Implementing Internal Reflux Control

26 Apr 2025 at 17:36 // under Chemical Engineering Process Control

Many years ago now, I implemented a significant process control improvement around the top of a debutaniser column and this significantly improved overall control of the column.

Background

The E101 debutaniser column is used to remove LPG and Fuel gas from a gasoline product. It is critical that the light LPG components are removed to ensure the liquid product is stable enough to be stored in tanks, while also ensuring that the LPG stream is not contaminated with heavier C5 components which would cause issues with the downstream LPG processing units.

Most distillation columns operate using a 'total condenser'. These condense all vapour leaving the top of the column and route place this back into a reflux drum before returning a portion taking some out as product

Unlike most distillation columns, the main top product does not leave the system directly from the reflux drum but a side stream 4 trays down the column. This allows a better separation from the offgas, the vapour taken off the top of the reflux drum, and the liquid LPG product.

As a result all the liquid in the reflux drum gets returned to the column.

The pressure in the column and reflux drum is maintained by adjusting the rate of off-gas production. This has a direct effect on the pressure and it is important to operate the column at a consistent pressure so there is no need to change this.

The only two 'handles' left to maintain the status at the top of the column are the reflux, flow from the reflux drum to the top of the column, and the LPG draw rate from the 6th tray.

Initial control scheme

In the original configuration:

• E101 (debutaniser) reflux was set at a fixed flow rate
• F120 (reflux drum) level was controlled by adjusting the LPG draw from tray 4

This setup presented several drawbacks:

• Indirect level control: Changes in LPG draw didn't directly affect the F120 level, as it relied on altering the overhead vapor flow, which is influenced by multiple variables.
• Fluctuating internal reflux: The fixed E101 reflux flow combined with variable LPG draw led to significant changes in the liquid flow down the column below tray 4, impacting separation efficiency.

These issues often resulted in C5 excursions in the LPG draw and challenges in reformate RVP control.

Internal reflux control scheme

I redeveloped the control schemes around the top of the column to address these challenges:

• Direct F120 Level Control: I reconfigured the control system to adjust the E101 reflux flow based on the F120 level. This provides a more direct and responsive level control mechanism.
• Steady Internal Reflux: I implemented a new control scheme for the LPG draw based on a calculated internal reflux. This calculation takes into account:
• E101 reflux flow
• LPG draw
• Internal condensation due to subcooling

Internal condensation occurs when vapour rising up the column encounters cooler conditions in the upper sections, causing some of the vapour to condense back into liquid within the column itself. This phenomenon is particularly pronounced when there's subcooling in the overhead system - where the reflux returning to the column is at a lower temperature than the vapour it contacts. The condensed liquid effectively acts as additional internal reflux, but unlike external reflux flow which is measured and controlled, this internal condensation is invisible to the control system and varies with operating conditions such as overhead cooling efficiency and reflux temperature. This creates an unmeasured disturbance that can significantly impact separation performance and makes precise level control challenging.

The internal reflux is now calculated as:

$$ \text{Internal Reflux} = \text{E101 Reflux} + \text{E101 internal condensation} - \text{LPG draw} $$

Where:

$$ \text{E101 internal condensation} = \frac{C_{p(\text{Rx})}}{H_{\text{vap}(\text{Rx})}} \times (T_{\text{Oheads}} - T_{\text{Rx}}) \times F_{\text{Rx}} $$

• $C_{p(\text{Rx})}$ = Reflux Heat Capacity = 2.816 kJ/kg·°C
• $H_{\text{vap}(\text{Rx})}$ = Reflux Heat of Vaporization = 376.3 kJ/kg
• $T_{\text{Oheads}}$ = Temperature of overheads leaving the tower
• $T_{\text{Rx}}$ = Temperature of reflux entering the tower
• $F_{\text{Rx}}$ = Flowrate of reflux entering the tower

This approach ensures that both the E101 reflux and LPG draw vary in tandem, maintaining a steady liquid return down the column.

Results and Benefits

The implementation of these changes has yielded several significant improvements:

• Enhanced Stability: The trend graph shows markedly reduced fluctuations in key parameters post-implementation, indicating a more stable operation of the E101 column.
• Improved Product Quality: With better control over the internal reflux, we've observed a reduction in C5 excursions in the LPG draw and improved reformate RVP control.
• Operational Flexibility: An additional benefit is the ability to operate the column under total reflux by simply closing the LPG draw valve, without additional manual intervention.
• Disturbance Rejection: The new control scheme helps mitigate the impact of external factors such as ambient temperature changes or rain showers on the column's performance.

Conclusion

This project demonstrates the significant impact that targeted process control improvements can have on refinery operations. By addressing the root causes of variability in the E101 column, we've not only enhanced product quality and consistency but also improved operational flexibility and efficiency.

Go Top