Modelling a counter-diffusive reactor for methane combustion

Abstract

The counter diffusive reactor containing a catalyst supported on a fibre pad is often used as a radiant heater. Fuel fed from the inlet of the pad is combusted by oxygen diffusing from the reactor exit. This paper describes the development of a two dimensional model of this heater with methane as the fuel. The detailed transport equations are solved using the finite element method. The correct implementation of the boundary conditions is emphasized. The effects of external mass transfer, reaction rate and feed flow rate on the reactor performance are studied. Comparisons are made to experimental results obtained in an earlier investigation. It is shown that the primary limiting step that controls the conversion of fuel is the rate of mass transfer of oxygen through the boundary layer that develops in front of the reactor. Combustion efficiencies approaching 100% can be achieved at sufficiently high rates of reaction.

Introduction

The combustion of fuel, also known as deep oxidation, can be achieved in two ways. The most common is homogeneous combustion, which proceeds at high temperature and (usually) in the presence of a flame. The alternative is catalytic combustion, where the presence of a catalyst initiates and sustains the combustion reaction. First reported by Davy (1818), catalytic combustion offers the advantage of low temperature reaction, thus avoiding NO_X formation, and the absence of flammability limits (Hayes & Kolaczkowski, 1997). There are many examples of catalytic combustion systems, and applications are often divided into primary combustion, where the objective is the generation of heat and/or power, and secondary applications, where the purpose is the destruction of pollutants. Sometimes an application achieves both purposes. The use of catalytic combustion for the abatement of various fugitive emission streams is of considerable interest (Hayes, 2004, Kiwi-Minsker et al., 1999). In these applications, the use of catalytic combustion often has the advantage of not requiring the addition of support fuel.

Catalytic combustors require the presence of sufficient oxygen, and in most cases the feed to the reactor is pre-mixed, that is, the fuel and oxygen are both supplied at the reactor inlet. However, there also exist reactors where only the fuel is supplied at the reactor inlet, and the necessary oxygen for combustion is supplied by other means, usually be mass transfer from the surrounding air. The most common example of this type of reactor is the so-called counter-diffusive reactor. These reactors are most often used for heating applications, where radiant heating is desired (Dongworth and Melvin, 1977, Hayes and Kolaczkowski, 1997, Radcliffe and Hickman, 1975). A typical design, which is basically a box, is shown in Fig. 1. At the back (reactor inlet) there is usually an arrangement of tubes, with appropriate flow and pressure regulators, for the fuel inlet, through which pure fuel is admitted. Directly in front of this open section is a porous insulation blanket, which acts to reduce energy loss through the back. After the insulation blanket is the catalyst pad, which is a porous matrix of fibres that contains the active catalyst. The catalyst in most commercial units is based on platinum, although palladium (Kiwi-Minsker, Yuranov, Slavinskaia, Zaikovskii, & Renken, 2000) and non-noble metals (Saracco, Cerri, Specchia, & Accornero, 1999) have also been reported. The catalyst pad is often held in place by a wire screen at the front surface. An electrical heating element is usually positioned between the catalyst pad and the insulation blanket. Its role is to heat the pad to a sufficiently high temperature to initiate reaction, after which the heater is switched off and the reactor should have auto-thermal operation. The oxygen required for reaction diffuses from the air. The oxygen must first diffuse through the natural convection boundary layer at the front of the catalyst pad, and then diffuse into the pad, against the direction of the flow. The main mechanism for heat transfer from the reactor is radiation from the front of the pad. Smaller amounts are lost by convection from the heater body, and also by bulk flow through the heater. Some of the more significant literature on the experimental and modelling aspects of these reactors is given in the following paragraphs.

Trimm and Lam, 1980a, Trimm and Lam, 1980b studied a convective–diffusive combustion system over Pt supported on alumina fibre pad. They studied the effect of methane flow rate on temperature distribution across and through the catalyst pad and radiant energy from the catalyst surface. By increasing methane flow rate, the maximum temperature moved towards the front of the pad and combustion efficiency decreased. Seo, Kang, and Shin (1999) used a counter-diffusive catalytic system with propane and air as fuel and oxidant was used. System performance was improved by applying the combustion air by force and pre-mixing air and fuel. In a later study, Seo, Cho, Song, and Kang (2002) studied the effect of space velocity on combustion efficiency and found that increasing space velocity has a negative effect on the combustion efficiency.

Trimm and Lam (1980b) conducted a modelling study on a counter-diffusive system, in which a one dimensional model was developed and the governing equations were solved by finite difference method. Model predictions were compared with experimental results from the convective–diffusive catalytic combustor. The model predicted a wider reaction zone and the position of maximum temperature moved to the front of the catalyst pad at higher fuel flow rate. However temperature prediction at the front and back of the pad was not accurate. Both model and experiments showed decreasing trend in combustion efficiency by increasing fuel inlet flow rate. It was discussed that at higher fuel flow rate the diffusion of oxygen to the catalyst pad becomes difficult and combustion efficiency decreases. They report that at higher fuel flow rate the reaction zone moves to the front and the low contact time leads to methane slippage.

Specchia, Sicardi, and Gianetto (1981) proposed a model for methane catalytic combustion assuming very high reaction rate. They regarded the panel as an isothermal tubular reactor with constant longitudinal dispersion through the panel. The modelling data were compared with the experimental results obtained by Dongworth and Melvin (1977). Two sets of data showed good agreement with each other in terms of components mole fraction. They report that the combustion takes place in a thin layer near the front surface of the catalyst panel. Both studies emphasize the oxygen diffusion into the catalyst panel to be the controlling factor in the diffusive catalytic combustion system.

Although there have been some modelling studies done on the radiant heater, we are not aware of any comprehensive two dimensional studies that included correctly all of the physical phenomena, especially the boundary conditions. The objectives of this paper, and hence its contribution, are to show clearly the effects of the operating parameters, the use of correct boundary conditions, and comparison to experimental results.