基于自回归模型的RBCC隔离段激波串位置识别与压力值预估  

RBCC isolator shock train location identification and pressure prediction based on Auto-Regressive model

作　　者：马文蕙 何国强[1] 王亚军[1] 王鹏飞 秦飞[1] 张铎[1] 朱韶华 党文娟 MA Wenhui;HE Guoqiang;WANG Yajun;WANG Pengfei;QIN Fei;ZHANG Duo;ZHU Shaohua;DANG Wenjuan(College of Astronautics,Northwestern Polytechnical University,Xi’an 710072,China;Xi’an Changfeng Electromechanics Institute,Xi’an 710065,China)

机构地区：[1]西北工业大学航天学院,陕西西安710072 [2]西安长峰机电研究所,陕西西安710065

出　　处：《推进技术》2024年第10期66-74,共9页Journal of Propulsion Technology

摘　　要：为了清楚客观地判断火箭基组合循环发动机(Rocket-based combined-cycle,RBCC)隔离段激波串位置,将Ma=6,4,3.5工况下直连试验中实测得到的RBCC隔离段测压点压力数据按照时间的先后顺序排列形成一时间序列,建立自回归(Auto-Regressive,AR)模型并计算赤池信息准则(Akaike information criterion,AIC)值,完成了不同工况下激波串前缘位置的识别。研究表明:当隔离段测压点没有受到激波串影响时,实时压力值仅存在微弱波动,模型AIC值变化较为平稳;当激波串运动至测压点处时,该点压力升高,振荡幅度明显增加,AIC值随之瞬时增大。取同一时间段内发动机沿程测压点中首个AIC值增加500以上,并在不改变工况的情况下始终保持较大值的测点位置为激波串前缘位置。与压比法相比,时间序列分析法能敏感监测到实时压力值的升高和振荡,激波串前缘位置识别更为准确。通过建立自回归模型还可以实现激波串内部压力值预估,记录连续160 ms内Ma=6,4,3.5工况下测压点压力数据,采样频率1 kHz,使用前80 ms数据建立自回归模型,完成后80 ms压力值预估及准确性检验,得到三个工况下预估平均误差分别为3.21%,7.68%,6.49%。In order to clearly and objectively identify the location of the shock train in the isolator of the Rocket-based combined-cycle(RBCC),the measured pressure data of the RBCC isolator in the direct connec⁃tion test under Ma=6,4,3.5 conditions were arranged according to the order of time to form a time series,Auto-Regressive(AR)model was established and the data of Akaike information criterion(AIC)was calculated.The location of the shock train was identified under different working conditions.The results show that when the pres⁃sure measurement point of the isolator is not affected by the shock train,the real-time pressure only fluctuates slightly,and the AIC changes steadily.When the shock train moves to the pressure measurement point,the pres⁃sure at the point increases,the oscillation amplitude increases obviously,then the AIC increases instantaneous⁃ly.The position where the first AIC of the measurement point along the engine increases by more than 500 in the same time period and maintains a larger value without changing the test condition is the location of the shock train.Compared with the pressure ratio method,the time series analysis method can sensitively monitor the rise and oscillation of the pressure,the shock train leading edge location identification is more accurate.The Auto-Regressive model can also be used to predict the internal pressure of the shock train.The pressure data of the measurement point under Ma=6,4,3.5 conditions within 160 ms were taken,and the sampling frequency was 1 kHz.The Auto-Regressive model is established using the first 80 ms data,and the pressure prediction and accu⁃racy test of the last 80 ms are completed.The average error of prediction under the three working conditions is 3.21%,7.68%and 6.49%,respectively.

关 键 词：火箭基组合循环发动机 激波串 自回归模型 赤池信息准则 时间序列 

分 类 号：V430[航空宇航科学与技术—航空宇航推进理论与工程]