基于Attention-LSTM的有载调容变压器运行方式优化研究  

Optimization of On-load Adjustable Capacity Transformer Operation Mode Based on Attention-LSTM

作　　者：杨景亮 齐林海 陶顺[2] 王红 ANG Jingliang;QI Linhai;TAO Shun;WANG Hong(School of Control and Computer Engineering,North China Electric Power University,Changping District,Beijing 102206,China;School of Electrical and Electronic Engineering,North China Electric Power University,Changping District,Beijing 102206,China)

机构地区：[1]华北电力大学控制与计算机工程学院,北京市昌平区102206 [2]华北电力大学电气与电子工程学院,北京市昌平区102206

出　　处：《电网技术》2020年第7期2449-2456,共8页Power System Technology

基　　金：国家自然科学基金项目(51777066)。

摘　　要：调容点和调容策略的制定决定了有载调容变压器的工作效率,但现行采用的升降容计时器控制方式仍存在很大弊端。提出了以序电流负荷为基础的有载调容变压器调容判据、引入注意力机制的长短期记忆网络的有载调容变压器运行方式优化方法,从调容点选择和运行方式规划两方面降低其在配电网中产生的损耗。通过对实际数据的分析,验证了文中所提出的方法的有效性,在实现有载调容变的节能降损和复杂环境下安全运行方面有明显提升。The formulation of the capacity adjustment point and the capacity adjustment strategy determine the working efficiency of the on-load adjustable capacity transformer,but the current control method of the lift-down timer still has a lot of drawbacks.In this paper,an optimization method for operation mode of on-load adjustable transformer is proposed with the long short-term memory network in attention mechanism based on the sequence current.It reduces the loss in the distribution network from the selection point and in the operation mode planning.Through analysis with the actual data,the validity of the proposed method is verified,and it has significant improvement in energy-saving and loss reduction and the safe operation under complicated environments.

关 键 词：电能替代 有载调容变压器 序电流负荷预测 长短期记忆网络 注意力机制 

分 类 号：TM721[电气工程—电力系统及自动化]