Etransportation最新文献

Data-driven approaches have gained increasing attention in the field of battery life-related prediction, as building a comprehensive mechanistic degradation model remains a challenge. Deep learning has emerged as a powerful data-driven fitting method for battery-related applications. However, interpretability remains an issue in this field, hindering the practical utilization of deep learning methods. With the development of interpretable techniques, deep learning methods not only can be conducted as black box tools for fitting, but also for exploring the relationship between external battery data and internal electrochemical changes. In this paper, an interpretable deep learning procedure is proposed and exemplified by accelerated fading point (knee-point) recognition based on an open battery dataset. The Gradient-weighted Class Activation Mapping (Grad-CAM) is conducted to explain the link between the input and output of the trained convolutional neural networks (CNN) model. The trained CNN model possesses deep insight into battery degradation, giving the very first warning when accelerated fading occurs. Through interpretability analysis, it is confirmed that the well-trained model can spontaneously focus on features associated with internal battery degradation and identify some additional features beyond existing human experience. The proposed method can be used to discover the relationship between battery data and degradation mechanism by artificial intelligence in the electric vehicles (EVs) field.

The capacity degradation of lithium-ion batteries (LIBs) will accelerate after long-term cycling, showing nonlinear aging features, which not only shortens the long-term life of LIBs, but also seriously endangers their safety. In this paper, by introducing the concept of nonlinear aging degree, a knee-point identification method based on the maximum distance method is established, and the nonlinear aging behavior of LIBs is identified and marked, so as to know whether the nonlinear aging phenomenon has occurred. Furthermore, two knee-point prediction methods have been proposed and compared. The direct knee-point prediction method based on stacked long short-term memory (S-LSTM) neural network and sliding window method is proposed for the scenarios of battery development, early performance evaluation and online application. For scenarios such as echelon utilization and post-safety evaluation, an indirect knee-point prediction method combining capacity prediction and knee-point identification algorithm is proposed. Through multi-dimensional comparison of the two methods, the strengths and weaknesses of their applicable scenarios are analyzed. Our work has guiding significance for finding the ideal replacement opportunity of LIBs in different scenarios, so that the user can be reminded whether to maintain or replace the battery, which greatly reduces the risk of battery safety problems.

To meet the increased requirements for efficiency and compactness of the Polymer Electrolyte Membrane Fuel Cell (PEMFC) system for heavy-duty vehicles (HDVs) application, PEMFC has to operate under high temperature and low humidity (HTLH) conditions to reduce the parasitic power consumption of radiators and the size of humidifier. However, HTLH would negatively affect the performance and durability of PEMFC. Through conducting in-situ current mapping, this paper found highly inhomogeneous current in-plane distribution during current cycling under HTLH conditions, which is attributed to the self-reinforced feedback of local current density and membrane water content. Instead, under the situations where current increase is prior to the temperature increase, PEMFC would have much more uniform in-plane current distribution and membrane water distribution, resulting in more efficient utilization of the produced water in a dry environment.

Automotive battery packs for electromobility applications consist of a large number of interconnected battery cells. Different cell-to-busbar joining techniques are utilized, with cylindrical cells frequently being contacted using wire bonding. Failure of individual connections can occur due to strong vibrations during operation and improper stress, making detection by the battery management system a necessity. This study investigates the identification of an electrical wire bond failure in a state-of-the-art electric vehicle module of a Lucid Air with 10 series-connected and 30 parallel-connected cells (10s30p). Four individual cells were characterized extensively in order to generate a simulation model taking into account parameter scatter. The failure case under investigation was simulatively incorporated in one parallel circuit and subsequently replicated in experimental validation measurements at the module level. The results show that this defect can be detected using pulses of C/3 or higher currents at various states of charge. An even more robust detection is achieved using differential voltage analysis of constant current C/20 discharge voltage trajectories. This defect identification method does not require any additional measurement sensors beyond the voltage taps and sensors provided by the manufacturer – with one voltage sensor per parallel circuit – and can therefore be implemented during electric vehicle usage, e.g. at dedicated service checks. A discussion on the applicability and scalability as well as the limitations of the method is provided. All measurement data of the state-of-the-art Lucid battery system is available as open source alongside the article.

