Queensland Center for Advanced Technology

Pullenvale, Australia

Queensland Center for Advanced Technology

Pullenvale, Australia
Time filter
Source Type

Howard D.,Queensland Center for Advanced Technology | Bull L.,University of the West of England | Lanzi P.-L.,Polytechnic of Milan
Neural Processing Letters | Year: 2015

Learning classifier systems (LCS) are population-based reinforcement learners that were originally designed to model various cognitive phenomena. This paper presents an explicitly cognitive LCS by using spiking neural networks as classifiers, providing each classifier with a measure of temporal dynamism. We employ a constructivist model of growth of both neurons and synaptic connections, which permits a genetic algorithm to automatically evolve sufficiently-complex neural structures. The spiking classifiers are coupled with a temporally-sensitive reinforcement learning algorithm, which allows the system to perform temporal state decomposition by appropriately rewarding “macro-actions”, created by chaining together multiple atomic actions. The combination of temporal reinforcement learning and neural information processing is shown to outperform benchmark neural classifier systems, and successfully solve a robotic navigation task. © 2015 Springer Science+Business Media New York

Meng J.,Julius Kruttschnitt Mineral Research Center | Xie W.,Julius Kruttschnitt Mineral Research Center | Brennan M.,Julius Kruttschnitt Mineral Research Center | Runge K.,Julius Kruttschnitt Mineral Research Center | And 2 more authors.
Minerals Engineering | Year: 2014

Turbulence and its distribution are of great importance in flotation processing and has been the subject of much research. However, there is no mature technique to measure turbulence in three phase (liquid-solid-gas) systems. In this research, the Piezoelectric Vibration Sensor (PVS) was developed, based on previous research, as a promising tool for turbulence measurements in industrial flotation environments. A frequency response model was established to calculate force applied to the sensor. Experimental results and comparison with Laser Doppler Anemometry (LDA) measurement data showed that the PVS can measure intensity of kinetic energy fluctuation (σv2), which has been found in experiments to correlate with turbulent kinetic energy (TKE), a parameter often related to flotation performance in the literature. The sensor was then applied to a 60 l laboratory batch cell running at different impeller speeds and air flow rates to obtain turbulence profiles. Results showed that the piezoelectric sensor is fully capable of measuring turbulence in a multi-phase environment. © 2014 Elsevier Ltd. All rights reserved.

Meng J.,Julius Kruttschnitt Mineral Research Center | Xie W.,Julius Kruttschnitt Mineral Research Center | Runge K.,Julius Kruttschnitt Mineral Research Center | Runge K.,Queensland Center for Advanced Technology | Bradshaw D.,Julius Kruttschnitt Mineral Research Center
Measurement Science and Technology | Year: 2015

Measuring turbulence in an industrial flotation environment has long been problematic due to the opaque, aggressive, and abrasive three-phase environment in a flotation cell. One of the promising measurement techniques is electrical resistance tomography (ERT). By measuring the conductivity distribution across a measurement area, ERT has been adopted by many researchers to monitor and investigate many processes involving multiphase flows. In the research outlined in this paper, a compact ERT probe was built and then used to measure the conductivity distribution within a 60 l flotation cell operated with water and air. Two approaches were then developed to process the ERT data and estimate turbulence-related parameters. One is a conductivity variance method and the other is based on the Green-Kubo relations. Both rely on and use the fluctuation in the ERT measurement caused by bubbles moving through the measurement area changing the density of the fluid. The results from both approaches were validated by comparing the results produced by the ERT probe in a 60l flotation cell operated at different air rates and impeller speeds to that measured using an alternative turbulence measurement device. The second approach is considered superior to the first as the first requires the development of auxiliary information which would not usually be known for a new system. © 2015 IOP Publishing Ltd.

Meng J.,Julius Kruttschnitt Mineral Research Center | Tabosa E.,Queensland Center for Advanced Technology | Xie W.,Julius Kruttschnitt Mineral Research Center | Runge K.,Julius Kruttschnitt Mineral Research Center | And 3 more authors.
Minerals Engineering | Year: 2016

Flotation is one of the most important primary separation processes in the minerals industry. As far as the mechanism of flotation is concerned, turbulence is one of the key parameters determining flotation performance because it affects three main processes: suspension of particles, air dispersion and particle-bubble collision, attachment and detachment. To study turbulence in industrial flotation cells, both numerical simulation and experimental measurement can be performed. Development of turbulence models and validation of Computational Fluid Dynamic (CFD) numerical simulation need experimental data obtained from turbulence measurement techniques that can be used in the three phase abrasive opaque environment present in a flotation cell. In this paper, the different techniques which have been used to characterise turbulence in the literature are reviewed in terms of their basic principles, system structure, range of application and limitations. Laser Doppler Anemometry (LDA), Particle Image Velocimetry (PIV), Constant Temperature Anemometer (CTA) and the Aeroprobe are all techniques that have been widely used to characterise the turbulence created in flotation machines operating with only fluid (or fluid and air). They cannot however be used when the concentration of solids is high as commonly occurs in a flotation machine. Techniques that have been identified that have the potential to be used to produce accurate measurements in three phase flows include Positron Emission Particle Tracking (PEPT), Piezoelectric Vibration Sensor (PVS) and Electrical Resistance Tomography (ERT). It is envisaged that applications of PEPT in three phase flotation cells will mostly be confined to studies at the laboratory scale. ERT has been tested in flotation cells filled with water and air but needs more development before it can be applied confidently in industrial scale flotation units. PVS, on the other hand, has been validated at laboratory scale and has been applied successfully for measuring turbulence in large scale operating flotation machines. © 2016 Elsevier Ltd. All rights reserved.

Loading Queensland Center for Advanced Technology collaborators
Loading Queensland Center for Advanced Technology collaborators