[author_cn_name].[cn_title][J].空军工程大学学报:自然科学版,[year_id],[volume]([issue]):[start_page]-[end_page] 基于随机森林的航空发动机工作状态识别-An Aero-Engine Working Condition Recognition Based on Random Forest
文章摘要
李鼎哲, 彭靖波, 赵泽平, 王玮轩, 赵彪.基于随机森林的航空发动机工作状态识别[J].空军工程大学学报:自然科学版,2020,21(1):15-20
基于随机森林的航空发动机工作状态识别
An Aero-Engine Working Condition Recognition Based on Random Forest
  
DOI:
中文关键词: 航空发动机  飞参数据  工作状态识别  随机森林  主成分分析  属性约简
英文关键词: aero-engine  flight data  working condition recognition  random forest  principal component analysis  attribute reduction
基金项目:国家自然科学基金(51506221);陕西省自然科学基础研究计划(2015JQ5179)
作者单位
李鼎哲, 彭靖波, 赵泽平, 王玮轩, 赵彪 空军工程大学航空工程学院 西安 710038 
摘要点击次数: 94
全文下载次数: 105
中文摘要:
      为解决人工识别航空发动机工作状态中存在的误判和耗时费力等问题,提高识别准确率,提出了一种基于主成分分析(PCA)的特征提取方法和随机森林(RF)的智能识别方法。首先对飞参数据进行预处理,利用PCA将数据降维进行属性约简,并根据发动机工作状态将样本分组,用随机森林方法训练获得分类器;然后将几种分类方法的识别效果进行对比;最后采用该方法对某一架次的发动机工作状态进行识别。结果表明,该方法能够准确快速地识别航空发动机的稳定工作状态,识别准确率达到97.89%。可应用于发动机工作状态的相关研究。
英文摘要:
      In order to solve the misjudgment and time consuming problem in manual identification of aero engine working condition, and to improve the accuracy of the recognition, an intelligent recognition method based on principal component analysis(PCA)and a random forest(RF)is proposed. Firstly, PCA is used to reduce the dimensionality of the original flight data preprocessed, and the processed data on the basis of aero engine working condition are grouped, and then RF are constructed. Secondly, the recognition effect of several classification methods is made in comparison with each other. At last, the method is used to recognize the working condition of one sort. The experiment results indicate that the recognition accuracy is 97.89% by this method. And this method is able to recognize the aero engine working condition fast and accurately, and simultaneously is able to apply to the research related to the aero engine working condition.
查看全文   查看/发表评论  下载PDF阅读器
关闭