Battery health prognosis with gated recurrent unit neural networks and hidden Markov model considering uncertainty quantification

Lin, Mingqiang; You, Yuqiang; Wang, Wei; Wu, J.

doi:10.1016/j.ress.2022.108978

Cited by 32 publications

(5 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [40] a method for predicting the SOH of lithium-ion batteries using gated recurrent unit (GRU) neural networks and hidden Markov model (HMM) with considering uncertainty quantification is proposed. The work decomposes the battery capacity into the global downward trend and the local fluctuations by using empirical mode decomposition (EMD), and trains a GRU network to fit the global trend and an HMM to fit the local fluctuations.…”

Section: B Battery Health Prediction Using Deep Learningmentioning

confidence: 99%

Comprehensive Review of Machine Learning, Deep Learning, and Digital Twin Data-Driven Approaches in Battery Health Prediction of Electric Vehicles

Renold,

Kathayat

2024

IEEE Access

View full text Add to dashboard Cite

This paper presents a comprehensive survey of machine learning, deep learning, and digital twin technology methods for predicting and managing the battery state of health in electric vehicles. Battery state of health estimation is essential for optimizing the battery usage, performance, safety, and cost-effectiveness of electric vehicles. Estimating the state of health of a battery is a complex undertaking due to its dependency on multiple factors. These factors include battery characteristics such as type, chemistry, size, temperature, current, voltage, impedance, cycle number, and driving pattern. There are drawbacks to traditional methods, such experimental and model-based approaches, in terms of accuracy, complexity, expense, and viability for real-time applications. By employing a variety of algorithms to discover the nonlinear and dynamic link between the battery parameters and the state of health, data-driven techniques like machine learning, deep learning, and data-driven digital twin technologies can get beyond these restrictions. Data-driven methods can also incorporate physics and domain knowledge to improve the explainability and interpretability of the results. This paper reviews the latest advancements and challenges of using data-driven techniques for battery state of health estimation and management in electric vehicles. The paper also discusses the future directions and opportunities for further research and development in this field. The survey scope spans publications from the year 2021 to 2023.INDEX TERMS Machine learning models; Deep learning models; Data-driven methods; Lithium-ion batteries; Digital twin technology

show abstract

Section: B Battery Health Prediction Using Deep Learningmentioning

confidence: 99%

Comprehensive Review of Machine Learning, Deep Learning, and Digital Twin Data-Driven Approaches in Battery Health Prediction of Electric Vehicles

Renold,

Kathayat

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Hidden Markov Model (HMM) is a machine learning model for time series or state series, which can mainly solve the problem of containing both types of sequences, namely observable data and unobservable states. HMM is able to statistically model information over a period of time and, at the same time, speculate on the transfer laws of the hidden states and the mapping relationship between the observed value variables and the state variables through the learning of this time-series information [18].…”

Section: Hidden Markov-based Semantic Segmentationmentioning

confidence: 99%

A Study of Differences in Sentential Characteristics of Japanese Language Teaching in Colleges and Universities Based on Semantic Contrastive Analysis

Tang

2024

Applied Mathematics and Nonlinear Sciences

View full text Add to dashboard Cite

Colleges and universities actively carry out Japanese language teaching courses to promote the friendly cooperation and communication between China and Japan. However, improving students’ Japanese intercultural communication ability still faces specific difficulties and practical obstacles due to cultural differences. In this paper, firstly, a Japanese language teaching corpus is collected based on Python technology to establish a Japanese language teaching corpus, based on which a Japanese language flipped classroom teaching model is constructed in colleges and universities. Secondly, to understand the semantic distribution of Japanese vocabulary in the Japanese teaching corpus, the Hidden Markov Model is used to semantically segment Japanese language, and the attention mechanism based on categories is used to extract and classify the textual features of the Japanese corpus. Finally, the model’s teaching effect and sentence features are analyzed using University of N as an example. The results show that after the teaching experiment, the average score of Japanese application proficiency of the experimental class is 86.57, which is 14.495% higher than that of the control class. 22.52% of the sentence length frequency ratios in spoken corpora were distributed in the [10,20] range, and the highest values of sentence length frequency ratios in written corpora, academic corpora, and policy corpora fell in the range of 20-30. Teaching Japanese language in higher education needs to target the lexical meanings of different sentence lengths to help students improve their Japanese language application skills.

show abstract

“…Therefore, the RUL prediction methods using point estimation are unsuitable for making maintenance decisions while considering risks, since risk is usually measured through uncertainty. The uncertainty quantification in RUL prediction mainly depend on statistical model-based methods, such as Gaussian process regression [32], hidden Markov model [33], etc. Only a few researchers have paid attention to uncertainty quantification of RUL prediction results based on machine learning in recent years.…”

Section: Introductionmentioning

confidence: 99%

A Novel Remaining Useful Life Prediction Approach Combined eXtreme Gradient Boosting and Multi-Quantile Recurrent Neural Network

Yin,

Zhao,

Zhang

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Remaining useful life (RUL) prediction is a key technology to ensure the reliability and safety of high-end equipment. Deep learning is widely used for RUL prediction due to the excellent feature extraction ability and nonlinear fitting ability. Traditional recurrent neural networks adopt recursive strategy, which easily lead to the problems of error accumulation and low stability. On the other hand, most deep learning methods are used for point prediction and cannot quantify uncertainty in the prediction results. Although some probability prediction methods based on deep learning can provide probability prediction results, it require the assumption that the prediction results follow a specific distribution in advance. However, the distribution of most prediction results does not match a suitable distribution function. To solve above problems, a novel RUL prediction approach based on eXtreme gradient boosting (XGBoost) and multiquantile recursive neural network (MQ-RNN) is proposed in this article. Specifically, XGBoost is used to select features closely related to RUL, and the selected features are fed into MQ-RNN to train the RUL prediction model. The advantage of MQ-RNN is that it has non-parametric framework, prediction results can be obtained by multi-quantile regression, which does not require prior assumptions about the distribution of predicted results. Furthermore, the proposed framework is verified by C-MAPSS dataset. Finally, comparative experiments are conducted. The experimental results show that the proposed method maintains good predictive performance in both point estimation and interval estimation.INDEX TERMS Remaining useful life, eXtreme gradient boosting, Multi-quantile recursive neural network, Quantile regression, Forking sequences training scheme.

show abstract

Battery health prognosis with gated recurrent unit neural networks and hidden Markov model considering uncertainty quantification

Cited by 32 publications

References 40 publications

Comprehensive Review of Machine Learning, Deep Learning, and Digital Twin Data-Driven Approaches in Battery Health Prediction of Electric Vehicles

Comprehensive Review of Machine Learning, Deep Learning, and Digital Twin Data-Driven Approaches in Battery Health Prediction of Electric Vehicles

A Study of Differences in Sentential Characteristics of Japanese Language Teaching in Colleges and Universities Based on Semantic Contrastive Analysis

A Novel Remaining Useful Life Prediction Approach Combined eXtreme Gradient Boosting and Multi-Quantile Recurrent Neural Network

Contact Info

Product

Resources

About