H.R. Seren, M. Deffenbaugh
Aramco Americas,
United States
Keywords: autonomous vehicles, harsh environments, navigation, location, magnetic sensors
Summary:
Automation in oil and gas operations became a prime target to increase efficiency and safety. Various efforts have been put forward to create autonomous downhole logging and pipe inspection robots. Navigation of these robots remains as one of the challenges as electromagnetic waves are highly attenuated within steel tubulars. Therefore, new ways to measure position are needed. We developed a new position sensing method based on 1-D feature matching of residual magnetic fields generated by the steel casings. The variation in the residual magnetic field along the length of steel tubulars provides recognizable landmarks. The developed method employs two magnetometers with a known separation along the direction of the motion. A correlation algorithm compares the magnetic logs from the two sensors and performs feature matching. Matched features indicate that the robot traveled a distance corresponding to the distance between the sensors. This provides an average speed of the robot based on the feature size, distance between the sensors, and the time of travel of the feature between the two sensors. By repeating the algorithm continuously, an average speed profile as a function of time can be found. Traveled distance is calculated by integrating the speed. This approach has an advantage over using accelerometers since acceleration needs to be integrated twice and it’s more prone to accumulated error. The method was demonstrated first by collecting residual magnetic field log from a steel cased well using a wireline tool. The collected signal was shifted in time to mimic a second magnetometer. The matching algorithm was used to find the time between matched feature in the two signals along the full log which provided speed of the tool. The speed was used to calculate the position. The data from this sensing method can be integrated to a sensor fusion algorithm along with data from other sensors such as an accelerometer, pressure sensor, or casing collar locator. Direct position sensing (without integrating speed) is possible by using an array of magnetometers that can capture a whole feature instantaneously which provides a time independent location sensing.