PDF(1163 KB)
Long-term prediction monitoring by principal component analysis（PCA） based on hybrid models
WANG GaoSheng; LIU ZhenJuan; LI HongGuang
Journal of Beijing University of Chemical Technology ›› 2014, Vol. 41 ›› Issue (2) : 104-108.
PDF(1163 KB)
PDF(1163 KB)
Long-term prediction monitoring by principal component analysis（PCA） based on hybrid models
Traditional principal component analysis (PCA)-based process monitoring technologies focus on the current states of industrial processes rather than their future behavior. Motivated by this observation, this paper introduces a PCA-based predictive model which implements prediction of the operating states of industrial processes. Monitoring models of a PCA combined statistical index are first established based on normal historical data for the process. A k-NN least squares support vector machine (LSSVM) combined with GM (1,1) is then adopted to establish a on-line combinatorial prediction model, which allows forecasting of future states of the industrial process. Simulations on a multi-variables dynamic process demonstrated that the proposed method is more accurate than either k-NN based LSSVM or GM(1,1) in long-term prediction monitoring. The results of this work prove that PCA based on hybrid models is especially suitable for the long-term prediction of slowly drifting faults.
[1]王雅琳, 欧文军, 欧春华, 等. 基于PCA和改进BP神经网络的浮选精矿品位在线预测[C]∥第29届中国控制会议, 北京, 2010: 2347-2353.
Wang Y L, Ou W J, Ou C H, et al. Online prediction of concentrate grade in flotation process based on PCA and improved BP neural networks[C]∥Proceedings of the 29th Chinese Control Conference, Beijing, 2010: 2347-2353. (in Chinese)
[2]He D, Liu R Y. Ultra-short-term Wind Power Prediction using ANN Ensemble Based on PCA[C]∥7th International Power Electronics and Motion Control Conference, Harbin, 2012: 2108-2112.
[3]赵珍, 王福利, 贾明兴, 等. 缓变故障的概率故障预测方法研究[J]. 控制与决策, 2010, 25(4): 572-576.
Zhao Z, Wang F L, Jia M X, at al. Research on probabilistic fault prediction of incipient fault[J]. Control and Decision, 2010, 25(4): 572-576. (in Chinese)
[4]杨珍贵, 郝永生, 赵双龙. 基于LS-SVM的多步故障预测模型[J]. 现代电子技术, 2011, 34(3): 25-27.
Yang Z G, Hao Y S, Zhao S L. LS-SVM-based Multi-step Failure Prediction Model[J]. Modern Electronics Technique, 2011, 34(3): 25-27. (in Chinese)
[5]Zhao Z, Wang F L, Jia M X, et al. Probabilistic Fault Prediction of Incipient Fault[C]∥Chinese Control and Decision Conferrence, Xuzhou Jiangsu, 2010: 3911-3915.
[6]王姝, 胡学发, 赵珍, 等. 基于MPCA-AR的故障预测方法研究[J]. 仪器仪表学报, 2009, 30(8): 1778-1782.
Wang S, Hu X F, Zhao Z, et al. Research on fault prediction method based on MPCA-AR[J]. Chinese Journal of Scientific Instrument, 2009, 30(8): 1778-1782. (in Chinese)
[7]Wang S, Zhao Z, Wang F L, et al. Multi-step Ahead Fault Prediction Method Based on PCA and EMD[C]∥Chinese Control and Decision Conferrence, Mianyang, Sichuan: 2011: 2874-2878.
[8]Yue H H, Qin S J. Reconstruction-based fault identification using a combined index[J]. Ind Eng Chem Res, 2001, 40(20): 4403-4414.
[9]Cover T M, Hart P E. Nearest neighbor pattern classification [J]. Transactions on Information Theory, 1967, 13(1): 21-27.
[10]Huang Z F, Shyu M L. k-NN based LS-SVM framework for long-term time series prediction[C]∥IEEE International Conference on Information Reuse and Integration, Las Vegas, USA, 2010: 69-74.
[11]Stefánsson A， KonAcˇGar N, Jones A J. A note on the Gamma test[J]. Neural Computing and Applications, 1997, 5(3): 131-133.
PDF(1163 KB)
/
| 〈 |
|
〉 |
