基于相对熵权重组合的多阶段贮存数据融合评估方法
A Multi-stage Storage Data Fusion Evaluation Method Based on Relative Entropy Weight Combination
张生鹏 1倪瑞政 2徐如远 2马小兵 3李宏民1
作者信息
- 1. 北京航空航天大学,北京 100191;航天科工防御技术研究试验中心,北京 100854
- 2. 航天科工防御技术研究试验中心,北京 100854
- 3. 北京航空航天大学,北京 100191
- 折叠
摘要
目的 针对目前仅使用加速贮存数据获取产品性能参数退化信息导致贮存寿命评估可信度低的问题,融合自然贮存数据和加速贮存数据 2 阶段的数据结果.方法 采用基于最小二乘和相对熵的方法,确定不同阶段数据的权重比例,最后形成融合 2 阶段数据的评估方法,并在扭杆和控制放大器 2 个产品上进行自然贮存数据与加速贮存数据的融合分析.结果 实现了 2 个产品贮存寿命的高置信度评估,扭杆自然贮存数据和加速贮存数据的权重比例分别为 0.263 和 0.737,控制放大器自然贮存数据和加速贮存数据的权重比例分别为 0.776 和 0.224.结论 融合评价的方法能够结合 2 部分数据的退化特点,给出一个更有信服力的评价结果.
Abstract
The equipment has been stored in a natural environment for certain years before the accelerated storage test.Its usage and detection information during different stages has different characteristics,and has irreplaceable value in determining the health status of the equipment.Natural storage data has high credibility but lacks data,while laboratory accelerated storage data has rich information but low credibility.This article aims to address the issue of low reliability in evaluating storage life due to the use of only accelerated storage data to obtain degradation information of product performance parameters.The results of natural storage data and accelerated storage data from two stages were integrated.The methods based on least squares and relative entropy were used to determine the weight ratios of data from different stages.Finally,an evaluation method that inte-grated two-stage data was formed.By conducting fusion analysis of natural storage data and accelerated storage data on the tor-sion bar and control amplifier products,a high confidence evaluation of the storage life of the two products was achieved.The weight ratios of natural storage data and accelerated storage data on the torsion bar were 0.263 and 0.737,respectively.The weight ratios of natural storage data and accelerated storage data on the control amplifier were 0.776 and 0.224,respectively.The data results show that the fusion evaluation method can combine the degradation characteristics of the two parts of the data to provide a more convincing evaluation result.
关键词
贮存数据/加速贮存数据/多源信息/相对熵/权重计算/融合评估Key words
storage data/acceleration data/multi source information/relative entropy/weight calculation/fusion evaluation引用本文复制引用
基金项目
装备预先研究共用技术(50904060202)
出版年
2024
NETL
NSTL
万方数据