电力系统及其自动化学报2025，Vol.37Issue(5)：113-120,8.DOI:10.19635/j.cnki.csu-epsa.001538
基于电池健康状态预测的储能系统功率分配
Power Allocation of Energy Storage System Based on State-of-health Prediction of Batteries
摘要
Abstract
The applications of battery energy storage systems(BESSs)in power system are increasingly widespread.To address the issue of capacity attenuation of battery packs,a power allocation method for energy storage is proposed based on the state-of-health(SOH)prediction of battery packs.First,a hybrid neural network model combining convo-lutional neural networks(CNNs)and long short-term memory(LSTM)networks is constructed,and the Bayesian algo-rithm is employed to fine-tune the network parameters,thereby enhancing the accuracy of SOH prediction of battery packs.Second,based on the proposed prediction model,a prioritization ranking rule for the charging/discharging of bat-tery packs is put forward using the entropy weight method.Third,a power allocation optimization model is established with the objective of achieving consistency in the state-of-charge(SOC)among battery packs.Finally,simulation vali-dation is conducted using the historical operation data from the BESS of one wind farm.Results demonstrate that the pro-posed method significantly extends the lifespan of used batteries,prevents the over-charging and over-discharging of in-dividual batteries,and achieves a balanced SOC across battery packs.关键词
健康状态/卷积-长短期记忆网络/雨流计数法/熵权法Key words
state-of-health(SOH)/convolutional-long short-term memory neural networks/rainflow counting algo-rithm/entropy weight method分类
信息技术与安全科学引用本文复制引用
徐珂,佘小平,李凯,曹玉杰,冯艳虹,方琰..基于电池健康状态预测的储能系统功率分配[J].电力系统及其自动化学报,2025,37(5):113-120,8.基金项目
中国电力工程顾问集团华北电力设计院有限公司科技项目(JBGS2023-01). （JBGS2023-01）
