DR Peng Lam is taking the guesswork out of mining by building smarter software for processing plants.

Working with colleagues at Murdoch’s School of Engineering, Curtin University and the AJ Parker CRC, Dr Lam is finding ways to automatically extract information about specific processes in the plant, through the use of mathematical models and the plant data knowledge base.

Studies have shown that process industries typically run at approximately 70 per cent of their maximum production rate, whereas continuous plants can run at up to 92 per cent.

The lower performance of process plants indicates that they are not running to their top capacity due to frequent shut downs or applications running at sub-optimal settings for lengthy periods.

Despite the sophisticated machinery used in the mining industry, operators still rely on gut feelings stemming from past experience to fix problems in the processing plant.

But as Dr Lam explained, this means that operations must rely heavily on the knowledge of experts, making the job harder for new employees and very time consuming.

“We are finding ways to capture knowledge of the experts within the plants,” said Dr Lam.

“Improving software systems will not replace operators, instead they will make their jobs easier.”

Dr Lam and her colleagues are developing interactive intelligent software systems that will automatically detect poorly performing applications and abnormal process situations by checking whether performance targets are being met.

Using these new software systems, operators will be provided with a list that ranks the importance of error messages on equipment, thus simplifying the process of tracing problems back to the original source.

By simulating errors, the interactive intelligent software systems support the industrial training of new operators through providing them with valuable experience and thus helping them avoid making costly mistakes.

The system will also cut costs by reducing the need for experts to be flown out to remote mining sites to fix problems.

Although the research is based on capturing plant specific data in gold leaching operations, Dr Lam hopes that that the system will be generic enough to work in a range of production plants with a few minor modifications.