Using Machine Learning to Optimize Resource Use in Batteries and Engines: A Review (original) (raw)
Related papers
Battery Management in Electrical Vehicles Using Machine Learning Techniques
Journal of Pharmaceutical Negative Results
Electric vehicle (EV) manufacturers are increasing production in order to keep up with the fast growing demand from customers. To meet the ever-increasing demand for electric vehicles (EVs), both well enough and up-and-coming manufacturers of automobiles will need to immediately boost their production of EVs while simultaneously lowering their prices in order to compete. However, the production of electric vehicles comes with its own set of challenges for automotive manufacturers who are looking to survive in the years to come. By generating digital twins of both their products and their production processes, businesses have the ability to virtually develop and test complete assembly processes as well as entire facilities. This speeds up the process of transitioning to high-volume production without compromising the product's quality. Battery management is one of the most challenging components of driving electric vehicles (EV). Batteries and the amounts of energy that they can ...
Battery optimization by machine learning algorithms: Research gap via bibliometric analysis
E3S web of conferences, 2023
Technological developments enable low-carbon transitions to be accelerated by conceptualization systems and innovations for research and development to generate clean energy. Batteries are becoming one of the essential parts of the science of electrical power sources. Lithium-ion batteries are part of the change and development factors in technologies that significantly impact the portable devices sector and the development of electric vehicles. Designing the material structure and composition of battery manufacturing with the help of engineering system design will form a much more optimal battery. Machine learning algorithms can easily optimize the battery's composition through battery experiment test data history to produce a more optimal battery configuration. This study is prepared to identify research gaps in topics related to machine learning for battery optimization. Related studies about machine learning for battery optimization are identified using bibliometric analysis and systematic literature review of the study search index through database Scopus-indexed publications. The results from this paper reveal energy management systems and strategies, hybrid vehicles, other optimization algorithms, battery electrodes, and the safety of batteries as the particular research gap according to machine learning for battery optimization. This paper expects research on battery optimization using machine learning methods will continue to be developed to maximize the potential of machine learning algorithms in helping the research process.
Overview of machine learning applications to battery thermal management systems in electric vehicles
Nucleation and Atmospheric Aerosols, 2022
Electric vehicle (EV) is increasingly becoming an alternative vehicle of choice to replace an internal combustion engine-powered car. EV concept is clearly linked to sustainable development. Generally, there are four types of EVs: hybrid, plug-in hybrid, battery, and fuel cell EVs. The form of energy source and storage plays a key role for all EVs. Mostly, a lithium-ion battery (high-voltage battery) is used as energy storage due to its high energy density and long-life cycle. But, high rates of charging and discharging bring about high temperatures of the lithium-ion battery, reducing its useful lifetime. A battery thermal management system (BTMS) is crucial in improving EV performance. Here, in this work, we presented an overview of BTMS employed in the EV development, as well as applications of machine learning techniques to predict and optimize BTMS performance based on fast-charging protocols. Additionally, BTMS based on tropical environmental conditions like in Thailand was also discussed.
Machine learning-based lithium-ion battery life prediction for electric vehicle applications
International Journal of Power Electronics and Drive Systems (IJPEDS), 2024
The actual and anticipated battlefield creates a model capable of accurately estimating the lifetime of lithium-ion batteries used in electric cars. This inquiry uses a technique known as supervised machine learning, more particularly linear regression. In lithium-ion batteries, modeling temperature-dependent per-cells is the basis for capacity calculation. When a sufficient quantity of test data is accessible, a linear regression learning method will be utilized to train this model, ensuring a positive outcome in forecasting battery capacity. The conclusions drawn in the article are derived from the attributes of the initial one hundred charging and discharging cycles of the battery, enabling the determination of its remaining power. This determination facilitates the swift identification of battery manufacturing procedures and empowers consumers to detect flawed batteries when signs of performance degradation and reduced longevity are observed. MATLAB simulations have demonstrated the accuracy of the projected results, exhibiting a margin of error of approximately 9.98%. With its capacity to provide a reliable and precise means of estimating battery lifespan, the developed model holds the potential to revolutionize the electric vehicle industry.
Energy management in electric vehicle using machine learning
Nucleation and Atmospheric Aerosols, 2023
In power system, the energy which is fed to the grid control and management is accomplished using various architectures system. The variation in the control and manage systems are based on the performance and features as well as the overall cost. HEVs were really high increased air and oil emissions control. HEVs of different designs are becoming increasingly common, as they can achieve high fuel savings and mitigate emissions through multiple storage systems. The most powerful approach is the battery-driven vehicles.This article examines the management of energy issue with the minimisation of the operating costs of the HEVs for hybrid electric vehicles (HEVs). The energy management strategy and hybrid performance enhancement are the key to ensuring the system functions properly. The battery energy is managed using the proposed GA approach based on reinforcement learning (RL) which is originated from the ML. The results are obtained and verified successfully.
E3S Web of Conferences
Recently, the popularity of li-ion batteries has attracted many researchers to carry out the battery’s maximum potential. Predicting batteries condition and behavior is part of the process that is considered challenging. ML algorithm is widely applied to overcome this challenge as it demonstrates a successful outcome in optimizing the complexity, accuracy, reliability, and efficiency of battery prediction. Yet, we believe there is a particular research area of battery prediction that can further be explored and enhanced with machine learning capability. Therefore, we perform a systematic literature review and bibliometric study to uncover the gap in the machine learning application in the battery prediction field. This study is divided into four stages: (1) literature search from the Scopus Database, (2) filtering the results based on keywords and prepared criteria using PRISMA method, (3) systematic review from filtered papers to provide further understanding, and (4) bibliometric ...
Learning Policies For Battery Usage Optimization in Electric Vehicles.
In In Proceedings of European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD), 2012
The high cost, limited capacity, and long recharge time of batteries pose a number of obstacles for the widespread adoption of electric vehicles. Multi-battery systems that combine a standard battery with supercapacitors are currently one of the most promising ways to increase battery lifespan and reduce operating costs. However, their performance crucially depends on how they are designed and operated. In this paper, we formalize the problem of optimizing real-time energy management of multi-battery systems as a stochastic planning problem, and we propose a novel solution based on a combination of optimization, machine learning and data-mining techniques. We evaluate the performance of our intelligent energy management system on various large datasets of commuter trips crowdsourced in the United States. We show that our policy significantly outperforms the leading algorithms that were previously proposed as part of an open algorithmic challenge.
The Rise of AI driven BMS Revolutionizing Li ion Battery Performance
IEEE, 2024
—Accurate measurement of the number of cycles of Li-Ion batteries is essential for the battery management system (BMS). Virtual State of Health (SOH) estimation is increasingly attracting attention because to the simplicity of its structure, its flexibility in adapting to online usage, and its independence on battery designs, in contrast to direct methods of measurement and model-based methods. At first, data-driven methods are used to derive a reliable health indicator (HI) that may be used to measure the status of health (SOH). Afterwards, a machine learning model is created to establish the relationship between the Health Index (HI) and the State of Health (SOH). This study proposes employing a partial charge and discharge current sequence based on data analytics of battery aging data. Next, the data is inputted into the KNN algorithm, known for its rapid learning speed and strong generalization capabilities. The estimation accuracy is fully tested by considering the selected charging and discharging capacity. The efficacy of the suggested strategy is confirmed by testing on an openly available dataset. This study offers an analysis of the methodologies used to estimate the state of health (SOH) of batteries, highlighting their primary benefits and identifying their limits in terms of real-time compatibility with automotive systems, particularly in hybrid electric applications.