Method and system for rapid estimation of residual nuclide concentration for radioactive liquid waste

Abstract

The present application relates to the technical field of concentration estimation, and discloses a method and system for quickly estimating the residual nuclide concentration of radioactive waste liquid, which comprises the following steps: obtaining the absorbance values of the radioactive waste liquid under multiple wavelength channels in the ultraviolet-visible continuous wavelength range, calculating the multiple absorbance values to obtain a hyperspectral absorbance vector representing the absorption characteristics of the radioactive waste liquid in the full spectral range; by generating the hyperspectral absorbance vector, combining spectral decomposition, correction and other steps, the matrix interference spectrum of the iron and chromium ion interference spectral characteristics can be accurately separated from the residual characteristic spectrum of the residual nuclide absorption spectral characteristics, the dynamically changing matrix color interference can be effectively stripped, and through the correction of multiple parameters such as spectral fidelity coefficient and concentration mapping confidence, the problem of false high absorbance caused by the fluctuation of iron ion concentration due to the change of waste liquid batch can be avoided, and the accuracy and stability of the residual nuclide concentration estimation are improved.

Description

Technical Field

[0001] This invention relates to the field of concentration estimation technology, and more specifically to a method and system for rapid estimation of residual nuclide concentration in radioactive waste liquid. Background Technology

[0002] Currently, when rapidly estimating the concentration of residual nuclides in radioactive waste liquids, spectrophotometry or colorimetric analysis is often used. This method involves reacting a chromogenic agent with a specific nuclide (such as uranium or plutonium) to form a complex, and then measuring the absorbance to estimate the concentration of the residual nuclide. This method has the advantages of being simple to operate and having a fast response.

[0003] However, the above estimation method still has the following drawbacks: For highly acidic solutions, there are high concentrations of corrosion product ions such as iron and chromium. These ions themselves have strong characteristic colors, which will overlap with the spectral products of the chromogenic reagent of the target nuclide. In order to pursue rapid detection, the existing estimation methods usually use single-wavelength or dual-wavelength measurement, which makes it difficult to effectively remove the interference of this dynamically changing matrix color. When the iron ion concentration fluctuates due to batch changes in waste liquid, the background absorption will be directly superimposed on the target signal, resulting in an artificially high absorbance reading, which causes the estimation of the concentration of residual nuclide to be inaccurate. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for rapid estimation of residual nuclide concentration in radioactive waste liquids, thus solving the aforementioned problems.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0006] Rapid methods for estimating residual nuclide concentrations in radioactive waste liquids include:

[0007] Step S1: Obtain the absorbance values ​​of radioactive waste liquid in multiple wavelength channels within the ultraviolet-visible continuous wavelength range, calculate the multiple absorbance values, and obtain the hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid in the full spectrum range.

[0008] Step S2: Decompose the hyperspectral absorbance vector to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, generate the initial concentration coefficient vector of the initial contribution degree of each spectral component.

[0009] Step S3: Calculate the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum. Correct the initial concentration coefficient vector according to the spectral overlap factor to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction.

[0010] Step S4: Calculate the residual nuclide concentration coefficient to obtain an estimated value of the residual nuclide concentration in the radioactive waste liquid.

[0011] Furthermore, multiple absorbance values ​​were calculated to obtain a hyperspectral absorbance vector representing the absorption characteristics of the radioactive waste liquid across the entire spectral range, including:

[0012] The absorbance values ​​under multiple wavelength channels are decomposed to obtain a sequence of singular values. The singular value sequence is then calculated to generate a spectral singular entropy that represents the complexity and noise level of the spectral data.

[0013] The truncation threshold is determined based on the spectral singular entropy. The singular value sequence is truncated using the truncation threshold to obtain the absorbance row vector. The absorbance row vector and absorbance value are analyzed to generate a spectral fidelity coefficient that represents the degree of signal preservation during the denoising process.

[0014] Furthermore, multiple absorbance values ​​are calculated to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectral range, which also includes:

[0015] The absorbance vector is corrected based on the spectral fidelity coefficient to obtain the hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectral range.

[0016] Furthermore, the hyperspectral absorbance vector is decomposed to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. Simultaneously, an initial concentration coefficient vector representing the initial contribution level of each spectral component is generated, including:

[0017] The hyperspectral absorbance vector is calculated to obtain the spectral curvature feature matrix representing the characteristics of absorption curvature variation between adjacent wavelengths;

[0018] The spectral curvature feature matrix is ​​screened, and the wavelength positions corresponding to the curvature extrema are extracted as feature bands to generate feature band marker vectors representing key waveforms of interference and residual spectra.

[0019] The characteristic band label vector and hyperspectral absorbance vector are analyzed to construct an initial spectral matrix. The initial spectral matrix is ​​then decomposed to obtain the initial spectral characteristic basis representing the basic spectrum initially separated based on the characteristic bands.

[0020] Furthermore, the hyperspectral absorbance vector is decomposed to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. Simultaneously, an initial concentration coefficient vector representing the initial contribution level of each spectral component is generated, which also includes:

[0021] Using the initial spectral feature basis as the starting point, the hyperspectral absorbance vector is iterated to generate a spectral convergence coefficient that represents the stability of the iteration.

