Rapid digital process scale-up applied to heterogeneously catalysed processes

Heterogeneous catalysis in which a fluid (typically a gas) reacts on a solid catalyst is used in a number of vital industrial processes including the Haber-Bosch for ammonia production or the Fischer-Tropsch (FT) synthesis to produce sustainable fuels from waste or biomass.

According to Avioxx Chief Technology Officer, Dr Steve Wilkinson, “Scaling up chemical reactions and reactors from gram-scale to industrial production scale is a complex engineering process that hasn’t changed for over 100 years.” The major challenges of scale-up are:

• Decreased surface area to volume ratio of gas to catalyst which effects heat and mass transfer
• Massively increased power requirements and longer times required for mixing
• Moving from batch chemistry (i.e. mixing all material at once) to continuous flow chemistry (i.e. reacting materials in a continuous flowing stream)

There are many historic case studies in which a failure to account for these challenges have resulted in unsafe or inefficient process designs which are inoperable. The traditional answer, therefore, has been to follow a careful series of steps to scale up equipment (or amount of material for reaction) by 1 or 2 orders of magnitude at a time in order to gain a better understanding of the reaction chemistry at increasing scales.

“The problem with this approach, however, is that it is time consuming and costly, typically taking many years and many millions of pounds spent before the production scale system is even designed, let alone built.” says Dr Wilkinson. “The long payback times make it very difficult to secure investment from the private sector, which leaves chemical process development in the hands of big corporations with deep pockets. This massively stifles competition and much-needed innovation in a sector that must rapidly re-invent its processes to decarbonise by 2050”.

By-passing the traditional process scale-up paradigm using a PAIME informed Digital Twin Reactor with Machine Learning

At Avioxx, we have a new approach which we call PAIME (‘Perturb All Inputs & Measure Everything) to rapidly scale up from a test system to a full-scale production plant, focusing on the FT synthesis to make sustainable jet fuel. This process involves reacting carbon monoxide and hydrogen gases over a metal catalyst to produce a range of hydrocarbons then separating out the ones used for aviation.

The central idea of our method is to make far more measurements at the smallest scale (the Bench Scale) than in the traditional process development method. These measurements are achieved by multiple sensors for temperature, pressure etc. combined with online analysis methods such as gas chromatography-mass spectrometry (GC-MS) to measure the distribution of hydrocarbons being produced in real time.

The PAIME approach can be considered to adopt a signal processing philosophy in which time varying inputs in feed composition and throughput (composition and amount of materials undergoing reaction, respectively) are applied and the outputs recorded. These data are fed into a dynamic model (digital twin) which has a detailed representation of the chemistry and hydrodynamics of the reactor. In particular, both spatial and temporal variations in the catalyst bed are modelled using partial differential equations and/or multi-physics approaches.

By thoroughly collecting large sets of data that represent all aspects of the FT reaction, we then use complementary ‘Black Box’ statistical data analysis and machine learning methods to optimise designs for large-scale plants from bench scale systems.

To learn more about the Avioxx Process and our areas of research and development, please get in touch at info@avioxx.com.

Sources

Peacock, M., et al., Innovation in Fischer–Tropsch: developing fundamental understanding to support commercial opportunities. Topics in Catalysis, 2020. 63(3-4): p. 328-339.

Gardezi, S.A., et al., Thermochemical biomass to liquid (BTL) process: Bench-scale experimental results and projected process economics of a commercial scale process. biomass and bioenergy, 2013. 59: p. 168-186.

Mitchell, S., N.-L. Michels, and J. Pérez-Ramírez, From powder to technical body: the undervalued science of catalyst scale up. Chemical Society Reviews, 2013. 42(14): p. 6094-6112.