一种基于人工免疫系统的尺度自适应免疫分类算法  

A Scale Adaptive Immune Classification Algorithm Based on Artificial Immune System

作　　者：任中元 崔继强[2] 刘洪斌 Ren Zhongyuan

机构地区：[1]上海大学机电工程与自动化学院,上海200444 [2]滨州学院,山东滨州256600 [3]国网湖北省电力有限公司五峰县供电公司,湖北宜昌443413

出　　处：《工业控制计算机》2020年第9期40-42,44,共4页Industrial Control Computer

基　　金：国家自然科学基金(Grants No.61603238)支持。

摘　　要：基于人工免疫系统的免疫分类算法能有效处理小样本故障诊断问题,但是对于复杂故障诊断问题中由于边界样本高度交叉而带来的分类困难目前却难以有效处理。针对此问题,基于人工免疫系统相关免疫机理,提出了尺度自适应B细胞检测器和尺度自适应免疫分类算法。算法通过异类抑制和同类抑制步骤有效实现了异类细胞间的边界竞争和同类细胞间的冗余删除,从而以较少的细胞有效解决了异类样本高度交叉的故障诊断问题。最后通过往复压缩机气阀故障诊断试验验证了算法的有效性。The immune classification algorithms based on artificial immune system(AIS)can deal with the fault diagnosis problems of small samples,but they cannot effectively deal with the classification difficulties caused by the high intersection of boundary heterogeneous samples in complex fault diagnosis problems.To solve this problem,based on the immune mechanism of AIS,this paper proposes a scale adaptive B-cell detector and a scale adaptive immune classification algorithm.The algorithm effectively realizes the boundary competition between heterogeneous cells and the redundancy deletion between homogeneous cells through the steps of heterogeneous suppression and homogenous suppression,so as to effectively solve the fault diagnosis problem of highly intersecting heterogeneous samples with only a few cells.Finally,the effectiveness of the algorithm is verified by the gas valve fault diagnosis test of a reciprocating compressor.

关 键 词：尺度自适应 克隆变异 种群亲和力 异类交叉 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]