Research On Radiation Detection Model Of Atmosphere Environment Monitoring Stations

In recent years,with the development of nuclear technology and its wide application in all walks of life,the exploitation of natural mineral resources has increased,resulting in changes in the concentration of radionuclides in the environment,so it is urgent to establish a corresponding mature radiation safety supervision system.Due to the complex types and low content of environmental nuclides,the current detection methods have the problems of high false negative rate and false positive rate.Therefore,this thesis studies the dose rate fitting method based on linear regression and the detection method of abnormal fluctuation of environmental energy spectrum.The main work and innovations are as follows:(1)In order to solve the problem that dose rate is easy to be interfered by external factors and lack of data,the method of dose rate estimation based on linear regression was studied.The correlation between air absorbed dose rate and weather parameters is analyzed,and weather data is introduced as supplementary information to reduce the false alarm of the dose rate;a linear regression model is used to determine the relationship between the dose rate and the energy spectrum count;the energy spectrum is used to fit the dose rate value as the complementary information of the measured value of the high-voltage ionization chamber,it can improve the reliability of the dose rate data and reduce the lack of data.(2)In order to solve the problem that the energy spectrum fluctuations in the natural environment will interfere with the energy spectrum analysis,an environmental energy spectrum clustering method based on the PCA-BIRCH algorithm is studied.A visual analysis method based on t-SNE is introduced to reveal the cluster distribution characteristics of the data set;based on the distribution characteristics of the data set in the low-dimensional space,a clustering model based on the PCA-BIRCH algorithm is established.This model can divide the data into different data sets according to clusters,and can suppress the influence of background energy spectrum fluctuations.(3)In order to solve the limitation of the existing energy spectrum anomaly analysis,the thesis studies the environmental energy spectrum anomaly detection method based on the local anomaly factor.On the basis of energy spectrum classification,an anomaly discrimination model is established by using the local density characteristics of the data set,and use the anomaly score to determine the outlier of the spectrum.Comparing the detection results of unclassified data sets,it is proved that energy spectrum cluster can improve the accuracy of the model;comparing the detection results based on the kNN model and the dimensionality reduction reconstruction model,it is verified that the method can effectively identify the abnormal energy spectrum in the test data set and reduce the false positives.(4)Investigate the needs and technical framework of the information system of the air radiation monitoring station.The online monitoring platform is built with browser/server structure,including client interface,background management module,data analysis module and database module.Finally,taking the historical data of an automatic station in Zhejiang Province as an example,verify the effectiveness of the data analysis and anomaly discrimination algorithms studied in this thesis.