SPWLA Thirty-Third Annual Logging Symposium, June 14-17, 1992        PAPER FF

 

Paper ff

 

SHELL CANADA EXPERIENCE WITH OXYGEN ACTIVATION IN THREE PHASE FLOW

 

Keith Boyle

Shell Canada Limited

 

ABSTRACT

 

It has long been recognized that defining a flowing profile in a well pro­ducing three phases from a small interval (less than 50m) is extremely difficult with current production logging tools. The diffi­culty arises from the limited number of measurements that can be made with the current spinner, gradiomanometer, capaci­tance, pressure and temperature tools. The recent introduction of the oxygen activation technique to detect water flow brought with it an independent measurement of water-phase velocity. This data, when combined with traditional production log interpreta­tion and surface measurements, can theoreti­cally be used to accurately define a three phase flowing profile. Unfortunately, .recent field experience has shown that the water phase velocity derived from an oxygen acti­vation log does not correlate to the tradi­tional spinner interpretation or to surface measured rates in the presence of multi-phase flow. Two wells were logged using an oxygen activation log and a conventional production logging suite (spinner, gradioma­nometer, pressure, temperature and, in one case, a water cut meter). The wells are part of a hydrocarbon miscible flood, which have recently experienced solvent breakthrough. The reservoir is a multi-layered reef, and the logs were run to identify which layers had solvent breakthrough. The wells pro­duce between 500 - 700 m3/d total fluids @ 90% water cut (W.C.) with a gas-oil-ratio of 2000 - 3000 m3/m3. In one well the oxygen

activation log results correlated very well with the production log interpretation in a low flow rate/single phase interval. In the higher flow rate/multi-phase section, how­ever, the water velocity from the oxygen ac­tivation log was substantially higher than that calculated from a traditional production log interpretation. In logging the second well the two techniques did not correlate well in the low flow/single phase interval, but there was a better relationship in the higher flow rate/multi-phase interval. In both instances there was not a consistent trend between the oxygen activation data and the traditional production log interpreta­tion. This disparity implies that the results are dependent upon the flow regime and flowing mechanics.