The CO2 Huff -Puff technique is becoming an attractive technique to stimulate a large variety of wells. This paper presents a new approach to the simulation of this technique, using the concept of relative permeability hysteresis. A newly developed software was used for the study. It is made of a general purpose compositional simulator system that displays numerous specific features arranged in a modular structure. For increased efficiency, numerical models adapted to given problems, can be automatically generated from this system, In the present study, a model was generated to include among other options the kr, Pc hysteresis and to simulate the Huff-Puff process. This study is mainly focused on low productivity reservoir for which CO2 stimulation could represent a viable alternative to being shut down. First the case of a watered out light oil reservoir was investigated. The model was validated by successfully history-matching production data from a typical huff-puff cycle. A performance prediction was performed for additional cycles and parametric runs on amounts of injected CO2 were performed. Then the case of a heavy oil reservoir with no water production was studied. A successful history match of three consecutive stimulation cycles was obtained. However it was found necessary to introduce a rate of gas dissolution in heavy oil, which is more representative of the physical phenomena. A performance prediction was conducted for additional cycles; and a number of parametric runs were performed. The results show that for heavy oil, gas injectivity and oil productivity may increase with the number of cycles in an unlimited medium, higher final oil productivity is obtained with short cycles than with long ones, the influence of up to 25% hydrocarbon gas in the CO2 is limited. Finally it appears that in both types of reservoirs, the relative permeability hysteresis is the main cause for incremental oil production. Oil viscosity plays a major role with the heavy oil only, as would normally be expected. Introduction With the success of large CO2 flood operations, particularly in West Texas, there has been a world wide drive for testing CO2 in many reservoirs. Often, it has been found easier to conduct a huff and puff test rather than a pilot test for a large number of reasons. Many operators have subsequently found that the CO2 stimulation process could become profitable and that the life of their wells could be extended without going to the trouble of putting much money into large investments. In the recent past, there has been several publications dealing with CO2 huff and puff. Most of these publications present field cases, while only two papers have tackled the problem of simulating the process. The first one by Patton, Coats and Spence deals with heavy oil and the second one by Hsu and Brugman with light oil only. P. 109^

L. Denoyelle, SPE, Inst. Fran<;:ais du Petrole C. Bardon, SPE, Inst. Fran<;:ais du Petrole E. Couve de Murville, SPE, Petrorep Summary. The Coulommes-Vaucourtois field operated by Petrorep and located in the Paris basin is near abandonment. During the 25-year production span, one-fourth of the original oil in place (OOlP) has been produced by a strong bottomwater drive. Reservoir engineering studies and laboratory measurements have shown that the reservoir characteristics were favorable for CO 2 injection.In 1983, an injection well was drilled between four existing production wells. During the spring and summer of 1984, 49 x 10 6 scf [1.4 x 10 6 std m 3 ] of CO 2 followed by 54 x 10 6 scf [1.5 x 10 6 std m 3 ] of N2 were injected. After 21/2 months of CO 2 injection, gas broke through into a well located far outside the pattern area. Shortly afterward, a producing well of the pattern showed a four-fold oil production increase as gas broke through. The three other wells exhibited smoother responses, with production of minute quantities of gas. Significant WOR decreases and oil production increases were observed. Gas analyses showed that the areal extent of the gas "bubble" was 15 to 20 times larger than the pattern area and that CO 2 and N2 did not spread in the same directions.This paper tries to match the test results with the geologic description, which shows the presence of fissures, and the production history of the reservoir, which exhibits a single-porosity behavior. It also stresses the difference between the incremental oil recovery resulting from the formation of an oil bank and that resulting from the gas-lift effect.

No abstract

No abstract