基于多源信息融合的航天阀门可靠性评估

Reliability Evaluation of Aerospace Valves Based on Multi-source Information Fusion

航天阀门在研制过程投入试验的样本量很小、样本极少失效甚至无失效。为解决其可靠性评估问题,需要充分利用现有的各类试验数据。用于阀门可靠性定量分析的试验主要包括温度试验、振动试验和动作试验。在对温度试验和振动试验数据进行预处理的基础上,融合成统一的先验分布,并将动作试验数据作为现场数据,根据Bayes原理完成阀门可靠性评估。研究结果表明,与单类试验评估结果、大样本评估结果比较,基于多源信息融合的方法评估结果准确,稳定性好,可信度高。

During the development process,the sample size of aerospace valves is small,and the sample rarely fails or even has no failure.In order to solve this reliability evaluation problem,it is necessary to make full use of various existing test data.The tests used for quantitative reliability analysis of valves mainly include temperature test,vibration test and action test.On the basis of preprocessing the temperature and vibration test data,the test data is fused into a unified prior distribution.Then taking the action test data as field data,the reliability evaluation of valves is completed according to the Bayes theory.The research shows that,compared with single-type test evaluation results and large sample evaluation results,the multi-source information fusion method is accurate,stable and credible.