基于数据挖掘的船舶主推进机械装置可靠性评估
Reliability assessment of ship propulsion machinery based on data mining
仲崇丽 1刘华2
作者信息
- 1. 西安明德理工学院信息工程学院,陕西西安 710100
- 2. 西安邮电大学网络信息中心,陕西西安 710100
- 折叠
摘要
为分析评估指标间的相关性,构建科学的评估指标体系,基于多维评估指标获取准确的可靠性评估结果,提出基于数据挖掘的船舶主推进机械装置可靠性评估方法.采集海量船舶主推进机械装置运行相关数据,获取初始评估指标;采用数据挖掘技术中的因子分析法分析初始指标间的相关性,选取相互独立的评估指标构建最终的船舶主推进机械装置可靠性评估指标体系.利用数据挖掘技术中的支持向量机构建可靠性评估模型,将多维评估指标数据输入评估模型内,通过核函数将多维评估指标数据映射至高维空间内进行分类,输出可靠性评估等级.实验结果显示:选择的评估指标的方差解释均在0.8以上,说明所选评估指标信息量保全较多;评估结果与其研究对象实际运行情况完全一致,说明该方法的有效性.
Abstract
To analyze the correlation between evaluation indicators,construct a scientific evaluation indicator system,obtain accurate reliability evaluation results based on multidimensional evaluation indicators,and propose a reliability evalu-ation method for ship main propulsion machinery based on data mining.Collect massive amounts of data related to the oper-ation of ship propulsion machinery and obtain initial evaluation indicators;The factor analysis method in data mining techno-logy is used to analyze the correlation between initial indicators,and independent evaluation indicators are selected to con-struct the final reliability evaluation index system for ship main propulsion machinery.Using support vector mechanism in data mining technology to build a reliability evaluation model,inputting multidimensional evaluation index data into the evaluation model,mapping the multidimensional evaluation index data to a high-dimensional space through a kernel func-tion for classification,and outputting the reliability evaluation level.The experimental results show that the variance explana-tions of the selected evaluation indicators are all above 0.8,indicating that the information preservation of the selected evalu-ation indicators is relatively high,the evaluation results are completely consistent with the actual operation of the research object,indicating the effectiveness of this method.
关键词
数据挖掘/推进机械装置/可靠性评估/因子分析法/指标体系/支持向量机Key words
data mining/promote mechanical devices/reliability assessment/factor analysis method/indicator system/support vector machine引用本文复制引用
出版年
2024
NETL
NSTL
万方数据