Fault detection by segment evaluation based on inferential statistics for asset monitoring

Abstract

Detection of unexpected events (e.g. anomalies and faults)from monitoring data is very challenging in machine healthassessment. Hence, abrupt or incipient fault detection fromthe monitoring data is very crucial to increase asset safety,availability and reliability. This paper presents a genericmethodology for abrupt and incipient fault detection andfeature fusion for health assessment of complex systems.Proposed methodology consists of feature extraction, featurefusion, segmentation and fault detection steps. First of all,different features are extracted using descriptive statistics.Secondly, based on linearly weighted data fusion algorithm,extracted features are combined to get the generic andrepresentative feature. Afterward, combined feature isdivided into homogeneous segments by sliding windowsegmentation algorithm. Finally, each segment is furtherevaluated by coefficient of variability which is used ininferential statistics, to evaluate health state changes thatindicate asset faults. To illustrate its effectiveness, themethodology is implemented on point machine and Li-ionbattery monitoring data to detect abrupt and incipient faults.The results show that proposed methodology can beeffectively used in fault detection for asset monitoring.