Research of Fan System Identification Based on Improved Bayes Algorithm

Abstract:

In every filed of natural science, more and more researchers attach importance to system quantitative analysis, control and prediction. In filed of automatic control, system identification is the extension of system dynamic characteristics testing. System modeling is the basis of system identification, non-parametric model can be obtained by means of dynamic characteristics testing, but parametric model must be established by means of parameter estimation algorithm, which is more prevalent than dynamic characteristics testing. Coal power plant produces more gas and dust, so how to control the fan system plays a very important role in environment protection. We must clarify the parameter of fan system before controlling it. The traditional Bayes identification algorithm is used widely in research and industry, and the effect is relatively good. The paper induces the concept of loss function based on traditional Bayes identification algorithm, and proposes an improved Bayes identification algorithm, which can be applied to fan system identification successfully.

• Periodical: Applied Mechanics and Materials (Volumes 66-68)
• Pages: 448-453
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.66-68.448
• Online since: July 2011
• Authors: Hai Tao Wang, Ze Zhang
• Keywords: Bayes Identification, Estimation, Fan System, System Modeling

References

[1] Karakos Damianos, Khudanpur Sanjeev. Error bounds and improved probability estimation using the maximum likelihood set[A]. IEEE International Symposium on Information Theory – Proceedings, 2007. 1851-1855.

DOI: 10.1109/isit.2007.4557150

[2] Singh, S. Ghosh, A.K. Parameter estimation from flight data of a missile using maximum likelihood and neural network method [A]. Atmospheric Flight Mechanics Conference. 2006. 714-734.

DOI: 10.2514/6.2006-6284

[3] Ferrari, Marco Bellini, Sandro. Maximum likelihood decoding of turbo codes on the binary erasure channel[A]. IEEE International Confe- rence on Communications, 2008. 1144-1149.

DOI: 10.1109/icc.2008.223

[4] Qi Yongqing, Jing Zhongliang, Hu, Shiqiang. Modified maximum likelihood registration based on information fusion [J]. Chinese Optics Letters, 2007, 11: 639-641.

[5] Klein V. Estimation of bias errors in measured airplane responses using maxium likelihood method. NA SA N 87-15977, (1987).

[6] FE IK R A. Determ ination of instrument erro rs by compatibility check ing results from dynamic fligh t test data. AD-A 159 427, (1985).

[7] HOU Yuan-bin, LI Ning. Identification of Random Noise for Nonlinear System Based on Improved Maximum-Likelihood Estimation [J]. Noise and Vibration Control, 2008. 12: 136-139.