At MOL Refineries we recognized the Power of Data - the Power of Big Data. Our Refinery Information System “products” generate ~80 billion data points/year. But how could we utilize these data points better?
With the Refinery team we first identified the most problematic areas and issues. We then set out to resolve these issues through advanced analytics:
- Coke yield
- Delay Coker Unit Steam Eruption
- White product yield
- Advanced Process Control Soft Sensor validation
- Statistical Process Control fine tuning
- Utilities (electricity , Natural Gas, Nitrogen, Oxygen ) forecasting
- Catalyst life cycle
- Integrity Operating Window parameters
- Alternative Crudes
Advanced analytics and visualization were performed with PI AF calculations, PI Coresight reports and deep analyses with Azure Machine Learning.
This presentation will include examples of the issues identified and resolved, and the results obtained. These results contributed significantly to MOL's Next Downstream Program's targets. An additional $500M in realized benefits is expected by the end of 2016 compared to 2014.
Tibor Komróczki has been working at MOL since 1987. He started his career at the DFCCU unit and has moved through different roles over time. In 2001 he joined the Process Information team as an application developer. Since 2009 he has lead the Process Information team and in 2013 he was appointed the leader of the Integrated Information and Automation team in MOL's Technology and Development department.Tibor holds an MSc in computer engineering and is a specialist in Microsoft Business Application Development. He and his team have created many applications, worked on many projects and provided services involving the PI Server's Asset Framework.