CN115759324A - Method for optimizing monitoring position of environmental aerosol after nuclear accident - Google Patents
Method for optimizing monitoring position of environmental aerosol after nuclear accident Download PDFInfo
- Publication number
- CN115759324A CN115759324A CN202211162817.XA CN202211162817A CN115759324A CN 115759324 A CN115759324 A CN 115759324A CN 202211162817 A CN202211162817 A CN 202211162817A CN 115759324 A CN115759324 A CN 115759324A
- Authority
- CN
- China
- Prior art keywords
- monitoring
- accident
- environmental
- nuclear
- aerosol
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Monitoring And Testing Of Nuclear Reactors (AREA)
Abstract
The invention relates to an environmental aerosol monitoring position optimization method after a nuclear accident, which preliminarily predicts the concentration distribution of radioactive nuclides in the environment by preliminarily estimating a release source item caused by the occurrence of the nuclear facility accident, combining meteorological parameters and topographic parameters of an accident occurrence place and based on an atmospheric diffusion model; correcting the concentration distribution of the radioactive nuclide in the environment, which is obtained based on the atmospheric diffusion model, based on the environmental aerosol monitoring data; and generating an emergency monitoring point position arrangement scheme with sampling representativeness by using a space simulation annealing algorithm so as to optimize sampling distribution of the aerosol monitoring position. By adopting the method disclosed by the invention, the predicted concentration distribution is systematically corrected based on the environmental aerosol monitoring data, the environmental monitoring position is optimized, the accuracy and the real-time performance of the estimation of the distribution of the radioactive nuclides in the environment after the nuclear facility accident occurs are improved, and the requirement of quickly and accurately estimating the environmental influence after the nuclear facility accident is met.
Description
Technical Field
The invention belongs to the field of radiation monitoring and evaluation after nuclear accidents, and particularly relates to an optimization method for an environmental aerosol monitoring position after a nuclear accident.
Background
After an accident occurs to the nuclear facility, the radioactive substances may be released into the external environment, which poses a threat to the environment and personnel safety. The rapid and reasonable outcome evaluation and emergency decision making are of great significance to the reduction of the consequences of the accident and the harm of the accident to personnel and environment.
And the rapid and accurate evaluation of the spatial distribution of the radionuclide in the external environment and the further calculation of the irradiation dose which may result in the evaluation are the basis for making a reasonable decision. Therefore, extensive research is carried out on the estimation aspect of the exposure dose of the personnel in nuclear emergencies at home and abroad, and great progress is made in the aspects of nuclear emergencies source item evaluation, pollution monitoring and dose estimation. The existing research mainly aims at the characteristic of rapid response of nuclear emergencies, and needs further research for improving the accuracy of an evaluation result on the basis of rapid evaluation.
The current radionuclide spatial distribution evaluation technology mainly comprises a predictive evaluation method based on an atmospheric diffusion model and a current state evaluation method based on environmental monitoring data. The predictive evaluation method based on the atmospheric diffusion model is based on known source items or source items predicted by other methods, combines data such as meteorological parameters and terrain parameters, uses diffusion models such as Gaussian models, lagrange models and Euler models to predict radionuclide concentrations at different positions and time points in the external environment, but because the source items of accidents are usually unknown under the actual condition, the predicted source items and the meteorological parameters have larger uncertainty, and the concentration distribution predicted by the algorithm possibly has a certain difference with the actual condition. By arranging the aerosol monitoring points in the external environment for sampling analysis, the radionuclide concentration of the acquired relevant positions can reflect the actual situation in the environment, but the radionuclide concentration cannot reflect the concentration information of the non-monitored positions.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a method for optimizing the monitoring position of the environmental aerosol after a nuclear accident.
In order to achieve the above purposes, the invention adopts the technical scheme that: a method for environmental aerosol monitoring location optimization after a nuclear accident, the method comprising the steps of:
s1, preliminarily estimating a release source item caused by an accident of a nuclear facility;
s2, preliminarily predicting the concentration distribution of the radioactive nuclide in the environment by using the estimated release source item, the meteorological parameters after the accident and the topographic parameters of the accident occurrence place based on an atmospheric diffusion model;
s3, correcting the concentration distribution of the radioactive nuclide in the environment, which is obtained based on the atmospheric diffusion model, based on the environmental aerosol monitoring data;
and S4, generating an emergency monitoring point position arrangement scheme with sampling representativeness by using a space simulated annealing algorithm so as to optimize sampling point arrangement of the actual aerosol monitoring position.
