CN114925615A - Radioactive leakage source positioning method and system - Google Patents
Radioactive leakage source positioning method and system Download PDFInfo
- Publication number
- CN114925615A CN114925615A CN202210607187.6A CN202210607187A CN114925615A CN 114925615 A CN114925615 A CN 114925615A CN 202210607187 A CN202210607187 A CN 202210607187A CN 114925615 A CN114925615 A CN 114925615A
- Authority
- CN
- China
- Prior art keywords
- leakage
- data
- leakage source
- atmospheric diffusion
- time sequence
- 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
- 238000000034 method Methods 0.000 title claims abstract description 70
- 230000002285 radioactive effect Effects 0.000 title claims abstract description 26
- 238000012544 monitoring process Methods 0.000 claims abstract description 67
- 238000009792 diffusion process Methods 0.000 claims abstract description 60
- 238000004088 simulation Methods 0.000 claims abstract description 56
- 238000012549 training Methods 0.000 claims abstract description 38
- 239000013598 vector Substances 0.000 claims abstract description 26
- 230000005855 radiation Effects 0.000 claims abstract description 23
- 230000004044 response Effects 0.000 claims abstract description 23
- 230000007613 environmental effect Effects 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 238000005457 optimization Methods 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 17
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000012216 screening Methods 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 9
- 230000003595 spectral effect Effects 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 7
- 238000002790 cross-validation Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000006870 function Effects 0.000 description 18
- 238000004590 computer program Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000013398 bayesian method Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 230000008030 elimination Effects 0.000 description 3
- 238000003379 elimination reaction Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000012614 Monte-Carlo sampling Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000003344 environmental pollutant Substances 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 231100000719 pollutant Toxicity 0.000 description 2
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000002904 solvent Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E30/00—Energy generation of nuclear origin
- Y02E30/30—Nuclear fission reactors
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Measurement Of Radiation (AREA)
Abstract
The invention relates to a radioactive leakage source positioning method and a system, comprising the following steps: determining the magnitude of unit leakage rate according to the atmospheric diffusion simulation result of the boundary point, and obtaining a unit leakage rate vector according to the magnitude of the unit leakage rate; inputting the unit leakage rate vector into an atmospheric diffusion model for simulation, and constructing an atmospheric diffusion simulation sample library according to the response result of each monitoring point; preprocessing sample data in an atmospheric diffusion simulation sample library; extracting the time sequence characteristics of the preprocessed sample data, and fusing the time sequence characteristics with the corresponding simulated leakage source position to form a training data set; establishing a leakage source positioning model, and fitting the model through data in a training data set; and (4) extracting the time sequence characteristics of the environmental radiation monitoring data, and inputting the time sequence characteristics into the fitted model to complete the estimation of the position of the leakage source. The leakage position estimation method does not depend on the initial value of the leakage source position, improves the utilization rate of monitoring data information, and can estimate the leakage position when the leakage rate is unknown.
Description
Technical Field
The invention relates to a radioactive leakage source positioning method and a radioactive leakage source positioning system, belongs to the technical field of tracing of pollutants, and particularly relates to the technical field of tracing of radioactive pollutants.
Background
In the fukushima nuclear accident, as tsunami and earthquake damage the power system of a plant area, the radioactive leakage source information cannot be directly measured through a detection device in a nuclear facility, so that the transport and diffusion prediction of radioactive nuclide has great uncertainty, and therefore, the radioactive leakage source release source item (namely, leakage rate) needs to be reconstructed based on environmental radiation monitoring data and an atmospheric diffusion model. In the event of nuclide leakage occurring in europe in 2017, the increase of the content of the ruthenium-104 isotope is only monitored in the european atmosphere, and the position of an actual leakage is not known, so that the positioning work of a radioactive leakage source is also a core problem of human nuclear emergency. In most nuclear accident situations, the source position and the leakage rate (collectively referred to as leakage source parameters) are often unknown, and many existing technologies are designed to estimate the two parameters. The methods for estimating the leakage rate are mature day by day, but the techniques for estimating the location of the source of the leakage are still in the continuous searching and trying stage, and most methods estimate the leakage rate and the location of the source of the leakage at the same time.
The existing leakage source parameter estimation method comprises an optimization method and a Bayesian method. (1) The optimization method comprises the following steps: the method aims to find leakage source parameters which minimize a cost function, the cost function is used for quantifying the difference between an atmospheric diffusion simulation result and an environment monitoring result, and the specific form of the cost function determines the convergence speed and the accuracy degree of the optimization method. The method mainly adopts an iterative process, and continuously iterates through a parameter updating rule until the algorithm converges so as to minimize a cost function. (2) The Bayes method comprises the following steps: the input and model forms of the algorithm are designated through a probability density function, information such as input data and uncertainty of an atmospheric diffusion mode is contained, the leakage source parameter space is randomly sampled, the posterior probability distribution of the parameters is estimated, and the leakage source parameter estimation result with the confidence level is generated. Common bayesian methods are the markov chain monte carlo sampling method (MCMC), the sequential monte carlo Sampling Method (SMC) and the differential evolution monte carlo method (DEMC).
