ceshi

  以实际所测随机中子脉冲数据的自相关函数为研究对象，借助仿真实验，开展Elman神经网络对不同浓度核材料进行识别的研究。在实测数据的基础上，通过叠加随机噪声，模拟产生了不同浓度核材料的相关函数样本用于神经网络的训练与测试，实验结果表明，(Calculating the stochastic neutron pulse autocorrelation function with the actually measured data as the research object, we carry out different simulate experiments in concentration identification of varied nuclear materials using Elman neural network. Superimposing random noise on the measured data, we get correlation function samples simulating the different concentrations of nuclear material for training and testing the neural network.)