Online Analysis of Coal on A Conveyor Belt by use of Machine Vision and Kernel Methods

• Published in: International journal of coal preparation and utilization 2010-11, Vol.30 (6), p.331-348
• Main Authors: Aldrich, C., Jemwa, G. T., van Dyk, J. C., Keyser, M. J., van Heerden, J. H. P.
• Format: Article
• Language: English
• Subjects: Algorithms; Analysis; Applied sciences; Belt conveyors; Coal; Coal and derived products; Coal gasification; Conveyors; Energy; Estimates; Exact sciences and technology; Fuel processing. Carbochemistry and petrochemistry; Fuels; Gasification; Illumination; Lighting; Mechanical preparation; Methods; On-line systems; Online; Online measurement; Particle size distribution; Processing; Reliability; Solid fuel processing (coal, coke, brown coal, peat, wood, etc.)
• ISSN: 1939-2699; 1939-2702
• DOI: 10.1080/19392699.2010.517486

The application of machine vision systems to measure particle size distributions has among other things been driven by sophisticated control systems used to monitor and control mills and other ore-processing systems. Machine vision is nonintrusive and offers reliable online measurements in potentially harsh environments. Although considerable advances have been made over the last decade, reliability of measurements with segmentation algorithms is still an issue, particularly where lighting conditions may vary, fines are present, or heterogeneous particle surfaces may result in irregular reflection of light. In practice the alternative to online measurement of particle size distributions is sieve analysis, which is slow and tedious and not suitable for control purposes. The efficient preparation and quality control of coal are important for stable and effective operation of the Sasol® FBDB™ Gasification Process. The operation of these gasifiers depend among other on melting properties and composition of the ash, thermal and mechanical fragmentation, and caking properties of the coal, as well as the particle size distribution of the coal. Although many of these properties can be assessed in some way to expedite process improvement, particle size distributions are difficult to estimate beforehand from feedstocks, since these distributions may change significantly during the feeding process, or by insufficient screening, resulting in an access/increase of fine coal to gasification. The ability to measure these distributions online would therefore play a crucial role in continuous process improvement and real-time quality control. The objective of this project is to explore the use of image analysis to quantify the amount of fines (