Further, after step S4, the method further comprises the step of:
and S5, correcting the concentration distribution of the radioactive nuclide in the environment by using the environmental aerosol monitoring data after the sampling and the distribution of the optimized actual aerosol monitoring position.
Further, after step S4, the method further comprises the step of:
and based on the corrected concentration distribution of the radioactive nuclide in the environment obtained in the step S5, generating an emergency monitoring point position arrangement scheme with sampling representativeness by using a space simulated annealing algorithm so as to optimize the sampling point arrangement of the actual aerosol monitoring position.
Further, the release source items in step S1 include the species of radionuclide released by the nuclear facility accident, the total release amount, the release amount of each nuclide, and the release duration.
Further, in the step S1, a release source item caused by an accident of the nuclear facility is estimated according to accident condition monitoring data, expert judgment or other source item inversion methods.
Further, the atmospheric diffusion model in step S2 includes a gaussian model, a lagrange model, and an euler model.
Further, in step S3, after a kriging interpolation result is obtained based on the environmental aerosol monitoring data, the obtained residual error term is input into a predicted concentration field based on an atmospheric diffusion model to correct the prediction concentration field.
Further, step S4 includes the following substeps:
s41, setting the regression Kriging variance in the step S3 as a first part of cost function, taking the weighted sum of the areas occupied by the false positive and the false negative of the emergency monitoring point location as a second part of cost function, and carrying out weighted sum on the two cost functions of the first part of cost function and the second part of cost function to construct a cost function so as to simultaneously optimize the accuracy of the accident influence range judgment and the corrected concentration distribution;
and S42, generating a more sampling representative emergency monitoring point position arrangement scheme by using a space simulated annealing algorithm.
Further, step S42 includes the following sub-steps:
a) From a random initial monitoring position S 0 Initially, the associated cost function value C is calculated (S) 0 )；
b) Given position S k Constructing a candidate new monitoring location S by moving the randomly selected monitoring location a distance h k+1 H is randomly chosen and has a length of a random number between zero and the maximum displacement, and a cost function C (S) for the new position is calculated k+1 )；
c) If C (S) k+1 )＜C(S k ) Accepting the new position, otherwise accepting the new position only with a certain probability, if the new position is accepted, then increasing k by 1, using the new position S k+1 As a starting point, go back to step b), if not, use the old position S k ；
d) After a number of iterations, or stopping when other stopping criteria are met, the monitored location that minimizes the cost function value is stored.
The invention has the beneficial technical effects that: by adopting the method for optimizing the environmental aerosol monitoring position after the nuclear accident, disclosed by the invention, a method for correcting the predicted concentration distribution based on the environmental aerosol monitoring data and a method for optimizing the environmental monitoring position are systematically integrated, so that the accuracy and the real-time performance of the evaluation of the distribution of the radioactive nuclides in the environment after the nuclear facility accident occurs are improved, and the requirement of quickly and accurately evaluating the environmental influence after the nuclear facility accident is met.
Drawings
Fig. 1 is a flowchart of an environmental aerosol monitoring location optimization method after a nuclear accident according to an embodiment of the present invention.
Detailed Description
The invention is further described with reference to the following figures and detailed description.
Example one
The embodiment of the invention provides a method for optimizing an environmental aerosol monitoring position after a nuclear accident, which comprises the following steps:
s1, preliminarily estimating release source items caused by accidents according to working conditions of the nuclear facilities when the accidents happen, wherein the release source items mainly comprise radionuclide species released by the accidents of the nuclear facilities, total release amount (Bq), release amount of each nuclide and release duration.
The released source items caused by the accidents can be estimated according to accident condition monitoring data, expert judgment or other source item inversion methods.
S2, preliminarily predicting the concentration distribution of the radionuclide in the environment, namely the radionuclide concentrations (Bq/m 3) at different positions in the environment based on an atmospheric diffusion model by using the estimated release source items caused by the accident, meteorological parameters after the accident occurs and topographic parameters of the accident occurrence place.
