MULTIPLE CRITERIA APPROACH TO MODEL CALIBRATION IN DEM SIMULATION OF COMPACTION PROCESSES

摘要

A 2-dimensional computational model based on the discrete element method (DEM) is utilized for modeling of stamp-charged coke making. In this process, a large volume of coal is compacted to one single coal cake before entering the coke oven. Cake densification and mechanical strength development are key factors for the subsequent production of a high-quality coke. An approach is introduced for model calibration utilizing multiple control parameters obtained from a laboratory-scale stamping device. Experimental data on stamper rebound amplitude, attenuation and peak force are compared with simulation output in a pairwise manner by constructing a criteria space, which allows identification of model settings best satisfying the control parameters. The calibration procedure carried out by multiple criteria comparison aided model parameter selection for reproducing stamper behavior in response to the properties of the cake subjected to compaction. This modeling and calibration procedure offers possibilities for numerically studying the densification and mechanical strength development of coal cakes. However, the approach is also more widely applicable to other compaction processes.