Using deep learning and an annular triboelectric sensor for monitoring downhole motor rotor faults

Abstract

The rotor, as one of the key components of a downhole motor, directly affects the safety, cost, and efficiency of the entire drilling operation. This paper proposes an annular triboelectric sensor (ATES) for monitoring rotor faults in downhole motors, marking an innovative application of triboelectric nanogenerators in the field of downhole fault monitoring. The ATES is characterized by its simple structure, long lifespan, and high-temperature resistance, making it particularly suitable for the complex conditions of downhole environments. The ATES can also monitor radial vibrations of downhole tools in real time and, when combined with the ResNet-18 algorithm, can accurately identify rotor imbalances, misalignments, and rubbing faults, achieving a classification accuracy of up to 100 %. Additionally, this paper presents an intelligent offline analysis system for downhole rotor fault diagnosis, which integrates deep learning and visualization techniques. This system efficiently identifies rotor faults and outputs visual results, providing drillers with intuitive diagnostic references, thereby significantly improving the efficiency and accuracy of fault diagnosis. Overall, the ATES offers a viable pathway for developing new downhole intelligent sensing devices and technologies.