Study On The Prediction Model For Stamping Coke Quality

Though our country is abundant in coal for the coking, the resources of those coals are unbalanced. Most part of our country is rich in high volatile coal resource but poor in fat coal and coking coal. In a background of unbalanced coal distribution and poor coking coal resource, how to instruct coking production accurately and scientifically become more and more important among coking producers. Coke predictive can solve this problem effectively The stamping coke making can enhance the quality of coke and at the same time, this technology not only decrease the amount of coking coal but also enlarge the scope of blend coal The predictive model is essential for coking plant.According to the situation of domestic coking plant's production, we respectively took representational blend coals of domestic coking plant and single coals of WISCO, which were blended in the laboratory. We chose 14 projects to carbonize by experimental oven, which included 7 industry blend coals and 7 blend coals mixed in the lab. (the bulk density was control led 0.7 t/m³¹.2 t/m³ ).After that, we analyzed the coal properties, which contain ash and sulfur content, mineral catalytic index, caking index and volatile content ,at the same time, we also analyzed the quality of stamping coke, which included ash and sulfur content, coke cold-state strength and coke thermal properties. After the analysis of the relationships between the bulk density and stamping coke quality, the paper chose the bulk density and coal properties as the predictive indexes for the foundation of stamping coke quality for the fist time.The paper established the predictive model for stamping coke ,which included ash and sulfur prediction, coke cold-state strength(abrasive strength and coke anti-broken strength) and coke thermal strength (reactivity and strength after reaction) prediction. The established models are down here as follows: A_d,j= 1.75911+ 1.16855A_d,m R²=0.82 S_td,j= 0.0967+ 0.81465S_td,m R²=0.83 M₁₀= 13.93258+ 4.05102d² -15.68347d+ 0.35044V_daf -0.05851G R²=0.90 M₂₅= 83.3746 -2.97067d² +8.11453d- 0.16047V_daf+ 0.02338G R²=0.72 CRI=100.72001+21.42092d²-51.15587d-0.5482V_daf-0.4657G+3.90987MCI R²=0.86 CSR=-46.6483-30.06633d²+79.44806d-0.11255V_daf+0.68577G-0.05048MCI R²=0.90The stamping coke predictive model were checked by the test data, and the results showed that the forecast bias were not huge, basically within the experimental error, meanwhile, positive and negative deviations were in line with normal distribution of random error. Because of these above, the method we chose is feasible, and the model established are reasonable. The models also reveal the relationship between the coal properties and stamping coke.