Real-time Transmission of High Resolution Images

Abstract

Abstract Recent advances in high-resolution imaging are well perceived by the oilfield industry as an aid to refine reservoir characterization. The availability of high resolution image data in real-time allows for improved decision making and more optimal wellbore placement when navigating through the reservoir. However, the transmission of high resolution image data to surface is limited by the bandwidth offered by mud pulse telemetry systems. This paper describes a new concept for transmitting imaging data to surface in real-time. The concept includes a highly flexible and effective compression scheme that can be used for all currently available and future telemetry systems. Image transmission parameters can be modified via downlink commands at any time without interruption of the drilling process. In this way, image parameters can be optimally adjusted to suit the available telemetry rates and proportion of the bandwith dedicated to image transmission. This flexibility allows customization of the real-time imaging service to optimize the geosteering process and the ability to provide high resolution data over specific zones of interest. The relationship of pixel size, telemetry rate and the real-time image parameters of latency and redundancy are discussed with reference to real-time resistivity images from field measurements transmitted at different telemetry rates and the corresponding memory images. Introduction Wireline, featuring a multi-phase electrical communication line, provides a real-time connection between the downhole sensors and the surface system. Despite the immediate availability of data at surface, measurements are made long after the formation has been drilled. As a consequence, all major oilfield service companies invested heavily in Logging-While-Drilling (LWD) technology, aiming to measure formation parameters shortly after drilling with the rock in as close to pristine condition as possible. The immediacy of the measurement combined with the availability of downhole data at surface in real-time revolutionized wellbore placement. While bulk measurements still represent the vast majority of LWD logging data, a variety of oriented geophysical measurements such as gamma, density, caliper and resistivity data are available, allowing refined structural reservoir analysis. The visualization of such data via image plots simplifies interpretation. The amount of information available increases with image resolution, which is limited by the intrinsic resolution of the geophysical sensor. High resolution resistivity sensors (e.g. [Ritter et al., 2005]) are providing the highest image definition in industry with an effective pixel size of ¼″ × ¼″ in memory. Such highly resolved images are a valuable aid for a wide variety of applications including:–structural & fracture system analysis–thin-bed analysis to determine net pay thickness–sedimentary feature analysis for input to a depositional environment model–core depth calibration, core orientation and an alternative to conventional coring over long intervals. However, the exploitation of LWD data is currently limited by the slow communication through the mud column. The introduction of azimuthal geophysical measurements with the associated exponential increase in data potentially available further increases the gap between achievable and required data rates. While the need for additional downhole memory space can be resolved by introducing more memory space, recent increases in telemetry speeds can not keep up with the required channel bandwidths. For example, a high resolution azimuthal logging services can provide more than one hundred sectored measurements around the borehole. In memory each of these values is represented by 8 bits. A 500 ms acquisition cycle results in at least 2 kbps required to transmit such uncompressed image data sets to surface. Even sophisticated telemetry systems using the drilling mud as the communication medium allow communication speeds of only a few bits per second. Thus, data reduction and compression techniques (e.g. [Li et al., 2001], [Li and Wang, 2005]) are required suited to decrease the amount of downhole data by the magnitude of 10[3].