基于灰色-神经网络的核动力设备运行趋势预测

根据核电设备运行参数的历史数据，利用灰色系统GM（1，1）预测模型建立动态微分方程，并预测其发展趋势。如果原始数据序列呈线性变化且还原值序列的相对误差平方和较大，则用BP神经网络对GM（1，1）的预测结果进行修正，以提高预测精度。文中以二回路辐射剂量率的预测为例，对该方法进行了仿真实验验证。验证结果表明，用BP 神经网络对GM（1，1）的预测结果进行修正相比较GM（1，1）预测模型，预测精度得到了显著提高。

Research on the operating trend prediction of the nuclear power plants based on gray-neural network

Based on the historical data of the nuclear power plants （NPPs） operating parameters, the gray system GM（ 1,1） prediction model is used to build the dynamic differential equations and predict their own growing trend . In order to improve the prediction accuracy , the back propagation （ BP ） neural network is used to revise the pre-diction results of the gray GM（1,1） prediction model, in case that raw data sequence changes linearly and restored value sequence ’ s square sum of relative error is large .For validating the method ,the radiation dose rate forecasting in the second loop of the nuclear power plant （ NPP ） is taken as an example and the result shows that when using back propagation（BP） neural network to correct the GM（1,1）prediction results, the prediction accuracy is signifi-cantly improved , compared to GM （ 1,1 ） prediction model .