Common atmospheric diffusion models comprise a Gaussian model, a Lagrange model, an Euler model and the like, mature atmospheric diffusion simulation programs such as AERMOD, CALPUFF and the like exist, a proper diffusion model can be selected for calculation according to actual conditions, and a specific atmospheric diffusion simulation method is not detailed in the embodiment of the invention.
And S3, correcting the concentration distribution of the radioactive nuclide in the environment, which is obtained based on the atmospheric diffusion model, based on the environmental aerosol monitoring data.
Conventional monitoring equipment is usually arranged around the nuclear facility, environmental aerosol monitoring data can be obtained after the nuclear facility accident, and preliminary prediction radionuclide concentration distribution obtained based on the atmospheric diffusion model in the step S2 is corrected based on a kriging interpolation method.
Kriging interpolation mathematically provides an optimal linear unbiased estimate (a definite value at a point) for the subject under study. It is defined as follows:
wherein z(s) is a target environment variable, namely the concentration of the radionuclide, x(s) is m environment covariates (position coordinates of monitoring points), s = (x, y) represents a two-dimensional space coordinate, beta is a coefficient to be estimated, epsilon(s) is a residual term after regression of the target variable and the covariates, and obeys normal distribution with the mean value of zero. Residual the spatial autocorrelation property of epsilon(s) can be quantitatively expressed by a covariance function or a variogram, subject to the statistical second-order stationary assumption of satisfaction. Writing equation (1) in matrix form:
z(s)＝x′β+ε(s) (2)
the target variable z(s) is composed of two parts, the first part is a regression term or a trend term, and the second part is a residual term. Firstly, assuming that a target variable and a covariate satisfy a certain regression relationship, estimating a regression coefficient beta by adopting a generalized least square method according to known n sample points,
where C is the variance-covariance matrix of the residuals of n × n, and X is the covariate matrix of the sample points of n × (m + 1). Finally, modeling is carried out on the residual error, and a point s to be estimated can be obtained 0 The optimal linear unbiased estimate of (d) is:
wherein x is 0 Vectors formed for values of covariates at points to be estimated, c 0 Vectors formed for the covariance of the sample point and the point to be estimated, C and C 0 Are obtained from the variation plot of ε(s).
Can obtain s 0 The variance of the regression kriging is:
σ 2 (s 0 )＝c(0)-c′ 0 C -1 c 0 +x′ a (X′C -1 X) -1 x a (4)
wherein x is a ＝x 0 -X′C -1 c 0 . Equation (4) can be decomposed into two parts, the first part (the first two terms) being the estimated error variance of the residual and the second part (the last term) being the estimated error variance of the trend term.
After a kriging interpolation result is obtained, the obtained residual error item is input into a predicted concentration field based on an atmospheric diffusion model to be corrected, the finally obtained corrected concentration distribution not only contains the concentration distribution data simulated according to the inversion source item and the meteorological parameters, but also contains the actual monitoring data of the monitoring point position, and the corrected concentration distribution is more in line with the actual situation.
And S4, generating a more sampling representative emergency monitoring point position arrangement scheme by using a space simulated annealing algorithm so as to optimize actual sampling point arrangement and further obtain representative aerosol monitoring data.
And optimizing the position arrangement of the emergency monitoring points. After the nuclear facility has an accident, emergency monitoring equipment can be additionally arranged, and more precise radionuclide concentration distribution conditions can be obtained.
The arrangement of the monitoring point positions can influence the result of correcting the concentration distribution, and a space simulated annealing algorithm is used for generating a representative monitoring position with enough value so as to meet the requirement of correcting the concentration distribution in a large-scale space range and ensure that the corrected concentration distribution can describe the actual concentration distribution condition of all the point positions in the space to the maximum extent.
Step S4 comprises the following two substeps:
s41, constructing a cost function
Part of the purpose of arranging the emergency monitoring points is to obtain a corrected concentration distribution as accurate as possible, and the first part of the cost function can be set to the regression kriging variance in equation (4) to obtain the optimal interpolation result.
For post-accident emergency monitoring, after setting a proper early warning concentration threshold, the position in the target area can be divided into four sub-areas: false positive, false negative, true positive and true negative.
