合肥工业大学校徽 合肥工业大学学报自科版

导航菜单

基于 GM-RBF 不定权组合模型的输电线塔杆倾斜预测分析与应用

下载全文

Prediction analysis and application of transmission line tower inclination using GM-RBF uncertain weight combination model

期刊信息

合肥工业大学（自然科学版），2023年6月，第46卷第6期：788-794

DOI: 10.3969/j.issn.1003-5060.2023.06.011

作者信息

王洪武 $ ^{1} $，李俊鹏 $ ^{1} $，张继伟 $ ^{1} $，黄然 $ ^{2} $，朱宇 $ ^{3} $，宋宝 $ ^{4} $

（1. 云南电网有限责任公司输电分公司，云南昆明 650033；2. 云南电网有限责任公司电力科学研究院，云南昆明 650217；3. 北京空间飞行器总体设计部，北京 100094；4. 安徽理工大学空间信息与测绘工程学院，安徽淮南 232001）

摘要和关键词

摘要: 为了解决目前输电线塔杆倾斜姿态监测中出现的预测不准确、不及时和预测误报率高等问题,文章提出一种基于灰色模型-径向基函数(grey model-radial basis function, GM-RBF)不定权组合模型的输电线塔杆倾斜姿态预测方法,对昆明市某地区一处输电塔杆200 d的北斗逆向网络载波相位差分技术(real-time kinematic, RTK)数据,使用GM-RBF不定权组合预测模型对铁塔姿态进行预测。该方法不仅能有效规避灰色模型(grey model, GM)自身误差大的缺点,减弱神经网络中训练样本随机性对建模精度的影响,还可以消除因最小二乘定权组合影响整体模型精度的问题。实验表明:对于短期塔杆倾斜预测,GM-RBF不定权组合预测模型在X、Y、P向和倾斜角的预测精度与GM预测精度相当,优于径向基函数(radial basis function, RBF)神经网络模型和GM-RBF定权组合模型的精度;对于长期塔杆倾斜预测,GM-RBF不定权组合模型在X、Y、P向和倾斜角的预测精度分别优于GM预测模型约57.28%、48.07%、43.02%、42.08%,优于RBF预测模型约2.04%、2.31%、3.60%、2.02%,优于GM-RBF最小二乘定权组合模型约2.97%、2.36%、6.23%、4.73%。

关键词: 北斗逆向网络载波相位差分技术(RTK)数据;输电线塔杆;倾斜姿态监测;GM-RBF不定权组合模型

Authors

WANG Hongwu $ ^{1} $, LI Junpeng $ ^{1} $, ZHANG Jiwei $ ^{1} $, HUANG Ran $ ^{2} $, ZHU Yu $ ^{3} $, SONG Bao $ ^{4} $

(1. Transmission Branch, Yunnan Power Grid Co., Ltd., Kunming 650033, China; 2. Electric Power Research Institute, Yunnan Power Grid Co., Ltd., Kunming 650217, China; 3. Beijing Institute of Spacecraft System Engineering, Beijing 100094, China; 4. School of Spatial Information and Geomatics Engineering, Anhui University of Science and Technology, Huainan 232001, China)

Abstract and Keywords

Abstract: In order to solve the problems of inaccurate prediction, untimely prediction and high false alarm rate in the current inclination attitude monitoring of transmission line tower, this paper proposes a method of inclination attitude prediction of transmission line tower based on grey model-radial basis function(GM-RBF) uncertain weight combination model. For the 200 d real-time kinematic(RTK) data of BeiDou reverse network of a tower in a certain area of Kunming City, the attitude of the tower is predicted based on the GM-RBF uncertain weight combination prediction model. This method can not only effectively avoid the shortcomings of GM itself, and reduce the influence of the randomness of training samples in neural network on the modeling accuracy, but also eliminate the problem that the accuracy of the whole model is affected by the least-squares fixed weight combination. The results show that in the short-term inclination prediction of transmission tower, the accuracy of GM-RBF uncertain weight combination prediction model is about the same as that of GM prediction model in the prediction of X, Y and P directions and overall inclination angle of transmission tower, which is better. than those of RBF neural network prediction model and fixed weight combination prediction model based on least squares method; in the long-term inclination prediction of transmission tower, the accuracy of GM-RBF uncertain weight combination prediction model is about 57.28%, 48.07%, 43.02% and 42.08% better than that of GM prediction model in the prediction of X, Y and P directions and overall inclination angle of transmission tower, respectively. It is about 2.04%, 2.31%, 3.60% and 2.02% better than that of RBF prediction model and about 2.97%, 2.36%, 6.23% and 4.73% better than that of fixed weight combination prediction model based on least squares method, respectively.

Keywords: BeiDou reverse network real-time kinematic (RTK) differential data; transmission line tower; inclination attitude monitoring; grey model-radial basis function (GM-RBF) uncertain weight combination model

基金信息

江苏省“六大人才高峰”高层次人才资助项目（XYDDX-045）；西宁市科技资助项目（2019-Y-12）和无锡市科技发展资金资助项目（N20201011）

前一篇 后一篇
返回期刊目录

个人中心

投稿登录 审稿登录 编辑登录