Accurate and reliable prediction of the future capacity degradation of lithium-ion batteries is crucial for their application in electric vehicles. Recent publications have highlighted the effectiveness of deep learning, in particular, in generating precise forecasts regarding the aging patterns. However, large quantities of training data covering various aging behaviors are required to train such models effectively. Collecting such a large database centrally is not feasible due to privacy and data communication restrictions of data owners, such as testing facilities or fleet operators. Federated learning provides a solution to this open issue. A framework, which incorporates federated learning into the training of a data-based battery aging model, is presented in this paper. The benefit of federated learning is that even data owners with sensible information can participate in a collaborative model training, since the model training is only conducted locally and all the data remains local and does not have to be disclosed. Thus, more data owners are likely to participate in this collaborative training. This will improve the prediction performance due to the enlarged dataset that can be utilized for the model training. This work shows that the prediction accuracy of the model trained with federated learning is only slightly worse than the prediction results obtained by the ideal case in which all aging data is stored in a central database. A sensitivity analysis is presented to prove the robustness of federated learning even if the datasets between participating data owners are highly imbalanced or exhibit different aging behaviors. Within exemplary scenarios, it is shown that individual data holders can reduce their prediction errors from $MAP{E}_{mean}=7.07\text{\%}$ to $MAP{E}_{mean}=0.91\text{\%}$ by participating in the proposed federated learning-based framework.

As the global popularity of electric vehicles (EVs) continues to grow, the demand for EV batteries has significantly increased. Unfortunately, when the battery state of health (SoH) reaches a certain value, the batteries may no longer meet the driving requirements of EVs, resulting in many retired batteries. Properly disposing of these batteries is a major challenge. In this paper, we will take fast charging stations (FCSs) as an example to evaluate the economic feasibility of reusing retired batteries, which is one of the most environmentally friendly methods of disposal. To make a fair comparison between fresh and retired batteries, a bi-level sizing framework is designed to determine the optimal sizes of fresh and retired batteries, as well as photovoltaic (PV) panels in FCSs, to minimize the total cost. To ensure the high fidelity of results, diverse EV travel patterns are considered to generate the charging power demand. Additionally, a dynamic battery degradation model focused on batteries that have not yet reached their turning point is developed. This model is designed to precisely quantify the gradual reduction in battery capacity. We conducted several representative case studies using real-world data, and the simulation results indicate that FCSs with fresh batteries can achieve 42.2 % cost savings compared to those without energy storage systems, while retired batteries can achieve an additional 5.41 %–11.79 % cost savings under different scenarios. These findings can be generalized that retired batteries are promising in small-scale applications due to the relatively low refurbishment cost, and a new market stream can be potentially generated in this direction.

The construction and development of the new power system with new energy sources as the main component will face significant challenges in terms of scarcity of flexible resources. User-side adjustable loads and energy storage, particularly electric vehicles (EVs), will serve as substantial reservoirs of flexibility, providing stability to the new power system. The rapid deployment of renewable energy and the surpassing of expectations in the penetration rate of EVs in China present opportunities for the significant growth of virtual power plants (VPPs) and vehicle-to-grid (V2G) interactions. The enormous potential and advantages of V2G as a primary user-side resource are further revealed. Under China's current electricity market policies, the pilot projects of user-side interactions are being analyzed. Furthermore, the prospects for the future development of VPPs and V2G, along with relevant recommendations and perspectives regarding business models, technologies, and policies, are highlighted. The development of VPPs and V2G interactions in China holds immense potential for leveraging adjustable resources, especially EVs and energy storages with lithium-ion batteries, to enhance grid flexibility and stability.

