An Examination of Thermal Features’ Relevance in the Task of Battery-Fault Detection

Abstract

Uninterruptible power supplies (UPS), represented by lead-acid batteries, play an important role in various kinds of industries. They protect industrial technologies from being damaged by dangerous interruptions of an electric power supply. Advanced UPS monitoring performed by a complex battery management system (BMS) prevents the UPS from sustaining more serious damage due to its timely and accurate battery-fault detection based on voltage metering. This technique is very advanced and precise but also very expensive on a long-term basis. This article describes an experiment applying infrared thermographic measurements during a long term monitoring and fault detection in UPS. The assumption that the battery overheat implies its damaged state is the leading factor of our experiments. They are based on real measured data on various UPS battery sets and several statistical examinations confirming the high relevancy of the thermal features with mostly over 90% detection accuracy. Such a model can be used as a supplement for lead-acid battery based UPS monitoring to ensure their higher reliability under significantly lower maintenance costs.

Ensemble Learning, Prediction and Li-Ion Cell Charging Cycle Divergence

“…Thus, PCA outlier detection is a better indicator of charging cycle divergence when normal versus full overcharging cycle divergence is considered. The amount of input current and duration of that input current is directly proportional to the amount of influence that the outlier data samples have on Si,j in (7). As mentioned earlier in Section II.B, the number of PCA outlier scores is not the only indicator of charging cycle divergence.…”

“…14 and 15, the mappings of features T2 and T3 onto principal components 2 and 1 are significantly different. Such changes in mapping although not measured in (7), can be used as secondary indictors of charging cycle divergence. Figs.…”

“…Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government. [4] or heterogeneous cell capacity, impedance or State-of-Charge (SoC) in battery stacks [5]- [7]. Failure in protection electronics or in cell balancing circuits might lead to failure in overcharge protection functions [4].…”

“…This method, however, often requires interrupting battery normal operation so that its resistance or impedance can be measured [9]. Long term monitoring of battery stacks using infrared images and voltage measurements has also been proposed as a means of assessing battery health and detecting overcharged batteries [7]. This method, however, requires thermal imaging of the batteries, which might be very hard to achieve in commercial batteries assembled inside of compact modules.…”

Special Issue on ‘Novel Ideas for Infrared Thermography and Its Application to Integrated Approaches’

“…The assumption that the battery overheats implied its damaged state was the leading factor of the experiments. They were based on real measured data on various UPS battery sets and several statistical examinations confirming the great relevance of thermal features with mostly over 90% detection accuracy [33].…”