Computer-aided thermal design of furnace

Abstract

An energy balance of a furnace had been performed to develop a comprehensive mathematical model for the prediction of its temperature and heat flux distributions. The energy sources under consideration included combustion of the air/fuel intake, soot formation during the combustion processes, heat content of the exhaust gas, heat transfer by convection and radiation to the load and furnace enclosure, and conduction in the load and the furnace walls. The non-dimensional equation emerged from the experimental work of Lebedev[16] had been adopted to calculate the convection heat transfer. The Hottel's zone method [6,7,8] had been used for the radiation heat transfer calculation. The major difficulty in linking up the convection and radiation models directly to produce the overall heat transfer is the determination of the total exchange areas, which are the most important inputs in performing gaseous radiation in a grey enclosure. In addition, integral equations involved in solving the direct exchange areas are incompatible with the equations used for convection heat transfer when it is determined numerically. On the other hand, such integral equations used to estimate direct exchange areas are only available for very simple geometry such as cube or cylinder. A numerical approach, the Monte Carlo method [2,3], had been applied to overcome this drawback and generate the total exchange areas (i.e. surface-surface, surface-gas and gas-gas) for the radiation calculations. However, the total exchange areas obtained by using the Monte Carlo method usually suffer from rather poor accuracy because of the random generation nature of the method. A least square smoothing technique [4,5], which is based on the concept of a better fulfillment of the Reciprocity Theorem, had been introduced to improve their accuracy. I had made a successful contribution to apply a new and better approach for the calculation of gaseous radiation in a grey enclosure. Based on the proposed model, I had also developed a computer programme, with the aid of Visual Basic to perform the following prediction for an oil-fired open flame furnace under both the transient and steady-state operations: 1. Combustion gas temperature distribution along axial length of the furnace; 2. Temperature distribution of the furnace enclosure; 3. Heat flux distribution in the furnace; 4. Soot amount inside the combustion gas; 5. Time required for a particular part of the furnace to reach a pre-set temperature. My other contribution was to integrate all energy sources in the above model to predict the thermal performance of a furnace, especially during the transient operation which has rarely been investigated before the present study. Results of the present study provided a very useful tool for the Design Engineers and Plant Engineers to predict and analyze the thermal performance of an oil-fired open flame furnace, including the start-up period, which is most critical for many industrial heating processes. A numerical prediction had been made to test the applicability of the proposed methods and the programme developed. The sample furnace had been divided into 18 surface zones and 4 gaseous zones. The surface-surface, surface-gas, gas-surface and gas-gas total exchange areas had been calculated using the proposed smoothing Monte Carlo method. After integrating the other thermal models, a transient response curves had been obtained by the programme. The proposed methods had been presented in local/international conferences [39,40]. A journal paper had been sent to The Journal of Heat and Mass Transfer [41], and was accepted for publication.