Another part of the objective of choosing the best contingency monitoring point is to minimize the costs associated with false positive and false negative decisions. When the set threshold is exceeded due to a wrong predicted concentration profile, false positive decisions may occur, leading to unnecessary measures, such as evacuating people from a practically safe area. In contrast, the costs associated with false negative decisions are that no measures are taken, which are necessary in fact, and the costs of false negatives may be greater than those of false positives, as this can seriously impact people's health and life. Therefore, another part of the optimization goal for the emergency monitoring point is to minimize the weighted sum of the areas occupied by the false positives and the false negatives, and the second part of the cost function is:
c = α · area (false positive) + (1- α) · area (false negative) (5)
Wherein alpha is a weight factor, and the value of alpha is less than 0.5 because the result of false negative is more serious.
According to the method, two cost functions of a formula (4) and a formula (5) are used as a final cost function after weighted summation, so that the accuracy of concentration distribution after accident influence range judgment and correction is optimized, and the optimal emergency monitoring point position is obtained.
And S42, generating a more sampling representative emergency monitoring point position arrangement scheme by using a space simulated annealing algorithm.
The spatial simulated annealing is the spatial extension of the simulated annealing algorithm, and comprises five main steps:
a) From a (random) initial monitoring position S 0 Initially, the associated cost function value C is calculated (S) 0 )；
b) Given position S k Constructing a candidate new monitoring location S by moving the randomly selected monitoring location a distance h k+1 . The direction of h is randomly chosen and is a random number with a length between zero and the maximum displacement. In the iterative process of the space simulated annealing, the maximum movement amount is gradually reduced;
c) Calculating a cost function C (S) for the new location k+1 ). If C (S) k+1 )＜C(S k ) Then accept the new bitOtherwise, only the new position is accepted with a certain probability (the purpose is to ensure that the algorithm can get rid of the locally optimal solution). If a new position is accepted, then k is increased by 1. With the iteration of the space simulation annealing, the probability of accepting the inferior position is gradually reduced;
d) Returning to step b), if the new location is accepted, the new location S is used k+1 If the old position S is not used as a starting point k ；
e) Stopping after a certain number of iterations, or when other stopping criteria are met. The monitoring position that minimizes the cost function value is stored.
Since the actual values of the concentration data of all locations in the external environment cannot be obtained in practice, many possible realistic situations can be simulated by adding random errors to the predicted concentration distribution based on the atmospheric diffusion model, the cost function value of the selected monitoring location is calculated for a large number of possible realistic situations, and if the number of simulated realities is sufficient, the calculated average cost will approach the expected cost associated with the monitoring location.
And S5, correcting the concentration distribution of the radioactive nuclide in the environment by using the environmental aerosol monitoring data after optimized monitoring distribution.
And arranging environmental aerosol monitoring equipment according to the optimized monitoring distribution scheme obtained in the step S4, obtaining environmental aerosol monitoring data, and correcting the concentration distribution of the radioactive nuclide in the environment by using the environmental aerosol monitoring data after optimized monitoring distribution based on a Krigin interpolation method.
The processes of the step S4 and the step S5 can be iteratively carried out according to actual needs, an environment monitoring point distribution scheme is optimized for multiple times based on a space simulated annealing method, emergency monitoring equipment is arranged, environment aerosol data are obtained, and the monitoring data are used for correcting the concentration distribution field of the radioactive nuclide in the environment based on a Krigin interpolation method, so that the environment aerosol concentration distribution field which is more in line with the actual situation is obtained.
According to the embodiment, the method for optimizing the environmental aerosol monitoring position after the nuclear accident is based on the Krigin interpolation, uses the environmental monitoring data to quickly correct the distribution of the radionuclide concentration field, and is combined with the method for optimizing the environmental monitoring position distribution based on the spatial simulation annealing algorithm, so that the accuracy and timeliness of the prediction of the distribution of the radionuclide environmental concentration after the nuclear facility accident are improved, and more accurate and real-time evaluation after the accident is realized.
The method of the present invention is not limited to the examples described in the specific embodiments, and those skilled in the art can derive other embodiments according to the technical solutions of the present invention, and also belong to the technical innovation scope of the present invention.