The optimization method is easily influenced by the initial value of the given leakage source parameter, so that the optimization method falls into local optimization, and many technologies adopt a hybrid algorithm to simultaneously draw the advantages of global search and local search. The likelihood function of the Bayes method is difficult to select under the actual nuclear emergency condition, and although many techniques can modify the form of the likelihood function by combining the distribution of observation errors and simulation errors, the likelihood function is usually only suitable for a single modeling scene. And no matter the optimization method or the Bayesian method, the method has high dependence on a monitoring network, and the parameter estimation result is greatly influenced by inaccuracy of an atmospheric diffusion mode or existence of monitoring data noise.
Disclosure of Invention
In view of the above problems, an object of the present invention is to provide a method and a system for positioning a radioactive leakage source, which improve the utilization rate of monitoring data information by using the time sequence characteristics of radiation monitoring data, improve the problem of insufficient information provided by a monitoring network, and estimate the position of the leakage source without depending on the initial value of the position of the leakage source.
In order to achieve the purpose, the invention provides the following technical scheme: a radioactive leakage source locating method, comprising the steps of: determining the magnitude of unit leakage rate according to the atmospheric diffusion simulation result of the boundary point, and obtaining a unit leakage rate vector according to the magnitude of the unit leakage rate; taking different grid positions in a calculation domain as simulated leakage sources, inputting the unit leakage rate vector as a source item into an atmospheric diffusion model for simulation, and constructing an atmospheric diffusion simulation sample library according to the response result of each monitoring point; preprocessing sample data in an atmospheric diffusion simulation sample library; extracting time sequence characteristics (including time domain characteristics and frequency domain characteristics) of the preprocessed sample data, and fusing the time sequence characteristics and corresponding simulated leakage source positions to form a training data set; establishing a leakage source positioning model by fitting data in the training data set; and (3) extracting the time sequence characteristics of the environmental radiation monitoring data, and inputting the time sequence characteristics into the fitted leakage source positioning model to complete the estimation of the position of the leakage source.
Further, the method for determining the magnitude of the unit leakage rate according to the simulation result of the atmospheric diffusion of the boundary point comprises the following steps: establishing gridding coordinates, inputting the gridding coordinates into a calculation domain, determining grid boundary point coordinates according to the calculation domain, and taking the grid boundary points as leakage sources; establishing an atmospheric diffusion model, inputting meteorological data and known unit leakage rate data with different magnitudes, and simulating by using the atmospheric diffusion model; and determining the magnitude of the unit leakage rate in the actual monitoring data by comparing the difference between the simulation data and the actual data.
Further, the unit leak rate vector q e T×1 Comprises the following steps:
q e T×1 =1×10 λ *ones(T,1)
where ones (T,1) represents a full 1 matrix of T × 1, where T refers to the number of analog time steps; λ is the order of the unit leak rate.
Further, the pretreatment comprises the following steps: carrying out sliding mean filtering on the sample data; selecting data corresponding to a time period of 'obvious response' in the filtered sample data, wherein the 'obvious response' means that the observed data is more than or equal to the average value of all data (different leakage source positions) simulated by adopting the unit leakage rate in the same time step; screening a sample library from the data obtained in the last step by a method of maximizing a correlation coefficient; and adding Gaussian random noise into the sample library screened in the last step to finish the pretreatment of the atmospheric diffusion simulation sample library.
Further, the time domain features include fluctuation rate, mean, median and sample entropy; and the frequency domain characteristics are obtained by performing fast Fourier transform on the preprocessed sample data to obtain power spectral density and then counting the amplitude statistical characteristics and the shape statistical characteristics of the power spectral density.
Further, the training data set generated by fusing the time domain characteristics and the frequency domain characteristics with the corresponding simulated leakage source positions is subjected to normalization and abnormal value zero filling processing.
Further, the leaky source positioning model is fitted through an XGboost algorithm based on Bayesian hyper-parameter optimization and feature screening.
Further, the fitting method comprises the following steps: inputting data of a training data set into an XGboost algorithm for model fitting, optimizing a hyperparameter based on a Bayesian optimization method, and training the result of five-fold cross validation of the data set (determining a coefficient R) 2 Average value of) as an evaluation index, iteratively updating the training hyper-parameter, and stopping optimization when the maximum iteration number is reached to obtain the optimal hyper-parameter combination. And then, screening the characteristics by a recursive characteristic elimination method with cross validation so as to obtain an optimal leakage source positioning model.
