带有不确定工况的切换系统的故障检测方法研究

For the switched systems with faults, if the switching signals of analytical mode cannot accurately characterize the practical operating status transformation due to the occurrence of the fault, the proper mode cannot be used to match the current operating status. In this case, the fault cannot be detected via the existing approaches due to the difference between the analytical mode and operating status. Noting that the above problem and the huge demand of the fault detection (FD) theory and technology for the advanced equipment, the fault of fuel flow supervisory component of Turbofan-engine is chosen as the research object and a novel FD method is proposed, which includes the uncertain operating status identification, switching signal correction and multi-mode FD scheme. Further, the following key scientific problems will be solved: the acquisition，analysis and filtrating for the information of uncertain operating status, the construction of frequency-dependent identification space, the establishment of the identification performance index and the system performance improvement based on the multi-mode FD scheme. We aim to obtain a novel universal FD method differing from the traditional framework and lay the solid foundations for further research.

对于发生故障的切换系统，如果受故障的影响使其解析模型的切换信号不能准确刻画实际工况之间的转换规律，将导致被切换信号调用的子系统模型与实时工况不再匹配。在这种情况下，现有方法由于高度依赖系统模型与工况的一致性而无法实现对故障的有效检测。针对这一问题并结合国家对高端设备故障检测理论及技术的重大需求，选择涡扇发动机进油量监控元件故障作为研究示范对象，提出新的故障检测方法：不确定工况的辨识，解析模型切换信号的校正及多模式故障检测。重点解决所遇到的一些关键科学问题：不确定工况特征信息的提取、分析与筛选，频域相关可辨识空间的构造，辨识效能指标的建立及多模式故障检测机制下的系统性能改进等。取得一些有别于传统框架且具有普适性的故障检测新方法并为进一步的研究打下坚实的理论基础。

对于发生故障的切换系统，如果受故障的影响使其解析模型的切换信号不能准确刻画实际工况之间的转换规律，将导致被切换信号调用的子系统模型与实时工况不再匹配。在这种情况下，现有方法由于高度依赖系统模型与工况的一致性而无法实现对故障的有效检测。针对这一问题并结合国家对高端设备故障检测理论及技术的重大需求，选择涡扇发动机进油量监控元件故障作为研究示范对象，提出新的故障检测方法：不确定工况的辨识，解析模型切换信号的校正及多模式故障检测。重点解决所遇到的一些关键科学问题：不确定工况特征信息的提取、分析与筛选，频域相关可辨识空间的构造，辨识效能指标的建立及多模式故障检测机制下的系统性能改进等。取得一些有别于传统框架且具有普适性的故障检测新方法并为进一步的研究打下坚实的理论基础。