高可靠嵌入式系统电源功率模块劣化过程的非线性时变行为研究

The highly reliable embedded systems have been widely applied in the fields of aerospace, nuclear power,high-speed rail, etc., which are related to national strategy and security. The reliability of the power supply directly influences the security of the embedded system, and has been the research focus of electronic information and energy. The power modules degradation occupies a dominant position among the key factors affecting the power supply reliability. It is critical how to dynamically get the degradation state and forecast the remaining useful life of the working power modules.Therefore, the project tries to build nonlinear time-varying behavior model of degradation process of power modules, analyze dynamic degradation evolutionary process based on functional theory of the time-varying state equation solution instability, which changes structural and statistical modeling of traditional degradation analysis into behavioral and analytical modeling. An online obtaining method of non-intrusive power module degradation state is proposed based on Volterra series. It uses self-driving signal as non-intrusive incentive, and extracts degradation characters of power modules by frequency-domain kernel of Volterra series, which changes the nonmeasurable into the measurable online. As a result, the dynamic prediction algorithm of remaining useful life of the power modules is put forward,which changes traditionally offline statistical analysis into online dynamic prediction. These present theoretical foundations for prediction and intelligent management of power health state of highly reliable embedded system, provide new ideas and ways for improving power system reliability and maintainability.

高可靠嵌入式系统广泛应用于航天、核电、高铁等关系国家战略与安全的领域中，其电源的可靠性直接影响系统的安全性，并成为电子信息和能源学科的研究热点。在影响电源可靠性的关键因素中，功率模块劣化占据主导地位，如何在运行中动态获取功率模块的劣化状态、预测剩余使用寿命已成为研究的焦点。因此，本研究拟建立功率模块劣化过程的非线性时变行为模型，基于时变状态方程解不稳定性的泛函理论探讨劣化动态演变过程，改传统劣化分析的结构性、统计性建模方式为行为性、解析性建模方式；提出基于Volterra级数的非侵入式功率模块劣化状态在线获取方法，以驱动信号为非侵入式激励，用Volterra级数频域核抽取劣化特征，变劣化状态在线不可测为可测；在此基础上，提出功率模块剩余使用寿命的动态预测算法，改传统离线统计分析为在线预测。为高可靠嵌入式系统电源健康状态预测与智能管理奠定理论基础，为提高其可靠性、可维护性提供新思路和新方法。

高可靠嵌入式系统是航天、核电、高铁等重要工程领域的关键设备，电源是这些设备的核心部分，其可靠性直接影响设备的安全性。据统计，在影响电源可靠性的关键因素中，功率模块劣化占据主导地位。因此，功率模块的剩余寿命预测是实现高可靠嵌入式系统预测维护与健康管理、保障其安全可靠长周期运行的关键。考虑到传统的离线、数学统计方式难以监测其劣化状态和估计剩余使用寿命，在运行中估计功率模块的劣化状态、预测剩余使用寿命已成为学科前沿问题。 本项目:（1）建立了功率模块劣化过程的非线性时变行为模型。从时变状态方程解的不稳定性剖析功率模块劣化行为，探讨了劣化动态演变过程，阐明了劣化过程中线性和非线性特征，改传统劣化分析的结构性、统计性建模方式为行为性、解析性动态建模方式。(2)提出一种基于Volterra 级数的非侵入式功率模块(MOSFET)劣化特征抽取方法。利用DC/DC 电源本身开关驱动信号为非侵入式激励，基于Volterra 级数分析MOSFET模块频响，使用频域核抽取MOSFET劣化线性、非线性特征，变MOSFET劣化状态在线不可测为可测。(3)考虑功率模块劣化行为模型和劣化特征抽取方法，提出了基于强跟踪滤波器、改进的强跟踪滤波器、改进的卡尔曼滤波等功率模块剩余使用预测算法，突破传统离线处理和统计分析估计劣化状态的局限，为传统电源系统的经验管理变为智能管理奠定基础。(4) 建立了电源系统健康状况预测与测试实验平台，为研究功率模块劣化过程提供了实验条件。 (5）建立了基于分数阶的电源系统中新型储能系统（超级电容）模型，提出了基于模糊的超级电容故障预测算法,构建了超级电容故障预测实验平台。本课题的研究成果具有重要的理论意义，将为高可靠嵌入式系统电源健康状态预测与智能管理奠定理论基础，并为提高电源系统的可靠性、可维护性和安全性提供新思路和新方法。