The advent of machine learning (ML) has led to an exponential rise in the exploration of new methods to estimate the states of lithium-ion batteries (LIBs) in electric vehicle (EV) applications. Data-driven methods involving ML are increasingly engaged to estimate the state-of-charge (SOC) and state-of-health (SOH) due to greater availability of battery datasets in the public domain; alongside improvements in computing system efficiency. At present, the battery management system (BMS) of EVs tend to face challenges in attaining highly accurate state estimation results using traditional methods, since the LIBs possess strong time-varying and non-linear traits, whilst remaining susceptible to the influence of external factors. Therefore, this paper provides a comprehensive review of the existing ML methods in SOC and SOH estimation of EV application-based LIBs. The insights gained from this review will contribute towards the future development of advanced LIB state estimation algorithms.

Data-driven methods have attracted significant interests to estimate the state-of-health (SOH) of lithium-ion batteries (LIBs). Yet, it is often laborious and computationally costly to robustly train and implement one machine learning model for different LIB chemistries across varied operations. In light of this issue, coupled with the recent technological breakthrough of large language models, this letter exploits the strong generalisation capability of the generative pre-trained transformer-4 (GPT-4) for SOH estimation. Since battery data usually arrives sequentially with varied distribution in the real world, the teacher-to-student model-based distillation of knowledge and lifelong learning are incorporated into GPT-4 to estimate SOH adaptively with minimal fine-tuning. Testing results of the proposed method on an embedded system show a very high estimation accuracy (mean RMSE of 0.64%) at low-compute cost.

This paper presents the design and simulation of a piezoelectric cantilever beam for energy harvesting based on ambient mechanical vibration. The energy harvesting process is facilitated by the forward piezoelectric effect, which converts mechanical energy into electrical energy during detection of ambient vibration by the cantilever beam. The cantilever beam model consists of two piezoelectric layers and an elastic layer clamped to a vibrating host structure at one end, as well as a tip mass mounted at the other end. Using COMSOL Multiphysics, a comparative analysis of the cantilever beam's simulated output characteristics was conducted. These characteristics were analyzed with respect to varying input parameters such as resonant frequency, load resistance and length-width ratio of the cantilever beam. Uased on the simulation results, an optimized model of the cantilever beam was eventually developed. A maximum output power of 4.4 mW and DC output voltage of10.1 V was generated by the model at 4 kΩ load resistance, at a mechanical acceleration of 1 g and a resonant frequency of 75 Hz.

DC microgrids have permeated the energy market in recent years due to the achievement of higher efficiency outputs during power distribution as compared to AC microgrids. Current DC microgrid technology relies on renewable energy sources (e.g. photovoltaic panels, wind turbines) and sub-systems to attain high efficiency while facilitating maximum power point tracking (MPPT) among charge controllers. The key component of a microgrid system is its energy storage system, which assists in mitigating power intermittency due to the deployment of renewables. This paper emphasizes on energy management and control of a DC microgrid system, whereby a simulation model of the proposed DC microgrid is developed in MATLAB/Simulink environment for electrification of a small town. The acquired simulation results have demonstrated feasibility of the proposed DC microgrid during operations.

The use of quasi-Z-source inverters (qZSIs) for DC-DC power conversion applications has gained much recognition when dealing with grid-tied renewable energy resource integrations. This paper proposes a novel self-powered dynamic system (SPDS) involving a piezoelectric vibration energy harvester (PVEH) using qZSI to establish interoperability with a DC load rated at 16.15 mW. Based on uncertain output performances from a piezoelectric cantilever beam (CB), the qZSI-based PVEH serves as a dynamic voltage restoration unit that establishes load-following synchronisation. It uses a proportional-integral based boost controller (PI-based BC) to generate strategic ordering of shoot-through voltage amplification into pulse-width modulation (PWM) gating sequences. The SPDS was modelled using two software based on commercially available product specifications: (i) COMSOL Multiphysics to mechanically design and optimise a CB. (ii) PSCAD/EMTDC to electronically design and integrate the qZSI with the optimised CB, while functioning as a testbed to model the SPDS against arbitrary wind speed and structural vibration frequency data collected from an above-ground mass rapid transit (MRT) train station in Khatib, Singapore. The acquired simulation results have depicted desirable transient responses at respective sub-systems, procuring fast settling-time responses, negligible steady-state error, as well as high efficiencies of 94.07% and 91.64% for the CB and SPDS respectively.