Speciality Chemicals Magazine JAN / FEB 2024 | Page 61

SUSTAINABILITY
AI and machine learning are rapidly emerging as tools that can greatly accelerate the ability to employ plant data , both to calibrate first principles models , and to create data-based models of phenomena and processes quickly . AI has the potential to lower the expertise bar towards modelling process systems , but it must be combined with domain expertise to create the real-world guardrails that make it work safely , reliably and intuitively .
Hybrid models
Hybrid models can deliver comprehensive and accurate models quicker , all without requiring significant deep process or AI expertise . Machine learning is used to create the model , leveraging simulation , plant or pilot plant data , while using domain knowledge including first principles and engineering constraints to build an enriched model . This all democratises the application of AI within hybrid models for the optimal design , operation and maintenance of assets – deploying them online and at the edge .
The advanced and proven hybrid models approach is already accelerating bio feedstock based process development , through better and faster design of the new processes and faster learning , from early operating data from the first generation of biobased chemical products .
Firstly , the variability of biobased feedstocks , which may come from diverse sources such as agricultural residue , forestry biomass or specialised energy crops , presents significant challenges . Each type possesses unique characteristics and complexity , and AI will be valuable in using data to fit the particular feedstock to the process .
Secondly , these hybrid models can enhance the sustainability of bio feedstocks by optimising resource use and minimising waste . By integrating AI and analytics with domain expertise , these models can identify the most efficient and sustainable pathways for feedstock processing , taking into account factors like energy use , water use , carbon footprint and byproduct creation .
This holistic approach ensures that the bio feedstocks are not only viable as alternatives to traditional fossil fuels but also align with broader environmental and sustainability goals . For example , researchers at the University of Malaysia have conducted optioneering using process models to improve the conversion of palm oil to oleochemicals , achieving 30 % energy savings .
Finally , the real-time and edge deployment capabilities of hybrid models make them invaluable for continuous process monitoring and optimisation in the bio feedstocks industry . These models can be used to make swift decisions on feedstock sourcing and processing , reacting to changing conditions such as feedstock availability , market demand , environmental considerations and regulatory changes .
This agility is key in managing the supply chains of bio-based feedstocks , which can be subject to seasonal and ecological variations . For example , AspenTech is working with several
European refiners who are seeking to integrate biofeedstocks into their existing refinery processes . Operating challenges include the variable viscosity and composition of the bio components , which create unpredictable performance in the key economic units of a refinery .
Here , hybrid models can take the operating data , quickly recalibrate the first principle models and provide an accurate digital twin picture of how the refinery is performing . This can be given to the refinery planners , to better manage the bio content in the context of carbon intensity and margin in the refinery . The economic and sustainability benefits are immediate .
Optimising green certificates
To monetise the move to bio feedstocks , companies need to understand how to ensure audited compliance , through digital tracking of the bio-certificates that are created when bio feedstocks move through a production process .
Under EU rules , green certificates can be optimally employed with selected produced products that can be sold for the highest premium when accompanied by those certificates . We are working collaboratively with several European refining companies to develop digital approaches that capitalise in an optimised way on bio feedstock incorporation .
Outlook
The journey towards economically viable and sustainable bio feedstocks may be challenging . However , by embracing advanced modelling , a harmonious fusion of AI and hybrid technologies , and planning and supply chain modelling , industries can conquer the variability of bio feedstock compositions and bolster performance into the bargain . ●
J j
Kate Jones
SENIOR PR MANAGER
ASPENTECH kate . jones @ aspentech . com www . aspentech . com
JAN / FEB 2024 SPECCHEMONLINE . COM
61