深度学习网络的微电网变压器电压波动抑制研究

Research on Voltage Fluctuation Suppression of Micro Grid Transformer Based on Deep Learning Network

为了解决微电网变压器电压波动的估计和抑制问题,以获得理想的微电网变压器电压为目标,提出了基于深度学习网络的微电网变压器电压波动估计与抑制方法。首先,分析了引起微电网变压器电压波动的主要原因。采用深度学习网络提取电压波动波形数据特征,并根据提取的特征实现电压波动估计。然后,在电压波动条件下,考虑变压器可用补偿容量。通过虚拟阻抗实时监控变压器剩余容量,对其进行补偿,以抑制电压波动。最后,与比例谐振控制器共同完成电压波动抑制。采用微电网变压器电压对估计与抑制效果进行了仿真测试。仿真测试结果表明:该方法可有效提取电压数据特征;相对于传统方法,提高了电压波动估计精度,电压波动抑制效果显著,波形畸变率下降大,具有比较明显的优势。

In order to obtain the ideal voltage of micro grid transformer,a method of estimating and suppressing the voltage fluctuation of micro grid transformer based on deep learning network is proposed.This method analyzes the main causes of voltage fluctuation of micro grid transformer,extracts the data characteristics of voltage fluctuation caused by different reasons by using the deep confidence network model,realizes the voltage fluctuation estimation according to the data characteristics,and then Under the condition of voltage fluctuation,considering the available compensation capacity of the transformer,the virtual impedance is used to monitor the residual capacity of the transformer in real time and compensate it to suppress the voltage fluctuation.Finally,the voltage fluctuation is suppressed together with PR controller.The experimental results show that this method can effectively extract the characteristics of voltage data,and the accuracy of voltage fluctuation estimation is high.