Soft sensor modeling based on Gaussian processes

Abstract

In order to meet the demand of online optimal running, a novel soft sensor modeling approach based on Gaussian processes was proposed. The approach is moderately simple to implement and use without loss of performance. It is trained by optimizing the hyperparameters using the scaled conjugate gradient algorithm with the squared exponential covariance function employed. Experimental simulations show that the soft sensor modeling approach has the advantage via a real-world example in a refinery. Meanwhile, the method opens new possibilities for application of kernel methods to potential fields.

Keywords

Gaussian processes / soft sensor / modeling / kernel methods

References

[1] MartinG. Consider soft sensors[J]. Chemical Engineering Progress, 1997, 66(7): 66-70

[2] MacKayD J CIntroduction to Gaussian processes[R], 1998, Cambridge, Cambridge University

[3] SeegerMGaussian processes for machine learning[R], 2004, Berkeley, University of California at Berkeley

[4] RasmussenC EBousquetO, von LuxburgU, RatschG. Gaussian processes in machine learning[A]. Advanced Lectures on Machine Learning [C], 2004, Heidelberg, Springer-Verlag: 63-71

[5] WilliamsC K I, RasmussenC ETouretzkyD C, MozerM C, HasselmoM E. Gaussian processes for regression[A]. Advances in Neural Information Processing Systems 8[C], 1997, Cambridge, MIT Press

[6] NealR MBayesian learning for neural networks[D], 1995, Toronto, Department of Computer Science, University of Toronto

[7] WahbaGSpline Models for Observational Data[M], 1990, Philadelphia, SIAM

[8] ScholkopfB, SmolaA JLearning with Kernels[M], 2002, Cambridge, MIT Press

[9] FengRui, ShenWei, ShaoHui-he. A soft sensor modeling approach using support vector machines[A]. Proceedings of the 22nd American Control Conference[C], 2003, Piscataway, IEEE: 3702-3707

[10] WangXu-dong, LuoRong-fu, ShaoHui-he. Designing a soft sensor for distillation column with the fuzzy distributed radial basis function neural networks[A]. Proceedings of the 35th IEEE Conference on Decision and Control [C], 1996, New York, IEEE: 1714-1719