Sensor Fault Diagnostics Using Physics-Informed Transfer Learning Framework

The field of smart health monitoring, intelligent fault detection and diagnosis is expanding dramatically in order to maintain successful operation in many engineering applications. Considering possible fault scenarios that can occur in a system, indicating the type of fault in a sensor is one of th...

Full description

Bibliographic Details
Main Authors: Chen, Y. (Author), Guc, F. (Author)
Format: Article
Language:English
Published: MDPI 2022
Subjects:
Online Access:View Fulltext in Publisher
LEADER 02622nam a2200373Ia 4500
001 0.3390-s22082913
008 220421s2022 CNT 000 0 und d
020 |a 14248220 (ISSN) 
245 1 0 |a Sensor Fault Diagnostics Using Physics-Informed Transfer Learning Framework 
260 0 |b MDPI  |c 2022 
856 |z View Fulltext in Publisher  |u https://doi.org/10.3390/s22082913 
520 3 |a The field of smart health monitoring, intelligent fault detection and diagnosis is expanding dramatically in order to maintain successful operation in many engineering applications. Considering possible fault scenarios that can occur in a system, indicating the type of fault in a sensor is one of the most important and challenging problems in the area of intelligent sensor fault diagnostics. Within this frame of reference, we extended the physics-informed transfer learning framework, first presented previously for a fault cause assignment, to the level of sensor fault diagnostics for a range of different fault scenarios. Hence, the framework is utilized to perform intelligent sensor fault diagnostics for the first time. The underlying dynamics of the reference system are extracted using a completely data-driven methodology and dynamic mode decomposition with control (DMDc) in order to generate time-frequency illustrations of each sample with continuous wavelet transform (CWT). Then, sensor fault diagnostics for bias, drift over time, sine disturbance and increased noise sensor fault scenarios are achieved using the idea of transfer learning with a pre-trained image classification algorithm. The classification results yields a good performance on sensor fault diagnostics with 91.5% training and 84.7% test accuracy along with a fair robustness level with a set of reference benchmark system parameters. © 2022 by the authors. Licensee MDPI, Basel, Switzerland. 
650 0 4 |a Benchmarking 
650 0 4 |a Data-driven approach 
650 0 4 |a data-driven approaches 
650 0 4 |a dynamic mode decomposition with control 
650 0 4 |a Dynamic mode decomposition with control 
650 0 4 |a Dynamic mode decompositions 
650 0 4 |a Dynamics 
650 0 4 |a Fault detection 
650 0 4 |a fault diagnostics 
650 0 4 |a Fault scenarios 
650 0 4 |a Faults diagnostics 
650 0 4 |a Health monitoring 
650 0 4 |a Intelligent control 
650 0 4 |a Intelligent sensors 
650 0 4 |a Learning frameworks 
650 0 4 |a Sensors faults 
650 0 4 |a transfer learning 
650 0 4 |a Transfer learning 
650 0 4 |a Wavelet decomposition 
700 1 0 |a Chen, Y.  |e author 
700 1 0 |a Guc, F.  |e author 
773 |t Sensors