Koszo, J.J., Smith, D.J.

Across the large-scale metals processing industry, plants are frequently faced with production issues capable of impacting profitability by the millions, forcing decision makers to act rapidly, often without all the necessary information. Especially true for alumina refineries, complexities within the process make it difficult to foresee the full implications of decisions without heavy reliance on historical experience and rules of thumb. As a result, common practice has teams comprised of highly experienced personnel from operations and engineering backgrounds work together to decide rapidly on how to handle these frequent issues. While this approach is very effective when economics and processes are consistent, this is rarely the case as even fundamental variables such as raw material quality, equipment performance and economics are constantly shifting.

At an alumina refinery, the question was posed whether a third component was missing from this team of operational and engineering personnel in the form of a program, one that the team could use which combines process engineering principles, live process data and programming to instantly calculate the profitability of decisions in real-time. This would synthesise the complex problems and provide the team with critical information about the economic viability of various options, allowing decision makers to quantitatively compare options leading to better decisions.

Taking this approach for two common production hurdles at the refinery, cleaning and maintenance cycles for classification area and production rate management resulted in total savings of at least AUD$6 million in the first year. Development and roll-out of each project was three and four months respectively, demonstrating that this new avenue for optimisation has both the high value opportunity and the low execution time required to make it a highly profitable opportunity at this refinery and likely many others.