航空发动机气路故障诊断的SANNWA-PF算法

许梦阳; 黄金泉; 鲁峰

doi:10.13224/j.cnki.jasp.2017.10.026

航空发动机气路故障诊断的SANNWA-PF算法

doi: 10.13224/j.cnki.jasp.2017.10.026

1.
南京航空航天大学能源与动力学院江苏省航空动力系统重点实验室,南京 210016
2.
南京航空航天大学能源与动力学院江苏省航空动力系统重点实验室,南京 210016；先进航空发动机协同创新中心,北京 100191

基金项目: 国家自然科学基金（51276087）；国家自然科学青年基金（61304133）；南京航空航天大学研究生创新基地（实验室）开放基金（kfjj20160211）

计量
- 文章访问数: 599
- HTML浏览量: 3
- PDF量: 431
- 被引次数: 0
出版历程
- 收稿日期: 2016-03-07
- 刊出日期: 2017-10-28

SANNWA-PF algorithm of aeroengine gas path fault diagnosis

摘要

摘要: 针对航空发动机非线性、非高斯的特点，提出一种用于航空发动机气路故障诊断的自适应神经网络权值调整粒子滤波（SANNWAPF）算法。该算法根据粒子分布情况确定分裂和调整的粒子数目，进而根据粒子权重采用正态分布的方式进行分裂，采用反向传插(BP)神经网络进行权值调整，缓解了粒子的退化和贫化，具有更强的自适应性能和跟踪能力。通过一维非线性跟踪模型和航空发动机气路故障诊断仿真研究表明：SANNWAPF算法具有良好的非高斯性能，相对粒子滤波一维非线性追踪模型估计精度提高约21%，航空发动机气路故障诊断在高斯噪声和非高斯噪声下分别提高约30%和26%，诊断速度分别提高约7倍和10倍。
- 航空发动机 /
- 故障诊断 /
- 粒子滤波 /
- 自适应 /
- 神经网络 /
- 非高斯噪声
Abstract: A selfadaptive neural network weight adjustment particle filter algorithm was proposed for aeroengine gas path fault diagnosis of the nonlinear and nonGaussian properties of aeroengine. Number of particles split and adjusted was determined by the distribution of particles. Then particles were spilt by the way of normal distribution and adjusted by back propagation (BP) neural network, which avoided the degradation and impoverishment of particles and had stronger selfadaptive and tracking ability. The simulation results of onedimensional nonlinear tracking model and aeroengine gas path fault diagnosis show that selfadaptive neural network weight adjustmentparticle filter (SANNWAPF) algorithm has a good nonGaussian performance. Compared with normal particle filter, SANNWAPF improved 21% in accuracy of onedimensional nonlinear tracking model, 30% with Gaussian noise and 26% with nonGaussian noise in aeroengine gas path fault diagnosis; and the diagnosis speed improved about 7 times with Gaussian noise and 10 times with nonGaussian noise.
- aero-engine /
- fault diagnosis /
- particle filter /
- selfadaptive /
- neural network /
- nonGaussian noise

HTML

参考文献(19)

[1]	MERCER C R,SIMON D L,HUNTER G W,et al.Fundamental technology development for gasturbine engine health management［R］.NASA/TM-2007-0022364,2007.
[2]	LITT J S,SIMON D L,GARY S,et al.A survey of intelligent control and health management technologies for aircraft propulsion systems［R］.NASA/TM-2005-213622,2005.
[3]	LUPPOLD R H,GALLOPS G W,KERR L J,et al.Estimating inflight engine performance variations using Kalman filter concepts［R］.AIAA 89-2584,1989.
[4]	SIMON D L,LITT J S.Application of a constant gain extended Kalman filter for inflight estimation of aircraft engine performance parameters［R］.NASA/TM-2005-213865,2005.
[5]	LU F,HUANG J Q,L Y Q.Gas path health monitoring for a turbofan engine based on a nonlinear filtering approach［J］.Energies,2013,6(1):492-513.
[6]	王修岩,李萃芳,高铭阳,等.基于SVM和SNN的航空发动机气路故障诊断［J］.航空动力学报,2014,29(10):2493-2498. WANG Xiuyan,LI Cuifang,GAO Mingyang,et al.Fault diagnosis of aeroengine gas path based on SVM and SNN［J］.Journal of Aerospace Power,2014,29(10):2493-2498.(in Chinese)
[7]	MERWE R V D,DOUCET A,FREITAS N D,et al.The unscented particle filter［R］.Cambridge,UK:Cambridge University,CUED/F-INPENG/TR-280,2000.
[8]	CHRISTOPHE B,LAURENT T,JEAN G.EKF and particle filter track to track fusion:a quantitative comparison from radar/lidar obstacle tracks［R］.Philadelphia,PA:8th International Conference on Information Fusion,2005.
[9]	JULIER S J,UHLMANN J K.Unscented filtering and nonlinear estimation［J］.Proceedings of the IEEE,2004,92(3):401-422.
[10]	HAMMERSLEY J M,MORTON K W.Symposium on Monte Carlo methods:poor mans Monte Carlo［J］.Journal of the Royal Statistical Society,1954,16(1):23-38.
[11]	GORDON N J,SALMOND D J,SMITH A F M.Novel approach to nonlinear/nonGaussian Bayesian state estimation［J］.Radar and Signal Processing,2002，140(2):107-113.
[12]	DOUCET A.On sequential Monte Carlo methods for Bayesian filtering［J］.Statistics and Computing,1998,10(3):197-208.
[13]	ARULAMPALAM M S,MASKELL S,GORDON N,et al.A tutorial on particle filters for online nonlinear/nonGaussian Bayesian tracking［J］.IEEE Transaction Signal Processing,2002,50(2):174-188.
[14]	ADALI T,HAYKIN S.Adaptive signal processing:next generation solutions［M］.New York:Wiley,2010.
[15]	朱志宇.粒子滤波算法及其应用［M］.北京：科学出版社,2010.
[16]	PANG Z,LIU D,JIN N,et al.Neural network strategy for sampling of particle filters on the tracking problem［R］.Orlando:International Joint Conference on Neural Networks,2007.
[17]	MAHTAB S S,ABDOLAH R,NASER P.Classifying the weights of particle filter in nonlinear systems［J］.Communications in Nonlinear Science and Numerical Simulation,2015,31(1):69-75.
[18]	BORGUET S,LONARD O.Comparison of adaptive filters for gas turbine performance monitoring［J］.Journal of Computational and Applied Mathematics,2010,234(7):2202-2212.
[19]	李文峰,王永生,杨纪明,等.航空发动机测试信号噪声特性分析［J］.航空动力学报,2005,20(5):900-904. LI Wenfeng,WANG Yongsheng,YANG Jiming,et al.Noise characteristics analysis of aeroengine testing signals［J］.Journal of Aerospace Power,2005,20(5):900-904.(in Chinese)