Fault detection framework using neural networks for condition monitoring of high voltage equipment in power grids swegrids-logo

SweGRIDS research area Controllable Power Components
SweGRIDS project code CPC18
Project type PostDoc
Status running
Researcher Yue Cui   (webpage)
University KTH (EPE)
Project period 2021-02-01 to 2021-12-31   
Project supervisor Lina Bertling Tjernberg   (webpage)
Industrial sponsors Svenska kraftnät

back to project index

Project abstract

The scientific challenge is to propose a fault detection framework to analyze operation data and log files in the supervisory control and data acquisition systems. In ongoing research such a framework has been developed for wind turbines. It uses log files and each event is mapped to an assembly based on the IEA Reliawind taxonomy. The generated results can help to understand when the turbine deviates from healthy behaviors. For the operation data, a semi-supervised approach is built based on recurrent neural networks to model the healthy behavior, which can learn the longtime temporal dependencies between various time series signals. Based on the estimation results, a two-stage threshold method is proposed to trigger alarms indicating potential operation risks. The method evaluates both the shift values away from the healthy behaviors and their duration time to determine the current health conditions. The assessment framework is validated with the data from a Swedish onshore wind park. The numerical results show that the framework can notify operators of potential operation risks in the near future and reduce false alarms. In this project the proposed framework would be applied to power grids. The project contributes to solutions for developing the future sustainable energy system. The proposed framework would support in lowering the cost for the operation and result in an overall lower cost for investments.

Summary of work


Event log


Project reference-group

Nilanga Abeywickrama,  Hitachi ABB Power Grids
Michele Luvisotto,  Hitachi ABB Power Grids
Jan-Henning Juergensen,  Hitachi ABB Power Grids
Cristian Rojas,  KTH
Göran N. Ericsson,  Svenska kraftnät
Jørn Egil Johnsen,  Statnett, Norway

Publications by this researcher

See alternatively the researcher's full DiVA list of publications, with options for sorting.
Publications in journals and conferences usually will not show until a while after they are published.

A Fault Detection Framework Using Recurrent Neural Networks for Condition Monitoring of Wind Turbines
Yue Cui.
2021,   Thesis (PhD), KTH Royal Institute of Technology, TRITA-EECS-AVL 2021:4

A fault detection framework using RNNs for condition monitoring of wind turbines
Yue Cui,   Pramod Bangalore,   Lina Bertling.
2021,   Wind Energy

Wind Turbine Health Assessment Framework Based on Power Analysis Using Machine Learning Method
Qiuyi Huang,   Yue Cui,   Lina Bertling,   Pramod Bangalore.
2019,   2019 IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2019, Bucharest, Romania, September 29 - October 2, 2019

An anomaly detection approach based on machine learning and scada data for condition monitoring of wind turbines
Yue Cui,   Pramod Bangalore,   Lina Bertling Tjernberg.
2018,   2018 International Conference on Probabilistic Methods Applied to Power Systems, PMAPS 2018, 24 June 2018 through 28 June 2018

An Anomaly Detection Approach Using Wavelet Transform and Artificial Neural Networks for Condition Monitoring of Wind Turbines' Gearboxes
Yue Cui,   Pramod Bangalore,   Lina Bertling Tjernberg.

Applying High Performance Computing to Probabilistic Convex Optimal Power Flow
Zhao Yuan,   Mohammad Reza Hesamzadeh,   Yue Cui,   Lina Bertling Tjernberg.
2016,   International Conference on Probabilistic Methods Applied to Power Systems (PMAPS), OCT 16-20, 2016, Beijing, PEOPLES R CHINA

Publication list last updated from DiVA on 2021-09-18 22:00.

back to project index

Page started: 2021-02-01
Last generated: 2021-09-18