Development of an Accurate State of Health Estimation Technique for Lithium-Ion Batteries

Development of an Accurate State of Health Estimation Technique for Lithium-Ion Batteries
The operational requirements that batteries have to satisfy in automotive and stationary applications are more demanding. Therefore, effective control and management of batteries is crucial for assuring the best battery performance, considering using the batteries in the safest and longest living way. Despite the fact that the diagnosis and prognosis of the state of health (SoH) are absolutely needed in practical applications, they are not yet effectively implemented. Extensive work is still required to develop an accurate and applicable diagnosis method.

In this PhD research, a thorough and extensive state-of-the-art study regarding SoH estimation techniques has been performed. The literature research highlighted the need for an accurate and implementable SoH algorithm, as well as the requirements that the algorithm needs to fulfil. According to these demands, this PhD dissertations developed different SoH estimation algorithms for two types of lithium-ion chemistries: Graphite based anode / Nickel Manganese Cobalt Oxide cathode and Graphite based anode / Lithium Iron Phosphate cathode type of batteries. All algorithms can be implemented in a battery management system , and are able to detect the SoH at cell-, module- and battery pack level. Precise and accurate estimations are obtained through partial charges or discharges of the batteries, with no interference with the application’s functionality.

Related to the SoH estimation, the path dependency is of great importance for the detection of degradation mechanisms. Path dependence is emerging as a key issue for how battery cells age in different conditions. Even though it is not easy to handle path dependence due to the numerous features that can vary in a cell, a novel and precise diagnosis model has been developed during this PhD dissertation. This online- and implementable diagnosis model was validated with different experimental tests. As future work this model will give the chance of developing thorough research regarding battery ageing, SoH prediction and even battery life extension.