Geosteering Improvement With A Telemetry Model
Andreas Hartmann, Oleg Akimov, Christian Fulda, Stephen Morris · SPE Middle East Oil and Gas Show and Conference · 2011
Abstract Introduction of new measurements constantly increases the amount of real-time data from logging-while-drilling (LWD) services. This is matched by telemetry technologies that provide data rates from bits to Mbits per seconds. During a deployment, the bandwidth is shared between several services requiring different data rates, leaving the user with the task to align telemetry technology and shared data rates to the application. High-volume services often allow for data compression. Lossless or near lossless compression provides best data quality but makes a service greedy for bandwidth. Lossy compression drastically reduces bandwidth usage, but may result in degraded data quality. Thus, planning the telemetry must be closely tied to the service objectives and expert advice is required. We present a planning approach where we model the real-time data compression of an LWD high-resolution imaging tool that uses a flexible compression algorithm. This system allows lossy compression for telemetry rates as low as 1 bit/s, while at the same time can deliver memory data quality using high-speed telemetry technology. The compression rate determines the image resolution and must be adjusted to the detail required in the transmitted image. Using offset well data or simulated memory data, the expected real-time image is simulated and its level of detail quantified. For instance, the presence and appearance of fractures in real time can be forecasted. This enables us to optimize telemetry usage and define compression and drilling parameters for a successful deployment of the imaging service. The paper will outline the methods and benchmark the technology using real-time data measured during field deployments. We will then use parameter variations to show how increasing net bandwidth improves the amount of detail in the images and how these are used for different levels of real-time data interpretation.