基于多源信息的可修长贮产品可靠性预测方法研究

Products, such as missiles, will remain in stockpile or in off-work state in most of their lifetimes. Such products are called long storage products. For complex long storage products, more than one performance parameter will usually degrade, which results in the deterioration of reliability. In order to prevent the failure of such product, it is desirable to predict the variance of its reliability precisely and taking proper maintenance measures. Therefore, this project makes use of multi-source information such as performance degradation data, maintenance data and reliability test data, to solve the reliability estimation problem for high reliable long storage product due to insufficient test data. To meet this end, according to the usual phenomena of multi-parameter degradations in complex products, the reliability model constructing methods for long storage products will be first studied, using the most widely used Wiener process and Gamma process as the research objects, based on degradation analysis on multi-parameter; As the maintenance's impacts on product's reliability should be taken into account, maintenance decisions method and the analysis methods for the impacts on the product reliability due to maintenance will be further studied, to update the models; Finally, based on foregoing studies, reliability estimation method for long storage products, which incorporate small sample size reliability test data and other information during storage, will be derived. This project uses certain type of long storage missiles as the application object to validate the proposed methods. The research result of this project can be applied to estimate the reliabilities of other highly reliable long storage products.

导弹等产品在寿命周期中大部分时间都处于贮存或非工作状态，这里统称为长贮产品。复杂长贮产品往往出现多元性能参数的退化，导致可靠性下降。只有准确预测可靠性变化并采取适当的维修，才能避免长贮产品的失效。因此，本项目通过对长贮产品的性能退化数据、维修数据、可靠性试验数据等多源信息的综合利用来解决高可靠长贮产品寿命试验数据不足难以开展可靠性预测的问题。为此，首先以最常见的Wiener过程和Gamma过程为对象，研究基于多元性能参数退化的长贮产品可靠性建模方法；考虑到贮存期间的维修措施对产品可靠性的影响，研究相应的维修决策方法和维修对产品可靠性的影响分析方法来更新模型；在此基础上研究结合长贮产品的小样本可靠性试验数据和其他信息来预测长贮产品可靠性的方法。本项目以某型长贮导弹为应用对象来验证所得到的方法，研究成果可应用于其他高可靠性长贮产品的可靠性预测。

导弹等长贮产品，由于贮存环境因素和检测维修措施的影响，部件的性能参数会出现退化甚至突发失效，导致可靠性下降。只有准确预测可靠性变化的趋势并采取适当的维修，才能避免长贮产品的失效。但是，导弹等长贮产品可靠性很高，在贮存期间失效很少，难以采用传统的基于失效数据的方法来预测其可靠性变化。因此，课题通过对长贮产品贮存过程中的性能退化数据、寿命数据、维修检测数据、相似产品数据、专家判断等多源信息的综合利用，来解决高可靠长贮产品寿命数据不足难以开展可靠性预测的问题。为此，首先根据复杂产品存在多个性能参数退化的情况，研究基于多个性能参数退化的长贮产品可靠性建模方法；考虑到贮存期间的维修措施对产品可靠性的影响，进一步研究相应的维修决策以及维修对部件和产品可靠性的影响分析方法；在此基础上研究结合长贮产品的性能退化数据、寿命数据、相似产品数据等多源信息来预测长贮产品可靠性的方法。通过课题研究，我们对Wiener过程和Gamma过程这两类典型退化过程的退化建模积累了经验，探索了在二维参数退化条件下的可靠性建模和求解方法，对多维性能参数退化情况下的计算复杂性问题，采用近似解析求解和基于智能方法的建模两种方式来解决；从维修对长贮产品的性能退化产生影响和维修对产品的失效率产生影响这两个角度进行了分析，针对定期检测维修采集到的性能数据或成败型数据来建立模型，在此基础上对长贮产品的可靠性和寿命进行预测；在长贮产品有极少失效甚至无失效的寿命数据、性能退化数据、相似产品数据等多源信息的场合，提出了基于寿命数据以及将其与其他多源信息融合来预测产品可靠性的方法。课题研究成果既可以应用于结构较简单的单机产品，也可以应用于导弹、卫星等结构复杂的系统。

内容获取失败，请点击重试