Hybrid: having or produced by a combination of two or more distinct elements 
 - Merriam Webster Dictionary

 

AspenTech’s process modeling software has been used to design at least three-quarters of the world’s upstream, midstream, downstream and chemical assets. AspenTech has introduced a long list of firsts to the world of chemical engineering and process manufacturing over the past 39 years.

 

A Historical Look

I first learned how to use Aspen Plus® in a course, alongside professionals from Dow, DuPont, Exxon and others, from Dr. Larry Evans (the founder of AspenTech) in 1987. It involved learning the details of the chemical equations inside the software, and what each change in a flowsheet might do to those equations. It involved learning a lot of details about the mathematics of solving many chemical equations at once. And, of course, it involved working out what you wanted to model on a piece of paper, writing down a series of computer commands, then typing them onto punch cards. You marched into the computer room with a deck of punch cards, handed them over to an operator, watched them fed into a reader, then waited while the computer crunched away, slowly solving your simulation.

Since that time, process simulation has changed the world of chemicals and oil. It has made possible the incredible innovations and complexity involved in making the high-performance materials used in products from jet planes to cellphones, from sporting gear to wind generators. 

And it spawned a craft: the craft of the process modeling expert, engineers with the experience and knowledge to expertly turn processes into accurate predictive models.

 

The Disappearing Expert

As a generation of these experts retires, process organizations face a gap in essential knowledge and a new generation of workers without the time to develop that critical expertise. Hybrid models, embedded with AI, address those gaps, creating immediate value for organizations and assets. All but those enterprises with the deepest pockets need the ability to build and deploy these models without scarce and expensive experts.

 

A New Day

 

Now, in 2020, AspenTech is ready to show the world a new generation of innovative process modeling, which again will change the course of the chemical and hydrocarbon industry.  

 

This new technology is called Aspen Hybrid Models. It takes advantage of the rapid development of analytics, machine learning and AI. Hybrid models combine AI and first principles modeling to deliver a comprehensive, accurate model more quickly, without requiring significant levels of expertise which are becoming more scarce in industry. 

 

Machine learning is used to create the model leveraging simulation or plant data, while using domain knowledge including first principles and engineering constraints to build an enriched model without requiring the user to have modeling expertise or become an AI expert. AspenTech is uniquely positioned to leverage over 40 years domain expertise to make AI applicable to process industries, delivering Industrial AI. This next generation of solutions will democratize AI within hybrid models to optimally design, operate and maintain assets.

 

Proven Benefits

 

With hybrid models, users can model processes and assets that cannot easily be modeled with first principles alone. You get the accuracy of an empirical models, the strength of first principles models, leveraging the power of AI, along with 40 years of domain expertise, to create a more predictive model. 

 

Hybrid models help organizations create and sustain better models, faster. Hybrid models provide a better representation of the plant, which keeps the model more relevant over a longer period of time. 

 

 

 

 

AspenTech has been working with over 60 companies to develop and test Aspen Hybrid Models. Here are a few examples that have emerged from this collaboration with customers:  

 

• Refining and olefins planning model updates: Embedding accurate, fast running reduced-order, nonlinear hybrid models of key economic units with planning LP models. Conservatively, these will create over $10 million USD annually for a typical 200,000 BOED refinery.
• Equipment monitoring: Unit and equipment level models, using AI analytics combined with first principles simulation, for easy to develop, update and run digital twin models.  Just the fouling monitoring use case can deliver millions of USD annually for a single heat exchanger train. Reactor fouling and catalyst monitoring can create value of $5-10 million USD annually per catalyst reactor unit.
• Specialty and high-performance polymer lines: Polymer production hybrid models predicting operating performance will create an estimated value of over $1 million USD per line, due to changeover efficiencies as well as reactor uptime.

Other applications include separation membranes, oils-to-chemicals asset models, and a wide range of others.

 

Read more in our white paper, "Hybrid Modeling: AI and Domain Expertise Combine to Optimize Assets."