基于模型的机电作动器故障预测与健康管理及自修复策略研究

The servo system of Electro-Mechanical Actuator (EMA) is applied widely in the aerospace field. In order to increase the reliability of EMA and on the backgroud of redundancy technique, the strategy of Prognostics and Health Management (PHM) and Self-Repairing is researched in this proposal. For reducing the volume of EMA servo motor and inhibiting the mutual inductance between windings, a novel dual-redundancy structure of servo motor is designed. To resolve the problem of torque unbalance, a novel torque balancing strategy in dual-redundancy servo system is proposed. This proposal aims at analyzing the coupling relation of real-time detecting data from multi-sensor and presenting new theoretical mapping model and extraction method of model attributes under varying loading conditions. In this proposal, the PHM strategy of servo system is presented. The on-line self-repair technology of Linear Variable Differential Transformer (LVDT), Rotary Variable Differential Transformer (RVDT) and their aided circuit is designed. Via the theoretical and experimental validation which provided by the observing platform of health management and self-repairing strategy and dual-redundancy EMA prototype respectively, the function of PHM based on EMA mapping-model is verified. To resolve the problem of multi-sensor aliasing and sensor noise in real-time system, a kind of high effective filter algorithm of self-adapting data fusion is given. Through research work presented in this proposal, the reliability, maintainability and lifespan of EMA will be enhanced effectively, the development demand of EMA servo system will be meted, and the theoretical research and the technical application of PHM in the field of flight control system will be promoted.

针对机载机电作动器伺服系统可靠性问题，以余度协调为背景，开展基于模型的机电作动器故障预测与健康管理及自修复策略研究。构建新型作动器无刷直流伺服电机余度结构，研究减小电机体积并抑制绕组间互感的有效方法，提出双余度"力平衡"新策略，解决余度间力纷争问题；在机电作动器健康状态的不同负载条件下，构建其理论映射模型，研究映射模型完全特征的提取方法，提出作动器伺服电机故障预测与健康管理策略，并建立测控平台验证其功能，同时完成线性可变差动变压器和旋转可变差动变压器及其辅助电路的故障在线自修复技术研究；剖析多传感器实时检测的数据间耦合关系，提出高效自适应数据融合滤波算法，实时处理机电作动器伺服系统中多传感器混叠和传感器噪声问题。本研究将有效提高现有机电作动器的可靠性、寿命和可维护性，满足机电作动器伺服系统的发展要求，推动其故障预测与健康管理的理论研究及技术应用进程。

机电伺服系统被广泛应用于航空航天领域的功率电传操纵中，针对其可靠性问题，项目以双余度机电作动器为研究背景，开展基于模型的机电作动器故障预测与健康管理及自修复策略研究。项目首先开展了新型作动器用无刷直流伺服电机余度结构的设计，提出了一种“并联半圆形”的绕组结构，两套定子绕组分别绕在圆形定子铁心的上下半圆，使得电机体积减小并可有效抑制两套绕组间互感。针对余度作动器的“力纷争”问题，项目提出了位置环交叉反馈控制方法，两套余度的最外环反馈信号交叉馈入对方余度，通过算法可有效解决双余度电机转矩纷争问题，实现余度作动器的协调控制。为了进一步提高机载作动系统电机转子位置检测的可靠性，项目提出了新的故障诊断方法，提高了机电作动器伺服系统的容错控制能力。针对多传感器实时检测的数据间耦合关系，项目提出了高效自适应数据融合滤波算法，采用平方根滤波的思想，对原理简单且实时性好的Sage-Husa自适应滤波算法进行改进，提出了一种新的线性自适应平方根卡尔曼滤波算法。该算法直接对系统状态方差阵和噪声方差阵的平方根进行递推与估算，确保了状态和噪声方差阵的对称性和非负定性，以满足机电伺服系统对精确性、稳定性、实时性和自适应性的要求。在研究故障诊断方法及自适应滤波算法的基础上，项目进一步提出了双余度机电作动器自修复策略，假如线性可变差动变压器（LVDT）失效，则根据旋转可变差动变压器（RVDT）信号及映射模型实时重构LVDT的反馈信号；若某一余度RVDT失效，则根据另一余度RVDT信号进行基于健康模型的转子位置信号重构；若两余度RVDT均失效，则根据无刷电机三相电流与滑动控制模型实时解算电机转子位置。本项目所涉及的故障预测与健康管理及自修复技术，可有效降低机载伺服系统自检测的虚警率，增加机电伺服系统智能化程度与可靠性指标，对电传操作系统可靠性技术进步及传感器高效融合问题研究具有推动作用。