Publication:
SOC Estimation Using Deep Bidirectional Gated Recurrent Units with Tree Parzen Estimator Hyperparameter Optimization

dc.citedby3
dc.contributor.authorHow D.N.T.en_US
dc.contributor.authorHannan M.A.en_US
dc.contributor.authorLipu M.S.H.en_US
dc.contributor.authorKer P.J.en_US
dc.contributor.authorMansor M.en_US
dc.contributor.authorSahari K.S.M.en_US
dc.contributor.authorMuttaqi K.M.en_US
dc.contributor.authorid57212923888en_US
dc.contributor.authorid7103014445en_US
dc.contributor.authorid36518949700en_US
dc.contributor.authorid37461740800en_US
dc.contributor.authorid6701749037en_US
dc.contributor.authorid57218170038en_US
dc.contributor.authorid55582332500en_US
dc.date.accessioned2023-05-29T09:40:40Z
dc.date.available2023-05-29T09:40:40Z
dc.date.issued2022
dc.descriptionBattery management systems; Charging (batteries); Computer architecture; Digital storage; Heuristic methods; Lithium-ion batteries; Long short-term memory; Temperature measurement; Trees (mathematics); Bidirectional gated recurrent unit; Computational modelling; Deep learning; GRU; Hyper-parameter; Learning models; Optimisations; Parzen estimators; State-of-charge estimation; States of charges; Optimizationen_US
dc.description.abstractState-of-charge (SOC) is a crucial battery quantity that needs constant monitoring to ensure cell longevity and safe operation. However, SOC is not an observable quantity and cannot be practically measured outside of laboratory environments. Hence, machine learning (ML) has been employed to map correlated observable signals such as voltage, current, and temperature to SOC values. In recent studies, deep learning (DL) has been a prominent ML approach outperforming many existing methods for SOC estimation. However, yielding optimal performance from DL models relies heavily on an appropriate selection of hyperparameters. At present, researchers relied on established heuristics to select hyperparameters through manual tuning or exhaustive search methods such as grid search and random search. This results in lengthy development time in addition to less accurate and inefficient models. This study proposes a systematic and automated approach to hyperparameter selection with a Bayesian optimization strategy known as tree Parzen estimator (TPE) in combination with the Hyperband pruning algorithm. The TPE optimization is run on various DL models such as BGRU, GRU, LSTM, CNN, FCN, and DNN to optimize for the best hyperparameter combination for each architecture. The Hyperband pruning algorithm is used to prune unpromising trials during the TPE run which results in time and computational resource savings. Experiment results show that the best performing model, BGRU-TPE incurs a low computation cost with no compromise in model accuracy, maintaining the best balance of both. The proposed model BGRU-TPE achieves the lowest RMSE and MAE error at 0.8% and 0.56%, respectively on various electric vehicle drive cycles at varying ambient temperatures and maintains a low computational cost at 15 600 FLOPS with the model size of only 193.3 KB. � 1972-2012 IEEE.en_US
dc.description.natureFinalen_US
dc.identifier.doi10.1109/TIA.2022.3180282
dc.identifier.epage6638
dc.identifier.issue5
dc.identifier.scopus2-s2.0-85131715742
dc.identifier.spage6629
dc.identifier.urihttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85131715742&doi=10.1109%2fTIA.2022.3180282&partnerID=40&md5=151429d136de49ff60ab6dd5a4a59d49
dc.identifier.urihttps://irepository.uniten.edu.my/handle/123456789/27186
dc.identifier.volume58
dc.publisherInstitute of Electrical and Electronics Engineers Inc.en_US
dc.sourceScopus
dc.sourcetitleIEEE Transactions on Industry Applications
dc.titleSOC Estimation Using Deep Bidirectional Gated Recurrent Units with Tree Parzen Estimator Hyperparameter Optimizationen_US
dc.typeArticleen_US
dspace.entity.typePublication
Files
Collections