J. Biedermann 1, L. Cromme1, L. Rhede2, K. Weber1 and S. Würll1
1Technische Universität Cottbus, Lehrstuhl Numerische und Angewandte Mathematik
2Lausitzer Braunkohle AG
To improve the output rate of a complex packing process for brown coal briquette bundles we designed a simulation system combining statistics, visualization and optimization routines using state-of-the-art object-oriented software design techniques. With the optimization part we found optimal settings of the numerous production parameters like weight limits, briquette and bundle transport velocities, sensor positions, and delays etc. for different production conditions. Comparing conventional techniques and AI methods for this kind of optimization problems we can demonstrate that conventional methods will either fail in practice or cannot be realized at reasonable cost. Due to the optimization the performance and output of the plant could be increased significantly. The whole system serves as a basis for an on-line process control. Furthermore the system allows to simulate and investigate changes and extensions to the plant at minimal costs and without disturbing the production process.