作者: CHRIS ALDRICH , DERICK W. MOOLMAN , JACQUES J. EKSTEEN , JANNIE S. J. VAN DEVENTER
DOI: 10.1080/00986449508936396
关键词: Process (engineering) 、 Engineering 、 Characterization (materials science) 、 Pattern recognition 、 Artificial intelligence 、 Structure (mathematical logic) 、 Cluster analysis 、 Operations research 、 Artificial neural network 、 Identification (information) 、 Classifier (linguistics)
摘要: Flotation processes are difficult to describe fundamentally, owing the stochastic nature of froth structures and ill-defined chemorheology froth. Considerable information on process is reflected by structure In previous work it has been shown that structural features extracted from flotation froths can be related behavior in a qualitative way through identification certain behavioral regimes or classes using supervised neural net as classifier. Although useful an aid control decisions, this method less suitable for quantitative dynamic analysis plants. paper new plants consequently proposed, based use order preserving maps digitized images phase. The construction these means self-organizing demonstrated examples concerning industri...