obtains its pulsed neutron signal produced under different-thickness for standard nuclear material and carries out correlation analysis in time domain, produce a series of neutron peaks nominal data, as the acquisition reference standard of material thickness and the part in nominal data storehouse;

The time that the neutron peak point of 4. accurately locating nuclear component occurs, by itself and the storehouse comparison of material thickness nominal data, obtain the thickness of nuclear component material;

(2) non-nuclear material is rejected and nuclear component material tentatively judges: by following formula, be namely called that neutron peak accounting method rejects non-nuclear material and the material of preliminary interpretation nuclear component;

$r=\frac{G_u}{G_d}=\frac{{\∫}_{{\mathrm{\τ}}_1}{\mathrm{Cov}}_{12}\left(\mathrm{\τ}\right)\mathrm{d\τ}}{{\∫}_{{\mathrm{\τ}}_2}{\mathrm{Cov}}_{12}\left(\mathrm{\τ}\right)\mathrm{d\τ}}$

G in formula _urising edge integrated value, G _dfor negative edge integrated value, and τ ₁represent the time interval of neutron peak rising edge, τ ₂represent neutron peak negative edge time interval, Cov ₁₂(τ) be the cross covariance function of source channels and probe access;

Standard nuclear component pulsed neutron signal for unlike material carries out correlation analysis in time domain, show that it accounts for ratio r at neutron peak, classifies according to r value, as the part in nominal data storehouse;

Gained r value and the comparison of nominal data storehouse will be calculated, reject the interference of non-nuclear material, and preliminary interpretation nuclear component material;

(3) the clear and definite and recognition result of nuclear component material is strengthened: identify by γ peak counting diagnostic method the material nuance that first two steps fail accurately to distinguish further, and strengthen recognition effect;

1. the cross covariance function of different-thickness material is classified, and therefrom extract the count value of γ peak point;

2. the standard nuclear component pulsed neutron signal for same thickness unlike material carries out correlation analysis in time domain, draws its count value at γ peak point, classifies, as the part in nominal data storehouse according to count value;

3. the count value of actual measurement γ peak point and nominal data storehouse are compared, obtain nuclear component material information;

4. the information obtained in material information and step (2) is verified, distinguish dissimilar, thickness nuclear component further.