A fault-tolerant FL-IIMM method for near-bit attitude signal extraction under sensor faults

In practical, the drilling tool will inevitably subjected to the fault or damage of the sensors after working for a long time due to disturbances under the downhole hostile environments. Aiming at this problem, this paper proposes a fault-tolerant method for near-bit attitude signal extraction, fuzzy logic combining with improved interactive multiple model (FL-IIMM), which fully utilizes the redundant measurement structure of multiple measurement short sections. Firstly, in the single measurement short sections, fuzzy logic (FL) is utilized to perform fault-tolerant fusion on the internal results of sensors to address sensors faults by setting fault weights and establishing a fuzzy inference system. Then in multiple measurement short sections, when there are sensor faults in multiple short sections or even when sensors in one section all have faults, the improved IMM algorithm (IIMM) is employed for fusion among different sections to eventually achieve fault-tolerant fusion of multiple measurement short sections by establishing models for each short section and adapting the likelihood function at the same time to adaptively update the model probabilities. Simulation experiments with sensor damage of harsh downhole environments scenarios show the fault-tolerant extraction accuracy of our proposed method with the inclination angle ±0.0163°, the toolface angle ±3.0993°, and the azimuth angle ±1.8774°.