冰蓄冷技术时间序列研究

为了解决传统的各种单一预测方法中存在的问题,运用一种BP神经网络的预测方法,将人工神经网络理论和算法应用于冰蓄冷技术的研究,并在正交实验66组数据中选择56组作为BP神经网络的输入样本,余下的10组为检测样本的输出样本。在对网络进行训练的基础上,建立了蓄冷时间与管壁温度之间的数学模型。

The Research about Time Sequence Model of the Ice-storage Technology

Aiming at solving the problems in each traditional forecasting method,an optimal forecasting method based on BP neural network was put forwards.The theory and the algorithm of the artificial neural network are applied in the research of the technique and the ice-storage.Using the 56 orthogonal data inputs for the BP neural and the surplus 10 factual values as the outputs,the time of cool storage as well as the atmosphere of pipe wall is built.