Nonlinear model predictive control for dividing wall columns

Ethyl acetate is a platform chemical conventionally obtained through fossil fuel routes, but more recently its production by fermentation from carbohydrates has been scaled up to a pilot scale. Yet, the complexity of downstream processing (low product concentrations in liquid broth and in off‐gas, azeotrope formation, and the presence of microorganisms) may complicate industrial application. This original theoretical study is the first to develop advanced downstream processing, based on process intensification principles, for large‐scale recovery (~10 kton/year) of ethyl acetate after fermentation. To minimize product losses, ethyl acetate is separated from both the liquid broth and off‐gas. The final purification is performed in a highly integrated azeotropic dividing‐wall column. The economic and sustainability analysis shows that using refrigeration for initial product separation from the gas phase is more cost‐effective (~0.61 $/kg) and less energy‐intensive (2.20–2.40 kWthh/kg) than compression combined with high‐pressure condensation using chilled water (1.09 $/kg and 9.98 kWthh/kg).

show abstract

Dividing‐wall column for separating wide boiling mixture: Optimal design and servo control strategy

Aurangzeb,

Banerjee,

Roy

et al. 2024

Asia-Pacific J Chem Eng

View full text Add to dashboard Cite

The dividing‐wall column (DWC) stands as an energy‐efficient distillation technology designed to efficiently separate ternary systems into their pure components within a single column. Despite its efficiency, integrating two towers into a DWC poses challenges in controllability. Published studies have focused on developing controllers for disturbance rejection but have not addressed the issue of servo control. Addressing this, our study proposes a servo controller problem and a methodical approach to strategically select controlled variables that undergo consecutive set point step changes, aiming to track the purity of key products obtained from the DWC. To achieve this goal, we developed a dynamic model specifically tailored to the DWC. Determining the design parameters involved employing a genetic algorithm optimization technique, minimizing the total annual cost. Subsequently, we implemented three servo parallel proportional‐integral (PI) feedback controllers for three distinct product streams. The fine‐tuning of their controller gain and reset time is carried out by minimizing the integrated square error while adhering to valve opening constraints. Our investigation revealed a limitation: the introduction of simultaneous step changes in all three controlled variables proved unfeasible with the three active PI controllers. As an alternative, our findings recommend prioritizing two variables: the lightest (distillate) and heaviest (bottom) key components, alongside the middle component (side stream) and either the lightest or heaviest key component. This research underscores the complexities of implementing step changes in multiple controlled variables within a DWC, offering insights into optimal control strategies for enhancing the efficiency of this innovative distillation technology.

show abstract

Nonlinear model predictive control for dividing wall columns

Cited by 10 publications

References 30 publications

Dynamic Simulation and Control of Intensified Chemical Processes

Dynamic Simulation and Control of Intensified Chemical Processes

Advanced downstream processing for sustainable ethyl acetate production by fermentation

Dividing‐wall column for separating wide boiling mixture: Optimal design and servo control strategy

Contact Info

Product

Resources

About