Jani Kaartinen

Optical Spectrum Based Estimation of Grades in Mineral Flotation

Mineral enrichment by notation is one of the hardest industrial processes to monitor and control. The main process measurements collected from a flotation circuit are the metal contents or grades in the different phases of the circuit. To improve the process monitoring capabilities, machine vision systems have previously been developed to characterize the flotation froth properties from digital images of the froth. This study discusses an on-line measurement of the optical froth spectrum in order to improve the slurry grade measurements obtained from the X-ray fluorescence based Courier analyzer. It is shown that the grades of the zinc flotation circuit concentrate can be estimated with a linear multiregression model using the measured froth spectra as input values. The model is then applied to estimate the grades of the concentrate between the originally sparse X-ray fluorescence measurements. Additionally, some preliminary results are presented which suggest that also the spectra measured directly from a slurry flow may correlate with the corresponding grade values.

Utilizing 3D Height Measurement in Particle Size Analysis

This paper introduces a novel approach for performing non-invasive particle size analysis for a material stream running on a standard conveyor belt. The measurements, carried out with a 3D laser scanner and a measuring belt weigher, are accurate, robust and real world physical measures. The 3D data obtained with the laser scanner enables more accurate analysis than the spatial monochrome or colour images that are commonly used in this field. In this paper the proposed analysis method is used in mineral processing application to get information about particle size distribution of the ore flow from the mine to a screening station at the surface. This information can be used to optimize operation of a semi autogenous grinding station used in Pyhasalmi. However, with certain limitations, the analysis method can be utilized in different kinds of applications.

Image analysis based control of zinc flotation - a multi-camera approach

In this paper a new multi-camera set-up is introduced to improve the controllability of a zinc flotation process. The multi-camera system can analyze the froth appearance simultaneously from six flotation cells. The froth appearance is characterized with twelve different parameters that are calculated separately for each cell, and these parameters are used to control the flotation process in real-time. The development of this system was motivated by a single-camera system that has been shown in practice to improve considerably both the grade and recovery of the zinc circuit.

Using a Virtual Flotation Process to Track a Real Flotation Circuit

Abstract This paper presents a novel way of using a dynamic flotation process simulation environment for concurrent simulation alongside a real flotation process. The simulation environment is actively adapted to a real flotation process. This yields new possibilities on the usability of the simulation model. For example, virtual measurements can be introduced that provide completely new variables or increased fault tolerance through redundancy. Another option is to provide feedback for the process operators in the form of textual reports. These reports contain e.g. shift-specific performance indices that can be compared against reference data. Furthermore, by speeding up the simulation, it is possible to make prognosis on the effects of different control actions before actually making them.

Parameter Estimation for a Flotation Process Tracking Simulator

Abstract A parameter estimation method for an online flotation process simulator is described. The applications of an online process simulator include soft sensor implementations, process trajectory predictions and advisory feedback to the operator, which have potential to improve the process efficiency and minimize the detrimental effect of disturbances on the process or the environment. The online simulator uses a detailed and dynamic model of an actual industrial flotation process, and therefore accurately corresponds to the process phenomena present at the plant. Parameter estimation is required for the flotation kinetics in order for the simulator to adapt to changes in the process conditions. A relatively generic parameter estimation algorithm is developed and tested with a dual simulator setup. The particular requirements and limitations of adapting an online simulator are discussed, and modifications to well-known estimation algorithms are presented as a possible method of meeting the requirements and overcoming the limitations. The results show that online parameter estimation and simulation model tracking is possible with the chosen method, and point out areas of further development for application when the simulator is used alongside the real process.

Machine Vision of Flotation Froths with a Rapid-Prototyping Platform

Abstract This paper presents generic, modular and scalable machine vision architecture for single or multi camera applications. It can be used for rapid-prototyping but, as shown in the text, it has also been successfully applied for on-line plant installation for several years without any problems. The key idea is to separate the algorithm development and testing from the rest of the infrastructure that is needed when on-line applications are considered. Automatic code generation and modular systems design methods are used when filling the gap between these two worlds. As an end result, a highly effective but generic machine vision rapid development environment has been implemented that can be applied both to lab-scale and to industrial machine vision applications.