Abstract A preliminary two-dimensional model for in-situ underground coal gasification is presented for discussion. The model describes cavity formation from the injection well end and link zone burn up, where the initial link zone geometry is assumed at this time. Cavity growth is predicted via an integral boundary layer analysis with boundary conditions determined by oxygen mass transfer to the wall for forced and natural convection. Preliminary results are based on carbon oxidation, the reduction of Co, and the water-gas reaction, but exclude reaction rate effects. The effect of water influx is discussed. A comparison between Eastern (Pricetown) and Western (Hanna) coals is presented, as well as dimensionless plots which indicate similarity solutions for various blast flow rates. Introduction Recent emphasis has shifted from one-dimensional to two-dimensional modeling efforts to describe the underground coal gasification process. Most of these two-dimensional models have sought to predict the extent to which the coal is gasified, as well as to predict the quality of the gasification products. Although progress is being made towards a better understanding of the UCG process, much remains to be learned about the detailed process, much remains to be learned about the detailed mechanism which governs the process. The modeling difficulties are compounded by the many aspects that must be considered, such as seam thickness, water inflow, type of coal, subsidence and seam angle. The primary difficulty in multi-dimensional modeling of the UCG process is the lack of adequate laboratory and field process is the lack of adequate laboratory and field data to aid in an accurate description of the gas flow patterns and reaction mechanisms taking place within a patterns and reaction mechanisms taking place within a coal seam. Until further progress is made along these lines, modeling efforts will include a good deal of speculation concerning the physical mechanisms involved in the UCG process. This stage of the model development is devoted to an investigation of the flow processes and overall reaction mechanisms which may significantly affect the growth of the exploited region in a UCG system. Concentration on the overall process defers the detailed reaction kinetics which are needed to accurately predict the product gas composition until the next stage. The product gas composition until the next stage. The specific mechanisms considered in the cavity growth and the first stage computer program of the model are described in this paper. PREVIOUS 2-D UCG MODELING WORK PREVIOUS 2-D UCG MODELING WORK Previous two-dimensional, UCG model studies describing the areal sweep of the gasification zone have usually treated the borehole method of gasification. These analyses have examined the growth of a borehole by starting with a small diameter horizontal hole embedded in the coal. This type of model in part describes a geometric arrangement which has been employed at Newman Spinney in Great Britain and in the Soviet Union. Warner and Szekely carried out one of the earlier analyses of the enlargement of the entrance section of a channel which simulated some of the British UCG field work. The model assumed that the process was mass transfer controlled in the combustion zone and that the wall and gas temperatures were uniform. The mass transfer coefficient was obtained by using the fully developed pipe flow equation for the Sherwood number. Little detail of the specific model equations were given and it is assumed that the analysis was a simplified first attempt at predicting channel growth. A second paper by Warner and Szekely describes an analysis that determines the limiting width of the reaction zone. A thermal analysis was carried out to show that this width was limited by heat transfer effects and that the maximum width was directly proportional to the blast intensity, the calorific value of the coal and the square root of the seam thickness. Dinsmoor, et al., recently developed a model of channel formation using essentially the same geometry assumed by Warner and Szekely. The model departs significantly from the earlier British model including a numerical analysis which considered temperature and specie concentration changes along the channel. Carbon consumption reactions were assumed to occur on the tube walls and heat transfer into the coke was considered.