推进技术 ›› 2015, Vol. 36 ›› Issue (8): 1242-1247.

• 控制 测量 故障诊断 • 上一篇    下一篇

基于MPC的民用涡扇发动机主动容错控制研究

杜 宪,郭迎清,陈小磊   

  1. 西北工业大学 动力与能源学院,陕西 西安 710072,西北工业大学 动力与能源学院,陕西 西安 710072,西北工业大学 动力与能源学院,陕西 西安 710072
  • 发布日期:2021-08-15
  • 作者简介:杜 宪(1989—),女,博士生,研究领域为航空发动机模型预测控制关键技术研究。

MPC Based Active Fault Tolerant Control of a Commercial Turbofan Engine

  1. School of Power and Energy,Northwestern Polytechnical University,Xi’an 710072,China,School of Power and Energy,Northwestern Polytechnical University,Xi’an 710072,China and School of Power and Energy,Northwestern Polytechnical University,Xi’an 710072,China
  • Published:2021-08-15

摘要: 在民用涡扇发动机模型预测控制器设计的基础上,提出一种隐性判断故障并调整控制律的主动容错控制方案。首先建立正常模式及各种已知故障模式的动态模型库,然后在每个采样周期判断当前发动机状态与动态模型库中各模型的匹配度,选择最佳匹配的预测模型,以其预测控制算法作为该时刻的子控制器,且保证各子控制器间能够平滑切换。最后以两种事先考虑的部件级故障,及某种未知故障的情况为例,进行了主动容错控制仿真,结果表明监控决策机制能够在0.5s内判断故障并给出切换指令,切换逻辑能保证不同子控制器间的平滑过渡,验证了该方法的有效性。

关键词: 模型预测控制;主动容错控制;民用涡扇发动机;动态模型库;模型匹配度;平滑切换

Abstract: On the basis of model predictive controller of a commercial turbofan engine,a novel active fault tolerant control scheme that can judge faults recessively and then adjust control laws accordingly is put forward. Firstly,a dynamic model library which contains normal mode models and all kinds of known fault models was established. Then at each sampling period,the match quality of current engine state and each model in dynamic model library was assessed,thus selecting the best matching corresponding prediction model. And the predictive control algorithm based on this selected model was used as the controller during this sample period. A switching logic was also designed to ensure the smooth transition among different controllers. Finally,fault-tolerant control simulations of two kinds of forethought component-level faults and a certain unknown fault used as an example were carried out. Results show that the selection part can sense the fault within 0.5 sec and give switch command,and the switch logic ensures the smooth transition,verifying the effectiveness of the proposed method.

Key words: Model predictive control(MPC);Active fault tolerant control;Commercial turbofan engine;Dynamic model library;Model matching quality;Smooth transition