A New Production Logging Tool for Determining Holdups

Abstract

Abstract Multi component flows tend to segregate in deviated and horizontal wells, making the interpretation of their production logs difficult. Logging tools that measure only on-axis are ill suited to accurately measuring flow components that are not axially symmetric, as in Figure 1. To address this problem of flow geometry, the Fluid Profiling Tool (FPT) can sense fluids not only at the wellbore center, but at any other point as well. The tool is implemented to be compatible with the Computalog FlexStak(R) multiplexing scheme. With three conductivity sensors that can be moved to any polar position within a plane normal to the tool axis, a conductivity map of the flow stream can be generated. Shown in Figure 2, the tool is currently 1–11/16" diameter, and when possible uses an optional body centralizer of 2–1/8" collapsed diameter. The tool consists of a lower caliper section containing the sensors, and an upper sub which rotates the caliper section below. The lower section housing freely rotates within the body centralizer collars. Centralizers above the rotator are locked and immobilize tools above the caliper section. The conductivity sensors have an almost binary response, giving absolute contrast between water and hydrocarbons. For a hydrocarbon droplet or bubble in water to be detectable, its diameter must equal or exceed that of a probe tip. For this reason, a complementary measurement is run with the FPT to detect and quantify the smaller hydrocarbon inclusions. Introduction As with most types of wireline logging, Production Logging has its regimen of measurements that is accepted as typical. Casing Collar Locators and Gamma Ray tools are run for depth correlation, while Temperature and Pressure tools are run to relate downhole and surface production rates. These measurements are generally simple to perform, and the accuracy of pressure measurement has greatly improved in recent years. Other measurements, those involving fluid volume and flow measurement, remain difficult to perform with conventional tools. Because well production techniques generally commingle fluids of different specific gravities, the buoyancy of lighter components complicates the geometry of the flowing mixtures. In fact, the flows may become so chaotic that source entries for the various fluids become difficult to identify. In such conditions, fundamentally linear devices like downhole flowmeters and holdup meters may appear to be nonlinear, giving seemingly impossible holdup and flow readings. Historically, a popular approach to linearizing the multi component flow response of sensors has been to characterize or calibrate them in a flow loop. A problem with this approach, though, is that even minor differences in flow loop and field conditions can greatly affect the usefulness of correction charts and graphs. For this reason, a principal focus in FPT system development was to derive simple, analytic expressions for reducing its data. A flow loop facility was then used as a prover for the measurement system. Responses to many flow rate combinations were recorded, and two or more actual holdups were determined for each rate combination. Accuracy of the measurement system was then determined in each case, the results for which are presented in this report. FPT Measurement System Figure 3 shows a polar sampling grid used for acquiring FPT data. In the tests performed here, a maximum sampling grid of three diameters at each of three angles was used, giving a total of 27 sample points. The FPT is the bottom tool in the string, allowing relatively undisturbed flow to impinge upon the sensors. As mentioned earlier, a supplementary measurement is required to detect hydrocarbons too finely dispersed to be seen by the FPT probes. Specifically, the instrument used is a FlexStak(R) Radioactive Fluid Density (RFD) tool, shown by Figure 4. This tool is run above the FPT, separated from it by a centralizer. While an RFD tool is used as the complementary measurement, its data are not reduced in a typical fashion. Existing methods simply use averaged counts to determine downhole densities. Next, a nonlinear mapping or correction chart maps the measured density into the actual holdup. A significant part of the present development has been to derive a better method of using the data from the RFD tool, the details for which cannot yet be disclosed. P. 91^