摘要
大容量电池储能系统可以为新能源并网稳定运行提供有效支撑,电池储能系统的大功率输入/输出加剧了电池单体的损耗,造成电池容量发生不同程度的衰减,容量衰减过程较为缓慢,且不易发现。为及时准确预测电池容量衰减情况,提出基于PSO-BP神经网络的大容量电池储能系统电池衰减容量预测方法,通过粒子群算法(PSO)对反向传播(BP)神经网络隐层权值阈值进行优化。仿真结果表明,PSO-BP神经网络具有较高的准确性,可对任意衰减程度进行准确预测。
A large capacity battery energy storage system can provide effective support for renewable energy,but the high power input and output of the battery energy storage system contributes to the capacity fade. Thecapacity fade process is relatively slow,and not easy to be found. To accurately predict the capacity fade,this paperproposed the capacity fade prediction model of large capacity battery energy storage system based on PSO-BP neuralnetwork. The method uses the PSO to optimize the hidden layer weights threshold of BP neural network. Thesimulation results show that the PSO-BP network has higher accuracy and arbitrary attenuation degree can beaccurately predicted.
出处
《电器与能效管理技术》
2016年第14期75-78,共4页
Electrical & Energy Management Technology
基金
国家自然科学基金(51577065)
国家电网公司项目(DG71-15-039)
关键词
大容量电池储能系统
衰减容量预测
粒子群算法
BP神经网络
large capacity battery energy storage system
capacity fade prediction
particle swarmoptimization (PSO)
back propagation (BP) neural network