Further, a specific method for completing the estimation of the position of the leakage source is as follows: firstly, the radiation monitoring data is subjected to sliding mean filtering, time sequence characteristics of the filtered data are extracted, the time sequence characteristics are fused to form a verification data set, and the verification data set is input into a fitted leakage source positioning model to obtain a leakage source positioning result.
The invention also discloses a radioactive leakage source positioning system, which comprises: the leakage rate magnitude acquisition module is used for determining the magnitude of the unit leakage rate according to the boundary point atmospheric diffusion simulation result and acquiring a unit leakage rate vector according to the magnitude of the unit leakage rate; the sample library establishing module is used for inputting the unit leakage rate vector into an atmospheric diffusion model for simulation, and establishing an atmospheric diffusion simulation sample library according to the response result of each monitoring point; the pretreatment module is used for pretreating sample data in the atmospheric diffusion simulation sample library; the characteristic extraction module is used for extracting time sequence characteristics (including time domain characteristics and frequency domain characteristics) of the preprocessed sample data and fusing the time sequence characteristics and the corresponding simulated leakage source position to form a training data set; the model fitting module is used for establishing a leakage source positioning model and fitting the leakage source positioning model through the data in the training data set; and the result output module is used for extracting the time sequence characteristics of the environmental radiation monitoring data, inputting the time sequence characteristics into the fitted leakage source positioning model and finishing the estimation of the position of the leakage source.
Due to the adoption of the technical scheme, the invention has the following advantages:
1. according to the scheme, the time sequence characteristics (including time domain characteristics and frequency domain characteristics) of the radiation monitoring data in the atmospheric radioactive leakage are used, the utilization rate of monitoring data information is improved, and the defect that a monitoring network is sparse or poor in distribution in the prior art can be overcome.
2. According to the scheme, the time sequence change condition of the leakage rate is not required to be known, the magnitude of the leakage rate is determined only by calculating the boundary condition of a domain, then the data sample is constructed by forward simulation based on the same unit leakage rate, a simulation sample library can be constructed in advance in a place needing to be monitored, a large amount of forward simulation is not required to be carried out again when a nuclear accident happens, and the rapidity of nuclear emergency response is improved.
3. According to the scheme, data preprocessing is carried out on the monitoring data and the simulation data at the same time, the goodness of fit of the monitoring data and the simulation data is improved, and the influence of data noise and inaccuracy of an atmospheric diffusion mode on a leakage source positioning result is weakened to a certain extent.
4. Most parameters related to data preprocessing in the invention, all training parameters and feature screening work in the XGboost algorithm are driven by data without manual selection, so that the method is suitable for various complex radioactive leakage scenes and has stronger universality.
5. The method screens the data set used for training the XGboost model, quantifies the difference between the atmospheric diffusion mode prediction result and the radiation monitoring result through the correlation coefficient, and further screens a part of samples with larger correlation coefficients for extracting the time sequence characteristics and constructing the XGboost training data set, so that the XGboost training precision is improved, and the training speed is accelerated.
Drawings
FIG. 1 is a flow chart of a method for locating a radioactive leakage source in accordance with an embodiment of the present invention.
Detailed Description
The present invention is described in detail with reference to specific embodiments in order to enable those skilled in the art to better understand the technical solutions of the present invention. It should be understood, however, that the detailed description is provided for a better understanding of the invention only and that they should not be taken as limiting the invention. In describing the present invention, it is to be understood that the terminology used is for the purpose of description only and is not intended to be interpreted as indicating or implying any relative importance.
The existing leakage source positioning method has four defects: (1) the initial value of the position of the leakage source has great influence on the positioning result; (2) the method highly depends on the density and distribution of the monitoring network, and the sparse or poorly distributed monitoring network can greatly influence the accuracy of the positioning of the leakage source; (3) the position and the leakage rate of the leakage source are estimated at the same time, so that uncertainty of leakage source parameter estimation influences each other, and because the leakage source parameter estimation is a pathological problem, namely, small data change can cause the calculation result to change violently, the larger the parameter space is, the larger the influence of the parameter estimation result on data disturbance is; (4) because the atmospheric diffusion mode is inaccurate or noise exists in the monitoring data, the difference between the atmospheric diffusion simulation result and the actual monitoring result is large, and the similarity between the atmospheric diffusion simulation data and the monitoring data cannot be accurately utilized to position the leakage source.
