ABSTRACT

Screening of coal is one of the processes carried out to produce clean coal suitable for the blast furnace. In this work, the screening of moist coal was carried out for different angles of the screen and frequencies. A 2 mm screen perforation was used to separate undersize coal of size +1 mm-2 mm from the +1 mm-3 mm coal samples. For each experimental condition, the screening efficiency was calculated. Maximum screening efficiency of 85.96%, 75.64%, and 63.46% was obtained at 4%, 6%, and 8% moisture content, respectively. As the moisture content of coal increases, the efficiency minimizes due to high screen clogging. After determining the screening efficiency, prediction was carried out using regression modeling. In this work, linear and second-order polynomial regression modeling was utilized to develop a prediction model for the experimental values. From the results, it was clear that the polynomial regression model has high regression coefficient (R²) percentage and low P-value in comparison with the linear regression model. After prediction, validation was carried out on the best fit model. The value of Variance Account For (VAF), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE) was in the acceptable range, which shows that the developed model was most effective.

摘要

煤的筛选是生产适用于高炉的洁净煤的过程之一。在这项工作中，针对不同的筛网角度和频率进行了湿煤的筛分。使用 2 mm 筛孔从 +1 mm-3 mm 煤样品中分离尺寸为 +1 mm-2 mm 的筛下煤。对于每个实验条件，计算筛选效率。在 4%、6% 和 8% 的水分含量下，分别获得了 85.96%、75.64% 和 63.46% 的最大筛选效率。随着煤的水分含量增加，由于筛网堵塞率高，效率会降低。在确定筛选效率后，使用回归模型进行预测。在这项工作中，利用线性和二阶多项式回归建模来开发实验值的预测模型。从结果可以看出，多项式回归模型具有较高的回归系数（R²）与线性回归模型相比的百分比和低P值。预测后，对最佳拟合模型进行验证。方差帐户 (VAF)、均方根误差 (RMSE) 和平均绝对百分比误差 (MAPE) 的值在可接受的范围内，这表明开发的模型是最有效的。