Fault detection in hot steel rolling using neural networks and multivariate statistics

The paper addresses the issue of maintaining consistent high quality production in the steel industry by extending techniques emanating from the fields of neural networks and multivariate statistics. Process diagnostic methodologies based on these tools were developed and applied to a six-stand hot rolling mill. The objective was to achieve better mill setup parameters so that the manufactured coils consistently meet the required customer specifications. A wavelet neural network was successfully used for modelling the mill parameters and for detecting errors in the rolling stand settings. Model prediction accuracy and robustness were enhanced through stacked generalisation. Multivariate statistical performance monitoring techniques were then applied on top of the mill control systems to provide early warning of strips being badly rolled. Both approaches yielded comparable results on monitored data from a hot strip mill and, in combination, provided enhanced manufacturing performance.

[1]  David H. Wolpert,et al.  Stacked generalization , 1992, Neural Networks.

[2]  Shang-Liang Chen,et al.  Orthogonal least squares learning algorithm for radial basis function networks , 1991, IEEE Trans. Neural Networks.

[3]  P. Miller,et al.  Contribution plots: a missing link in multivariate quality control , 1998 .

[4]  Bhavik R. Bakshi,et al.  Empirical Learning Through Neural Networks: The Wave-Net Solution , 1995 .

[5]  Jie Zhang,et al.  Process performance monitoring using multivariate statistical process control , 1996 .

[6]  Robert Tibshirani,et al.  A Comparison of Some Error Estimates for Neural Network Models , 1996, Neural Computation.

[7]  Yagyensh C. Pati,et al.  Analysis and synthesis of feedforward neural networks using discrete affine wavelet transformations , 1993, IEEE Trans. Neural Networks.

[8]  B. Kowalski,et al.  Partial least-squares regression: a tutorial , 1986 .

[9]  C. Kiparissides,et al.  Inferential Estimation of Polymer Quality Using Stacked Neural Networks , 1997 .

[10]  Babu Joseph,et al.  Wavelet applications in chemical engineering , 1994 .

[11]  Qinghua Zhang,et al.  Wavelet networks , 1992, IEEE Trans. Neural Networks.

[12]  S. Wold.  Cross-Validatory Estimation of the Number of Components in Factor and Principal Components Models , 1978 .

[13]  John F. MacGregor,et al.  Multivariate SPC charts for monitoring batch processes , 1995 .

[14]  N. Draper,et al.  Applied Regression Analysis , 1966 .

[15]  Theodora Kourti,et al.  Statistical Process Control of Multivariate Processes , 1994 .

[16]  I. Jolliffe.  Principal Component Analysis , 2002 .