This article presents a double-layer-integrated charging equalizer based on magnetic coupling to address battery module imbalances. The proposed architecture consists of a main transformer and parallel transformers. The main transformer ensures the overall module balance, while the parallel transformers achieve individual cell balancing. The system eliminates the need for additional magnetic components and operates independently. The experimental results on a nine-cell (three modules) setup demonstrate its reduced voltage variation during charging mode, with an efficiency of 90.1% at 35-W output. The proposed solution offers an effective and practical approach to address battery module imbalances in a concise manner.
25 December 2023 in IEEE Journal of Emerging and Selected Topics in Power Electronics
Using a multi-winding transformer to drive voltage multipliers (VM) for balancing ultra-capacitors (UC) string can improve the scalability of the equalizer. However, each winding of the multi-winding transformer differs in the ratio of turns, leakage inductance, and bypass resistance, resulting in an imbalance between UC modules. Therefore, a modular equalization system based on double-layer voltage multipliers for series-connected ultra-capacitors string is proposed, which uses VMs to balance the UC cells and modules, respectively. The converter’s energy transfer inductor acts as the voltage equalizer’s driver. An experimental charging test for two UC modules was performed. Cell and module voltages gradually became consistent.
This thesis proposes a charging equalizer for hybrid energy storage systems, incorporating the Zeta converter, voltage multiplier, and multi‐winding transformer. The objective is efficient charging and equalization for battery and supercapacitor strings. The Zeta converter serves to charge the battery string, while the transformer with multi‐winding shares the inductor of Zeta converter as the primary side, facilitating battery string balancing through the utilization of current ripple in the inductor. Additionally, the voltage ripple generated by the Zeta converter is utilized to drive the voltage multiplier, enabling the charging and equalization of the supercapacitor string. The system's operation halts upon reaching the predefined threshold voltage for the supercapacitor strings. The proposed circuit transforms into a Zeta converter by employing a single‐pole double‐throw switch, facilitating the charging of the battery strings. Experimental validation uses four series‐connected supercapacitors and four series‐connected batteries, demonstrating successful equalization and charging for both energy storage components.
Groundwater level forecasting is essential for groundwater monitoring and sustainable water resource management in urban regions. This study proposes a graph-based spatiotemporal forecasting framework for regional multi-well groundwater prediction using a Graph Convolutional Network Long Short-Term Memory (GCN-LSTM) model. Daily sensor-derived observations from 56 monitoring wells in Chengdu, China, collected from 2019 to 2023, are used to construct a graph representing inter-well spatial relationships. The GCN component captures spatial dependencies among wells, while the LSTM component models temporal variations in groundwater levels. Two data-availability scenarios are evaluated: a reduced-input setting and a multi-factor setting incorporating rainfall and water temperature. Results demonstrate that the proposed framework provides reliable groundwater level forecasts across multiple wells under both scenarios, and the inclusion of additional temporal variables improves the prediction of groundwater fluctuations. A network-wide controlled comparison under the S+W setting further shows that GCN-LSTM outperforms the conventional LSTM baseline across the 56-well monitoring network, supporting the contribution of inter-well spatial modeling to regional groundwater level forecasting. The findings highlight the potential of graph-based deep learning for regional groundwater forecasting and intelligent sensor-based monitoring networks.
Accurate groundwater level forecasting is essential for groundwater monitoring, risk assessment, and sustainable water resource management in rapidly urbanizing regions. This study proposes a graph-based spatiotemporal forecasting framework for regional multi-well groundwater level prediction using a Graph Convolutional Network–Long Short-Term Memory (GCN–LSTM) model. Daily monitoring data from 56 state-owned observation wells in Chengdu, China, collected from 2019 to 2023, are used to construct a graph that represents inter-well spatial relationships under an irregular monitoring layout. In the proposed framework, the GCN component captures spatial dependence and heterogeneous interactions among wells, while the LSTM component models temporal evolution in groundwater level fluctuations. Two data-availability scenarios are further considered to reflect practical monitoring conditions: a groundwater-level-only setting and a multi-factor setting incorporating rainfall and water temperature. Results show that the proposed framework can achieve accurate multi-step forecasting for multiple wells simultaneously under both scenarios, while the inclusion of additional temporal drivers further improves the prediction of fluctuation trends and turning points. The results also indicate that explicitly modeling inter-well spatial relationships is beneficial for regional groundwater forecasting in heterogeneous urban environments. Overall, the proposed framework provides an effective and transferable approach for groundwater level prediction in irregular monitoring networks and has practical implications for urban groundwater management and early warning.