Module-Level Modelling Approach for Li-Ion Batteries: a Cloud-based Digital Twin Simulation Platform
The adoption of large-scale Lithium-ion Batteries (LIBs) has been growing steadily and evolving. These installations involve the interconnection of multiple batteries to form larger and more powerful systems capable of providing megawatt-hours (MWh) of stored energy. LIBs have emerged as a promising...
| Idioma: | English |
|---|---|
| Publicación: |
Mondragon Unibertsitatea. Goi Eskola Politeknikoa
2024
|
| Materia: | |
| Acceso electrónico SOLREBILTEGIA: |
https://hdl.handle.net/20.500.11984/6350 https://doi.org/10.48764/argh-0r92 https://katalogoa.mondragon.edu/janium-bin/janium_login_opac.pl?find&ficha_no=176406 |
| Resumen: |
The adoption of large-scale Lithium-ion Batteries (LIBs) has been growing steadily and evolving. These installations involve the interconnection of multiple batteries to form larger and more powerful systems capable of providing megawatt-hours (MWh) of stored energy. LIBs have emerged as a promising solution for electrical energy storage due to their decreasing prices and improved manufacturing efficiency. This combination has made LIBs more accessible, and their demand has rapidly increased in key applications such as electric vehicles and stationary applications.
In the context of LIBs, specifically in the case of modules, individual heterogeneities and imbalances among the different cells that compose the module pose a significant technological challenge. In fact, these disparities can compromise the energy efficiency and overall lifespan of the battery module. While numerous studies have been conducted on individual cells, there is a significant gap in understanding and adequately considering the effects and complexities at the module level.
In this thesis, an innovative methodology is proposed to develop module-level battery models that include thermal and electrical components, as well as a State of Charge (SoC) estimator. These module-level models are based on equivalent circuits extrapolated from widely-used cell-level models. A detailed thermal model is proposed to capture the interactions between each cell within the battery system, and an electrical model is developed to simulate the behavior of individual cells through co-simulation or parallel execution. Additionally, an approach to implement these models in a cloud-based simulation platform is presented, enabling estimations of each cell’s performance, identification of potential issues, and providing sufficient computational capacity.
The proposed methodology has been validated at laboratory level by means of a prototype specifically built for this purpose. The correct operation of the thermal and electrical model and the SoC estimator at the cell level has been demonstrated by means of a series of laboratory tests. These models have then been adapted at module level, taking into account the electrical and geometrical characteristics of the module. By means of a series of laboratory tests carried out on the module prototype, the correct extrapolation of the cell models to the module level has been demonstrated. In addition, and with the aim of evaluating the heterogeneity and imbalance detection capacity of the developed models, two case studies have been conducted. In them, certain anomalies have been introduced in the laboratory prototype, and it has been proved that the models exhibit these functionalities.
In particular, two types of anomalies have been introduced: a) the first one consists of a voltage unbalance between the cells of the module and b) the second one consists of a thermal unbalance in the module by means of a thermal blanket. In both case studies, the ability to detect irregularities in the module has been demonstrated. The proposed methodology has been validated at the laboratory level using a specifically designed prototype. The correct operation of the thermal and electrical models, as well as the SoC estimator at the cell level, has been demonstrated through a series of laboratory tests. Subsequently, these models have been adapted to the module level, taking into account the corresponding electrical and geometric characteristics. Through a series of laboratory tests conducted on the module prototype, the proper extrapolation of cell-level models to the module level has been demonstrated. Furthermore, two case studies have been conducted to evaluate the capability of the developed models to detect heterogeneities and imbalances. These case studies involved the introduction of anomalies in the laboratory prototype, such as voltage imbalances between module cells and thermal imbalances using a thermal blanket. In both cases, the models showed the ability to detect irregularities in the module.
In general, the methodology proposed in this thesis allows to have a holistic model of a LIB at module level, which represents the electrical and thermal behaviour of each of the cells that compose the module, thus contributes to a better understanding allowing an adequate monitoring of the system. |
|---|