中文摘要：

多数用于航天、国防的密封式电磁继电器长期处于贮存状态，为保证其寿命剖面各阶段始终保持在备用激活状态，有必要对继电器贮存寿命进行准确预测。以密封式电磁继电器的触点接触压降为研究对象，采用集合经验模态分解（EEMD）和相空间重构理论对信号进行处理，建立分解序列的 RBF 神经网络预测模型，根据接触压降的预测值实现继电器的贮存寿命预测。预测结果表明，该方法具有较高的预测精度。

英文摘要：

The majority of sealed electromagnetic relays used in aerospace,national defense are left to storage state for long periods.It is necessary to predict the storage life accurately to ensure the relay remains active in standby at each stage of life profile.The contact resistances of sealed electromagnetic relay contacts were taken as research object,and ensemble empirical mode decomposition（EEMD）and phase space reconstruction（PSR）were applied for signal processing,then RBF network prediction model of each decomposed sequence was established. Finally,the storage life prediction was realized based on prediction value of the contact resistance.The prediction results show that this method is of high precision.