A strong tracking adaptive fading‐extended Kalman filter for the state of charge estimation of lithium‐ion batteries

Takyi‐Aninakwa, Paul; Wang, Shunli; Zhang, Hongying; Appiah, Emmanuel; Bobobee, Etse Dablu; Fernandez, Carlos

doi:10.1002/er.8307

Cited by 47 publications

(13 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Firstly, it is difficult to obtain the initial value of the battery SOE accurately [56]. In the power integration method, an inaccurate initial value is utilized to predict the SOE, which easily causes the accumulation of errors [45]. Secondly, the method is sensitive to the accuracy of the current sensor.…”

Section: Power Integration Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A Critical Review of Improved Deep Convolutional Neural Network for Multi-Timescale State Prediction of Lithium-Ion Batteries

et al. 2022

Self Cite

View full text Add to dashboard Cite

Lithium-ion batteries are widely used as effective energy storage and have become the main component of power supply systems. Accurate battery state prediction is key to ensuring reliability and has significant guidance for optimizing the performance of battery power systems and replacement. Due to the complex and dynamic operations of lithium-ion batteries, the state parameters change with either the working condition or the aging process. The accuracy of online state prediction is difficult to improve, which is an urgent issue that needs to be solved to ensure a reliable and safe power supply. Currently, with the emergence of artificial intelligence (AI), battery state prediction methods based on data-driven methods have high precision and robustness to improve state prediction accuracy. The demanding characteristics of test time are reduced, and this has become the research focus in the related fields. Therefore, the convolutional neural network (CNN) was improved in the data modeling process to establish a deep convolutional neural network ensemble transfer learning (DCNN-ETL) method, which plays a significant role in battery state prediction. This paper reviews and compares several mathematical DCNN models. The key features are identified on the basis of the modeling capability for the state prediction. Then, the prediction methods are classified on the basis of the identified features. In the process of deep learning (DL) calculation, specific criteria for evaluating different modeling accuracy levels are defined. The identified features of the state prediction model are taken advantage of to give relevant conclusions and suggestions. The DCNN-ETL method is selected to realize the reliable state prediction of lithium-ion batteries.

show abstract

Section: Power Integration Methodsmentioning

confidence: 99%

“…After discharging, the voltage-current fitting curve is used to find the highest charging voltage and the lowest discharging voltage of the battery at different depths of discharge. Then, the peak power of the battery is obtained by substituting the corresponding calculation formula [45].…”

Section: Overview Of the State-of-power Predictionmentioning

confidence: 99%

A Critical Review of Improved Deep Convolutional Neural Network for Multi-Timescale State Prediction of Lithium-Ion Batteries

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Accurate variants of Kalman filter-based approaches are widely employed for many commercial rechargeable batteries. 31 Deep Neural Networks (DNNs) have been found to be very useful in precise estimation of State-of-Charge and State-of-Health metrics of batteries. 32 Domain adaptation along with robust transfer learning capabilities of DNNs have been exploited for enhancing estimation generalizability.…”

Section: Recent Machine Learning Methodsmentioning

confidence: 99%

“…Accurate variants of Kalman filter‐based approaches are widely employed for many commercial rechargeable batteries 31 . Deep Neural Networks (DNNs) have been found to be very useful in precise estimation of State‐of‐Charge and State‐of‐Health metrics of batteries 32 .…”

Section: Introductionmentioning

confidence: 99%

Contrastive learning for practical battery synthetic data generation using seasonal and trend representations

Gowda

Channegowda

2022

Intl J of Energy Research

View full text Add to dashboard Cite

Summary A vast majority of human population around the planet has limited availability to electricity. Inexpensive energy storage systems are crucial for solving this critical issue. Rapid rise in global pollutants has also sparked widespread efforts to limit fossil fuelled automobiles. Lithium‐ion batteries continue to be a commanding force in the energy storage domain. Lithium‐ion batteries have been consistently used in distributed solar‐battery modules as well. Precise prognostics of battery health and remaining shelf life of the battery is a challenging task. Sufficient access to high‐quality battery dataset is important to compute accurate State‐of‐Charge estimations and also to predict end‐of‐life of the storage system. A data‐driven approach is the need of the hour to derive useful insights into battery health and cycle life. Researchers have limited access to battery data chiefly due to privacy concerns. Commercial manufacturers rarely share data, which leads to data scarcity issues. Lack of data leads to ineffective State‐of‐Charge evaluations as well. This paper aims to provide reasonable solutions to resolve the limited data difficulty faced by battery researchers. The primary contributions of our paper include: (a) we employ a seasonal‐trend contrastive learning approach to generate reliable battery synthetic data, (b) heterogeneity of the synthetic data is preserved which helps to improve battery health forecasting. We also release the code implementation details relevant to this research work to enable battery scientists to reproduce and develop upon our work.

show abstract

“…Moreover, because of its less computational cost and simple architecture, it is often used in high real-time applications, such as rocket position estimation, vehicle motion control and other fields [29][30][31][32][33]. Non-linear Kalman filter retains the advantage of Kalman filter and adds support for non-linear optimization, which greatly improves its application range [34][35][36].…”

Section: Introductionmentioning

confidence: 99%

Instantaneous grating signal subdivision system with non-linear Kalman filters

Wang

Shi

Li³

et al. 2022

Optical Design and Testing XII

View full text Add to dashboard Cite

Grating encoder is a sinusoidal encoder based on grating diffraction principle, which is currently utilized in many high-precision displacement systems because of its advantageous characteristics: low cost, simple structure, works in harsh environments, high reliability, and so on. The output signal of grating encoder usually contains noise interference error, amplitude inconsistency error, DC bias error, harmonic error and quadrature phase error. These non-ideal factors are the main reasons for affecting the precision of subdivision. In the traditional signal subdivision system, it is usually necessary to compensate each kind of error separately, which will consume many hardware and computing resources and cause a significant output latency, especially in the filtering section and normalization section. In this paper, a non-linear Kalman filter-based sin-cos wave subdivision method is proposed. Compared with the traditional filtering methods, non-linear Kalman filter has higher dynamic response and can provide instantaneous phasor estimation. In addition, it can simultaneously achieve filtering, amplitude normalization, decoupling DC bias, harmonic suppression, and phase compensation functions, which significantly reduces the computational burden and facilitates the implementation on low-cost processors. In this study, a non-linear Kalman filter-based signal segmentation system is implemented on an FPGA platform and verified on a six-degree-of-freedom grating ruler platform. The results show that the single-channel output delay is only 1.8us at a 50MHz clock, which has a very high real-time ability. When the frequency and amplitude of the input signal varies, the non-linear Kalman filter can track instantaneously and has high dynamic characteristics. Experimental results show the effectiveness of this method.

show abstract

A strong tracking adaptive fading‐extended Kalman filter for the state of charge estimation of lithium‐ion batteries

Cited by 47 publications

References 59 publications

A Critical Review of Improved Deep Convolutional Neural Network for Multi-Timescale State Prediction of Lithium-Ion Batteries

A Critical Review of Improved Deep Convolutional Neural Network for Multi-Timescale State Prediction of Lithium-Ion Batteries

Contrastive learning for practical battery synthetic data generation using seasonal and trend representations

Instantaneous grating signal subdivision system with non-linear Kalman filters

Contact Info

Product

Resources

About