惯性仪器故障诊断模型设计与实现-《计算机技术与发展》

文章信息/Info

Title:: Design and Implementation of Inertial Apparatus Fault Diagnosis Model

Author(s):: LI Bi-feng; LI Fu-rong; YU Jian-li; QIN Hao; Qingdao Branch of Naval Aeronautical and Astronautical University

Keywords:: inertial apparatus ; two-step clustering ; k-means clustering; C5.0

摘要:: 为了挖掘隐藏在惯性仪器测试数据背后的信息知识，解决数据丰富而知识贫乏的问题，运用数据挖掘技术筛选出典型的故障测试数据，借鉴CRISP—DM行业标准并以Clementinel2．0为平台进行惯性仪器故障诊断模型的设计与实现。提出一种基于两阶段聚类的C5．0算法，即在两步聚类和k-means聚类的基础上使用C5．0算法，与传统C5．0算法相比，提高了预测精度和普适能力。结果表明，基于两阶段聚类的C5．0模型具有较好的分类能力和较强可解释性，为建立基于数据挖掘技术的惯性仪器故障诊断系统提供了研究基础

Abstract:: In order to tap the information and knowledge hidden behind the test data of inertial apparatus, solving the problem that data is rich but information is poor, apply data mining technology to select typical fault test data, referring to CRISP-DM and taking Clemen- tinel2.0 as the platform to design and implement the model of inertial instrument fault diagnosis. Propose C5.0 algorithm based on two -stage clustering that is using C5.0 based on two-step and k-means clustering, compared with traditional C5.0 algorithm, the prediction accuracy and universal capacity has been improved. The results prove that the model of C5.0 based on two-stage clustering has good classification capacity and strong interpretability, which provides a research base for fault diagnosis system of inertial instrument based on data mining technology