中文摘要：

针对传统的故障诊断模型无法跟踪故障的动态变化问题,提出了故障诊断模型自动更新的概念,建立了一种基于异常点检测的人工免疫故障诊断模型自动更新算法。该算法以人工免疫算法为基础,通过引入异常点检测、故障类别使用感知器、增量式单粒子浓度算法感知器来感知故障特征数据的动态变化,并设计了新故障类型的插入、故障抗体删除和特征中心的修订等操作,实现了故障诊断模型的自动更新。实验仿真及应用结果显示,该算法具有较强的自适应性和记忆能力,能够有效实现新故障类型的产生和故障特性中心漂移等问题的识别和处理。

英文摘要：

To solve the problem that traditional fault diagnosis model could not track dynamic data,the concept of automatic update for fault diagnosis model was proposed,and an automatic model updating algorithm for fault diagnosis based on artificial immune was developed.Based on the artificial immune algorithm,outlier detection algorithm,the counting sensor for fault occurrence and the incremental algorithm for single-particle density were employed to track the dynamic changes of fault feature data.And operations such as fault insertion,fault removal and fault identification center revision were designed to realize automatic update of the diagnosis model.Simulation and application results showed that the algorithm could identify the new class of fault and fault feature center drifting with strong self-adaptation and memory capacity.