论文部分内容阅读
可靠性定量设计的关键是建立可靠性定量模型。现有的可靠性定量模型建模方法主要基于设计人员对产品对象故障规律的知识,包括故障模式、环境扰动、故障机理等。但知识固有的有限性和不完整性必然会给可靠性定量模型带来模型误差和输入参数的不确定性。针对这个问题,提出了基于贝叶斯理论融合知识和数据的可靠性定量模型建模方法,量化并更新模型误差和输入参数的不确定性。为此,首先说明了知识与数据融合的可靠性定量模型建模工作,建立了知识与数据融合的可靠性定量模型建模框架;接着阐述了基于贝叶斯理论的知识与数据融合原理;然后介绍了基于贝叶斯理论融合知识与数据的通用方法,并分别针对性能波动数据和性能退化数据2种常见数据类型进一步详细讨论了各自适用的贝叶斯融合方法;最后通过机载轴向柱塞泵的案例验证了前述方法的可行性和有效性。
The key to reliability quantitative design is to establish a quantitative model of reliability. The existing reliability quantitative model modeling method is mainly based on the designer’s knowledge of the fault rule of the product object, including fault modes, environmental disturbances and fault mechanisms. However, the inherent limitation and incompleteness of knowledge inevitably lead to model errors and uncertainty of input parameters in the quantitative model of reliability. In response to this problem, a quantitative modeling method of reliability based on Bayesian theory, which combines knowledge and data, is proposed to quantify and update the uncertainty of model errors and input parameters. To this end, the reliability quantitative model modeling of knowledge and data fusion is first described, and a reliability quantitative model modeling framework of knowledge and data fusion is established. Then the principle of knowledge and data fusion based on Bayesian theory is described. Then, The general method of integrating knowledge and data based on Bayesian theory is introduced. Two common data types of performance fluctuation data and performance degradation data are respectively discussed in detail, and their respective Bayesian fusion methods are discussed in detail. Finally, The case of a plug pump validates the feasibility and effectiveness of the above method.