The electrification of the transportation sector leads to an increased deployment of lithium-ion batteries in vehicles. Today, traction batteries are installed, for example, in electric cars, electric buses, and electric boats. These use-cases place different demands on the battery. In this work, simulated data from 60 electric cars and field data from 82 electric buses and six electric boats from Germany are used to quantify a set of stress factors relevant to battery operation and life expectancy depending on the mode of transportation. For this purpose, the open-source tool SimSES designed initially to simulate battery operation in stationary applications is extended toward analyzing mobile applications. It now allows users to simulate electric vehicles while driving and charging. The analyses of the three means of transportation show that electric buses, for example, consume between 1 and 1.5 kWh/km and that consumption is lowest at ambient temperatures around 20 °C. Electric buses are confronted with 0.4–1 equivalent full cycle per day, whereas the analyzed set of car batteries experience less than 0.18 and electric boats between 0.026 and 0.3 equivalent full cycles per day. Other parameters analyzed include mean state-of-charges, mean charging rates, and mean trip cycle depths. Beyond these evaluations, the battery parameters of the transportation means are compared with those of three stationary applications. We reveal that stationary storage systems in home storage and balancing power applications generate similar numbers of equivalent full cycles as electric buses, which indicates that similar batteries could be used in these applications. Furthermore, we simulate the influence of different charging strategies and show their severe impact on battery degradation stress factors in e-transportation. To facilitate widespread and diverse usage, all profile and analysis data relevant to this work is provided as open data as part of this work.

In this study, we carry out a fundamental and modeling study to investigate, for the first time, the gas coverage at the catalyst surface and its impacts on performance loss in polymer electrolyte membrane electrolysis cells (PEMECs). Oxygen, produced in the anode catalyst layer (CL) through the oxygen evolution reaction (OER), is removed via the pore network of the anode CL and porous transport layer (PTL) to the flow field. Oxygen gas bubbles can cover the catalyst surface and reduce the area for catalyst OER activity and hence cell performance. To investigate the oxygen bubbles’ impact, we consider various degrees of gas coverage and temperatures (25 °C, 80 °C, and 95 °C) in the range of current density from 0 to 7 A/cm². We also, for the first time, elucidate the impacts of CL’s material properties on gas coverage morphology in the nano/micropores of CLs. Analytical solutions are derived for the gas fraction and gas composition at different temperatures and pressures. It was found that the gas fraction can be as high as 85% with water vapor contributing to 71% of the total gas coverage when operating at 95 °C and 1 atm. The modeling results indicate the gas coverage can contribute 57% of the total overpotential at 95 °C, 7 A/cm², and a coverage coefficient of 7. The work contributes to a fundamental understanding of the impacts of two-phase phenomena on PEMEC performance and is valuable for catalyst layer design and optimization.

The traction substations of urban electric transport grids are oversized and underutilized in terms of their capacity. While their over-sizing is an unfortunate waste, their under-utilization creates the major hurdle for the integration of renewables into these grids due to the lack of a base load. Therefore, integrating smart grid loads such as EV chargers is not only an opportunity but a necessity for the sustainable transport grid of the future.

This paper examines six methods for increasing the potential of EV chargers in three case studies of a trolleygrid, namely a higher substation no-load voltage, a higher substation power capacity, a smart charging method, adding a third overheard parallel line, adding a bilateral connection, and installing a multi-port converter between two substations. From the case studies, the most promising and cost-effective method seems to be introducing a bilateral connection, bringing a charging capacity for up to 175 electric cars per day. Meanwhile, other costly and complex methods, such as smart charging with grid state sensors and communication, can offer charging room for over 200 electric cars per day. Furthermore, using solar PV systems to power the grid showed a more than doubling of the directly utilized energy by installing a 150kW charger, from 19% to 41%. This reduces the power mismatch between the trolleygrid and the PV system from 81% to 59% and thereby reduces the severe economic need for storage, AC grid power exchange, or PV power curtailment while allowing a high penetration of renewables.