In view of the defects of the prior art, the invention provides a radioactive leakage source positioning method and a radioactive leakage source positioning system, and the method comprises the steps of firstly simulating and calculating the response results of monitoring points generated by leakage at different positions in a domain under the condition of unit leakage rate by utilizing an atmospheric diffusion mode, and constructing an atmospheric diffusion simulation sample library; then, preprocessing a sample library, extracting time sequence characteristics (including time domain characteristics and frequency domain characteristics) of all monitoring point data, and fusing to form a training data set; and then fitting the training data set by adopting an XGboost algorithm, and inputting the time sequence characteristics of the radiation monitoring data to obtain an estimation result of the position of the leakage source. The XGboost training super-parameter is optimized based on a Bayesian method, the data features are screened by adopting a recursive feature elimination method with cross validation, overfitting is avoided while training precision is improved, parameters are driven by data, manual participation is not needed, and universality of the method is high. The positioning requirement of the leakage source does not depend on the initial value selection of the position of the leakage source, but the result is obtained only through the environmental radiation monitoring data and the atmospheric diffusion mode; the time sequence characteristics of the radiation monitoring data are fully utilized to improve the utilization rate of monitoring data information and make up for the problem of insufficient information provided by a monitoring network; the space-time coupling relation between the leakage source positioning and the leakage rate estimation is weakened to a certain extent, namely the leakage position is estimated by using the unit leakage rate under the condition that the leakage rate is unknown; the method utilizes the relation between radiation monitoring data and atmospheric diffusion simulation data, provides a proper preprocessing method to reduce the difference between an atmospheric diffusion mode and actual diffusion, and improves the accuracy of leakage source positioning. The present invention will be described in detail below by way of examples with reference to the accompanying drawings.
Example one
The embodiment discloses a radioactive leakage source positioning method, as shown in fig. 1, including the following steps:
s1, determining the magnitude of the unit leakage rate according to the boundary point atmospheric diffusion simulation result, and obtaining the unit leakage rate vector according to the magnitude of the unit leakage rate.
Specifically, establishing networked coordinates according to the experimental site condition, inputting the networked coordinates into a calculation domain, determining grid boundary point coordinates according to the calculation domain, and taking the grid boundary points as leakage sources; establishing an atmospheric diffusion model, and inputting meteorological data and known unit leakage rate data with different magnitudes to simulate the atmospheric diffusion model; and then, the magnitude of the unit leakage rate in the actual monitoring data is determined by comparing the difference between the simulation data and the actual data. It is noted that the "specific leak rate" herein does not mean 1Bq/s, and is defined as 1X 10 λ Bq/tau. So unit leakage rate vector q e T×1 The calculation formula of (2) is as follows:
q e T×1 =1×10 λ *ones(T,1)
where ones (T,1) represents a full 1 matrix of T × 1, where T refers to the analog time step; λ is the order of the unit leak rate. The upper right corner of the vector, T x 1, represents the dimension size, assuming that the magnitude of the leak rate is approximately constant over time interval τ.
S2, inputting the unit leakage rate vector into the atmospheric diffusion model for simulation, and constructing an atmospheric diffusion simulation sample library according to the response result of each monitoring point.
Specifically, grid points are uniformly sampled in a calculation domain, grid coordinates of sampling points are used as coordinates of a simulated leakage source, and a unit leakage rate vector q constructed in S1 is calculated on the assumption that N sampling points exist e T×1 The leakage source items and meteorological data are input into an atmospheric diffusion model together for simulation, and time sequence simulation results of the leakage sources at different positions are obtained at monitoring points. If n monitoring points are provided, the simulation result of a single sampling point is used Is shown in whichSimilarly, the monitoring results of all n monitoring points are obtained byIs shown in which
S3, sample data in the atmospheric diffusion simulation sample library is preprocessed.
The pretreatment comprises the following steps:
s3.1, performing sliding mean filtering on the sample data.
To pairAnd (5) performing sliding average filtering processing, wherein TL is taken as the step length of a sliding window. It is noted that this step is unified with the processing of the radiation monitoring data, and the purpose is to reduce the positioning error caused by the time evolution component of the leakage rate, thereby weakening the spatio-temporal coupling relationship between the leakage source positioning and the leakage rate estimation. The realization method comprises the following steps: modifying each number from the TL-th data to the average value of the first TL numbers (j-TL, TL +1,.., T) including the first TL-1 data without changing the first TL-1 data of the original data, and filteringThe results of (a) are as follows:
and S3.2, selecting data of the time period of the obvious response in the filtered sample data.
In practical situations, as the nuclide plume does not reach the monitoring sensor network, the acquired data has more 0 elements in an initial time period, which is "no obvious response", but the monitoring station always has response in the simulation data, so that the monitoring data at the time step where the "obvious response" exists needs to be selected for leakage source positioning. The criteria for the presence of an "apparent response" at a single time step are: the observed data is more than or equal to the average value of all the N groups of data simulated by the unit leakage rate, namely:
Since the leak source location requires the use of time series data characteristics, it is ensured that the selected time period is continuous.So the time period is determined by selecting the starting time step T 0 And end time step T end To determine the starting time step T 0 Is the first time step, T, at which there is an "apparent response end Is the last time step for which there is an "apparent response". In order to not lose effective information of the data, T is respectively judged for the data of different monitoring points 0 And T end Finally, the minimum T is selected 0 And maximum T end And (3) namely ordering:
will [ T 0,min ,T end,max ]The data in the time period serves as the last data for leak source localization.
