Real Time Measurement of Borehole Shape and Diameter by Using Ultrasonic Caliper in Different Drilling Fluid and Wellbore Conditions for Detecting Instabilities

Abstract

Abstract Drilling programs continue to push into new and more complicated environments. As a result, accurate measurements of drilling data in real time are becoming more critical by means of minimizing the risks as well as the costs. An ultrasonic caliper sensor is a key measurement for determining the borehole diameter in MWD and LWD tools. Important is the use of ultrasonic caliper tools to offer a method for calculating borehole volumes, on the final bit run, the sensor collects data while tripping out of the hole for determining borehole size and furthermore required cement volumes. Real-time applications of ultrasonic caliper measurements also strongly support the early detection of borehole instability. This paper describes the experiments related to the accuracy of the ultrasonic sensor measurements for estimation of the wellbore diameter. A fully automated test robot has been designed and tests have been performed in different fluids and geometrical conditions. That test robot allows emulating vertical as well as lateral movements of a sensor head in an artificial wellbore which can be run with different fluids. The results of the tests were compared and the weak points and problems of the sensor for detecting the echoes were determined. Numerical simulation of the ultrasonic measurements and comparison of the simulated results to the recorded data gives estimates about the accuracy dependency to different drilling conditions. Tests with different decentralized positions of the ultrasonic caliper tool inside the wellbore give estimates about the accuracy dependence with respect to the decentralization of the tool. Finally measurements were performed in wellbores with geometrical anomalies like washouts and squeezing formations. It is shown that such anomalies can be detected in an appropriate accuracy if circle fitting methods like the Kasa method in combination with robust error models are applied.