基于组合优化相关向量机的航空发动机性能参数概率预测方法
收稿日期: 2012-11-19
修回日期: 2013-06-04
网络出版日期: 2013-06-09
基金资助
省部级项目
Probabilistic Prediction Method for Aeroengine Performance Parameters Based on Combined Optimum Relevance Vector Machine
Received date: 2012-11-19
Revised date: 2013-06-04
Online published: 2013-06-09
Supported by
Ministry Level Project
针对航空发动机性能参数预测过程中存在的不确定因素,提出一种基于组合优化相关向量机(CORVM)的概率预测方法。首先,通过正交小波变换将性能参数序列分解为具有不同特征尺度的随机分量和趋势分量,并分别建立各分量的相关向量机(RVM)回归预测模型。然后,以留一交叉验证误差最小作为优化目标,采用量子粒子群优化(QPSO)算法实现RVM核参数和嵌入维数的自适应优化选择。最后,组合各RVM回归预测模型得到最终预测均值及方差,进而得到预测值的概率分布。通过航空发动机排气温度变化量和滑油金属含量预测实例验证了方法的有效性,实验结果表明:该方法能够有效避免传统点预测方法可能导致的不可靠结果,并具有比单一模型更高的预测精度。
范庚 , 马登武 . 基于组合优化相关向量机的航空发动机性能参数概率预测方法[J]. 航空学报, 2013 , 34(9) : 2110 -2121 . DOI: 10.7527/S1000-6893.2013.0295
To cope with the uncertainties in the prediction process of aeroengine performance parameters, a probabilistic prediction method is proposed based on a combined optimum relevance vector machine (CORVM). Firstly, the performance parameter sequence is decomposed into sub-sequences in different frequency bands by orthogonal wavelet transform, and the prediction models of these sub-sequences based on relevance vector machine (RVM) regression are established respectively. Secondly, the quantum-behaved particle swarm optimization (QPSO) algorithm is employed to optimize the kernel parameters and embedding dimensions, which uses the minimum leave-one-out cross-validation error as the optimization target. Finally, all the prediction models are combined to obtain the final prediction values and variances. Thus the probabilistic distributions of prediction values are obtained. The validity of the proposed method is proved by experiments on aero-engine delta exhaust gas temperature prediction and lubrication metal content prediction. The experimental results show that the proposed method can avoid unreliability results and has better performance in prediction accuracy than the single prediction model.
[1] Jiang C H, Sun Z Y, Wang X. Critical technologies for aero-engine prognostics and health management systems development. Journal of Aerospace Power, 2009, 24(11): 2589-2594. (in Chinese) 姜彩虹, 孙志岩, 王曦. 航空发动机预测健康管理系统设计的关键技术. 航空动力学报, 2009, 24(11): 2589-2594.
[2] Wei X K, Feng Y, Liu F, et al. Development strategy and key prognostics health management technologies for military aero-engine in China. Journal of Aerospace Power, 2011, 26(9): 2107-2115. (in Chinese) 尉询楷, 冯悦, 刘芳, 等. 军用航空发动机PHM发展策略及关键技术. 航空动力学报, 2011, 26(9): 2107-2115.
[3] Li X B, Cui X L, Lang R L. Forecasting method for aeroengine performance parameters. Journal of Beijing University Aeronautics and Astronautics, 2008, 34(3): 253-256. (in Chinese) 李晓白, 崔秀伶, 郎荣玲. 航空发动机性能参数预测方法. 北京航空航天大学学报, 2008, 34(3): 253-256.
[4] Xu Z P, Lang R L, Deng X L. Uncertainty analysis of aircraft flight parameters prediction. Acta Aeronautica et Astronautica Sinica, 2012, 33(6): 1100-1107. (in Chinese) 许喆平, 郎荣玲, 邓小乐. 飞机性能参数预测的不确定性处理. 航空学报, 2012, 33(6): 1100-1107.
[5] Gan M L, Yang Z. Time-series modeling and fault forecast study spectral analysis of lubricating oil. Transactions of Nanjing University of Aeronautics & Astronautics, 2001, 18(1): 86-90.
[6] Yang Y W, Chen G. Modeling of spectrometric oil analysis data based on structure self adaptive RBF neural network. Chinese Journal of Scientific Instrument, 2006, 27(1): 98-101. (in Chinese) 杨虞微, 陈果. 基于结构自适应径向基神经网络的油样光谱数据建模. 仪器仪表学报, 2006, 27(1): 98-101.