Claims (9)
1. A method for environmental aerosol monitoring location optimization after a nuclear accident, the method comprising the steps of:
s1, preliminarily estimating a release source item caused by an accident of a nuclear facility;
s2, preliminarily predicting the concentration distribution of the radioactive nuclide in the environment by using the estimated release source item, the meteorological parameters after the accident and the topographic parameters of the accident occurrence place based on an atmospheric diffusion model;
s3, correcting the concentration distribution of the radioactive nuclide in the environment, which is obtained based on the atmospheric diffusion model, based on the environmental aerosol monitoring data;
and S4, generating an emergency monitoring point position arrangement scheme with sampling representativeness by using a space simulated annealing algorithm so as to optimize sampling point arrangement of the actual aerosol monitoring position.
2. A method for optimizing the location of environmental aerosol monitoring after a nuclear event as set forth in claim 1, wherein after step S4 the method further comprises the steps of:
and S5, correcting the concentration distribution of the radioactive nuclide in the environment by using the environment aerosol monitoring data after the sampling and the stationing of the optimized actual aerosol monitoring position.
3. A method for optimizing a post-nuclear event environmental aerosol monitoring location according to claim 2, wherein after step S4 the method further comprises the steps of:
and based on the corrected concentration distribution of the radioactive nuclide in the environment obtained in the step S5, generating an emergency monitoring point position arrangement scheme with sampling representativeness by using a space simulated annealing algorithm so as to optimize the sampling point arrangement of the actual aerosol monitoring position.
4. A method of optimizing the location of environmental aerosol monitoring after a nuclear accident as set forth in claim 3, wherein: the release source items in the step S1 comprise the radionuclide species released by the nuclear facility accident, the total release amount, the release amount of each nuclide and the release duration.
5. The method for optimizing the monitoring position of the environmental aerosol after the nuclear accident as recited in claim 4, wherein: in the step S1, release source items caused by accidents of nuclear facilities are estimated according to accident condition monitoring data, expert judgment or other source item inversion methods.
6. The method of claim 5 for optimizing the location of environmental aerosol monitoring after a nuclear accident, wherein: the atmospheric diffusion model in the step S2 comprises a Gaussian model, a Lagrange model and an Euler model.
7. The method of claim 6 for optimizing the location of environmental aerosol monitoring after a nuclear accident, wherein: and S3, based on the environmental aerosol monitoring data, after a kriging interpolation result is obtained, inputting the obtained residual error item into a predicted concentration field based on an atmospheric diffusion model to correct the residual error item.
8. The method for optimizing the monitoring position of the environmental aerosol after the nuclear accident according to claim 7, wherein the step S4 comprises the following substeps:
s41, setting the regression Kriging variance in the step S3 as a first part of cost function, taking the weighted sum of the areas occupied by the false positive and the false negative of the emergency monitoring point location as a second part of cost function, and carrying out weighted sum on the two cost functions of the first part of cost function and the second part of cost function to construct a cost function so as to simultaneously optimize the accuracy of the accident influence range judgment and the corrected concentration distribution;
and S42, generating a more sampling representative emergency monitoring point position arrangement scheme by using a space simulated annealing algorithm.
9. The method as claimed in claim 8, wherein the step S42 comprises the following sub-steps:
a) From a random initial monitoring position S 0 Initially, the associated cost function value C is calculated (S) 0 )；
b) Given position S k Constructing a candidate new monitoring location S by moving the randomly selected monitoring location a distance h k+1 H is randomly chosen and has a length of a random number between zero and the maximum displacement, and a cost function C (S) for the new position is calculated k+1 )；
c) If C (S) k+1 )＜C(S k ) Accepting the new position, otherwise accepting the new position only with a certain probability, if the new position is accepted, then increasing k by 1, using the new position S k+1 As a starting point, return to step b), if not use the old position S k ；
