Creating SoC algorithms for Li-ion batteries based on neural networks requires a large amount of training data, since it is necessary to test the batteries under different conditions so that the algorithm learns the relationship between the different inputs and the output. Obtaining such data through laboratory tests is costly and time consuming; therefore, in this article, a neural network has been trained with data generated synthetically using electrochemical models. These models allow us to obtain relevant data related to different conditions at a minimum cost over a short period of time. By means of the different training rounds carried out using these data, it has been studied how the different hyperparameters affect the behaviour of the algorithm, creating a robust and accurate algorithm. To adapt this approach to new battery references or chemistries, transfer learning techniques can be employed.

Getting accurate lifetime predictions for a particular cell chemistry remains a challenging process, largely dependent on time and cost-intensive experimental battery testing. This paper proposes a transfer learning (TL) method to develop LIB ageing models, which allow for the leveraging of experimental laboratory testing data previously obtained for a different cell technology. The TL method is implemented through Neural Networks models, using LiNiMnCoO₂/C laboratory ageing data as a baseline model. The obtained TL model achieves an 1.01% overall error for a broad range of operating conditions, using for retraining only two experimental ageing tests of LiFePO₄/C cells.

Conventional Li-ion battery ageing models require a significant amount of time and experimental resources to provide accurate predictions under realistic operating conditions. Furthermore, there is still an uncertainty on the validity of purely laboratory data-based ageing models for the accurate ageing prediction of battery systems deployed in field.At the same time, there is significant interest from industry in the introduction of new data collection telemetry technology. This implies the forthcoming availability of a significant amount of in-field battery operation data. In this context, the development of ageing models able to learn from in-field battery operation data is an interesting solution to mitigate the need for exhaustive laboratory testing, reduce the development cost of ageing models and at the same time ensure the validity of the model for prediction under real operating conditions.In this paper, a holistic data-driven ageing model developed under the Gaussian Process framework is validated with experimental battery ageing data. Both calendar and cycle ageing are considered, to predict the capacity loss within real EV driving scenarios. The model can learn from the driving data progressively observed, improving continuously its performances and providing more accurate and confident predictions.

Background The increasing integration of renewable energy sources (RES) in power systems introduces operational challenges due to their intermittent and difficult-to-predict generation. Hybrid Energy Storage Systems (HESS) can mitigate these issues by providing flexibility and stability to microgrids. However, efficient coordination between storage systems, renewable generation, and the utility grid requires advanced Energy Management Systems (EMS). This work presents a hierarchical EMS developed within the HAVEN project to optimize energy flow, reduce operational costs, and extend battery lifetime in microgrid environments. Methods The proposed EMS follows a three-layer hierarchical architecture. The High-Level EMS, deployed in the cloud, uses artificial intelligence algorithms to forecast renewable generation, load demand, and battery Remaining Useful Life (RUL). The Medium-Level EMS performs optimal energy management within the microgrid through a two-layer control strategy that separates long-term planning and short-term adjustments. It determines optimal power setpoints for the grid and the HESS, which consists of two lithium-ion batteries with different characteristics (high-energy and high-power). The Low-Level EMS, embedded in power converters through edge devices, supervises individual energy storage systems, monitors system states, and ensures the execution of the setpoints while maintaining converter stability. The system performance is evaluated through simulation using realistic seasonal profiles of load demand, RES generation, and time-varying electricity tariffs. Results Simulation results demonstrate that the EMS effectively coordinates renewable generation, grid interaction, and battery operation under both summer and winter conditions. The high-power battery compensates short-term fluctuations between forecasted and real profiles, while the high-energy battery manages longer-term energy balancing. Annual extrapolation of the results shows a reduction of 31.7% in electricity costs and grid energy imports, as well as the complete elimination of renewable energy curtailment. Conclusions The proposed hierarchical EMS improves the operational efficiency and economic performance of microgrids integrating hybrid energy storage systems. Its modular multi-layer design enables scalable deployment and efficient coordination between forecasting, optimization, and real-time control. The approach demonstrates significant potential for supporting renewable integration, reducing energy costs, and improving battery lifetime in microgrid applications.