(95e) Octane Boosting In a Membrane Assisted Radial Flow Naphtha Reactor, Using DE Optimization Technique

M.R. Rahimpour[1], D. Iranshahi, E. Pourazadi

School of Chemical and Petroleum Engineering, Department of Chemical Engineering,

Shiraz University, P.O. Box 71345, Shiraz, Iran

Abstract:

The significant impact of catalytic naphtha reactors on the total benefits from the refinery complexes has been road toward a continuous evolution and significant advances in naphtha reactors and its operational conditions. Regarding this, the present study proposes a tubular membrane reactor with the radial-flow patterns of the sweeping gas and the naphtha feed named RF-TMR as a novel configuration instead of a conventional axial-flow tubular packed bed naphtha reforming process. The cross section area of the reactor is divided into some subsections. The walls of the gaps are coated by a Pd-Ag membrane layer to increase the hydrogen permeation rate. The membrane layer enhances pure hydrogen while boosts the octane number of reformates by increasing the aromatic production. The DE optimization technique is utilized here to find proper operational and dimensional parameters. Operating pressures (reaction and permeation sides), catalyst mass distribution in each reactor, total number of subsections, the ratio of length to diameter (LOD) for each reactor and the proper ratio of sweep to catalytic side angles are considered as the most important decision variables. The results show successful aromatic production in addition an increase about 27 kmol/hr in the hydrogen production rate compared with the conventional naphtha reactor. The idea is capable to apply and revamp the fixed and moving bed radial-flow naphtha reactors. Mathematical modeling and also optimization technique should be apply for such a new ideas before commercialization and making decision for set up of a new pilot plants.

Keywords: Chemical Reactions; Membrane concept; Radial flow reactor; catalytic naphtha reforming; Octane boosting; Hydrogen production.

Conventional Axial-flow tubular packed bed reactor (CTR)

A simplified schematic of CTR is explained in Fig.1.

Figure.1

The naphtha feed is mixed with the recycled gas containing 60-90% (by mole) hydrogen and preheated before entering the 1^st reactor. Reactors are packed with catalysts and the chemical reactions take place on catalysts' surfaces. Since naphtha reforming is an endothermic process, the outlet stream must be preheated before entering the following reactor by inter-stage furnaces. In order to stabilize the liquid and separate the gaseous product, the effluent from the 3^rd reactor is cooled and directed into the separators. The liquid product is called reformate which mainly consists of aromatics (60–70 mass% of naphtha feed) and saturates in the C₅–C₉ carbon range. The main reactions in the first reactor are dehydrogenation and isomerization, in the second reactor are dehydrogenation, isomerization, cracking and dehydrocyclization and in the third one are cracking and dehydrocyclization [1].

Conceptual for proposed configuration (RE-TMR)

A schematic of proposed configuration is illustrated in Fig.2.

Figure.2

The internal design is such that the flow enters the vessel across the entire cross section of the vessel, first proceeding downwards near the wall of the vessel and then radially inwards, through the bed of catalytic particles and finally in a downwards direction via an axial collector. The reformed naphtha is collected by the axial collector pipe. The walls of the gaps between these subsections are coated by the Pd-Ag membrane layer, thus hydrogen permeates through the membrane layer and enters the shell side. Moreover, the sweeping gas flows radially in the shell side and carries the permeated hydrogen. Due to negligible pressure drop, smaller catalysts' particles with minor internal mass transfer resistance can be used.

In order to investigate the reactor behavior, four dominant reactions are considered based on the Smith's model to predict the outlet compositions of paraffin, naphthene and aromatics which are known as PNA. It is valuable to mention that a dynamic homogeneous one-dimensional model is used. All the physical properties are considered to be variable along the reactor. Orthogonal collocation method is used, as a powerful method in solving the PDEs, to achieve dynamic results of modeling. The validity of model is examined by comparison of daily report of conventional process during 800 days of operation.

Optimization results

Differential evolution (DE) is a heuristic approach based on optimization evolutionary algorithms and random search methods. This technique was presented by Storn and Price (1996) [2]. A Complete list of optimized parameters with their values is presented in Table 1.

Lower operational pressure in the permeation side is better from the economical point of view. However, it is worth mentioning that reducing the sweep gas pressure provides a stronger driving force for hydrogen permeation (lower H₂/HC) which in turn causes rapid catalysts deactivation. Membranes thickness is a key parameter for hydrogen permeation. Thickness reduction means less resistance for hydrogen permeation. However, as previously discussed, only the excessive amount of hydrogen needed for catalyst deactivation prevention (H₂/HC=4.73) must be removed from the reaction side and therefore it is not possible to reduce the thickness beyond a certain value and an optimum value should be determined. Sweep gas molar flow rate is also investigated in this work. Since the naphtha reforming reaction is endothermic, the temperature decreases along the reactor bed which is undesirable and it should be avoided. Therefore, the catalysts are distributed in three consecutive reactors with intermediate furnaces. In this work, the optimum catalysts mass distribution for each reactor is obtained. The hydrogen mole fraction in the sweep gas stream needs to be optimized. Higher hydrogen content of sweep gas reduces the need for high operating pressure and related cost .In this work, the value of hydrogen mole fraction in the recycle stream is optmized. Increasing the number of subsections provide higher heat transfer area to the reaction side and increases the products’ yields. On the other hand, more subsections cause more hydrogen extraction from the reaction side and reduces the H₂/HC. Consequently the proper number of gaps should be defined.  ratio is also optimized for all three reactors. Finally the fraction of sweep gas allocated to each reactor is optimized.

According to the achieved results from the optimization step, hydrogen and aromatic production rates are increases by 10.0 and 5.4 % compared with the non-optimized RF-TMR. The optimized reactor is capable to increase the reformate production rate by 3.9 % more than the CTR and 1.62 % boosting in aromatic content.

References:

[1] Khosravanipour Mostafazadeh A, Rahimpour MR. A membrane catalytic bed concept for naphtha reforming in the presence of catalyst deactivation. Chem Eng Process 2009; 48: 683–94.

[2] Storn R, Price K, Differential Evolution – A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces. J Global Optim 1997; 11: 341–359.