How To Increase Plant Performance With Artificial Intelligence And Expert Systems - Preprint 09-074

Hales, L. B.
Organization: Society for Mining, Metallurgy & Exploration
Pages: 3
Publication Date: Jan 1, 2009
Expert control of grinding and flotation plants has been successfully used in the minerals industry since the 1970?s. The earliest of these systems were written in a hard-coded fashion in FORTRAN, BASIC or Pascal. Second generation systems were built using the first experimental expert system shells that were being developed in the artificial intelligence community. Later systems were deployed in expert system shells designed for real-time processing plants that also included the ability to model the process with neural network models and optimize setpoint selection through the use of genetic algorithms. Significant performance increases have been achieved using these systems, but in general, they suffer from the static nature of their rules and to a degree the process models also. Typical improvements quoted in the literature suggest throughput rate increases of 4 to 8 percent and recovery increases of 1 to 4 percentage points. In over 30 years of designing and installing expert systems we have never seen performance numbers less that these but have often seen numbers in excess of the high values. However, once the system has been installed and tuned up the performance usually does not improve much thereafter. Granted, the initial increases in throughput rate and recovery are significant but a question always lingers, that is, can the expert strategies be modified further to squeeze a little more improvement out of the system. The answer to this question is a resounding yes, however a complete understanding of what we do well as designers and operators of expert control systems is required as well as how we can improve upon our best practices is required to move us from the performance plateau we have been on for a number of years to a new higher level of performance.
Full Article Download:
(251 kb)