Horizontal well production logging (PL) measurements, especially center-sample devices, have been misleading in evaluation of the flowing stream. These inaccuracies are due to the natural segregation of the fluids owing to the difference in the phase densities. A new multiphase holdup tool and interpretation method has been developed that provides accurate determination of holdups and flowrates in horizontal wells. The multiphase holdup tool uses 12 capacitance-sensing circuits with multiple arm arrangements to provide an excellent image of the holdup pattern. Since these sensors are on the same cross-sectional plane, depth inconsistencies are not a factor in the interpretation. Each sensor responds to the permittivity (dielectric constant) of the surrounding fluids, and the response can be converted to a phase holdup by applying the known sensor calibration. Consideration of each sensor's position relative to the wellbore allows for prediction of the total phase holdup across the entire wellbore cross sectional plane. Advanced analysis packages allow the user to interact with the computed holdups, providing an easy method of interpretation. Different views of the wellbore, cross-sectional displays, and displays incorporating the wellbore trajectory allow accurate and detailed analysis that is critical for understanding flow in horizontal wells. Once the holdup of each phase (gas, oil, and water) is determined, an additional interpretation package allows a complete production log analysis including flowrates in a very difficult environment. The results of these analysis packages allow an operator to understand, modify, and improve the productivity of a horizontal well. Tool Design and Operations Phase separation makes it extremely difficult to get a good view of holdup in horizontal wells by using standard center-sampling tools. The Capacitance Array Tool (CAT™) solves this problem by using a circular array of 12 micro-capacitance sensors. They are radially distributed in the wellbore to measure accurately phase holdups. Fig. 1 is a picture of the tool showing the sensors with the arms in the out position. These cylindrical sensors have a diameter of 0.157 inches and are 0.055 inches long located 0.35 inches from the end of the motorized arm as shown in Fig. 2 and Fig. 3. Each sensor forms part of a circuit that resonates at different frequencies in gas, oil, and water. This variation in resonance allows the tool to determine what phase exists at a given region across the wellbore. Responding to the capacitance value around the probe, the oscillator circuits produce a low frequency in water, high frequency in oil, and a higher frequency in gas. Sensor frequencies are typically sampled 72 times per second (dependent upon the telemetry) and relayed to surface where they can be processed for presentation.1 The sensors have a radius of investigation of about 0.01 inches. Sensors are electrically isolated from other CAT components so that they only register capacitance values from their immediate environment. These micro-capacitance sensors are expected to measure the individual segregated phases (water, oil, or gas) that are assumed to surround them. However, due to the sensor and calibration response the holdup mixtures of any two phases can be accurately determined. Therefore, each of the 12 sensors will measure gas, oil, or water, or a mixture of two phases depending upon software parameter selection. The CAT's geometry makes it especially well suited to measuring holdup near the top and bottom of any cross section normal to the axis of the wellbore since all sensors are in this plane. The array provides full coverage across the wellbore, making it possible to accurately identify fluid in horizontal or highly deviated wells as shown in Fig. 2.

Subsurface analysis-driven field development requires quality data as input into analysis, modelling, and planning. In the case of many conventional reservoirs, pay intervals are often well consolidated and maintain integrity under drilling and geological stresses providing an ideal logging environment. Consequently, editing well logs is often overlooked or dismissed entirely. Petrophysical analysis however is not always constrained to conventional pay intervals. When developing an unconventional reservoir, pay sections may be comprised of shales. The requirement for edited and quality checked logs becomes crucial to accurately assess storage volumes in place. Edited curves can also serve as inputs to engineering studies, geological and geophysical models, reservoir evaluation, and many machine learning models employed today. As an example, hydraulic fracturing model inputs may span over adjacent shale beds around a target reservoir, which are frequently washed out. These washed out sections may seriously impact logging measurements of interest, such as bulk density and acoustic compressional slowness, which are used to generate elastic properties and compute geomechanical curves. Two classifications of machine learning algorithms for identifying outliers and poor-quality data due to bad hole conditions are discussed: supervised and unsupervised learning. The first allows the expert to train a model from existing and categorized data, whereas unsupervised learning algorithms learn from a collection of unlabeled data. Each classification type has distinct advantages and disadvantages. Identifying outliers and conditioning well logs prior to a petrophysical analysis or machine learning model can be a time-consuming and laborious process, especially when large multi-well datasets are considered. In this study, a new supervised learning algorithm is presented that utilizes multiple-linear regression analysis to repair well log data in an iterative and automated routine. This technique allows outliers to be identified and repaired whilst improving the efficiency of the log data editing process without compromising accuracy. The algorithm uses sophisticated logic and curve predictions derived via multiple linear regression in order to systematically repair various well logs. A clear improvement in efficiency is observed when the algorithm is compared to other currently used methods. These include manual processing by a petrophysicist and unsupervised outlier detection methods. The algorithm can also be leveraged over multiple wells to produce more generalized predictions. Through a platform created to quickly identify and repair invalid log data, the results are controlled through input and supervision by the user. This methodology is not a direct replacement of an expert interpreter, but complementary by allowing the petrophysicist to leverage computing power, improve consistency, reduce error and improve turnaround time.

The techniques used to process open-and cased-hole image logging and other oriented tools that are run in offshore exploration and development wells require highly accurate navigation logs. Navigation packages usually consist of triaxial accelerometers and magnetometers that require calibration of the offset (bias), gain, and tool axis alignment for each sensor. Despite pre-or postacquisition surface calibrations, the downhole environment will alter calibrations of offset and gain as a result of many factors, such as temperature and electrical noise. These factors are impractical to characterize or predict, and it is impossible to completely isolate the navigation package from these factors. Experience shows that the offsets and gains of the sensors may change appreciably over time scales much shorter than the total logging time.This paper provides four detailed examples of possible navigation logs with various problems to develop a methodology for evaluating their quality and proposing corrective action. The goals of this study are to determine which navigation sensors to adjust or which sensors are not responding to the appropriate Earth field and need reconstruction. The paper proposes several quality measures for identifying, from the navigation log, drift in sensor response, including measured total field variations and correlations of tool deviation to tool rotation.When the methods described in this paper are used to align images with respect to a geographical reference (in cases where the navigation was affected by sensor failure or external factors, e.g., permanently magnetized formation layers, that alter the sensor calibrations), they significantly enhance the navigation quality. For example, image logs that originally wobbled erratically in ferrous formations can be oriented to the north allowing dipping-bed orientation to be measured. In addition, magnetometer reconstruction and accelerometer correction can easily manage oriented navigation in metal casing. In general, the algorithms developed in this study assume that a rough surface calibration is available to obtain an initial solution, but time is no longer required for meticulous wellsite surface calibration.