[0022] The spectral convergence coefficients are analyzed to generate the final spectral base. The final spectral base includes the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, the hyperspectral absorbance vector is calculated based on the final spectral base to generate the initial concentration coefficient vector representing the initial contribution of each spectral component.

[0023] Furthermore, the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum is calculated. The initial concentration coefficient vector is then corrected based on the spectral overlap factor to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction. This coefficient includes:

[0024] The similarity between the residual characteristic spectrum and the matrix interference spectrum is calculated to generate a spectral overlap factor that represents the overall degree of overlap between the two in terms of waveform and intensity.

[0025] The spectral overlap factor is analyzed to generate an interference contribution coefficient representing the degree of interference component encroachment on the residual signal.

[0026] Furthermore, the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum is calculated. The initial concentration coefficient vector is then corrected based on the spectral overlap factor to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction. This also includes:

[0027] Based on the interference contribution coefficient, the interference contribution of the initial concentration coefficient vector is removed, and the transition concentration coefficient vector representing the residual contribution after the interference is removed is obtained.

[0028] The transition concentration coefficient vector is corrected to generate a residual nuclide concentration coefficient that reflects the degree of contribution of the residual nuclide after interference correction.

[0029] Furthermore, the concentration coefficient of residual nuclides is calculated to obtain an estimated concentration of residual nuclides in the radioactive waste liquid, including:

[0030] Obtain the standard response curve established for the residual nuclide, substitute the residual nuclide concentration coefficient into the standard response curve for mapping transformation to obtain the initial concentration mapping value, calculate the deviation between the residual nuclide concentration coefficient and the standard response curve, and generate the concentration mapping confidence level that represents the reliability of the current mapping relationship.

[0031] Furthermore, the concentration coefficient of residual nuclides is calculated to obtain an estimated concentration of residual nuclides in the radioactive waste liquid, which also includes:

[0032] The initial concentration mapping value is corrected based on the concentration mapping confidence level to generate a concentration estimate representing the concentration of residual nuclides in the radioactive waste liquid.

[0033] Furthermore, a rapid estimation system for residual nuclide concentrations in radioactive waste liquids, applied to the aforementioned estimation method, includes:

[0034] The analysis unit is used to obtain the absorbance values ​​of radioactive waste liquid in multiple wavelength channels within the ultraviolet-visible continuous wavelength range, and to calculate the multiple absorbance values ​​to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid in the full spectrum range.

[0035] The decomposition unit is used to decompose the hyperspectral absorbance vector to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, it generates the initial concentration coefficient vector of the initial contribution degree of each spectral component.

[0036] The correction unit is used to calculate the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum. Based on the spectral overlap factor, the initial concentration coefficient vector is corrected to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction.

[0037] The estimation unit is used to calculate the residual nuclide concentration coefficient and to calculate the estimated concentration of residual nuclide in radioactive waste liquid.

[0038] In summary, the present invention has the following main beneficial effects:

[0039] By acquiring the absorbance values ​​of radioactive waste liquid in multiple wavelength channels within the continuous ultraviolet-visible wavelength range, a hyperspectral absorbance vector that comprehensively represents the absorption characteristics of radioactive waste liquid across the entire spectrum is generated. This avoids the limitations of existing methods that rely on single-wavelength or dual-wavelength measurements. Combined with key steps such as spectral decomposition and spectral overlap factor correction, the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ions can be accurately separated from the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. This effectively removes interference from dynamically changing matrix colors. Furthermore, through the coordinated correction of multiple parameters such as spectral fidelity coefficient, spectral convergence coefficient, and concentration mapping confidence level, the problem of artificially high absorbance readings caused by iron ion concentration fluctuations due to batch variations in waste liquid is effectively avoided. This improves the accuracy and stability of residual nuclide concentration estimation and enhances estimation precision. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating the steps of the rapid estimation method for residual nuclide concentration in radioactive waste liquid according to the present invention.

[0041] Figure 2This is a schematic diagram of the rapid estimation system for residual nuclide concentration in radioactive waste liquid according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] refer to Figure 1 and Figure 2 Methods for rapid estimation of residual nuclide concentrations in radioactive waste liquids include:

[0044] Step S1: Obtain the absorbance values ​​of radioactive waste liquid in multiple wavelength channels within the ultraviolet-visible continuous wavelength range, calculate the multiple absorbance values, and obtain the hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid in the full spectrum range.

[0045] Step S2: Decompose the hyperspectral absorbance vector to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, generate the initial concentration coefficient vector of the initial contribution degree of each spectral component.

[0046] Step S3: Calculate the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum. Correct the initial concentration coefficient vector according to the spectral overlap factor to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction.

[0047] Step S4: Calculate the residual nuclide concentration coefficient to obtain an estimated value of the residual nuclide concentration in the radioactive waste liquid.

