摘要
针对传统故障模式识别方法不能区别不同误判所造成损失不同的问题,提出了可变风险支持向量机(SVM)模型,对传统SVM模型的最优分类面进行重新设计,在利用实际数据识别故障的同时融入专家经验,使故障识别结果更具可靠性,该方法已成功应用于柴油机故障诊断.
Due that the traditional fault pattern recognition methods cannot distinguish different losses from misjudgments, the model of variable-risk support vector machines (SVM) is proposed. By redesigning the optimal classification via traditional SVM models, the expert experiences are integrated with the actual data for fault recognition to obtain more feasible fault recognition results. So far, this method has been suc- cessfully applied for fault diagnosis on diesel engines.
出处
《中国工程机械学报》
2012年第2期216-221,227,共7页
Chinese Journal of Construction Machinery
基金
国家自然科学基金资助项目(51075396)