Characterization of Flotation Froth Structure and Colour by Machine Vision - ChaCo The primary source of the disturbances in flotation is the changing ore quality which, except the metal grades, cannot be measured at the plant. The metal contents of different pulp streams are measured by X-ray fluorescence analysers typically in 5…10-minutes sampling intervals. However, measurement of certain variables related to the type and movement of the flotation froth is based on the visual ob-servation of the human operator.
Therefore a new vision system for monitoring and control of flotation processes has been developed in an ESPRIT Project: The Characterization of Flotation Froth Structure and Colour by Machine Vision (ChaCo). The project was coordinated by Helsinki University of Technology, Systems Engineering Laboratory and the partners were: Kungliga Tekniska Högskolan, Dept. of Civil and Environmental Engineering, Università di Roma "La Sapienza" Dipartimento di Ingegneria Chimica (MMPM), University of Oulu, Dept. of Process Engineering, Boliden Mineral AB, Mineral Processing Dept., Outokumpu Mining Oy, Pyhäsalmi Mine, and Technical Research Centre of Finland, Information Technology Institute. A froth analyser set-up consisting of a camera, a grabber card, a spectrophotometer and a PC-computer has been installed at the Pyhäsalmi flotation plant. The user interface is based on LabVIEW© environment. The system includes on-line analysis software producing measurements charac-terizing, for example, the structural properties, velocity, stability and colour of the froth to the process automation system PROSCON©.
For the classification of the froth there are several methods. The estimator contains also models for estimating concen-trations on the basis image variables. The parameters of the models can be tuned by statistical multivariate methods.
The principal goal of the control is to keep the system on a suitable operation area and the final optimisation is based on learning during the normal operation. The analyser gives a number of output variables which can be used for control. On the basis of operator’s knowledge and experimental work some of them have been chosen for controlled vari-ables. The controller design is partly based on a priori knowledge of the operators and partly on knowledge ob-tained during the project. The control has been realized by PROSCON.
Contact information: Heikki Koivo
.
|
