When conducting quality giveaway optimization projects for our oil and gas customers, we noticed that on average, refineries are giving away twice as much as they practically can. Another observation from these projects was that the value of performance optimization is usually underestimated initially. In contrast, physical (or structural) constraints are commonly perceived as the prime contributors to quality giveaway and capital projects are usually proposed as the solutions for this loss of opportunity.

Based on our work on reducing quality giveaway with more than 20 clients over the last several years, we want to summarize several “true” constraints as the theoretical giveaway target that refineries should consider. In other words, all refineries have the potential to improve their performance without investing heavily in CAPEX.

What are the initial constraints?

The first constraint is the measurement constraint. This is the most common constraint across all refineries due to the limitation because of the current state of the technology in terms of the measurement uncertainty introduced by the physical method. The second constraint is the logistic constraint – whether a refinery can flexibly adjust its octane and volatility pool through procurement and sales is critical in determining what should be the real giveaway. Last but not least, the degree of freedom in component production helps refinery to avoid getting into excess ON pool, limitation in high RVP materials, etc.

During our engagements, we assisted our customers in eliminating unnecessary waste by setting the correct giveaway target through careful evaluation of these constraints. For example, at one of our recent engagements on the US west coast the refinery was able to reduce its gasoline blending targets and recover more than 70% of the quality giveaway.

Use of Data in Achieving Gasoline Blending Quality Targets

Another inefficiency we observed during these projects is the under-utilization of data. With the advancement of technology, a tremendous amount of data is now available for the refinery personnel to make better decisions and respond quicker to operational events. However, most of this data was collected primarily for historian purposes but used for analytics at a very basic level. With the vast amount of data available, a lot of new data concepts can be deployed to improve the operation efficiency of the refinery. For example, in one of the projects we conducted with a mid-west refiner in the United States, the use of advanced modeling concepts and real-time data analysis allowed us to reduce the giveaway to the theoretical limit, saving millions of dollars.

This blog is authored by Anas Dabbakh and Kai Wan. Anas is a Senior Engagement Manager and Kai is an Engagement Manager at Trindent Consulting.