[0048] In one embodiment, multiple absorbance values ​​are calculated to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectral range, including:

[0049] The absorbance values ​​of multiple wavelength channels are decomposed to obtain a singular value sequence. This singular value sequence is then calculated to generate a spectral singular entropy representing the complexity and noise level of the spectral data. Specifically, this involves: arranging the absorbance values ​​of multiple wavelength channels in ascending order of wavelength to form an absorbance vector; setting the delay step size as the sampling interval between adjacent wavelength channels; setting the embedding dimension as the length of the subsequence truncated from the original sequence; starting from the first element of the absorbance vector, continuously truncating subsequences of length equal to the embedding dimension as the first row of the matrix; then moving the starting point backward by one delay step and truncating subsequences of the same length as the second row; repeating this process until no complete subsequence can be truncated, thus obtaining a two-dimensional matrix, the trajectory matrix. Singular value decomposition is performed on the trajectory matrix to obtain a set of non-negative singular values ​​and corresponding left and right singular matrices. This set of non-negative singular values ​​is then arranged in descending order to form a singular value sequence.

[0050] Normalize the sequence of singular values ​​so that the sum of all singular values ​​is 1 to obtain the probability distribution of each singular value. Calculate the entropy of this probability distribution, which is the sum of the negative products of each probability value and its natural logarithm. This entropy represents the complexity and noise level of the spectral data. The smaller the spectral singular entropy, the simpler the spectral components and the higher the signal-to-noise ratio, and vice versa.

[0051] The truncation threshold is determined based on the spectral singular entropy. The singular value sequence is truncated using the truncation threshold to obtain the absorbance row vector. The absorbance row vector and absorbance values ​​are analyzed to generate a spectral fidelity coefficient representing the degree of signal preservation during the denoising process. Specifically, this includes: calculating the proportion of the sum of the first N singular values ​​in the singular value sequence to the total sum of all singular values ​​to obtain the cumulative contribution ratio sequence, where N is the number of singular values; dividing the spectral singular entropy by the natural logarithm of N and normalizing the calculation result to the 0-1 interval to obtain the relative entropy value; setting the truncation position to the smallest integer position that satisfies the cumulative contribution ratio ≥ 1 minus the relative entropy value; retaining the singular values ​​before the truncation position; and setting all singular values ​​after the truncation position to 0.

[0052] The trajectory matrix is ​​reconstructed based on the retained singular values ​​and the corresponding left and right singular vectors. The reconstructed trajectory matrix is ​​then subjected to diagonal averaging, which involves calculating the arithmetic mean of all elements on each anti-diagonal line in the reconstructed trajectory matrix. These averages are then arranged in anti-diagonal index order to obtain the denoised absorbance row vector. Each element in the absorbance row vector corresponds to the absorbance value of each original wavelength channel.

[0053] Calculate the difference between the original absorbance value and the absorbance row vector: Calculate the square of the difference between the original absorbance value and the corresponding absorbance value in the absorbance row vector for each wavelength channel. Calculate the arithmetic mean of these squared values ​​for all wavelength channels, and then take the square root of this mean to obtain the root mean square difference. Calculate the difference between the maximum and minimum values ​​of the original absorbance values ​​to obtain the absorbance range. If the absorbance range is 0, it indicates that there is no spectral fluctuation, and the difference value is taken as 0; otherwise, divide the root mean square difference by the absorbance range and normalize the calculation result to the 0-1 interval to obtain the difference value. Subtract the difference value from 1 to obtain the spectral fidelity coefficient, which represents the degree of signal preservation during the denoising process. The closer the spectral fidelity coefficient is to 1, the more complete the signal preservation during the denoising process.

[0054] In one embodiment, the calculation of multiple absorbance values ​​to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the full spectral range further includes:

[0055] The absorbance row vector is corrected based on the spectral fidelity coefficient to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectral range. Specifically, this involves: for each wavelength channel in the absorbance row vector, obtaining the denoised absorbance value and the original absorbance value for that wavelength channel; multiplying the spectral fidelity coefficient by the denoised absorbance value to obtain the first product term; subtracting the spectral fidelity coefficient from 1 and then multiplying it by the original absorbance value to obtain the second product term; adding the first product term and the second product term to obtain the corrected absorbance value for that channel, and thus obtaining the corrected absorbance value for each wavelength channel.

[0056] These corrected absorbance values ​​are arranged in ascending order of wavelength to form a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectrum.

[0057] By calculating the absorbance values ​​of multiple wavelength channels, a hyperspectral absorbance vector is generated, effectively solving the shortcomings of existing single-wavelength and dual-wavelength measurements that cannot remove matrix color interference. By constructing a trajectory matrix and performing singular value decomposition to obtain spectral singular entropy, the complexity and noise level of spectral data can be accurately represented. Combined with the spectral singular entropy, a cutoff threshold is determined to achieve absorbance denoising. Then, the spectral fidelity coefficient reflects the degree of signal preservation, ensuring the integrity of the signal after denoising. This effectively counteracts the spectral overlap interference of high-concentration iron, chromium and other corrosion product ions in high-acidity waste liquid, avoids the falsely high absorbance caused by background absorption superposition, reduces the estimation deviation of residual nuclide concentration, and improves the accuracy of rapid estimation of residual nuclide concentration in radioactive waste liquid.

