Establishment and optimization of heating furnace billet temperature model

In the metallurgical industry, measuring the temperature distribution directly and accurately in billet heating process is a well-known difficult work. To improve the quality of heating billet, a billet temperature prediction model of heating furnace is necessary. Based on the characteristics of furnace section partition control, this paper firstly established a billet temperature prediction model with three serial neural networks as foundation, then optimized this model with the improved dynamically self-adaptive PSO. The simulation indicated that the establishment of this model is easy, the forecast precision and speed are obviously improved, and the match degree of prediction curve and actual curve is highly increased. All of these proved the effectiveness of this model.