[7] Ding G, Fu X Y, Zhong S S. Aeroengine performance parameters prediction based on process neural network. Computer Integrated Manufacturing Systems, 2011, 17(1): 198-207. (in Chinese) 丁刚, 付旭云, 钟诗胜. 基于过程神经网络的航空发动机性能参数预测. 计算机集成制造系统, 2011, 17(1): 198-207.
[8] Wei X K, Li Y H, Wang S, et al. Aeroengine lubrication monitoring analysis via support vector machines. Journal of Aerospace Power, 2004, 19(3): 392-397. (in Chinese) 尉询楷, 李应红, 王硕, 等. 基于支持向量机的航空发动机滑油监控分析. 航空动力学报, 2004, 19(3): 392-397.
[9] Li B W, Zhang Y, Sun T. Design of forecasting model for aero engine lubrication debris support vector machines based on immune-particle swarm optimization algorithm. Journal of Aerospace Power, 2009, 24(7): 1639-1643. (in Chinese) 李本威, 张赟, 孙涛. 基于免疫粒子群算法的滑油屑末支持向量机预测模型设计. 航空动力学报, 2009, 24(7): 1639-1643.
[10] Zhang X, Wang H L. Condition time series prediction using least squares support vector machine with adaptive embedding dimension. Acta Aeronautica et Astronautica Sinica, 2010, 31(12): 2309-2314. (in Chinese) 张弦, 王宏力. 嵌入维数自适应最小二乘支持向量机状态时间序列预测方法. 航空学报, 2010, 31(12): 2309-2314.
[11] Xu K, Xu J W, Ban X J. Forecasting of some non-stationary time series based on wavelet decomposition. Acta Electronica Sinica, 2001, 29(4): 566-568. (in Chinese) 徐科, 徐金梧, 班晓娟. 基于小波分解的某些非平稳时间序列预测方法. 电子学报, 2001, 29(4): 566-568.
[12] Zhang X, Wang H L. Failure rate time series prediction based on support vector empirical mode decomposition. Acta Aeronautica et Astronautica Sinica, 2011, 32(3): 480-487. (in Chinese) 张弦, 王宏力. 基于支持向量经验模态分解的故障率时间序列预测. 航空学报, 2011, 32(3): 480-487.
[13] Tipping M E. The relevance vector machine//Advances in neural information processing systems 12. Cambridge, USA: MIT Press, 2000: 652-658.
[14] Tipping M E. Sparse Bayesian learning and the relevance vector machine. Journal of Machine Learning Research, 2001, 1(3): 211-244.
[15] Tipping M E, Faul A C. Fast marginal likelihood maximisation for sparse Bayesian models. Proceedings of the Ninth International Workshop on Artificial Intelligence and Statistics, 2003: 1-13.
[16] Gong Z Q, Zou M W, Gao X Q, et al. On the difference between empirical mode decomposition and wavelet decomposition in the nonlinear time series. Acta Physica Sinica, 2005, 54(8): 3947-3957. (in Chinese) 龚志强, 邹明玮, 高新全, 等. 基于非线性时间序列分析经验模态分解和小波分解异同性的研究. 物理学报, 2005, 54(8): 3947-3957.
[17] Refaeilzadeh P, Tang L, Liu H. Cross-validation//Liu L, Ozsu M T. Encyclopedia of database systems. New York: Springer, 2009: 532-538.
[18] Chen G, Yan F, Gong X, et al. State estimate based on parameter-optimized least squares support vector machines. Power System Protection and Control, 2011, 39(19): 83-88. (in Chinese) 陈刚, 闫飞, 龚啸, 等. 基于参数优化的最小二乘支持向量机状态估计方法. 电力系统保护与控制, 2011, 39(19): 83-88.
[19] Sun J. Particle swarm optimization with particles having quantum behavior. Wuxi: School of Internet of Things Engineering, Jiangnan University, 2009. (in Chinese) 孙俊. 量子行为粒子群优化算法研究. 无锡: 江南大学物联网工程学院, 2009.
[20] Luan S G. Aeroengine condition monitoring technique and system based on gas path parameter sample. Harbin: School of Mechanical and Electrical Engineering, Harbin Institute of Technology, 2008. (in Chinese) 栾圣罡. 基于气路参数样本的航空发动机状态监视方法与系统研究. 哈尔滨: 哈尔滨工业大学机电工程学院, 2008.
[21] Yang Y W, Chen G. Artificial neural network forecasting method in monitoring technique by spectrometric oil analysis. Spectroscopy and Spectral Analysis, 2005, 25(8): 1339-1343. (in Chinese) 杨虞微, 陈果. 光谱油样分析监测技术中的神经网络预测方法. 光谱学与光谱分析, 2005, 25(8): 1339-1343.
/
| 〈 |
|
〉 |
CJA