[0058] In one embodiment, the hyperspectral absorbance vector is decomposed to obtain a matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and a residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. Simultaneously, an initial concentration coefficient vector representing the initial contribution level of each spectral component is generated, including:

[0059] The hyperspectral absorbance vector is calculated to obtain a spectral curvature feature matrix representing the absorption curvature variation characteristics between adjacent wavelengths. Specifically, this involves: setting the length of a sliding window, which is less than the total number of wavelength channels and is an odd number; starting from the third wavelength channel and ending at the third-to-last wavelength channel, each wavelength channel is sequentially used as a center point. At each center point, a local window sequence is constructed using the absorbance values ​​within the range of that center point and half the window length to its left and right. This local window absorbance sequence is then fitted with a quadratic polynomial to obtain the quadratic coefficients of the fitted polynomial. Twice these quadratic coefficients are used as the absorption curvature value at that center point.

[0060] After traversing all center points, the absorption curvature values ​​calculated at each center point are arranged in order of the corresponding wavelength channel index to form a curvature row vector.

[0061] The hyperspectral absorbance vector is subjected to first-order difference calculation, which involves subtracting the absorbance value of the previous wavelength channel from the absorbance value of the next wavelength channel to obtain the absorbance change rate, thus yielding a first-order difference row vector. The elements in the first-order difference row vector are the absorbance change rates. The length of the first-order difference row vector is one less than the original number of wavelengths. To ensure that the first-order difference row vector corresponds to the curvature row vector in the same wavelength channel, the first-order difference row vector is regarded as the change rate at the next wavelength, and a 0 is added to the beginning of the first-order difference row vector to indicate that there is no change rate at the first wavelength, resulting in a first-order difference row vector of the same length as the curvature row vector. The curvature row vector is used as the first row, and the first-order difference row vector with the 0 padding is used as the second row, together forming a two-row matrix. Each column of this matrix corresponds to a wavelength channel, storing the absorption curvature value and the corresponding absorbance change rate for that channel. The absorbance change rate for the first channel is 0, resulting in a spectral curvature feature matrix representing the absorption curvature change characteristics between adjacent wavelengths.

[0062] The spectral curvature feature matrix is ​​filtered, and the wavelength positions corresponding to the curvature extrema are extracted as feature bands. Feature band marker vectors representing key waveforms of interference and residual spectra are generated. Specifically, this includes: creating an empty candidate position list; for each internal wavelength channel of the curvature row vector in the spectral curvature feature matrix (i.e., from the second channel to the penultimate channel), the following judgments are performed sequentially: if the absorption curvature value of the channel is greater than the absorption curvature values ​​of both the preceding and following channels, then the index of the channel and its absorption curvature value are recorded in the candidate position list; if the absorption curvature value of the channel is less than the absorption curvature values ​​of both the preceding and following channels, then the index of the channel and its absorption curvature value are recorded in the candidate position list; if neither of the above conditions is met, the judgment continues for the next channel.

[0063] After traversing all internal channels, check if the candidate position list is empty. If the list is not empty, calculate the absolute value of all absorption curvature values ​​in the list, and then calculate the arithmetic mean of these absolute values ​​to obtain the extreme value amplitude threshold. Compare the absolute value of each absorption curvature value in the candidate position list with the extreme value amplitude threshold to filter out candidate positions whose absolute values ​​are greater than the extreme value amplitude threshold. These positions are the finally determined characteristic band positions.

[0064] If the candidate position list is empty, it means that there are no local maxima or minima in the current spectral curvature row vector. In this case, the mean of the entire curvature row vector is calculated, and the indices of all channels with absorption curvature values ​​greater than the mean are taken as feature band positions. A zero vector with a length equal to the number of original wavelength channels is created, and the elements at the determined feature band positions are assigned a value of 1, while the other positions are kept at 0. This generates a feature band marker vector representing the key waveforms of the interference and residual spectra. The wavelength channels corresponding to the elements with a value of 1 in the feature band marker vector are the key band positions where the waveform changes most significantly in the interference and residual spectra.

[0065] The characteristic band label vector and hyperspectral absorbance vector are analyzed to construct an initial spectral matrix. The initial spectral matrix is ​​then decomposed to obtain the initial spectral feature base representing the basic spectrum initially separated from the characteristic bands. Specifically, this includes: extracting the wavelength channel index corresponding to all elements with a value of 1 from the characteristic band label vector to obtain a characteristic band index set; calculating the difference between two adjacent indices in the characteristic band index set to obtain an adjacent characteristic band interval sequence; and taking the minimum value in the adjacent characteristic band interval sequence as the minimum interval.

[0066] The neighborhood radius is set to the minimum interval minus 1, divided by 2 and rounded down. The neighborhood radius can ensure that the neighborhood windows of each feature band do not overlap. For each feature band index, the absorbance values ​​of the index and its left and right neighborhood radii are obtained from the hyperspectral absorbance vector to form a local absorbance sub-vector. The length of the local absorbance sub-vector is twice the neighborhood radius plus 1.

