Operation Status Prediction Based on Top Gas System Analysis for Blast Furnace

Top gas is a critical reference for scheduling workers to analyze the operation status of a blast furnace (BF). An accurate prediction of top gas indices can help to forecast the operation status of BF. In this brief, an operation status analytical method is proposed based on top gas indices forecast for BF. The Hammerstein system is used to approximate the actual model of top gas indices series. Bayesian technique and reversible jump Markov Chain Monte Carlo method are adopted for parameter and model order identification in the proposed prediction algorithm. In the system, model order, parameters, and regularization parameters cannot be obtained directly and we compute these parameters based on sampling theory. The operation status of BF is reflected by indices such as the contents of CO and CO2, temperature, and pressure. The confidence regions are computed based on large amounts of top gas indices data, in which the prediction of the indices is considered to be normal. To verify the effectiveness of the proposed method, the top gas of #2 BF from Liuzhou Iron and Steel is employed in the experiments. A series of experiments are reported and the confidence regions of top gas indices are presented for the operation status prediction of BF.

[1]  R. Luus,et al.  A noniterative method for identification using Hammerstein model , 1971 .

[2]  Abdul Rahman Mohamed,et al.  Neural networks for the identification and control of blast furnace hot metal quality , 2000 .

[3]  G. U. Yule,et al.  The Foundations of Econometric Analysis: On a Method of Investigating Periodicities in Disturbed Series, with Special Reference to Wolfer's Sunspot Numbers ( Philosophical Transactions of the Royal Society of London , A, vol. 226, 1927, pp. 267–73) , 1995 .

[4]  Jan-Willem van Wingerden,et al.  Global Identification of Wind Turbines Using a Hammerstein Identification Method , 2013, IEEE Transactions on Control Systems Technology.

[5]  W. K. Hastings,et al.  Monte Carlo Sampling Methods Using Markov Chains and Their Applications , 1970 .

[6]  Simon J. Godsill,et al.  Bayesian model selection of autoregressive processes , 2000 .

[7]  K. Narendra,et al.  An iterative method for the identification of nonlinear systems using a Hammerstein model , 1966 .

[8]  S. Wu,et al.  Identification of multiinput-multioutput transfer function and noise model of a blast furnace from closed-loop data , 1974 .

[9]  P. Green.  Reversible jump Markov chain Monte Carlo computation and Bayesian model determination , 1995 .

[10]  Fumio Kojima,et al.  Inverse Problem for Electromagnetic Propagation in a Dielectric Medium using Markov Chain Monte Carlo Method (Preprint) , 2011 .

[11]  J. Besag.  A candidate's formula: A curious result in Bayesian prediction , 1989 .

[12]  E. Bai.  An optimal two stage identification algorithm for Hammerstein-Wiener nonlinear systems , 1998 .

[13]  A. Doucet,et al.  Reversible Jump Markov Chain Monte Carlo Strategies for Bayesian Model Selection in Autoregressive Processes , 2004, Journal of Time Series Analysis.

[14]  Yusheng Liu,et al.  Nonlinear Identification of Laser Welding Process , 2007, IEEE Transactions on Control Systems Technology.

[15]  Ligang Wu,et al.  Quantized Control Design for Cognitive Radio Networks Modeled as Nonlinear Semi-Markovian Jump Systems , 2015, IEEE Transactions on Industrial Electronics.

[16]  Sean R. Anderson,et al.  Computational system identification for Bayesian NARMAX modelling , 2013, Autom..

[17]  Hiroshi Nogami,et al.  Prediction of Blast Furnace Performance with Top Gas Recycling , 1998 .

[18]  Hannu Helle,et al.  Multi-objective Optimization of Ironmaking in the Blast Furnace with Top Gas Recycling , 2010 .

[19]  R. Stephenson.  A and V , 1962, The British journal of ophthalmology.