Comparative performance evaluation of blast furnace flame temperature prediction using artificial intelligence and statistical methods

The blast furnace (BF) is the heart of the integrated iron and steel industry and used to produce melted iron as raw material for steel. The BF has very complicated process to be modeled as it depends on multivariable process inputs and disturbances. It is very important to minimize operational costs and reduce material and fuel consumption in order to optimize overall furnace efficiency and stability, and also to improve the lifetime of the furnace within this task. Therefore, if the actual flame temperature value is predicted and controlled properly, then the operators can maintain fuel distribution such as oxygen enrichment, blast moisture, cold blast temperature, cold blast flow, coke to ore ratio, and pulverized coal injection parameters in advance considering the thermal state changes accordingly. In this paper, artificial neural network (ANN), multiple linear regression (MLR), and autoregressive integrated moving average (ARIMA) models are employed to forecast and track furnace flame temperature selecting the most appropriate inputs that affect this process parameter. All data were collected from Erdemir Blast Furnace No. 2, located in Ere˘gli, Turkey, during 3 months of operation and the computational results are satisfactory in terms of the selected performance criteria: regression coefficient and root mean squared error. When the proposed model outputs are considered for the comparison, it is seen that the ANN models show better performance than the MLR and ARIMA models.

PDF

___

[1] Mroz J. Current Situation and Predictions further development of Blast Furnace Technology. Journal of Achievements in Materials and Manufacturing Engineering 2012; 55: 889-894.
[2] Hellberg P, Jonsson TLI, Jonsson PG, Sheng DY. A Model of Gas Injection into a Blast Furnace Tuyere. Fourth International Conference on CFD in the Oil and Gas Metallurgical & Process Industries: SINTEF/NTNU Norway 2005; 1; 1.
[3] Hooney PL, Boden A, Wang C, Grip C, Jansson B. Design and application of a spreadsheet based model of the blast furnace factory. ISIJ International 2010; 50: 924-930.
[4] Asl ZM, Salem A. Investigation of the flame temperature for some gaseous fuels using artificial neural network. International Journal of Energy and Environmental Engineering 2010; 1: 57-63.
[5] Ustaoglu B, Cigizoglu HK, Karaca M. Forecast of daily mean, maximum and minimum temperature time series by three artificial neural network methods. Meteorological Applications 2008; 15: 431-445.
[6] Kirk O. Encyclopedia of Chemical Technology. 13th ed. New York, NY, USA: John Wiley and Sons, 1981.
[7] Garcia FA, Campoy P, Mochon J, Ruiz-Bustinza I, Verdeja LF, Duarte RM. A new user-friendly blast furnace advisory control system using a neural network temperature profile classifier. ISIJ International 2010; 50: 730-737.
[8] Danforth GW. An Elementary Outline of Mechanical Processes. 2nd ed. USA: The United States Naval Institute, 1917.
[9] Ghosh SK, Pal S, Roy SK, Pal SK, Basu D. Modelling of flame temperature of solution combustion synthesis of nanocrystalline calcium hydroxyapatite material and its parametric optimization. Bull Mater Sci 2010; 33: 339-350.
[10] Bidabadi M, Rahbari A. Novel analytical model for predicting the combustion characteristics of premixed flame propagation in lycopodium dust particles. J Mech Sci Technol 2009; 23: 2417-2423.
[11] Radhakrishnan VR, Mohamed AR. Neural Networks for the Identification and Control of Blast Furnace Hot Metal Quality. J Process Contr 2010; 10: 509-524.
[12] Wang Y, Liu X. Prediction of Silicon Content in Hot Metal Based on SVM and Mutual Information for Feature Selection. Journal of Information & Computational Science 2011; 8: 4275-4283.
[13] Arzuman S. Comparison of Geostatistics and Artificial Neural Networks in Reservoir Property Estimation. PhD Thesis, Middle East Technical University, Ankara, Turkey, 2009.
[14] Daliri MR, Fatan M. Improving the Generalization of Neural Networks by Changing the Structure of Artificial Neuron. Malayas J Comput Sci 2011; 24: 195-204.
[15] Yaakob SN, Saad P. Generalization performance analysis between Fuzzy Artmap and Gaussian Artmap neural network. Malayas J Comput Sci 2007; 20: 13-22.
[16] Dalgakıran I, Danı¸sman K. Artificial neural network based chaotic generator for cryptology. Turk J Elec Eng & Comp Sci 2010; 18: 225-240.
[17] Perez-Cruz JH, Alanis AY, Rubio JJ, Pacheco J. System identification using multilayer differential neural networks: a new result. J Appl Math 2012; Article ID 529176; 20.
[18] Shi SM, Xu LD, Liu B. Improving the accuracy of nonlinear combined forecasting using neural networks. Expert Syst Appl 1999; 16: 49-54.
[19] Holder RL. Multiple Regression in Hydrology. 1st ed. Wallingford, UK: Institute of Hydrology Press, 1985.
[20] Alhadidi B, Al-Afeef A, Al-Hiary H. Symbolic regression of crop pest forecasting using genetic programming. Turk J Elec Eng & Comp Sci 2012; 20: 1332-1342.
[21] Box GEP, Jenkins G. Time Series Analysis. Forecasting and Control. San Francisco, CA, USA: Holden-Day, 1976.
[22] Al-Wadi S, Ismail MT. Selecting wavelet transforms model in forecasting financial time series data based on ARIMA model. Applied Mathematical Sciences 2011; 5: 315-326.
[23] Huwiler M, Kaufmann D. Combining Disaggregate Forecasts for Inflation: The SNBs ARIMA model, 7th edition: Swiss National Bank Economic Studies, 2013.
[24] Otsuka K, Matoba Y, Kajiwara Y, Kojima M, Yoshida M. A hybrid expert system combined with a mathematical model for blast furnace operation. ISIJ International 1990; 30: 118-127.
[25] Jimenez J, Mochon J, Ayala JS, Obeso F. Blast furnace hot metal temperature prediction through neural networksbased models. ISIJ International 2004; 44: 573-580.
[26] Moriasi DN, Arnold JG, Van Liew MW, Bingner RL, Harmel RD, Veith TL. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. T ASABE 2007; 50: 885-900.