Aspen Plus Simulation Issues

Simulation of reactive distillation using the standard models in Aspen Plus has some significant problems. The reactions can be specified to be either kinetic or equilibrium. In the former case, the choices of concentration units are limited to mole fraction, molarity, or partial pressure. Unfortunately, activity is not on this list. This is a problem because activities are frequently used to describe kinetic data. 

In addition, the reaction rate expression is limited to a power law in reactive distillation. Reaction rate expressions such as Langmuir-Hinshelwood-Hougen-Watson (LHHW) cannot be used. These restrictions make the use of Aspen Technology simulations tools for reactive distillation somewhat inconvenient. 

In the MTBE case in which equilibrium can be assumed, this problem would seem to be of no consequence. In the steady-state design using Aspen Plus, the chemical equilibrium model can be used. However, a serious limitation arises when one attempts to export the file 

One solution to this problem is to develop a special user-generated subroutine for the reaction. Aspen Technology has such a subroutine in its library for MTBE (Program Files AspenTech AMSystem 2004.1 Help ADExamples.pdf). It is written in Fortran and is called RAMTBE.f. We will use this subroutine in both the steady-state design in this chapter and in the dynamic control discussed in Chapter 15. 

In the ETBE case considered later in this chapter, a user-supplied kinetic subroutine had to be developed. The efforts of Bobby Hung of the National Taiwan University are gratefully acknowledged in its development. 

---

The stream conditions shown in Figure 14.1 are from the dynamic simulation of the process at steady-state conditions with the recycle of solvent loop closed. This loop did not converge in the steady-state Aspen Plus simulation. Other simulation issues are discussed in the next section. 

Before we leave this example, let us take a look at the issue of heat transfer. In setting up the simulation, we have specified the reactor temperature (430 K) and volume (100 m3) but have said nothing about how the heat of reaction is removed. The simulation calculates a heat removal rate of 12.46 x 106 W. If the aspect ratio of the vessel is 2, a 100-m3 vessel is 4 m in diameter and 8 m in length, giving a jacket heat transfer area of 100.5 m2. If we select a reasonable 30 K differential temperature between the reactor and the coolant in the jacket, the jacket temperature would be 400 K. Selecting a typical overall heat transfer coefficient of 851 W K-1 m-2 gives a required heat transfer area of 488 m2, which is almost 5 times the available jacket area. Aspen Plus does not consider the issue of area. It simply calculates the required heat transfer rate. 

The dynamic simulation file prepared in Aspen Plus is exported in Aspen Dynamics [10]. We select the flow-driven simulation mode. Aspen Dynamics files have already implemented the basic control loops for levels and pressures. Units with fast dynamics, such as the evaporator or some heat exchangers, may be handled as steady state. The implementation of control loops for the key operational units, chemical reactor and distillation columns, take into account some specific issues from the plantwide perspective, which are developed in detail in Luyben et al. [8]. 

In order to focus on the main issues of process integration, we disregard the distillation column for heavies, as well as the transalkylation section. A preliminary simulated flowsheet in Aspen Plus [9] is shown in Figure 6.8, with values of temperatures, pressures and heat duties. The fresh feed of propylene is llOkmol/h. Note that design specifications are used for the fine tuning of the simulation blocks. The fresh benzene is added in the recycle loop as makeup stream so as to keep the recycle flow rate constant. This approach makes the convergence easier. 

Several important issues arose in attempting to put together both the steady-state and the dynamic simulations. The AMINES physical property package was used in both simulations and gave reasonable results in Aspen Plus. The only simulation issue in Aspen Plus was failure of the solvent recycle loop to converge. This was solved by exporting the file... 

Most of the treatments in the above books are qualitative and concepmal in namre, emphasizing VLLE issues and alternative configurations. Few of these books present in-depth rigorous designs that achieve optimum economic criteria. None of these books deal with the control and operation of azeotropic distillation systems. Detailed discussions of these two areas are the main contribution of this book. Rigorous steady-state and dynamic simulation tools (Aspen Plus and Aspen Dynamics) are used for design calculations and rigorous dynamic simulations. 

---