And S3.3, screening a sample library from the data obtained in the last step by a method for maximizing the correlation coefficient.
Many leakage source localization techniques employ a method that minimizes a cost function to perform the sample library screening, where the cost function is used to describe the difference between the atmospheric diffusion simulation data and the radiation monitoring data. Although this method is not accurate, it is helpful to narrow down the location area to use its location result as a priori information for subsequent methods. In this embodiment, the cost function is selected as the correlation coefficient multiplied by-1, that is, the problem of minimizing the cost function is converted into the problem of maximizing the correlation coefficient, and the cost function between the response data generated by the leakage source at the monitoring point in a single sample and the actual environmental monitoring data is as follows:
wherein,<a>represents that the column vector a ═ (a) 1 ,a 2 ,…,a i ,…a n ) By taking the average of the elements of (i.e.In the above formula
According to the method, a part of samples with the maximum correlation coefficient can be selected for subsequent feature extraction and are fused into a training data set, the training precision and the calculation speed can be effectively improved, the number of samples cannot be too small, model overfitting is easily caused by too small number of samples, and the accuracy of positioning a leakage source is reduced.
And S3.4, adding Gaussian random noise into the sample library screened in the last step to finish the pretreatment of the atmospheric diffusion simulation sample library.
After the steps from S3.1 to S3.3, the difference between the atmospheric diffusion simulation data and the radiation monitoring data is significantly reduced, but in order to further simulate the radiation monitoring response under the real condition and improve the accuracy of the leakage source positioning, it is very necessary to add gaussian random noise to the simulation data. Mean μ of gaussian random noise is 0 and variance σ i =|σ oi -σ mi 1,2, …, n, where σ oi For standard deviation, sigma, of monitored data at a single monitoring point mi The standard deviation of the simulated data for a single monitoring point, the probability density function of gaussian noise is expressed as:
the formula for adding random gaussian noise is as follows:
wherein,
and adding random noise to the simulation data of all monitoring points corresponding to all simulated leakage sources to finish the pretreatment work of the atmospheric diffusion simulation sample library.
S4, extracting the time sequence characteristics (including time domain characteristics and frequency domain characteristics) of the preprocessed sample data, and fusing the time sequence characteristics and the corresponding simulated leakage source positions to form a training data set.
The time domain features include the wave rate, average, median and sample entropy; the frequency domain characteristics are obtained by performing Fast Fourier Transform (FFT) on the preprocessed sample data to obtain power spectral density, and then counting the amplitude statistical characteristic (FFT _ mean) and the shape statistical characteristic (FFT _ shape _ mean) of the power spectral density.
And carrying out normalization and outlier zero filling processing on the training data set generated by fusing the time domain characteristics and the frequency domain characteristics with the corresponding simulated leakage source positions.
Simulated concentration data for single measurement point with single sample For example, the calculation formula of each feature is as follows:
(1) fluctuation rate (wave _ rate)
The calculation method of the fluctuation rate comprises the following steps: firstly, the first step is toNormalizing, namely obtaining a 90 quantile value and a 10 quantile value of the data, wherein the difference value of the two quantile values is the fluctuation rate of the line of data, namely:
wherein quantile is used for solving a quantile value, and norm is used for normalizing data. The fluctuation rate can reflect the fluctuation of the data amplitude.
(2) Mean (average) and median (mean)
The average number and the median number reflect the central moment characteristics of data, belong to the statistical characteristics of time sequence data, and the calculation formula is as follows:
(3) sample entropy (SampEn)
first, time-seriesExtracting the 1 st data to the T th data in sequence every T data to form a vector, and then obtaining a group of vector groups taking T as dimension:wherein
Second step of definingIs any two vectors in the vector groupAndthe distance between them. Will be provided withAnd (4) making a difference between elements at the same position in the two vectors, and then taking the maximum value of absolute values to be the distance between the vectors. The calculation is as follows:
third step, for a given C mi Statistics ofThe number of j (j is more than or equal to 1 and less than or equal to T-T +1, j is not equal to i) is recorded as B i . For 1. ltoreq. i.ltoreq.T-T +1, the definition:
fifthly, increasing the dimension number to t +1, and countingThe number of j (j is more than or equal to 1 and less than or equal to T-T, j is not equal to i) is recorded as A i . For 1. ltoreq. i.ltoreq.T-T, definitionAnd has the following components:
thus, B (t) (r) is the probability that two sequences match t points with a similarity tolerance of r, and A (t+1) (r) is the probability that two sequences match at t +1 points with a similarity tolerance r. The sample entropy is defined as:
when T is finite, it can be estimated by:
in particular, the amount of the solvent to be used,std represents the standard deviation of the fetched data. The sample entropy characteristic is mainly used for measuring the time series complexity of monitoring point response data generated by atmospheric diffusion simulation, and the complexity is represented by A (t+1) (r) and B (t) The magnitude of the ratio of (r), i.e., the probability of generating a new pattern, is quantified. The smaller the sample entropy, the smaller the probability that the sequence generates a new pattern, the higher the similarity of the sequence itself, and the smaller the randomness, otherwise, the higher the complexity of the sequence, and the larger the randomness.