[0067] All local absorbance sub-vectors corresponding to the characteristic bands are stacked in rows, with each local absorbance sub-vector serving as a row of a matrix, forming a two-dimensional matrix with the number of rows equal to the number of characteristic bands and the number of columns equal to the length of the local window. This is the initial spectral matrix. Since the first column vector of the right singular matrix corresponds to the largest singular value in the singular value sequence, representing the most important spectral component in the initial spectral matrix, namely the basic spectral morphology of iron and chromium ion matrix interference, the first column vector of the right singular matrix is ​​extracted as the initial spectral feature basis representing the basic spectrum initially separated from the characteristic bands. Each element in the initial spectral feature basis corresponds to the spectral weight at the relative position within the local window, reflecting the basic spectral morphology initially separated from the characteristic bands.

[0068] In one embodiment, the hyperspectral absorbance vector is decomposed to obtain a matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and a residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. Simultaneously, an initial concentration coefficient vector representing the initial contribution level of each spectral component is generated. The method further includes:

[0069] Using the initial spectral feature base as the starting point, the hyperspectral absorbance vector is iterated to generate a spectral convergence coefficient representing the stability of the iteration. Specifically, the iteration number is set to the number of elements with a value of 1 in the feature band marker vector; the initial spectral feature base is expanded into an initial matrix interference spectrum with a length equal to that of the hyperspectral absorbance vector. The expansion method is to use the initial spectral feature base as a template, fill each feature band position according to the template shape, and fill the remaining positions with 0.

[0070] Using the initial matrix interference spectrum as the starting point of the iteration, the difference between the hyperspectral absorbance vector and the initial matrix interference spectrum is used to obtain the residual vector. The residual vector is then subjected to non-negative constraint processing, that is, all negative values ​​are set to zero, and non-negative values ​​are retained. The processed vector is used as the initial residual feature spectrum. Alternating non-negative least squares iteration is performed with the initial matrix interference spectrum and the initial residual feature spectrum as the starting point. Each iteration includes two steps: First, the current initial matrix interference spectrum is fixed, and the coefficients corresponding to the initial residual feature spectrum are solved using the non-negative least squares method to minimize the sum of squares of the residuals after subtracting the product of the initial matrix interference spectrum and the coefficients from the hyperspectral absorbance vector. The solved coefficients are then multiplied with the initial residual feature spectrum to obtain the updated residual feature spectrum.

[0071] The second step involves fixing the updated residual characteristic spectrum and using the non-negative least squares method to solve for the coefficients corresponding to the initial matrix interference spectrum. This minimizes the sum of squares of the residuals obtained by subtracting the product of the initial residual characteristic spectrum and the coefficients from the hyperspectral absorbance vector. The solved coefficients are then multiplied by the initial matrix interference spectrum to obtain the updated matrix interference spectrum. After one iteration, the difference vector between the updated and unupdated matrix interference spectra is calculated. The Euclidean distance of this difference vector is then calculated and divided by the Euclidean distance of the matrix interference spectrum before the previous iteration to obtain the relative change in this iteration. This iterative process is repeated until the set number of iterations is reached.

[0072] Calculate the arithmetic mean of the relative changes over all iterations, and subtract the arithmetic mean from 1 to obtain the spectral convergence coefficient, which represents the stability of the iteration. A spectral convergence coefficient close to 1 indicates that the changes in the matrix interference spectrum during the iteration process are smaller, and the stability of the iteration process is higher.

[0073] The spectral convergence coefficient is analyzed to generate the final spectral base. The final spectral base includes the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, the hyperspectral absorbance vector is calculated based on the final spectral base to generate the initial concentration coefficient vector representing the initial contribution of each spectral component. Specifically, the standard deviation of the relative change under all iterations is normalized to the 0-1 interval as the termination judgment threshold.

[0074] The spectral convergence coefficient is compared with the termination judgment threshold: if the spectral convergence coefficient is ≥ 1 minus the termination judgment threshold, it is determined that the iteration process has reached a stable state, and the matrix interference spectrum and residual feature spectrum updated in the last iteration are taken as the final spectral basis; if the spectral convergence coefficient is < 1 minus the termination judgment threshold, it is determined that the iteration process has not yet fully converged. At this time, the iteration result that minimizes the sum of squared residuals of the hyperspectral absorbance vector and the sum of the matrix interference spectrum and residual feature spectrum is selected from all iterations, and the corresponding updated matrix interference spectrum and residual feature spectrum are taken as the final spectral basis.

[0075] Using the matrix interference spectrum and residual characteristic spectrum in the final spectral base as two independent variables and the hyperspectral absorbance vector as the dependent variable, the non-negative least squares method is used to solve for the matrix interference coefficient and the residual nuclide coefficient. These two coefficients are arranged in the order of matrix interference spectrum first and residual characteristic spectrum second to form a two-dimensional vector, which is the initial concentration coefficient vector representing the initial contribution of each spectral component. The first element in the initial concentration coefficient vector corresponds to the contribution of the matrix interference component, and the second element corresponds to the contribution of the residual nuclide component.

