作者: Xi Lin , Zhigang Zhang , Derong Liu
DOI: 10.1109/IJCNN.2007.4370980
关键词:
摘要: In cement manufacturing, pre-calcining process (PCP) contains an added precalcinator between the preheater and kiln. Since raw meal decomposition rate in kiln is over 90%, optimal control necessary for keeping temperature stable to ensure normal decomposition. Due diversity of composition consistent thermal transformation, a complex nonlinear problem. Many factors may influence it, there are couplings various uncertainties. It difficult solve dynamic stability problem precalcinator. This paper addresses using dual heuristic programming (DHP). The approach mainly composed three parts: model network, action network critic network. All parts use neural networks (NN) simulate trend. self-learning adaptive abilities were analyzed MATLAB environment simulation results demonstrate advantage adaptation optimization solutions compared conventional control.