(4) Frequency domain characteristics: firstly, toPerforming Fast Fourier Transform (FFT) to obtain power spectral density, and then performing statistics on amplitude statistical characteristics and shape statistical characteristics of the power spectral density, wherein a data expression after FFT is as follows:
wherein c is mk The frequency amplitude value of the kth window is obtained, the number of the windows is the time step T, the amplitude statistical characteristic (fft _ mean) and the shape statistical characteristic (fft _ shape _ mean) of the power spectral density are respectively counted, and the calculation formula is as follows:
and performing normalization and abnormal value zero filling processing on the training data set generated by fusing the time domain characteristics and the frequency domain characteristics with the corresponding simulated leakage source positions.
S5, a leakage source positioning model is established, and the leakage source positioning model is fitted through the data in the training data set.
And fitting the leakage source positioning model by an XGboost algorithm based on a Bayesian optimization method and feature screening. The fitting method comprises the following steps: inputting data of a training data set into an XGboost algorithm for model fitting, optimizing the hyper-parameters based on a Bayesian optimization method, and performing five-fold cross validation on the training data set to obtain a result (determining coefficient R) 2 Average value of) as evaluation index, and iteratively updating the training hyperparameter when the maximum iteration number N is reached max And then, the optimization can be stopped to obtain the optimal hyper-parameter combination. In particular, N max Is set to 2000, N max The larger the optimization, the better the optimization, but the longer the calculation time. The feature screening is based on a recursive feature elimination method with cross validation, and the two parts of work are completely driven by data without manually selecting parameters.
S6, extracting the time sequence characteristics of the environmental radiation monitoring data, inputting the time sequence characteristics into the fitted leakage source positioning model, and finishing estimation of the position of the leakage source.
The specific method comprises the following steps: firstly, the radiation monitoring data isAnd the sliding mean filtering treatment is performed, so that the influence of leakage rate time evolution components on the positioning of a leakage source is favorably weakened. After filteringThe results are shown below:
where j ═ TL, TL + 1. And extracting the time sequence characteristics of the filtered data, fusing the time sequence characteristics to form a verification data set, and inputting the verification data set into the fitted leakage source positioning model to obtain a leakage source positioning result.
Example two
Based on the same inventive concept, the embodiment discloses a radioactive leakage source positioning system, which comprises:
the leakage rate magnitude acquisition module is used for determining the magnitude of the unit leakage rate according to the boundary point atmospheric diffusion simulation result and acquiring a unit leakage rate vector according to the magnitude of the unit leakage rate;
the sample library establishing module is used for inputting the unit leakage rate vector into an atmospheric diffusion model for simulation, and establishing an atmospheric diffusion simulation sample library according to the response result of each monitoring point;
the pretreatment module is used for pretreating sample data in the atmospheric diffusion simulation sample library;
the characteristic extraction module is used for extracting time sequence characteristics (time domain characteristics and frequency domain characteristics) of the preprocessed sample data and fusing the time sequence characteristics and the corresponding simulated leakage source position to form a training data set;
the model fitting module is used for establishing a leakage source positioning model and fitting the leakage source positioning model through the data in the training data set;
and the result output module is used for extracting the time sequence characteristics of the environmental radiation monitoring data, inputting the time sequence characteristics into the fitted leakage source positioning model and finishing the estimation of the position of the leakage source.
In conclusion, the XGboost model can be fitted by fully utilizing the time sequence characteristics of the radiation monitoring data in the time domain and the frequency domain, then the monitoring data is input to obtain the positioning result of the leakage source, the verification is carried out by utilizing the Belgian field release experiment, the estimated position of the leakage source has small difference with the true value, and is smaller than the preset minimum grid unit, and the limitation of the physical resolution is broken through. Particularly, the specific size of the leakage rate does not need to be determined in advance, the leakage source can be accurately positioned only by knowing the magnitude, and assistance can be provided for decoupling and rebuilding of leakage source parameters.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims. The above disclosure is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A radioactive leakage source positioning method is characterized by comprising the following steps:
determining the magnitude of unit leakage rate according to the atmospheric diffusion simulation result of the boundary point, and obtaining a unit leakage rate vector according to the magnitude of the unit leakage rate;
inputting the unit leakage rate vector into an atmospheric diffusion model for simulation, and constructing an atmospheric diffusion simulation sample library according to the response result of each monitoring point;
preprocessing sample data in the atmospheric diffusion simulation sample library;
extracting time sequence characteristics of the preprocessed sample data, and fusing the time sequence characteristics and the corresponding simulated leakage source position to form a training data set;
establishing a leakage source positioning model, and fitting the leakage source positioning model through the data in the training data set;
and extracting the time sequence characteristics of the environmental radiation monitoring data to be detected, and inputting the time sequence characteristics into the fitted leakage source positioning model to complete the estimation of the position of the leakage source.