d) After a number of iterations, or stopping when other stopping criteria are met, the monitored location that minimizes the cost function value is stored.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211162817.XA CN115759324A (en) | 2022-09-23 | 2022-09-23 | Method for optimizing monitoring position of environmental aerosol after nuclear accident |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211162817.XA CN115759324A (en) | 2022-09-23 | 2022-09-23 | Method for optimizing monitoring position of environmental aerosol after nuclear accident |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115759324A true CN115759324A (en) | 2023-03-07 |
Family
ID=85351839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211162817.XA Pending CN115759324A (en) | 2022-09-23 | 2022-09-23 | Method for optimizing monitoring position of environmental aerosol after nuclear accident |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115759324A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116739151A (en) * | 2023-05-23 | 2023-09-12 | 中国电器科学研究院股份有限公司 | Offshore area salt spray prediction correction and accuracy assessment method |
| CN116976202A (en) * | 2023-07-12 | 2023-10-31 | 清华大学 | Fixed complex source item distribution inversion method and device based on deep neural network |
| CN116973523A (en) * | 2023-09-22 | 2023-10-31 | 深圳市智芯微纳科技有限公司 | MEMS gas sensor array-based gas detection method and system |
| CN116739151B (en) * | 2023-05-23 | 2024-05-31 | 中国电器科学研究院股份有限公司 | Offshore area salt spray prediction correction and accuracy assessment method |
-
2022
- 2022-09-23 CN CN202211162817.XA patent/CN115759324A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116739151A (en) * | 2023-05-23 | 2023-09-12 | 中国电器科学研究院股份有限公司 | Offshore area salt spray prediction correction and accuracy assessment method |
| CN116739151B (en) * | 2023-05-23 | 2024-05-31 | 中国电器科学研究院股份有限公司 | Offshore area salt spray prediction correction and accuracy assessment method |
| CN116976202A (en) * | 2023-07-12 | 2023-10-31 | 清华大学 | Fixed complex source item distribution inversion method and device based on deep neural network |
| CN116976202B (en) * | 2023-07-12 | 2024-03-26 | 清华大学 | Fixed complex source item distribution inversion method and device based on deep neural network |
| CN116973523A (en) * | 2023-09-22 | 2023-10-31 | 深圳市智芯微纳科技有限公司 | MEMS gas sensor array-based gas detection method and system |
| CN116973523B (en) * | 2023-09-22 | 2023-12-15 | 深圳市智芯微纳科技有限公司 | MEMS gas sensor array-based gas detection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115759324A (en) | Method for optimizing monitoring position of environmental aerosol after nuclear accident | |
| Li et al. | Deep learning for high-dimensional reliability analysis | |
| Burges et al. | PROBABILISTIC METHODS IN STREAM QUALITY MANAGEMENT 1 | |
| Mu et al. | Ground motion prediction equation development by heterogeneous Bayesian learning | |
| Vesely et al. | Uncertainties in nuclear probabilistic risk analyses | |
| CN110321603A (en) | A kind of depth calculation model for Fault Diagnosis of Aircraft Engine Gas Path | |
| BROWN et al. | Randomness and local regularity of points in a plane | |
| Zio et al. | How to effectively compute the reliability of a thermal–hydraulic nuclear passive system | |
| Wang et al. | A Bayesian framework for estimating fragility curves based on seismic damage data and numerical simulations by adaptive neural networks | |
| CN115545334B (en) | Land utilization type prediction method and device, electronic equipment and storage medium | |
| Burr et al. | Revisiting statistical aspects of nuclear material accounting | |
| Lee et al. | An online operator support tool for severe accident management in nuclear power plants using dynamic event trees and deep learning | |
| Ling et al. | Multi-robot collaborative radioactive source search based on particle fusion and adaptive step size | |
| Li et al. | Deep reliability learning with latent adaptation for design optimization under uncertainty | |
| Kennedy et al. | Bayesian analysis of computer code outputs | |
| Ling et al. | Improving the estimation accuracy of multi-nuclide source term estimation method for severe nuclear accidents using temporal convolutional network optimized by Bayesian optimization and hyperband | |
| Kang et al. | Deep neural network‐based regional seismic loss assessment considering correlation between EDP residuals of building structures | |
| JP2014229283A (en) | Device and method for correction of design calculation result | |
| CN103197564B (en) | Digitizing man-machine interface monitor unit quantity optimization method and system | |
| CN110043808B (en) | Water supply network leakage monitoring and early warning method based on time series analysis | |
| CN110400042B (en) | Method for determining emergency state of nuclear fuel circulation facility based on Markov process | |
| Drábek et al. | Statistical-based control points selection for indoor illuminance measurement | |
| CN114139431A (en) | Shielding fast calculation method based on particle sampling position real-time optimization | |
| CN111104298A (en) | LSTM-based power grid server running state prediction device | |
| CN116300480B (en) | Radioactive source searching method based on improved particle filtering and biological heuristic neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |