A review of froth flotation control

Machine learning applications in minerals processing: A review

Flotation froth image recognition with convolutional neural networks

https://www.cec.uchile.cl/~clperez/papers/Ore%20grade%20estimation%20by%20feature%20selection%20and%20voting%20using%20boundary%20detection%20in%20digital%20image%20analysis.pdf

Ore grade estimation by feature selection and voting using boundary detection in digital image analysis

Prediction of the metallurgical performances of a batch flotation system by image analysis and neural networks

Online monitoring and control of froth flotation systems with machine vision: A review

Automatic control of industrial flotation cells and circuits presents a set of significant challenges due to the number of variables, the sensitivity of flotation cells to variation in these variables and the complexity of predicting flotation performance and/or developing a strategy for optimisation. Air recovery, a measure of froth stability, has been shown to pass through a peak as flotation cell aeration increases. Furthermore, the air rate at which the peak air recovery (PAR) is obtained results in optimal flotation performance, whether improved concentrate grade, recovery or both grade and recovery. Peak air recovery, therefore, presents a clear optimising control strategy for the operation of flotation cells which is generic to all flotation cells regardless of position in the flotation circuit. In this study, a novel control system based on PAR is developed and demonstrated using a large continuous laboratory flotation cell. In this study, a direct search optimisation algorithm based on the GSS (generating set search) methodology was developed using a 70 l continuous flotation cell operating with a two-phase system (surfactant solution and air only). Characterisation of the laboratory system showed that it was stable for up to 6 h and exhibited a reproducible peak in air recovery. A dynamic model of the response of the system with regards to changes in air recovery was developed that allowed simulations of the proposed optimising control system to be carried out. The optimisation algorithm was then applied to the experimental system. The trialled GSS algorithm was shown to find the PAR air rate when starting above, below and at the PAR air rate, and additionally with a disturbance introduced into the system. While the direct search approach can be slow, it is simple and robust. This demonstrates an innovative approach to optimising control for froth flotation and is the first application of froth stability maximisation for flotation control.

/pdf/a-flotation-control-system-to-optimise-performance-using-485cw80g4o.pdf

B. Shean

Papers

Online monitoring and control of froth flotation systems with machine vision: A review

A review of froth flotation control

A flotation control system to optimise performance using peak air recovery

A dynamic model for level prediction in aerated tanks