ETBE Synthesis via Reactive Distillation. 2. Dynamic Simulation and Control Aspects

Abstract

Ethyl tert-butyl ether (ETBE) is growing in importance as a gasoline oxygenate and octane enhancer. Its gasoline blending properties are superior to methyl tert-butyl ether (MTBE), and its semirenewability is attracting subsidies from many governments. Synthesis of ETBE via reactive distillation offers advantages of higher conversion, improved energy efficiency, and lower capital costs. A dynamic simulation, based on the MESH equations with supplementary equations to model the main chemical reaction, was developed using SpeedUp. The simulations were then utilized for the study of transient open-loop responses and for control system design. The control of a reactive distillation column presents several difficulties not normally associated with distillation, and dynamic simulation proved to be the ideal tool for the study and resolution of these problems. Some general recommendations for the control of reactive ETBE columns are made, including the need to address control issues early in the design process to recognize implications on process equipment design.

Pattern-based predictive control for ETBE reactive distillation

“…Measurement points are also indicated in Figure 1 for temperature, flow rate, pressure, and level variables. More information about the the architecture of the RD process can be found in [3,4,7].…”

“…LV, LB, D/V, (L/D)(V/B), etc. Among all possible control structures including ratio schemes and those with consideration of feed flow rate F f , non-ratio schemes LV and LB has been shown to be preferred for the RD column under consideration [4]. This work will consider the LV configuration.…”

“…Although the relationship between T 7 and the ETBE purity is non-linear, T 7 determines the purity uniquely. Another advantage of using T 7 is that the sensitivity of T 7 to the purity is high [4].…”

“…While more complex expressions of T 7 dynamics can be precisely established from the first principles of mass balance and energy balance, e.g. [3,4], these types of models are difficult to use for control design.…”

“…Although it is possible to develop a reasonable model for specific RD processes [3,4], such a model is too complex for control design. Furthermore, RD processes contain a large degree of uncertainties that cannot be well described in process modelling.…”