Electric vehicle (EV) powertrains consist of power electronic components as well as electric machines to manage the energy flow between different powertrain subsystems and to deliver the necessary torque and power requirements at the wheels. These power subsystems can generate undesired electrical harmonics on the direct current (DC) bus of the powertrain. This may lead to the on-board battery being subjected to DC current superposed with undesirable high- and low- frequency current oscillations, known as ripples. From real-world measurements, significant current harmonics perturbations within the range of 50 Hz to 4 kHz have been observed on the high voltage DC bus of the EV. In the limited literature, investigations into the impact of these harmonics on the degradation of battery systems have been conducted. In these studies, the battery systems were supplied by superposed current signals i.e., DC superposed by a single frequency alternating current (AC). None of these studies considered applying the entire spectrum of the ripple current measured in the real-world scenario, which is focused on in this research. The preliminary results indicate that there is no difference concerning capacity fade or impedance rise between the cells subjected to just DC current and those subjected additionally to a superposed AC ripple current.

Existing validation methods for equivalent circuit models (ECMs) do not capture the effects of operating lithium-ion cells over legislative drive cycles at low ambient temperatures. Unrealistic validation of an ECM may often lead to reduced accuracy in electric vehicle range estimation. In this study, current and power are used to illustrate the different approaches for validating ECMs when operating at low ambient temperatures (−15 °C to 25 °C). It was found that employing a current-based approach leads to under-testing of the performance of lithium-ion cells for various legislative drive cycles (NEDC; FTP75; US06; WLTP-3) compared to the actual vehicle. In terms of energy demands, this can be as much as ~21% for more aggressive drive cycles but even ~15% for more conservative drive cycles. In terms of peak power demands, this can range from ~27% for more conservative drive cycles to ~35% for more aggressive drive cycles. The research findings reported in this paper suggest that it is better to use a power-based approach (with dynamic voltage) rather than a current-based approach (with fixed voltage) to characterise and model the performance of lithium-ion cells for automotive applications, especially at low ambient temperatures. This evidence should help rationalize the approaches in a model-based design process leading to potential improvements in real-world applications for lithium-ion cells.

Previous research has focused on the use of heat pumps in electric vehicles, with the focus on recuperating heat from, normally, ambient and one thermal source on the vehicle. Here 5 potential thermal sources on a vehicle have been identified and thorough testing on the benefit of each source has been performed. The results presented suggest the motor, a thermal storage device, and cabin exhaust extraction should be used >80% of the time according to the scenarios tested, while battery heating and transmission heat extraction should be used subject to conditions on the ambient temperature and drive cycle.

Lithiumion (Li-ion) batteries have become increasingly useful within the automotive industry and modern life applications due to high energy and power densities. However, these batteries suffer capacity loss due to different ageing mechanisms in various applications. Despite several existing models, lack of accurate predictability of capacity degradation limits the advancement of Li-ion batteries. The present work focuses on prediction of battery useful capacity degradation using long-short term memory (LSTM) transfer learning neural network model. At first, a base model was developed and trained using all the (100 ^% ) degradation data available at 0°C and 10°C environmental temperatures. Thereafter, the training of the base model was fixed, and additional hidden layers were added on top of the base model to further fine tune it with only the initial 30% degradation data available at 25°C environmental temperature. The remaining (70 ^% ) data of the 25°C case was tested for model prediction. To decide the number of fixed hidden layers to be transferred from base model to transfer model and the number of additional hidden layers on top, an optimization for minimum cross validation error was performed. It was found that the resulting model was able to forecast the remaining battery degradation with 96% accuracy. The model prediction was also compared with LSTM deep learning architecture without using transfer learning. The LSTM with transfer learning model was found to be 17% higher in prediction accuracy than that without utilizing transfer learning.

Today’s market leading electric vehicles, driven on typical UK motorways, have real-world range estimation inaccuracy of up to 27%, at around 10 °C outside temperature. The inaccuracy worsens for city driving or lower outside temperature. The reliability of range estimation largely depends on the accuracy of the battery’s underlying state estimators, e.g., state-of-charge and state-of-energy. This is affected by accuracy of the models embedded in the battery management system. The performance of these models fundamentally depends on experimentally obtained parameterisation and validation data. These experiments are mostly performed within thermal chambers, which maintain pre-set temperatures using forced air convection. Although these setups claim to maintain isothermal test conditions, they rarely do so. In this paper, we show that this is potentially the root-cause for deterioration of range estimation at low temperatures. This is because, while such setups produce results comparable to isothermal conditions at higher temperatures (25 °C), they fail to achieve isothermal conditions at sub-zero temperatures. Employing an immersed oil-cooled experimental setup, which can create close-to isothermal conditions, we show battery state estimation can be improved by reducing error from 49.3% to 11.7% at −15 °C. These findings provide a way forward towards improving range estimation in cold weather conditions.