A General Framework for Aliasing Corrections of Inductive Oil Debris Detection Based on Artificial Neural Networks

Wei Hong, Tongyang Li*, Shaoping Wang and Zebing Zhou*

Induction-based oil debris detection methods have shown a great potential for providing non-invasive monitoring and measurement to prolong the life of precise machinery. However, the superimposition of the induced voltages by the multiple debris particles prevents these methods from being more accurate. An artificial neural network is employed in this work to establish a general corrective framework aiming at solving the modeling and adaptability problems for different sensors, and a hybrid detection strategy is proposed to further reduce the detecting error under different aliasing conditions. A simulative test using two sinusoidal waveforms is conducted to validate the performance of the proposed method. Finally, an experiment is carried out with known oil debris concentrations ranging from 5 mg/L to 100 mg/L, and the linearity of the detected signals under the given concentrations is used to evaluate the performances of different methods. The results indicate that the maximum error by the proposed measurements is less than 20%, while for the non-corrected measurements, the maximum error is over 40%.

期刊名：IEEE SENSORS JOURNAL

期/卷：VOL. 20, NO. 18

页码：10724 - 10732

发表时间：2020年9月

DOI: https://10.1109/JSEN.2020.2994458