Multiple-Sensor Fault Isolation Based on Spatial Distances for Accelerometer System in Drilling Tools

Multiple-sensor fault isolation (FI) is essential for the reliable and efficient operation of the accelerometer system with bounded uncertainties in drilling tools. However, existing methods for the system mainly concern a single fault and are of poor performance for multiple faults. In this article, a sensor error system is derived with unknown time-varying uncertainty bounds. An adaptive Zonotopic Kalman Filter (ZKF) is developed to estimate fault directions. Then, based on spatial distances, an FI logic is proposed and the fault isolability is analyzed to determine faulty sensors and fault types. Finally, the effectiveness of the proposed method is demonstrated by simulation and experiment results. Compared with existing methods, the proposed method isolates more faulty sensors and performs better in state estimation and FI.