基于CNN和LSTM混合网络的直流充电桩功率模块优化  

Optimization of DC Charging Pile Power Module Based on CNN and LSTM Hybrid Network

作　　者：何培东[1] 陈维民[1] 白佳灵 邓舒予 谢智[1] 丁熠辉 HE Pei-dong;CHEN Wei-min;BAI Jia-ling;DENG Shu-yu;XIE Zhi;DING Yi-hui(Metering Center of Sichuan Electric Power Company of State Grid,Chengdu 610045 China)

机构地区：[1]国网四川省电力公司计量中心,四川成都610045

出　　处：《自动化技术与应用》2024年第11期51-55,共5页Techniques of Automation and Applications

基　　金：国网四川省电力公司科技项目(52199722001B)。

摘　　要：直流充电桩内部功率变化动态性过强,不易捕捉及计算误差过大,提出基于CNN和LSTM混合网络的功率模块优化方法。根据充电桩电路工作原理,得到电路内部各项参数间存在的直接关联,根据该点采用三相电网平衡原理设计电力稳定控制函数,计算满足电压、电流以及功率最稳的调节比例系数,并作为调节点,建立CNN和LSTM混合网络多层次模型,计算有序、无序充放电过程的电力负荷变化,利用调节比例系数按照充电、放电所需功率需求完成优化。实验结果表明,所提方法的充电桩功率稳定,对充电桩的电压偏移量实现有效改善。The internal power change of DC charging pile is too dynamic,difficult to capture and the calculation error is too large.A power module optimization method based on CNN and LSTM hybrid network is proposed.According to the working principle of the charging pile circuit,the direct correlation between various parameters in the circuit is obtained.According to this point,the three-phase power grid balance principle is adopted to design the power stability control function,calculate the adjustment proportion coefficient that satisfies the most stable voltage,current and power,and as the adjustment point,establish the CNN and LSTM hybrid network multi-level model to calculate the power load changes in the orderly and disordered charging and discharging processes.The power demand for charging and discharging is optimized by adjusting the proportional coefficient.The experimental results show that the power of the charging pile is stable and the voltage offset of the charging pile is effectively improved.

关 键 词：CNN和LSTM混合网络 三相电网平衡原理 电力稳定控制函数 模型 直流充电桩 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]