基于数据的故障诊断更新问题分析及应用

机内测试被广泛应用于故障诊断，装备健康管理与预测等领域。本文针对机内测试设备在设计和升级时遇到的分类器更新，样本数量不平衡，硬件条件限制问题，提出了初步解决方案。首先，利用基于密度的聚类和人工免疫的方法处理原始数据，然后提出了基于代表样本点的混合学习方法，最后利用支持向量机和仿真案例验证了本文所提方法。结果表明本文所提方法能够解决上述问题，有助于基于数据的机内测试设备设计与升级。

Data based fault diagnosis update problem analysis and its application

Built-in test equipment (BITE) is widely used in the fields of fault diagnosis, prognosis and health management for various systems. In this paper, the problems encountered in the process of BITE design and update including the classifiers update, samples imbalance and hardware limitation are analyzed, and the initial solution is proposed. Firstly, the density based cluster and artificial immune system are applied to process the raw data. Then the delegates based hybrid learning methods are proposed. We also show the evaluation of the solution using the numerical and experiment examples with support vector machine at last. Results show that the proposed solution could solve the mentioned problems well. It is helpful for data based fault diagnosis design and update in the process of BITE maturation.