【作者】 黄伟；

【导师】 黄向华；

【作者基本信息】 南京航空航天大学 ， 航空宇航推进理论与工程， 2013， 博士

【摘要】 传统的防喘控制属于被动控制方法，利用保守的喘振裕度保证了航空发动机安全工作的同时，也极大地牺牲了发动机的性能。然而，先进飞行器却对推进系统提出更高负荷、更高效率、更高推重比的要求，高稳定性控制成为先进航空发动机的必备技术之一。因此，本文在总结了国内外在该领域研究进展的基础上，围绕压气机失稳模型、压气机喘振主动控制、压气机失稳预测等高稳定性控制关键技术开展研究，并研究了高稳定性控制在涡扇发动机加速优化中的应用。推导了一个MG模型形式的轴流式压气机变转速过失速瞬态模型。新模型考虑了压气机转子动态和旋转失速高阶分量对压缩系统稳定性的影响。仿真结果表明，压气机转子转速的变化相当于系统内容扰动，有可能使系统进入气动失稳过程。为了从机理上解释压气机喘振与旋转失速现象，基于分岔理论，对MG模型开展了非线性动力学分析。旋转失速的迟滞现象由发生在压气机稳定特性线压升系数最大值点的亚临界音叉分岔引起。而喘振现象与系统的Hopf分岔相关，当参数小于某一临界值时，系统不会发生Hopf分岔。提出了喘振及旋转失速输出反馈控制、喘振及旋转失速双执行机构控制、基于二阶滑模的喘振控制、基于FLC的喘振主动/被动混合控制四种轴流式压气机喘振主动控制方案。输出反馈控制器利用流量估计器对压气机流量系数进行估计，只需要易于采集的压力信号就能实现主动控制。双执行机构控制器使用节流阀和紧连阀控制阀（Close Coupled Control Valve,CCV）同时作为主动控制的执行机构，具有比单执行机构更好的控制效果。二阶滑模喘振主动控制器利用二阶滑动模态的特性在扩展了压缩系统稳定工作范围的同时，保证了控制器对未建模动态及系统扰动等不确定性具有较强的鲁棒性。喘振主动/被动混合控制器使用放气阀门作为执行机构，喘振主动控制器作为控制系统的主模式可扩大压气机的稳定工作范围，而将防喘系统作为备份模式负责在主动控制模式失效时发挥避免压气机失稳的作用。混合控制器基于模糊逻辑设计，简化了控制设计过程及形式。提出了基于MG模型及混沌序列的压气机失稳先兆信号的模拟方法。失稳先兆信号的构造有利于开展高稳定性控制仿真试验研究。在缺少发动机喘振实验条件的情况下，可以利用构造的失稳先兆信号验证失稳预测算法的有效性。提出了基于时频分析、基于数学形态学分形维数及基于时间序列分析三种压缩系统失稳预测方案。基于时频分析的失稳预测方案利用形态滤波器等包络检测算法对隐藏在压气机测量参数轮廓中的低频扰动信号进行检测。数学形态学分形维数分析是一种先进的非线性信号处理方法，可以从信号复杂度的变化中探测系统状态的变化。复杂系统的状态变化，通常反映在系统输出信号中。对发动机稳定状态下的传感器输出信号进行时间序列建模，使用所建立的模型对传感器信号进行在线预测，模型预测误差的均方可作为发动机稳定性状况的度量。提出了一种基于SQP算法与稳定性寻优相结合的航空发动机加速优化控制方案。总结了发动机加速过程的特点及各种主要限制因素，开展了基于SQP算法的涡扇发动机加速过程优化控制研究。将SQP优化过程中的喘振裕度限制降至最低，根据失稳预测系统的输出对发动机工作点做出适当调整，实现在加速过程中剩余喘振裕度的充分利用。本文的研究成果可为喘振主动控制、稳定性寻求控制等高稳定性控制技术研究提供理论参考。更多还原

【Abstract】 Traditional surge avoidance control belongs to passive methods. The safety of the aero-engine isguaranteed by a conservative surge margin, but the performance of the engine is greatly sacrificed.Meanwhile, higher requirements are put forward on the propulsion system in moden advanced air-crafts, such as higher load, higher efficiency and higher thrust-weight-ratio. And high stability controlhas become one of key technologies in the advanced aero-engine. Therefore, following a review ofdevelopments in this field domestic and abroad, researches on key technologies of high stability con-trol are carried out, such as compression system modeling, compressor surge active control and com-pressor instability prediction. The application of the high stability control in turbofan engine accelera-tion optimization process is also researched.The compression system model of Moore Greitzer has been extended to include the spool dy-namic and higher harmonics of rotating stall. Simulation results test the influences of the spool dy-namic on the compression, which may lead to instability like the system inner disturbances.To explain the mechanism of the surge and rotating stall phenomena, nonlinear dynamic analysishas been done on MG model, based on bifurcation theory. It can be seen from the analysis results thatthe hysteresis loop of rotating stall is caused by a subcritical fork bifurcation at the peak point of thecompressor characteristic map, and the surge is related to a Hopf bifurcation, which will not occur inthe system with a low B-parameter.Four compressor surge active control strategies have been proposed, which are named surge androtating stall output feedback control, surge and rotating stall control via closed coupled valve (CCV)and throttle, surge control based on second order sliding mode and surge active/passive hybrid controlbased on fuzzy logic. A state estimator is employed to estimate the flow coefficient in the output con-troller, the pressure signal, which is easily to collect, is the only feedback parameter to achieve activecontrol. Compared to single actuator scheme, the advantages of using CCV and throttle as actuatorsimultaneously are demonstrated. Benefit from the character of the second order sliding mode, thesurge active controller possesses a better robustness on uncertainties such as un-modeled system dy-namic and disturbances. The active/passive hybrid control system uses a bleed valve as actuator, andcontains a surge active controller as main operating mode to enlarge compressor workingrange, and asurge avoidance controller as back-up mode to guarantee the reliability of the whole system. Thefuzzy logic simplifies the design process and the structure of the hybrid control system. Compressor instability precursor signal simulation methods based on MG model and chaotictime series have been studied. The construction of instability precursor signal is critical to compressorstability management simulation tests. The constructed signal also can be used to verify the effective-ness of compressor instability prediction algorithms with the absence of true experimental signal.Three compressor instability prediction algorithms have been proposed respectively based ontime-frequency analysis, mathematical morphology fractal dimension, and time series modeling.Envelope detection methods such as mathematical morphology filtering are used to detect low fre-quency instability precursors, which are hided in high frequency signals. Mathematical morphologyfractal dimension analysis is one advanced nonlinear signal processing technology, which can extractthe changes of compressor stability through quantization of the signal complexity. Since the variationsof complex system states are usually reflected in output signals, modelling the time series of systemoutput, prection the output by the model, the mean square of prediction error can be used as compres-sor stability index.A new aircraft engine acceleration optimization strategy which combined with SQP algorithmand stability management has been proposed. Based on summarization of the features and restrictivefactors of turbofan engine acceleration process, SQP based acceleration optimization is researched. Tomake full use of the remaining surge margin, the boundary condition of surge margin is minimized,and then, the engine operating point is adjusted realtime according to the output of the engine instabil-ity prediction system.All of the research results in this dissertation lay a solid theoretical reference on high stabilitycontrol technologies, including surge active control and stability seeking control.更多还原