{"title":"基于健康指数提取和QHDBO-BiTCN-BiGRU的锂离子电池SOH估计混合数据驱动方法","authors":"Rui Quan, Yulong Zhou, Wen Li, Hang Wan","doi":"10.1016/j.est.2025.116788","DOIUrl":null,"url":null,"abstract":"<div><div>To further enhance the forecasting exactness of the state of health (SOH) regarding lithium-ion batteries, a hybrid data-driven prediction model that integrated a bidirectional temporal convolution network (BiTCN) with the bidirectional gated recurrent unit (BiGRU) was presented, and its hyperparameters were optimized using quantum computing and a hybrid dung beetle optimization algorithm with multiple strategies (QHDBO). Firstly, the time interval of an equal discharging voltage difference (TIEDVD) method was used to extract the health indicator (HI) of batteries as the input sequence, and the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) was used to perform soft thresholding reconstruction and denoising on the health indicator sequence. Next, BiTCN learned the hidden information in the sequence, extracted the time series characteristics, and input them to BiGRU for SOH prediction. Finally, eight different initial SOH-based batteries from NASA and CALCE battery aging datasets were used for experimentation comparison, with eight other data-driven models. Results demonstrate that the novel QHDBO-BiTCN-BiGRU method achieves superior prediction performance, evidenced by the lowest mean absolute percentage error (MAPE) of 0.4 % and root mean square error (RMSE) of 0.64 %, alongside an exceptionally high decision coefficient (R<sup>2</sup>) of 99.8 %. The proposed approach offers a more accurate SOH estimation guide for lithium-ion batteries.</div></div>","PeriodicalId":15942,"journal":{"name":"Journal of energy storage","volume":"125 ","pages":"Article 116788"},"PeriodicalIF":8.9000,"publicationDate":"2025-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A hybrid data-driven approach for SOH estimation of lithium-ion batteries based on health index extraction and QHDBO-BiTCN-BiGRU\",\"authors\":\"Rui Quan, Yulong Zhou, Wen Li, Hang Wan\",\"doi\":\"10.1016/j.est.2025.116788\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>To further enhance the forecasting exactness of the state of health (SOH) regarding lithium-ion batteries, a hybrid data-driven prediction model that integrated a bidirectional temporal convolution network (BiTCN) with the bidirectional gated recurrent unit (BiGRU) was presented, and its hyperparameters were optimized using quantum computing and a hybrid dung beetle optimization algorithm with multiple strategies (QHDBO). Firstly, the time interval of an equal discharging voltage difference (TIEDVD) method was used to extract the health indicator (HI) of batteries as the input sequence, and the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) was used to perform soft thresholding reconstruction and denoising on the health indicator sequence. Next, BiTCN learned the hidden information in the sequence, extracted the time series characteristics, and input them to BiGRU for SOH prediction. Finally, eight different initial SOH-based batteries from NASA and CALCE battery aging datasets were used for experimentation comparison, with eight other data-driven models. Results demonstrate that the novel QHDBO-BiTCN-BiGRU method achieves superior prediction performance, evidenced by the lowest mean absolute percentage error (MAPE) of 0.4 % and root mean square error (RMSE) of 0.64 %, alongside an exceptionally high decision coefficient (R<sup>2</sup>) of 99.8 %. The proposed approach offers a more accurate SOH estimation guide for lithium-ion batteries.</div></div>\",\"PeriodicalId\":15942,\"journal\":{\"name\":\"Journal of energy storage\",\"volume\":\"125 \",\"pages\":\"Article 116788\"},\"PeriodicalIF\":8.9000,\"publicationDate\":\"2025-05-08\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Journal of energy storage\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2352152X25015014\",\"RegionNum\":2,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENERGY & FUELS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of energy storage","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2352152X25015014","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENERGY & FUELS","Score":null,"Total":0}
A hybrid data-driven approach for SOH estimation of lithium-ion batteries based on health index extraction and QHDBO-BiTCN-BiGRU
To further enhance the forecasting exactness of the state of health (SOH) regarding lithium-ion batteries, a hybrid data-driven prediction model that integrated a bidirectional temporal convolution network (BiTCN) with the bidirectional gated recurrent unit (BiGRU) was presented, and its hyperparameters were optimized using quantum computing and a hybrid dung beetle optimization algorithm with multiple strategies (QHDBO). Firstly, the time interval of an equal discharging voltage difference (TIEDVD) method was used to extract the health indicator (HI) of batteries as the input sequence, and the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) was used to perform soft thresholding reconstruction and denoising on the health indicator sequence. Next, BiTCN learned the hidden information in the sequence, extracted the time series characteristics, and input them to BiGRU for SOH prediction. Finally, eight different initial SOH-based batteries from NASA and CALCE battery aging datasets were used for experimentation comparison, with eight other data-driven models. Results demonstrate that the novel QHDBO-BiTCN-BiGRU method achieves superior prediction performance, evidenced by the lowest mean absolute percentage error (MAPE) of 0.4 % and root mean square error (RMSE) of 0.64 %, alongside an exceptionally high decision coefficient (R2) of 99.8 %. The proposed approach offers a more accurate SOH estimation guide for lithium-ion batteries.
期刊介绍:
Journal of energy storage focusses on all aspects of energy storage, in particular systems integration, electric grid integration, modelling and analysis, novel energy storage technologies, sizing and management strategies, business models for operation of storage systems and energy storage developments worldwide.