442612 Big Process Data Analysis Helps Improve Operational Efficiency in Ethylene Plants

Tuesday, April 12, 2016: 8:00 AM
335A (Hilton Americas - Houston)
Pratap Nair, Ingenero Inc., Mumbai, India

An ethylene facility in the Middle East was able to double its furnace run lengths from its design value, while simultaneously increasing yields and time between turnarounds, without impacting tube life and without any additional capital expenditure.  An ethylene facility in the USA has been achieving record production year after year.  A furnace tube leak was proactively detected very early, to prevent an unforseen shutdown at another ethylene facility in the USA. A facility in the Far East was able to increase feed to a furnace capacity constrained plant by 10% while maintaining the optimum cracking severity and obtaining the best yields of ethylene.

Ethylene plants today are by and large well instrumented and have robust advance control systems in place.  So much so that the volume of available data, the velocity at which this data is collected and the sheer variety of data available from the plant and the supply chain, makes it impossible for  unaided operating personnel to assimilate and get his/her mind around this Data overload.  As a result more than 80% of the data collected is just archived and never used.  Plants end up being run on experience based intuition, a minimal level of data or software driven analysis undertaken by limited resources with limited time and old Standard Operating Procedures prescribed by the process licensor.  This often results in anomalies going undetected, potential reliability problems and missing opportunities for improvement.

A unique model driven methodology for big data analysis that allows transformation of this Data available from a typical ethylene operation, into actionable information was the key to the results observed by the ethylene producers in the case examples being quoted here.  This methodology allows operating personnel to drive operations based on data-enhanced insights made readily available to them.  It enables operations personnel and managers with the ability to uncover truths and insights that aren’t readily obvious or don't allign with conventional wisdom or intuition.  The analysis from this methodology allows predictions and provides better prescriptive insights to better determine the best course of action, while continuously growing the experience “database” of managers and ethylene operating personnel at a faster pace over time.  The methology allows the improvements to be tracked, quantified and controlled, so that it can be sustained and improved continuously.

This paper covers ethylene producer cases where this methodolgy has been implemented and provides a description of the methodology.

Extended Abstract: File Uploaded
See more of this Session: Big Data Analytics and Smart Manufacturing I
See more of this Group/Topical: Topical A: 2nd Big Data Analytics