Detection of Gross Errors in Wellbore Directional Surveying for Petroleum Production with emphasis on Reliability Analyses

Abstract

This paper presents a method for the detection of gross errors in wellbore directional survey data. The method is general, and is applicable for both magnetic and gyroscopic surveys. However, this study will only examine magnetic surveys. The method is based on statistical tests and can be applied both in the single station estimation case or in connection with multi-survey processing techniques. In contrast to other methods for error detection commonly used in the petroleum industry, which compare the measured gravity field strength, magnetic field strength and dip angle with values predicted from independent sources, this test is capable of detecting gross errors in any single sensor reading or reference component. By rejecting the corrupted measurements only, the new method brings an improvement to the quality of inclination and magnetic azimuth estimates. The validity of such tests is shown to be dependent on several factors: Wellbore geometry, noise level in the measurements, estimation techniques, statistical significance and power. Based on synthetic data this is demonstrated by several examples to address two issues: 1. What is the critical error value, for each particular measurement, that may pass the test? 2. How much do potential undetected outliers affect the inclination and azimuth estimates? These methods can detect errors that are 4 to 5 times greater than the nominal measurement noise levels. They are shown to be sharpest in connection with multi-survey processing techniques. Key words: Wellbore surveying, Gross errors, Hypothesis testing, Data-snooping, Reliability analyses, Multi-station estimation, Single-station estimation.

Detection of Gross Errors in Wellbore Directional Surveying for Petroleum Production with emphasis on Reliability Analyses