Fault detection in satellite power system using convolutional neural network
- PDF / 1,030,580 Bytes
- 7 Pages / 595.276 x 790.866 pts Page_size
- 6 Downloads / 282 Views
Fault detection in satellite power system using convolutional neural network M Ganesan1
· R Lavanya1 · M Nirmala Devi1
Accepted: 18 September 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Satellite failures account for heavy, irreparable damages, especially when associated with the Power System which is the heart of a satellite. Anomalies in Satellite Power System (SPS) can lead to complete failure of the mission. This demands the need to understand the causes of power system related failures. Huge number of sensors installed in a satellite system conveys information regarding the health of the system. The conventional manual level checking of sensors can be augmented with data driven fault diagnosis approach to reduce the false alarm and burden on operating personnel. The latter has the advantage of exploiting the interrelationship between sensor measurements for fault diagnosis. In this work, Convolutional Neural Network (CNN) is trained on satellite telemetry data for sensor fault detection in SPS. Various processing schemes in time and frequency domains were explored to process the input data to CNN. Promising results were obtained with combination of Stockwell transform (S-transform) and CNN for data processing and classification, respectively. Advanced Diagnostics and Prognostics Testbed (ADAPT), a publicly-available dataset was analysed and used for validating the proposed algorithm, yielding an accuracy as high as 96.7%, precisison of 0.9, F1 score of 0.95 and AUC equal to 0.976. Keywords Satellite power system · Fault diagnosis · Stockwell transform · Convolutional neural network
1 Introduction Satellite in-orbit failures cause huge and irrecoverable losses if not detected early. Nearly half of the failures of satellites in orbit are power-system related [1]. Degradation of sensors and solar arrays can drastically affect the satellite system, triggering satellite shutdown and unavailability of critical services. The health of the satellite system is typically checked frequently using housekeeping telemetry data, in order to detect any failure or malfunctioning of sensors and subsystems. The general approach used for failure detection is limit checking, where the satellite operators manually check whether the sen-
B
M Ganesan [email protected] R Lavanya [email protected] M Nirmala Devi [email protected]
1
Department of Electronics and Communication Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India
sor measurements are within the predefined ranges. However, it is a difficult task to know the limit values appropriately for all the sensors in advance, especially in modern day satellite systems with increasing number of sensors. Satellite telemetry data involves multiple sensory measurements, each represented as a time stamped discrete series. In general, time series data can be classified into two types: univariate and multivariate; the former is a time-ordered measurement of a single parameter, while the la
Data Loading...