Rob Faessen's research sits at the intersection of physics, mathematics, and computer science. He extended mathematical theory to check physical equations that are used in designing chemical plants for all kinds of errors. To make this practical, he created a special language to express equations in a machine readable way and developed software to automatically verify them. Along the way, his work also highlighted shortcomings in the International System of Units (SI).
Computer models are essential for designing and running chemical plants. They calculate important properties such as pressures, temperatures, and flow rates, guiding decisions about equipment and plant operation. Even small mistakes in these equations can have serious consequences, including incorrectly sized equipment, costly delays, or inefficient processes.
PhD researcher Rob Faessen explored how such mistakes can be detected more reliably and developed methods to make process models safer and more robust. He defended his PhD thesis on Tuesday, October 7. Interestingly, Faessen is in his sixties and has almost the same age as his promotors. He currently works at ASML.
A smarter way to check equations
Most software used by chemical engineers can handle large systems of equations, but they offer limited help in checking whether the equations themselves are correct. Traditional checks, like making sure meters are not added to kilograms, catch only basic errors.
Rob Faessen introduced a more thorough approach that looks at five aspects of equations: chemical content, differential behavior, geometrical structure, relationships between variables that follow a consistent pattern, and a new concept called tensiveness.
Tensiveness extends the usual distinction between intensive properties (which remain unchanged when the amount of substance increases or decreases, such as temperature), and extensive properties (which increase or decrease with the amount of material, such as mass), allowing engineers to systematically check whether variables are combined correctly.
Turning ideas into software
To make these ideas practical, Faessen developed a modeling language called Equate and a software system called Equator. Equate helps engineers write process models in a clear and structured way, while Equator automatically checks them for internal consistency.
Unlike general-purpose tools like Julia or Mathematica, which mainly check unit compatibility, Equator goes deeper, analyzing the types of quantities and how they relate. This provides engineers with more accurate and useful feedback on potential errors.
Demonstrating effectiveness
Faessen first tested his methods on simple, deliberately flawed models, showing that Equator could find every mistake. He then applied it to a more complex, real-world example: a two-stage oil separation plant with hundreds of equations and variables. Equator successfully confirmed when the model was correct and flagged errors when they were introduced, demonstrating that the approach works in practice.
Safer and more efficient plants
The research also highlighted a bigger challenge: the International System of Units (SI) is not fully defined in a way that supports consistent computer-based modeling.
By refining these foundations, Faessen's work makes process modeling more reliable. In the future, these methods could be added to commercial software, giving engineers greater confidence in their designs, reducing costly mistakes, and improving both efficiency and safety in chemical plants.
Article written by Martijn Luyk
