Effective, data-driven process control can help conventional oil refineries to adopt sustainable feedstocks to produce greener fuels
contain a wide variety of impurities in high concentrations. Moreover, these are usually very similar to the main components, making it particularly difficult to remove them without incurring thermal degradation of key chemicals. Consequently, feedstock pre-treatment and product purification can play an essential role in determining the end quality, processing time, energy efficiency and ultimately environmental impact of decarbonising strategies for the oil and gas sector. Similarly, the reaction conditions applied during fermentation, transesterification and pyrolysis can greatly influence a business’s profitability. While under-processing can compromise end product quality, over-processing is energy and time ineffective. A data-driven system that can monitor the physical, chemical and/or biological properties of materials in real-time; determine the desired product quality and use this information to swiftly adjust process conditions is crucial to addressing the aforementioned challenges. Process analytical technology (PAT) is a framework that offers all these features. Data-driven process control at the core of sustainable processing PAT is based on Quality by Design (QbD) principles, which aim to build quality into products
from the earliest design phase rather than testing them at the end of the manufacturing process. According to this approach, a design space of operating critical process parameters (CPPs) is defined to ensure specified critical quality attributes (CQAs) are delivered in the end material. During processing activities, PAT utilises a network of sensors and analytical instruments to observe the relevant material properties via on-line, in-line or at-line measurements. The univariate and multivariate data obtained, such as near-infrared (NIR) or UV-vis spectra, are sent to key software equipment, such as chemometric and predictive modelling tools for data processing, analytics and visualisation, e.g. multivariate analysis (MVA) and chemometric modelling tools. By using these software solutions, refineries can generate a comprehensive knowledge of the materials and processes used. In particular, it is possible to characterise the feedstock, rather than relying on abstract, pre-set properties. For example, PAT has been successfully applied to determine lignin content in plant-based materials used for bioethanol production. More precisely, infrared spectroscopy combined with advanced chemometrics models can help biorefineries adjust their pre-treatment, fermentation and pyrolysis activities on the fly to address the
www.decarbonisationtechnology.com
56
Powered by FlippingBook