基于RVM的卫星蓄电池性能退化建模

Performance Degradation Modeling for Satellitic Battery Based on RVM

由于工作环境的特殊性,传统的基于物理退化的建模方法无法建立精确的卫星蓄电池性能退化模型。因此,在卫星蓄电池失效物理分析的基础上,提出了一种基于相关向量机回归的建模方法。介绍了相关向量机的基本原理,详细推导了相关向量机回归建模过程,并结合卫星蓄电池测试数据,进行了实例建模。结果显示,模型具有良好的准确性和稀疏性。

Because of the special working environment, the traditional physical degradation models of the satellitic battery can not suit accurately. In this case, a modeling method using RVM was proposed on the basis of the physical failure analyse of the satellitic battery. The rationale of relevance vector machine （RVM） is introduced,detailedly ratiocinated the process of RVM regress, and built a practical model integrated with the testing data of batteries. The result show that the model is veracious and sparse.