[0076] By analyzing the hyperspectral absorbance vector, the key waveforms of interference and residual spectra are accurately located, achieving effective separation of matrix interference spectra and residual characteristic spectra. Furthermore, the final spectral base is generated through iterative optimization, and the stability of the iteration is ensured by the spectral convergence coefficient. At the same time, the initial concentration coefficient vector can reflect the contribution of each spectral component. It can accurately remove the dynamic spectral interference of corrosion product ions such as iron and chromium in high-acidity waste liquid, and avoid the limitations of single and dual-wavelength measurements. It also avoids background absorption superposition and falsely high absorbance caused by batch changes in waste liquid, thus improving the accuracy of residual nuclide concentration estimation.

[0077] In one embodiment, the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum is calculated. The initial concentration coefficient vector is then corrected based on the spectral overlap factor to obtain a residual nuclide concentration coefficient reflecting the degree of contribution of the residual nuclide after interference correction. This includes:

[0078] The similarity between the residual characteristic spectrum and the matrix interference spectrum is calculated to generate a spectral overlap factor that represents the overall degree of overlap between the two in terms of waveform and intensity. Specifically, the residual characteristic spectrum and the matrix interference spectrum are treated as two numerical vectors of equal length. The sum of the squares of the absorbance values ​​of all wavelength channels of the residual characteristic spectrum is calculated, and then the square root is taken to obtain the modulus of the residual characteristic spectrum. Similarly, the modulus of the matrix interference spectrum is calculated, and the smaller of the two modulus values ​​is divided by the larger value to obtain the intensity similarity factor. Then, the inner product of the residual characteristic spectrum and the matrix interference spectrum is calculated, that is, the absorbance value of the residual characteristic spectrum and the absorbance value of the matrix interference spectrum under each wavelength channel are multiplied, and the products of all channels are summed to obtain the inner product value.

[0079] Divide the inner product value by the product of the residual characteristic spectral modulus and the matrix interference spectral modulus to obtain the waveform similarity factor. Multiply the waveform similarity factor by the intensity similarity factor and normalize the calculation result to the 0-1 interval to obtain the spectral overlap factor, which represents the degree of overall overlap between the two in terms of waveform and intensity. The larger the value of the spectral overlap factor, the higher the degree of overall overlap between the residual characteristic spectrum and the matrix interference spectrum in terms of waveform and intensity, that is, the more serious the spectral interference between the two.

[0080] Analyzing the spectral overlap factor generates an interference contribution coefficient representing the degree of interference by interfering components on the residual signal. Specifically, this involves: calculating the ratio of the matrix interference coefficient to the sum of the matrix interference coefficient and the residual nuclide coefficient in the initial concentration coefficient vector to obtain the matrix interference percentage; if the sum is zero, the matrix interference percentage is set to zero; multiplying the spectral overlap factor by the matrix interference percentage and normalizing the result to the 0-1 interval to obtain the interference contribution coefficient representing the degree of interference by interfering components on the residual signal. The larger the value of the interference contribution coefficient, the more severe the interference by interfering components on the residual signal.

[0081] In one embodiment, the calculation of the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum, and the correction of the initial concentration coefficient vector based on the spectral overlap factor to obtain the residual nuclide concentration coefficient reflecting the degree of contribution of the residual nuclide after interference correction, further includes:

[0082] Based on the interference contribution coefficient, the interference contribution of the initial concentration coefficient vector is stripped to obtain the transition concentration coefficient vector representing the remaining residual contribution after interference removal. Specifically, this includes: adding the matrix interference coefficient and the residual nuclide coefficient in the initial concentration coefficient vector to obtain the overall coefficient; multiplying the interference contribution coefficient by the overall coefficient to obtain the amount of interference to be stripped; and subtracting this amount of interference from the residual nuclide coefficient to obtain the difference.

[0083] If the difference is greater than zero, the difference is used as the residual nuclide coefficient in the transition concentration coefficient vector; if the difference is less than or equal to zero, the residual nuclide coefficient in the transition concentration coefficient vector is set to 0; at the same time, the matrix interference coefficient is directly used as the matrix interference coefficient in the transition concentration coefficient vector, thus obtaining a two-dimensional vector, which is the transition concentration coefficient vector representing the contribution of the remaining residual after interference removal. The first element in this vector is the matrix interference coefficient, and the second element is the corrected residual nuclide coefficient.

[0084] The transition concentration coefficient vector is corrected to generate a residual nuclide concentration coefficient that reflects the contribution of residual nuclides after interference correction. Specifically, this involves: extracting the matrix interference coefficient and the corrected residual nuclide coefficient from the transition concentration coefficient vector, denoted as the first element and the second element, respectively; obtaining the interference contribution coefficient and the spectral overlap factor, multiplying them to obtain the comprehensive interference intensity, multiplying the comprehensive interference intensity by the first element to obtain the residual interference amount; subtracting the residual interference amount from the second element to obtain the preliminary correction value. If the preliminary correction value is greater than zero, it is used as the residual nuclide concentration coefficient; if the preliminary correction value is less than or equal to zero, the residual nuclide concentration coefficient is set to 0, thereby generating a residual nuclide concentration coefficient that reflects the contribution of residual nuclides after interference correction.

