The on-board dynamic model is the basis of numerous advanced control technologies for modern aero-engine, and its accuracy and real-time performance are two crucial indicators. Since the simulation accuracy declines with the deviation from the ground point of the conventional compact propulsion system dynamic model (CPSDM), an improved CPSDM (ICPSDM) based on deep neural network is proposed in this paper. First, the K-means algorithm is utilized to get the sampling point, and the engine simulation data is collected in the cluster centers. Then, the batch normalize deep neural network (BN-DNN) is applied to build the ICPSDM, which can shorten the training time tremendously. The simulation results show that, in the same simulation environment, the accuracy of turbine inlet temperature T4, fan surge margin SMc, thrust F, and specific fuel consumption SFC calculated by ICPSDM are 11.04, 3.17, 1.89, and 1.98 times of CPSDM, respectively, at off-design point, and the real-time performance of ICPSDM is more than 30 times faster than the component-level model.
HuangWHuangJ. Onboard self-tuning model for aero-engine fault diagnostics. J Aerosp Power2008; 03: 580–584.
2.
WangZ.ZaraderJ. L.ArgentieriS.A novel aircraft engine fault diagnostic and prognostic system based on svm, in condition monitoring and diagnosis (CMD). In: 2012 international conference on IEEE, Bali, Indonesia, 24 January 2013.
3.
YaoYSunJ. Aeroengine direct thrust control based on neural network inverse control. J Aerosp Power2008; 29: 249–252.
4.
ZhengQPangSZhangH, et al.A study on aero-engine direct thrust control with nonlinear model predictive control based on deep neural network. Int J Aeronaut Space Sci2019; 20: 1–7.
5.
YaoWSunJ. nonlinear model predictive control for turboshaft engine. Acta Aeronaut Astronaut Sinica2008; 04: 15–19.
6.
WangJZhangHHuangX, et al.Hybrid predictive control for turbo-shaft engine based on online sliding LS-SVR. Acta Aeronaut Astronaut Sinica2012; 33: 1755–1764.
7.
WangJZhangHHuangX, et al.nonlinear model predictive control for the engine based on an integrated helicopter/turbo-shaft engine simulation platform. Acta Aeronaut Astronaut Sinica2012; 33: 402–411.
8.
YinLTuressonGTunestalP, et al.Model predictive control of an advanced multiple cylinder engine with partially premixed combustion concept. IEEE/ASME Trans Mechatron2020; 25: 804–814.
9.
SunJ. Prospects of the aeroengine control development in the early time of the 21st century. J Aerosp Power2001; 16: 97–102.
10.
XiaFHuangJZhouW. Modeling of and simulation research on turbofan engine based on MATLAB/SIMULINK. J Aerosp Power2007; 12: 2134–2138.
11.
HFishbachL. Computer simulation of engines systems. In: Aerosp sci. meeting, Pasadena, CA, 1 January 1980.
12.
CurlettB. P.FelderJ. L., Object-oriented approach for gas turbine engine simulation, In NASA-TM-106970, 1995.
13.
YangGSunJHuangX, et al.A non-iterative method of aero-engine modeling using complementary variables. J Aerosp Power2003; 2: 289–294.
14.
YinKZhouWQiaoK, et al.Research on methods of improving real-time performance for aero-engine component-level model. J Propul Technol2017; 01: 199–206.
15.
ZhouWHuangJHuangK. Real-time simulation system for aeroengine based on simplified model. J Nanjing Univ Aeronaut. Astronaut2005; 37: 251–255.
16.
LiSLSunJG. Application of genetic algorithm to solving nonlinear model of aeroengines. Chin J Aeronautics2003; 16(2): 69–72.
YangWFengLPengJ. An intelligent algorithm for solution of nonlinear mathematical model for aeroengine. J Propul Technol2008; 29: 614–616.
19.
LuJGuoYZhangS. Research on the iteration methods in aero-engine non-linear model real-time computation. J Aerosp Power2010; 025: 681–686.
20.
FengZSunJHuangJ, et al.A newmethod for establishinga state variable model of aeroengine. J Aerosp Power1998; 04: 435–438.
21.
LiJGuoYYuanX. Turbofan engine modeling of piece-wise linear model. Computer Simulation2007; 24: 37–40.
22.
YuXLiYWangJ, et al.A novel aero-engine identification model based on support vector machines. J Aerosp Power2004; 19: 684–688.
23.
XuLHuangJ. Dynamic identification with neural networks for aircraft engines in the full envelope. J. Nanjing University of Aeronaut. & Astronaut, 04, 2001, 334–337.
24.
YaoYSunJ. Aeroengine direct thrust control based on neural network inverse control. J Propul Technol2008; 04: 15–19.
25.
ZhengQFangJHuZ, et al.Aero-engine on-board model based on batch normalize deep neural network. IEEE Access2019; 7: 54855–54862.
26.
LecunYBottouLBengioY, et al.Gradient-based learning applied to document recognition. Proc IEEE1998; 86(11): 2278–2324.
27.
NobbsSJacobsSDonahueD. Development of the full-envelope performance seeking control algorithm. In: Joint propulsion conference & exhibit, Nashville, TN, USA, 06–08 July 1992.
28.
StewartJF, Integrated flight propulsion control research result using the NASA F-15 HIDEC flight research facility, In AIAA-92-4106, 1992.
29.
OrmeJSGilyardGB. Preliminary Supersonic Flight Test Evaluation of Performance Seeking Control. AIAA-93-1821, 1993.
30.
JainAK. Data clustering: 50 years beyond K-means. Pattern Recognition Lett2010; 31(8): 651–666.
31.
IoffeSSzegedyC. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. JMLR.org, 2015.
32.
SanturkarSTsiprasDIlyasA, et al.How does batch normalization help optimization. In: 32nd Conference on Neural Information Processing Systems; 2018.
33.
ShangY. Aviation Gas Turbine Engine. Beijing: Aviation industry press, 1995.
