Abstract：The accurate predictability of converter steelmaking oxygen consumption is important for the optimization scheduling energy of iron and steel enterprises, therefore, it is put forward prediction model for converter oxygen consumption, and it is the BP neural network which based on grey system and genetic algorithm to optimiztion the combination.First of all, according to the extract of converter smelting history data, it uses grey correlation method to ensure the main factor sequence of converter steelmaking oxygen consumption.Secondly, the method that is used to predict the selected sample data sequence is grey correlation method and GABP neural network.Finally, it is concluded that the optimal weight coefficient of combination model.Based on the minimum of combination forecasting error square sum.The simulations reveal that this method is effective on the reducing the prediction error, increasing prediction accuracy, and enhancing the generalization ability.Therefore, compared with the other models, the combined forecast model is more suitable for prediction of converter steelmak oxygen consumption.

Keyword：oxygen converter;grey correlation;analysis of gray system;GA-BP neural network;combination forecast model;