Dynamic control of reactive pressure-swing distillation process for separating tetrahydrofuran/methanol/water

Abstract

BACKGROUND

Reactive pressure-swing distillation (RPSD) is an innovative and promising technology for the separation of azeotropic systems, offering high integration and efficiency. However, the dynamic control of RPSD processes presents significant challenges due to complex coupling and nonlinearity. This paper investigates the practical control structures for the RPSD processes, and the dynamic control structure was first designed for the RPSD system for separating tetrahydrofuran/methanol/water.

RESULT

Three practical control structures based on PID control were developed for existing separation processes, both with and without heat integration. The results demonstrated that all designed control structures effectively withstand disturbances of ±20% in feed flow rate and composition, exhibiting strong anti-interference capabilities. Furthermore, an advanced intelligent control strategy was designed, integrating PID control with Back Propagation Neural Networks (BPNN) to enhance the performance of the product composition controllers.

CONCLUSION

The results indicated that the BPNN can accurately predict temperature setpoints using easily measurable variables, facilitating stable control of product concentrations without the necessity for direct composition measurements. This proposed control strategy offers an efficient and reliable alternative for the control of RPSD processes. © 2025 Society of Chemical Industry (SCI).