数据同化在核事故辐射场评估中的应用研究

发布时间：2018-05-12 16:57

本文选题：数据同化 + 核事故　；参考：《中国安全科学学报》2015年05期

【摘要】：为提高核事故辐射场评估的准确性和可靠度,建立一种基于数据同化理论的评价方法。以高斯烟团模式和核素沉降模式联合组成预测模型,以集合卡尔曼滤波(En KF)为同化算法,结合观测剂量率数据,构建核事故辐射场评估同化系统。在Matlab软件平台上,研究不同气象条件和不同核事故释放情形时,同化结果与模型预测结果的异同。试验结果表明,当气象条件和放射性物质释放率恒定时,同化场的相对均方根误差比模型预测场约低25%;当气象条件和释放率都变化时,同化场的相对均方根误差比模型预测场约低20%。通过有效利用观测数据,数据同化方法能及时调整模型,从而减小模型预测误差。
[Abstract]:In order to improve the accuracy and reliability of radiation field assessment of nuclear accident, a method based on data assimilation theory was established. The prediction model is composed of Gao Si smoke cluster model and nuclide sedimentation model. The assimilation system for radiation field assessment of nuclear accident is constructed by using the set Kalman filter en KF as assimilation algorithm and the observed dose rate data. On the platform of Matlab software, the similarities and differences between the assimilation results and the model prediction results are studied under different meteorological conditions and different nuclear accident release cases. The experimental results show that the root mean square error of the assimilation field is about 25% lower than that of the predicted field when the meteorological conditions and the radioactive release rate are constant, and when the meteorological conditions and the release rate vary, the relative root mean square error of the assimilation field is about 25% lower than that of the predicted field. The relative root mean square error of assimilation field is about 20% lower than that of model prediction field. By using the observed data effectively, the data assimilation method can adjust the model in time and reduce the prediction error of the model.
【作者单位】：防化学院核防护系;中国科学院大气物理研究所;
【分类号】：TL73

【参考文献】