Laboratory Steamfloods in Half of a Five-Spot

American Institute of Mining, Metallurgical, and Petroleum Engineers Inc. Abstract Eight crudes were steamflooded at different temperatures and pressures in a 3-dimensional model of a five-spot pattern. Oil production and steam sweep data are fitted with exponential curves whose parameters are correlated with oil and steam characteristics as are steam sweep and fraction of ultimate recovery occurring at steam breakthrough. In several correlations, crude viscosity as a correlator is replaced by distillation residue as determined by a simple test. Implications of the correlations are discussed. Introduction Enhanced recovery of oil by steam injection is practiced widely by the huff and puff method of stimulation. Several large-scale projects, among them Getty's huge Kern River, are being converted to steamflooding. However, few thermal projects are operated from the start as a projects are operated from the start as a steamflood. This is caused in part by the lack of information on how efficient and economic a continuous steam drive will be in a given field, and in part by the greater payout time that steamflooding part by the greater payout time that steamflooding necessitates. Furthermore, much of the oil needing enhanced recovery schemes falls in the category of heavy oil, and it is thought that oils of several thousand mPa-s (cp) will not allow a satisfactory steam sweep to be made. To obtain data on the applicability of steamflooding, eight crudes were selected from formations whose pressure was below 6.89 MPa (1000 psia)(at a depth of 2250 ft), thus allowing psia)(at a depth of 2250 ft), thus allowing real-pressure floods to be performed in a 3-dimensional model of half of a 5-spot pattern. To render the problem tractable, influence of gravity segregation is not taken into account, thus excluding modeling of steam override. For the same reason, inhomogeneities due to stratification are not examined, nor are pemeability variations. Moreover, steam injection rate is constant from test to test so as to keep the number of tests manageable. Even so, statistical analysis of the results yields useful correlations.