End-point Prediction of BOF Steelmaking Based on KNNWTSVR and LWOA
- 24 Downloads
Basic oxygen furnace (BOF) steelmaking plays an important role in steelmaking process. Therefore, research on BOF steelmaking modeling is very necessary. In this paper, a novel combination prediction model has been proposed, which consists of a time series prediction model and a compensation prediction model. Both models are established by k-nearest neighbor-based weighted twin support vector regression (KNNWTSVR) algorithm. By introducing Lévy flight algorithm and inertia weight, an improved algorithm of whale optimization algorithm (WOA) called Lévy flight WOA has been initially proposed to solve the optimization problem in the objective function of KNNWTSVR. The simulation results show that the proposed models are effective and feasible. Within different error bounds (0.005% for carbon content model and 10 °C for temperature model), the strike rates of carbon content and temperature both achieve 93%, and a double strike rate of 86% is obtained, which can provide a significant reference for real BOF applications, and the proposed method is also appropriate for the prediction models of other metallurgical applications.
KeywordsBasic oxygen furnace Twin support vector regression Whale optimization algorithm Prediction model
This work was supported by the National Natural Science Foundation of China under Grants 61403177.
- 13.Rastogi R, Anand P, and Chandra S, Appl Intell 46 (2016) 1.Google Scholar
- 14.Xu Y, Li X, and Pan X, Neural Comp Appl 2 (2017) 1.Google Scholar
- 23.Kennedy J, and Eberhart R, IEEE International Conference on Neural Networks (1995) p 1942.Google Scholar
- 24.Marco D, Montes O M A, and Sabrina O, Comp Intell Mag IEEE 1 (2003) 28.Google Scholar
- 25.Pinto P C, Runkler T A, and Sousa J M C, International Conference on Adaptive and Natural Computing Algorithms (2007) pp 350–357.Google Scholar
- 32.Yang X S, and Deb S, Int J Math Modelling Numer Opt 1 (2010) 330.Google Scholar
- 33.Shi Y, and Eberhart R, Proc. of IEEE ICEC conference (1998) p 69.Google Scholar