Wind power prediction based on High-frequency SCADA data along with isolation forest and deep learning neural networks

Zi Lin, Xiaolei Liu, Maurizio Collu

Research output: Contribution to journalArticle

Abstract

Wind power plays a key role in reducing global carbon emission. The power curve provided by wind turbine manufacturers offers an effective way of presenting the global performance of wind turbines. However, due to the complicated dynamics nature of offshore wind turbines, and the harsh environment in which they are operating, wind power forecasting is challenging, but at the same time vital to enable condition monitoring (CM). Wind turbine power prediction, using supervisory control and data acquisition (SCADA) data, may not lead to the optimum control strategy as sensors may generate non-calibrated data due to degradation. To mitigate the adverse effects of outliers from SCADA data on wind power forecasting, this paper proposed a novel approach to perform power prediction using high-frequency SCADA data, based on isolate forest (IF) and deep learning neural networks. In the predictive model, wind speed, nacelle orientation, yaw error, blade pitch angle, and ambient temperature were considered as input features, while wind power is evaluated as the output feature. The deep learning model has been trained, tested, and validated against SCADA measurements. Compared against the conventional predictive model used for outlier detection, i.e. based on Gaussian Process (GP), the proposed integrated approach, which coupled IF and deep learning, is expected to be a more efficient tool for anomaly detection in wind power prediction.
Original languageEnglish
Article number105835
Number of pages10
JournalInternational Journal of Electrical Power & Energy Systems
Volume118
Early online date10 Jan 2020
DOIs
Publication statusE-pub ahead of print - 10 Jan 2020

Fingerprint

Wind power
Data acquisition
Neural networks
Wind turbines
Offshore wind turbines
Condition monitoring
Deep learning
Degradation
Carbon
Sensors
Temperature

Keywords

  • wind power prediction
  • deep learning neural networks
  • isolation forest
  • outlier detection
  • offshore wind turbines

Cite this

@article{b3fd433f5be14e9b8313225c52569f15,
title = "Wind power prediction based on High-frequency SCADA data along with isolation forest and deep learning neural networks",
abstract = "Wind power plays a key role in reducing global carbon emission. The power curve provided by wind turbine manufacturers offers an effective way of presenting the global performance of wind turbines. However, due to the complicated dynamics nature of offshore wind turbines, and the harsh environment in which they are operating, wind power forecasting is challenging, but at the same time vital to enable condition monitoring (CM). Wind turbine power prediction, using supervisory control and data acquisition (SCADA) data, may not lead to the optimum control strategy as sensors may generate non-calibrated data due to degradation. To mitigate the adverse effects of outliers from SCADA data on wind power forecasting, this paper proposed a novel approach to perform power prediction using high-frequency SCADA data, based on isolate forest (IF) and deep learning neural networks. In the predictive model, wind speed, nacelle orientation, yaw error, blade pitch angle, and ambient temperature were considered as input features, while wind power is evaluated as the output feature. The deep learning model has been trained, tested, and validated against SCADA measurements. Compared against the conventional predictive model used for outlier detection, i.e. based on Gaussian Process (GP), the proposed integrated approach, which coupled IF and deep learning, is expected to be a more efficient tool for anomaly detection in wind power prediction.",
keywords = "wind power prediction, deep learning neural networks, isolation forest, outlier detection, offshore wind turbines",
author = "Zi Lin and Xiaolei Liu and Maurizio Collu",
year = "2020",
month = "1",
day = "10",
doi = "10.1016/j.ijepes.2020.105835",
language = "English",
volume = "118",
journal = "Electrical Power and Energy Systems",
issn = "0142-0615",

}

TY - JOUR

T1 - Wind power prediction based on High-frequency SCADA data along with isolation forest and deep learning neural networks

AU - Lin, Zi

AU - Liu, Xiaolei

AU - Collu, Maurizio

PY - 2020/1/10

Y1 - 2020/1/10

N2 - Wind power plays a key role in reducing global carbon emission. The power curve provided by wind turbine manufacturers offers an effective way of presenting the global performance of wind turbines. However, due to the complicated dynamics nature of offshore wind turbines, and the harsh environment in which they are operating, wind power forecasting is challenging, but at the same time vital to enable condition monitoring (CM). Wind turbine power prediction, using supervisory control and data acquisition (SCADA) data, may not lead to the optimum control strategy as sensors may generate non-calibrated data due to degradation. To mitigate the adverse effects of outliers from SCADA data on wind power forecasting, this paper proposed a novel approach to perform power prediction using high-frequency SCADA data, based on isolate forest (IF) and deep learning neural networks. In the predictive model, wind speed, nacelle orientation, yaw error, blade pitch angle, and ambient temperature were considered as input features, while wind power is evaluated as the output feature. The deep learning model has been trained, tested, and validated against SCADA measurements. Compared against the conventional predictive model used for outlier detection, i.e. based on Gaussian Process (GP), the proposed integrated approach, which coupled IF and deep learning, is expected to be a more efficient tool for anomaly detection in wind power prediction.

AB - Wind power plays a key role in reducing global carbon emission. The power curve provided by wind turbine manufacturers offers an effective way of presenting the global performance of wind turbines. However, due to the complicated dynamics nature of offshore wind turbines, and the harsh environment in which they are operating, wind power forecasting is challenging, but at the same time vital to enable condition monitoring (CM). Wind turbine power prediction, using supervisory control and data acquisition (SCADA) data, may not lead to the optimum control strategy as sensors may generate non-calibrated data due to degradation. To mitigate the adverse effects of outliers from SCADA data on wind power forecasting, this paper proposed a novel approach to perform power prediction using high-frequency SCADA data, based on isolate forest (IF) and deep learning neural networks. In the predictive model, wind speed, nacelle orientation, yaw error, blade pitch angle, and ambient temperature were considered as input features, while wind power is evaluated as the output feature. The deep learning model has been trained, tested, and validated against SCADA measurements. Compared against the conventional predictive model used for outlier detection, i.e. based on Gaussian Process (GP), the proposed integrated approach, which coupled IF and deep learning, is expected to be a more efficient tool for anomaly detection in wind power prediction.

KW - wind power prediction

KW - deep learning neural networks

KW - isolation forest

KW - outlier detection

KW - offshore wind turbines

U2 - 10.1016/j.ijepes.2020.105835

DO - 10.1016/j.ijepes.2020.105835

M3 - Article

VL - 118

JO - Electrical Power and Energy Systems

JF - Electrical Power and Energy Systems

SN - 0142-0615

M1 - 105835

ER -