Temperature monitoring and diagnostic system for a forced-cooled electrochemical energy storage string

Abstract

Operating temperature is one of the most critical factors affecting performance and life expectancy of a device. Hence, temperature monitoring and management has been an important area widely researched in various applications, including electrochemical energy storage devices such as batteries and supercapacitors (SCs). In an energy storage string or module consisting of a number of cells, a significant variation in temperature distribution could exist. However, monitoring the whole module temperature is often hindered by hardware and cost limitations, and typically, only a limited number of temperature sensors are employed.

This thesis proposes a model-based temperature monitoring and diagnostic system for a forced-cooled electrochemical energy storage string using a limited number of sensors. While the lumped thermal model for a single cylindrical battery or SC cell has been covered in literature, the multicell model has not been investigated in detail or experimentally validated. Hence, the lumped multicell thermal model of an energy storage string was developed and experimentally validated. The multicell model was subsequently used to study, compare, and assess different observability criteria based on observability Gramian in order to find optimal placement of a limited number of sensors for temperature monitoring.

A multiple model estimator (MME) is also proposed as a viable solution to monitor temperature of all cells as well as to detect and localise an abnormally overheating cell, with the limited number of temperature sensors. Performance of the proposed MME is analysed and experimentally evaluated using an SC string setup.

Lastly, the disturbance detectability criteria based on cross-Gramian matrix were used to analyse optimal placement of a limited number of sensors for the MME-based abnormal overheating diagnostic system. Performance evaluation of the estimation and detection deduced the suitability of particular observability and abnormal overheating detectability criteria for sensor placement in the string.