基于改进孪生神经网络的低频低压减载装置在线解耦控制方法
On Line Decoupling Control Method of Low Frequency and Low Voltage Load Shedding Device Based on Improved Siamese Neural Network
于立强 1刘佳言 1李露 1霍晓燕 1李世杰2
作者信息
- 1. 国网冀北承德供电公司,河北 承德 067000
- 2. 北京科东电力控制系统有限责任公司,北京 100192
- 折叠
摘要
低频低压减载装置的过度应用导致装置系统出现负荷大的问题,降低了低频低压减载装置的实用性.为此,提出了基于改进孪生神经网络的低频低压减载装置在线解耦控制方法.该方法通过分析低频低压减载装置频率特性,为低频低压减载装置在线解耦控制提供了重要信息基础;在此基础上构建低频低压减载装置混合模型,从该模型中生成可训练的数据集,并依据潮流雅可比矩阵的减载特征最小值灵敏度,确定低频低压减载装置系统减载量;为提升系统的控制精度,将获取结果输入至改进孪生神经网络中,训练出最佳减载量,并将其分配到装置系统减载节点中,达到切除负荷的目的,实现低频低压减载装置在线解耦控制.通过对该方法开展减载效率对比测试、控制性能对比测试,实验结果验证了该方法具有较强的实用性.
Abstract
The excessive application of low frequency and low voltage load shedding devices leads to the problem of large load in the device system,which reduces the practicability of low frequency and low voltage load shedding devices.In order to solve this problem,an online decoupling control method of low frequency and low voltage load shedding devices based on improved twin neural network is proposed.By analyzing the frequency characteristics of low frequency and low voltage load shedding devices,this method provides important information for online decoupling control of low frequency and low voltage load shedding devices;On this basis,a hybrid model of low frequency and low voltage load shedding device is constructed,from which a trainable data set is generated,and the load shedding amount of low frequency and low voltage load shedding device system is determined according to the minimum sensitivity of load shedding characteristics of power flow Jacobian ma-trix;In order to improve the control accuracy of the system,the obtained results are input to the improved twin neural net-work,the optimal load shedding amount is trained,and distributed to the load shedding node of the device system,so as to achieve the purpose of cutting off the load,and realize the online decoupling control of low-frequency and low-voltage load shedding devices.The experimental results show that the method has strong practicability through the comparative test of load shedding efficiency and control performance.
关键词
改进孪生神经网络/低频低压减载装置/在线解耦控制/潮流雅可比矩阵Key words
improved Siamese neural network/low frequency and low voltage load shedding device/on line decoupling control/power flow Jacobian matrix引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据