[0085] The initial concentration coefficient vector reflects the contribution of each spectral component, and the spectral overlap factor is calculated to accurately represent the overall overlap between the two. Combined with the interference contribution coefficient, the degree of interference on the residual signal is reflected, and finally the residual nuclide concentration coefficient is obtained. The spectral interference of iron and chromium ions is removed, and the artificially high absorbance caused by batch variation of waste liquid is avoided. This further improves the accuracy of residual nuclide concentration estimation and enhances the detection anti-interference ability and stability.

[0086] In one embodiment, the residual nuclide concentration coefficient is calculated to obtain an estimated concentration of residual nuclides in the radioactive waste liquid, including:

[0087] Obtain a standard response curve established for the residual nuclide, substitute the residual nuclide concentration coefficient into the standard response curve for mapping transformation to obtain the initial concentration mapping value, and calculate the deviation between the residual nuclide concentration coefficient and the standard response curve to generate a concentration mapping confidence level representing the reliability of the current mapping relationship. Specifically, the standard response curve refers to the concentration-absorbance relationship curve established in advance by measuring a residual nuclide standard sample with a known concentration. The absorbance value corresponding to the standard response curve is calculated according to steps S1 to S3 to obtain the standard concentration coefficient. At this time, the standard response curve is the standard concentration coefficient corresponding to the concentration. Substituting the residual nuclide concentration coefficient into the curve, the corresponding initial concentration mapping value can be obtained.

[0088] Calculate the absolute value of the difference between the residual nuclide concentration coefficient and the mean of all standard concentration coefficients, then divide it by the range of the standard concentration coefficients, and normalize the result to the 0-1 interval to obtain the relative deviation. Add 1 to the relative deviation and divide by 2 to obtain the relative uncertainty. Subtract the relative uncertainty from 1 and normalize the result to the 0-1 interval to obtain the concentration mapping confidence level, which represents the reliability of the current mapping relationship. The larger the concentration mapping confidence level, the more reliable the mapping relationship.

[0089] In one embodiment, the calculation of the residual nuclide concentration coefficient to obtain an estimated concentration of residual nuclide in the radioactive waste liquid further includes:

[0090] The initial concentration mapping value is corrected based on the concentration mapping confidence level to generate an estimated concentration value representing the concentration of residual nuclides in the radioactive waste liquid. Specifically, this includes: when establishing the standard response curve, using the concentration values ​​corresponding to the residual nuclide standard samples with known concentrations, i.e., the concentration values ​​of the standard samples themselves, calculating the arithmetic mean of the concentration values ​​of all standard samples to obtain the standard concentration mean; and subtracting the initial concentration mapping value from the standard concentration mean to obtain the concentration deviation.

[0091] Multiply the concentration deviation by 1 and subtract the concentration mapping confidence to obtain the correction offset. Subtract the correction offset from the initial concentration mapping value to obtain the preliminary correction value. If the preliminary correction value is less than 0, set it to zero; otherwise, keep the original value. Use the preliminary correction value as the concentration estimate representing the concentration of residual nuclides in the radioactive waste liquid.

[0092] By calculating the concentration coefficient of residual nuclides and mapping it to the standard response curve, an initial concentration mapping value is obtained. At the same time, the confidence level of the concentration mapping reflects the reliability of the mapping and corrects the initial concentration mapping value, thus obtaining an estimated value of the concentration of residual nuclides in radioactive waste liquid. This effectively eliminates the spectral interference of iron and chromium ions, avoids the limitations of single and dual wavelength measurements, avoids falsely high absorbance, and improves the accuracy of residual nuclide concentration estimation.

[0093] In one embodiment, the rapid estimation system for residual nuclide concentration in radioactive waste liquid, applied to the above-described estimation method, includes:

[0094] The analysis unit is used to obtain the absorbance values ​​of radioactive waste liquid in multiple wavelength channels within the continuous ultraviolet-visible wavelength range, and to calculate the multiple absorbance values ​​to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid in the full spectrum range.

[0095] The decomposition unit is used to decompose the hyperspectral absorbance vector to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, it generates the initial concentration coefficient vector of the initial contribution degree of each spectral component.

[0096] The correction unit is used to calculate the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum. Based on the spectral overlap factor, the initial concentration coefficient vector is corrected to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction.

[0097] The estimation unit is used to calculate the residual nuclide concentration coefficient and to calculate the estimated concentration of residual nuclide in radioactive waste liquid.

[0098] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A rapid method for estimating the residual nuclide concentration in radioactive waste liquid, characterized in that, include: Step S1: Obtain the absorbance values ​​of the radioactive waste liquid in multiple wavelength channels within the continuous ultraviolet-visible wavelength range, calculate the multiple absorbance values ​​to obtain a hyperspectral absorbance vector representing the absorption characteristics of the radioactive waste liquid across the entire spectrum, including: The absorbance values ​​under multiple wavelength channels are decomposed to obtain a sequence of singular values. The singular value sequence is then calculated to generate a spectral singular entropy that represents the complexity and noise level of the spectral data. The truncation threshold is determined based on the spectral singular entropy. The singular value sequence is truncated using the truncation threshold to obtain the absorbance row vector. The absorbance row vector and absorbance value are analyzed to generate a spectral fidelity coefficient that represents the degree of signal preservation during the denoising process. Step S2: Decompose the hyperspectral absorbance vector to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, generate the initial concentration coefficient vector of the initial contribution degree of each spectral component. Step S3: Calculate the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum. Correct the initial concentration coefficient vector according to the spectral overlap factor to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction. Step S4: Calculate the residual nuclide concentration coefficient to obtain an estimated value of the residual nuclide concentration in the radioactive waste liquid.

2. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 1, characterized in that, Calculations were performed on multiple absorbance values ​​to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectral range, which also includes: The absorbance vector is corrected based on the spectral fidelity coefficient to obtain the hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid across the entire spectral range.

3. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 2, characterized in that, The hyperspectral absorbance vector is decomposed to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. Simultaneously, an initial concentration coefficient vector representing the initial contribution level of each spectral component is generated, including: The hyperspectral absorbance vector is calculated to obtain the spectral curvature feature matrix representing the characteristics of absorption curvature variation between adjacent wavelengths; The spectral curvature feature matrix is ​​screened, and the wavelength positions corresponding to the curvature extrema are extracted as feature bands to generate feature band marker vectors representing key waveforms of interference and residual spectra. The characteristic band label vector and hyperspectral absorbance vector are analyzed to construct an initial spectral matrix. The initial spectral matrix is ​​then decomposed to obtain the initial spectral characteristic basis representing the basic spectrum initially separated based on the characteristic bands.

4. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 3, characterized in that, The hyperspectral absorbance vector is decomposed to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. Simultaneously, an initial concentration coefficient vector representing the initial contribution level of each spectral component is generated, which also includes: Using the initial spectral feature basis as the starting point, the hyperspectral absorbance vector is iterated to generate a spectral convergence coefficient that represents the stability of the iteration. The spectral convergence coefficients are analyzed to generate the final spectral base. The final spectral base includes the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, the hyperspectral absorbance vector is calculated based on the final spectral base to generate the initial concentration coefficient vector representing the initial contribution of each spectral component.

5. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 4, characterized in that, The spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum is calculated. The initial concentration coefficient vector is then corrected based on the spectral overlap factor to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction. This includes: The similarity between the residual characteristic spectrum and the matrix interference spectrum is calculated to generate a spectral overlap factor that represents the overall degree of overlap between the two in terms of waveform and intensity. The spectral overlap factor is analyzed to generate an interference contribution coefficient representing the degree of interference component encroachment on the residual signal.

6. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 5, characterized in that, The calculation of the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum, and the correction of the initial concentration coefficient vector based on the spectral overlap factor, yields the residual nuclide concentration coefficient reflecting the degree of contribution of the residual nuclide after interference correction. This also includes: Based on the interference contribution coefficient, the interference contribution of the initial concentration coefficient vector is removed, and the transition concentration coefficient vector representing the residual contribution after the interference is removed is obtained. The transition concentration coefficient vector is corrected to generate a residual nuclide concentration coefficient that reflects the degree of contribution of the residual nuclide after interference correction.

7. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 6, characterized in that, The residual nuclide concentration coefficient is calculated to obtain an estimated value of the residual nuclide concentration in the radioactive waste liquid, including: Obtain the standard response curve established for the residual nuclide, substitute the residual nuclide concentration coefficient into the standard response curve for mapping transformation to obtain the initial concentration mapping value, calculate the deviation between the residual nuclide concentration coefficient and the standard response curve, and generate the concentration mapping confidence level that represents the reliability of the current mapping relationship.

8. The method for rapid estimation of residual nuclide concentration in radioactive waste liquid according to claim 7, characterized in that, The concentration coefficient of residual nuclides is calculated to obtain an estimated concentration of residual nuclides in radioactive waste liquid, which also includes: The initial concentration mapping value is corrected based on the concentration mapping confidence level to generate a concentration estimate representing the concentration of residual nuclides in the radioactive waste liquid.

9. A rapid estimation system for residual nuclide concentration in radioactive waste liquid, applied in the estimation method described in any one of claims 1-8, characterized in that, include: The analysis unit is used to obtain the absorbance values ​​of radioactive waste liquid in multiple wavelength channels within the ultraviolet-visible continuous wavelength range, and to calculate the multiple absorbance values ​​to obtain a hyperspectral absorbance vector representing the absorption characteristics of radioactive waste liquid in the full spectrum range. The decomposition unit is used to decompose the hyperspectral absorbance vector to obtain the matrix interference spectrum reflecting the spectral characteristics of iron and chromium ion interference and the residual characteristic spectrum reflecting the absorption spectral characteristics of residual nuclides. At the same time, it generates the initial concentration coefficient vector of the initial contribution degree of each spectral component. The correction unit is used to calculate the spectral overlap factor between the residual characteristic spectrum and the matrix interference spectrum. Based on the spectral overlap factor, the initial concentration coefficient vector is corrected to obtain the residual nuclide concentration coefficient, which reflects the degree of contribution of the residual nuclide after interference correction. The estimation unit is used to calculate the residual nuclide concentration coefficient and to calculate the estimated concentration of residual nuclide in radioactive waste liquid.

Images