2. The method of locating a radioactive leak source of claim 1, wherein the determining the magnitude of the unit leak rate from the boundary point atmospheric diffusion simulation result is by:
establishing networked coordinates, inputting a calculation domain, determining grid boundary point coordinates according to the calculation domain, and taking the grid boundary points as leakage sources;
establishing an atmospheric diffusion model according to the position of the leakage source, and inputting meteorological data and known unit leakage rate data with different magnitudes to simulate the atmospheric diffusion model;
and determining the magnitude of the unit leakage rate in the actual monitoring data by comparing the difference between the simulation data and the actual data.
3. The method of claim 2, wherein the unit leak rate vector q is e T×1 Comprises the following steps:
q e T×1 =1×10 λ *ones(T,1)
where ones (T,1) represents a full 1 matrix of T × 1, where T refers to the analog time step; λ is the order of magnitude of the unit leak rate.
4. The method of locating a radioactive leak source of any one of claims 1 to 3, wherein the preprocessing comprises the steps of:
performing sliding mean filtering on the sample data;
selecting data of a time period of obvious response in the filtered sample data, wherein the obvious response refers to that the observed data is larger than or equal to the average value of different leakage source positions simulated by adopting the unit leakage rate;
screening a sample library from the data obtained in the last step by a method of maximizing a correlation coefficient;
and adding Gaussian random noise into the sample library screened in the last step to finish the pretreatment of the atmospheric diffusion simulation sample library.
5. The method of any of claims 1-3, wherein the timing characteristics comprise time domain characteristics and frequency domain characteristics, the time domain characteristics comprising fluctuation rate, mean, median, and sample entropy; and the frequency domain characteristics are obtained by performing fast Fourier transform on the preprocessed sample data to obtain power spectral density and then counting the amplitude statistical characteristics and the shape statistical characteristics of the power spectral density.
6. The method of claim 5, wherein the time domain features and frequency domain features are normalized and zero-filled with outliers for a training data set generated by fusing the time domain features and frequency domain features with corresponding simulated leakage source locations.
7. The radioactive leakage source locating method of any one of claims 1 to 3, wherein the leakage source locating model is fitted by an XGboost algorithm based on Bayesian hyper-parametric optimization and feature screening methods.
8. The method of locating a radioactive leakage source of claim 7, wherein the fitting is performed by: inputting the data of the training data set into an XGboost algorithm to perform model fitting, optimizing the hyper-parameters based on a Bayesian optimization method, iteratively updating the training hyper-parameters by taking the result of five-fold cross validation of the training data set as an evaluation index, and stopping optimization when the maximum iteration times is reached to obtain the optimal hyper-parameter combination.
9. The method of any of claims 1-3, wherein the estimating the location of the source of the leakage is performed by: firstly, performing sliding mean filtering processing on radiation monitoring data, extracting time sequence characteristics of the filtered data, fusing the time sequence characteristics to form a verification data set, and inputting the verification data set into the fitted leakage source positioning model to obtain a leakage source positioning result.
