Abnormal Diagnosis of Dam Safety Monitoring Data Based on Ensemble Learning

2019 3rd International Workshop on Renewable Energy and Development

能源学;生态环境科学

Jun, Zhang^1^2 ; Jiemin, Xie^1^2 ; Pangao, Kou^1

State Grid Hunan Electric Power Company Limited Research Institute, Changsha

410007, China^1

Hunan Xiangdian Test and Research Institute, Changsha

410007, China^2

关键词: Abnormal data;    Base learners;    Ensemble learning;    Gross errors;    Measured values;    Real-time data;    Stepwise regression;    Time points;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/267/6/062027/pdf
DOI  :  10.1088/1755-1315/267/6/062027

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

Screening out the gross errors and systematic errors of dam safety monitoring data by theoretical hypothesis will lead to the risk of misjudgment of abnormal data. In order to reduce this risk, based on the ensemble learning method in machine learning, this article extracts and integrates multiple base learners from the stepwise regression model, and proposes a matrix of abnormal indexes based on real-time data update, and analyzes the abnormal diagnosis of the measured data subsequently. The results show that the abnormal indexes have a strong practicability, which don't need to screen out the data with systematic errors and gross errors, and can effectively identify the abnormal time points and the degree of interference between the measured values.