TY - JOUR
T1 - Data-driven trajectory prediction of grid power frequency based on neural models
AU - Chamorro, Harold R.
AU - Orjuela-Cañón, Alvaro D.
AU - Ganger, David
AU - Persson, Mattias
AU - Gonzalez-Longatt, Francisco
AU - Alvarado-Barrios, Lazaro
AU - Sood, Vijay K.
AU - Martinez, Wilmar
N1 - Funding Information:
Funding: This work was supported by by the Laboratorio de Simulación Hardware-in-the-loop para Sistemas Ciberfísicos (LaSSiC). Código: 350202.
Publisher Copyright:
© 2021 by the authors.
Copyright:
Copyright 2021 Elsevier B.V., All rights reserved.
PY - 2021/1/12
Y1 - 2021/1/12
N2 - Frequency in power systems is a real-time information that shows the balance between generation and demand. Good system frequency observation is vital for system security and pro-tection. This paper analyses the system frequency response following disturbances and proposes a data-driven approach for predicting it by using machine learning techniques like Nonlinear Autoregressive (NAR) Neural Networks (NN) and Long Short Term Memory (LSTM) networks from simulated and measured Phasor Measurement Unit (PMU) data. The proposed method uses a horizon-window that reconstructs the frequency input time-series data in order to predict the frequency features such as Nadir. Simulated scenarios are based on the gradual inertia reduction by including non-synchronous generation into the Nordic 32 test system, whereas the PMU collected data is taken from different locations in the Nordic Power System (NPS). Several horizon-windows are experimented in order to observe an adequate margin of prediction. Scenarios considering noisy signals are also evaluated in order to provide a robustness index of predictability. Results show the proper performance of the method and the adequate level of prediction based on the Root Mean Squared Error (RMSE) index.
AB - Frequency in power systems is a real-time information that shows the balance between generation and demand. Good system frequency observation is vital for system security and pro-tection. This paper analyses the system frequency response following disturbances and proposes a data-driven approach for predicting it by using machine learning techniques like Nonlinear Autoregressive (NAR) Neural Networks (NN) and Long Short Term Memory (LSTM) networks from simulated and measured Phasor Measurement Unit (PMU) data. The proposed method uses a horizon-window that reconstructs the frequency input time-series data in order to predict the frequency features such as Nadir. Simulated scenarios are based on the gradual inertia reduction by including non-synchronous generation into the Nordic 32 test system, whereas the PMU collected data is taken from different locations in the Nordic Power System (NPS). Several horizon-windows are experimented in order to observe an adequate margin of prediction. Scenarios considering noisy signals are also evaluated in order to provide a robustness index of predictability. Results show the proper performance of the method and the adequate level of prediction based on the Root Mean Squared Error (RMSE) index.
UR - http://www.scopus.com/inward/record.url?scp=85099418991&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85099418991&partnerID=8YFLogxK
U2 - 10.3390/electronics10020151
DO - 10.3390/electronics10020151
M3 - Research Article
AN - SCOPUS:85099418991
SN - 2079-9292
VL - 10
SP - 1
EP - 21
JO - Electronics (Switzerland)
JF - Electronics (Switzerland)
IS - 2
M1 - 151
ER -