10. A radioactive leakage source positioning system, comprising:
the leakage rate magnitude acquisition module is used for determining the magnitude of the unit leakage rate according to the boundary point atmospheric diffusion simulation result and acquiring a unit leakage rate vector according to the magnitude of the unit leakage rate;
the sample library establishing module is used for inputting the unit leakage rate vector into an atmospheric diffusion model, simulating under different leakage source positions, and establishing an atmospheric diffusion simulation sample library according to the response result of each monitoring point;
the pretreatment module is used for pretreating the sample data in the atmospheric diffusion simulation sample library;
the characteristic extraction module is used for extracting the time sequence characteristics of the preprocessed sample data and fusing the time sequence characteristics and the corresponding simulated leakage source position to form a training data set;
the model fitting module is used for establishing a leakage source positioning model and fitting the leakage source positioning model through the training data set data;
and the result output module is used for extracting the time sequence characteristics of the environmental radiation monitoring data, inputting the time sequence characteristics into the fitted leakage source positioning model and finishing the estimation of the position of the leakage source.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210607187.6A CN114925615A (en) | 2022-05-31 | 2022-05-31 | Radioactive leakage source positioning method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210607187.6A CN114925615A (en) | 2022-05-31 | 2022-05-31 | Radioactive leakage source positioning method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114925615A true CN114925615A (en) | 2022-08-19 |
Family
ID=82812953
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210607187.6A Pending CN114925615A (en) | 2022-05-31 | 2022-05-31 | Radioactive leakage source positioning method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114925615A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113484198A (en) * | 2021-06-30 | 2021-10-08 | 重庆建安仪器有限责任公司 | Radiation smoke cloud diffusion prediction system and method |
CN115526011A (en) * | 2022-10-10 | 2022-12-27 | 北京维禹特科技发展有限公司 | Layout method and device for VOCs (volatile organic Compounds) leakage monitoring points, electronic equipment and medium |
CN116718330A (en) * | 2023-08-09 | 2023-09-08 | 江西强普瑞石化设备科技有限公司 | Leakage monitoring method and leakage monitoring system for container |
CN117129637A (en) * | 2023-10-25 | 2023-11-28 | 北京英视睿达科技股份有限公司 | Urban NO2 monitoring method based on mobile unmanned aerial vehicle nest |
-
2022
- 2022-05-31 CN CN202210607187.6A patent/CN114925615A/en active Pending
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113484198A (en) * | 2021-06-30 | 2021-10-08 | 重庆建安仪器有限责任公司 | Radiation smoke cloud diffusion prediction system and method |
CN115526011A (en) * | 2022-10-10 | 2022-12-27 | 北京维禹特科技发展有限公司 | Layout method and device for VOCs (volatile organic Compounds) leakage monitoring points, electronic equipment and medium |
CN116718330A (en) * | 2023-08-09 | 2023-09-08 | 江西强普瑞石化设备科技有限公司 | Leakage monitoring method and leakage monitoring system for container |
CN116718330B (en) * | 2023-08-09 | 2023-10-13 | 江西强普瑞石化设备科技有限公司 | Leakage monitoring method and leakage monitoring system for container |
CN117129637A (en) * | 2023-10-25 | 2023-11-28 | 北京英视睿达科技股份有限公司 | Urban NO2 monitoring method based on mobile unmanned aerial vehicle nest |
CN117129637B (en) * | 2023-10-25 | 2024-01-12 | 北京英视睿达科技股份有限公司 | Urban NO2 monitoring method based on mobile unmanned aerial vehicle nest |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114925615A (en) | Radioactive leakage source positioning method and system | |
CN111651941B (en) | Global ionosphere electron total content prediction algorithm | |
CN109740588B (en) | X-ray picture contraband positioning method based on weak supervision and deep response redistribution | |
CN112508105B (en) | Fault detection and retrieval method for oil extraction machine | |
CN113379107B (en) | Regional ionosphere TEC forecasting method based on LSTM and GCN | |
CN110608982A (en) | Detection method, detection device, mobile equipment, electronic equipment and storage medium | |
Kaloop et al. | Multi input–single output models identification of tower bridge movements using GPS monitoring system | |
CN105573329A (en) | Attitude and orbit control data analysis method based on decision tree | |
CN111695473B (en) | Tropical cyclone strength objective monitoring method based on long-short-term memory network model | |
CN110119853B (en) | Water supply network leakage alarm threshold value selection method based on time series monitoring data analysis | |
CN113687433A (en) | Bi-LSTM-based magnetotelluric signal denoising method and system | |
CN113449254B (en) | Method for analyzing monitoring stability of any net-type deformation and method for determining position of monitoring point | |
García-Alfonso et al. | Comparison of uncertainty analysis of the Montecarlo and Latin Hypercube algorithms in a camera calibration model | |
CN112767219B (en) | Post-earthquake disaster population rapid space assessment method and system based on machine learning | |
CN112329535B (en) | CNN-based quick identification method for low-frequency oscillation modal characteristics of power system | |
CN116255896A (en) | Landslide early warning method and device based on satellite and electronic equipment | |
CN111695501B (en) | Equipment soft fault detection method based on operating system kernel calling data | |
CN110083849A (en) | Flexible measurement method, computer and the computer-readable medium of semiconductor manufacturing facility | |
CN115310472A (en) | Nuclear pulse peak sequence-based one-dimensional convolution neural network nuclide identification method | |
CN113190985A (en) | Method and system for analyzing data errors in layered heterogeneous region | |
Hite | Bayesian Parameter Estimation for the Localization of a Radioactive Source in a Heterogeneous Urban Environment | |
CN114861800B (en) | Model training method, probability determining device, model training equipment, model training medium and model training product | |
CN116541726B (en) | Sample size determination method, device and equipment for vegetation coverage estimation | |
Khadanovich | Guaranteed Estimation in the Problem of Transformer Windings DC Resistance Measuring | |
Geng et al. | Research on Performance Evaluation Method of UAV Multi-Source Data Fusion Based on Credibility |
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 |