The Application Of Fractal In Ball Mill Process For Blast Furnace Slag

The particle size distribution and particle image outline are all satisfied with self-similarity in Statistics and can be characterized with fractal dimension for crushing particle materials.The phenomenon of particle agglomeration has attracted researchers' attention for a long time.In this paper,blast furnace slag produced by Panzhihua is chosen to study the relationship between processing parameters of planetary ball milling and volume fractal dimension for the blast furnace slag,and the effect of aggregation on the volume fractal dimension with laser particle size analyzer measuring the particle size distribution and Scanning electron microscope shooting pictures.The main work is as following:1.The application of fractal theory in ore crushing,particle size characterization and crushing process is summarized.The milling time,milling speed and the ratio of the ball to feed are selected to study the effect on particle size distribution for milled blast furnace slag.GBL-101 BI laser particle size analyzer was used to measure the original and milled blast furnace slag samples under 53 groups of processing parameters and obtained 211 particle size testing reports;Scanning electron microscope was used to shoot pictures of the original blast furnace slag and 5 groups of milled blast furnace slag.372 single particle images were processed with the ImageJ software.2.The Gauss Newton method was developed with MATLAB software to fit the Rosin-Rammler distribution function.Six kinds of materials,in which three are measured with the laser particle size analyzer,including magnesite,the other three are from reference,are chosen to verify the Gauss Newton method is superior to the other three fitting methods mentioned in the references.3.Six groups of blast furnace slag from the orthogonal experiment are chosen to verify the rationality of the relationships among three fractal dimensions proposed by research group.The volume fractal dimension is calculated with the particle size distribution function method on the basis of results by the Gauss Newton method;The surface fractal dimension is calculated with the power spectrum method by selecting appropriate cut-off frequency;And the line fractal dimension is calculated with the yardstick method,by selecting proper step length.4.Based on the three factor three level orthogonal analysis results,the regression equation between the volume fractal dimension and the milling parameters was established with regression orthogonal combination design.The regression equation was used to predict volume fractal dimension and the maximum relative error is 2.76%.5.Based on the fractal dimension results for 6 groups of blast furnace slag,the effect of agglomeration on the volume,surface and line fractal dimensions were analyzed,and the results were compared with SEM images.The effects of agglomeration on the results of the surface and line fractal dimensions were analyzed for single particle with amplification factor of 2.0k to 10.0k obtaining single particle SEM image.The results showed that the volume fractal dimension is 2.40,the surface fractal dimension is 2.30 when the blast furnace slag is more serious agglomerated.Further study on6.BP neural network was established with the ball milling parameters as input,and the fractal dimension as the output.180 fractal dimensions and the corresponding parameters were taken as training data and the other 31 were taken as forecast data.The forecast results were compared with the measured results and the results showed that the relative errors were less than 2%.