A hybrid machine learning framework for enhancing PMU-based event identification with limited labels

Haoran Li, Yang Weng, Evangelos Farantatos, Mahendra Patel

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The energy industry is experiencing rapid and dramatic changes on both the generator side and the load side, necessitating faster, more accurate, and robust event detection methods for situational awareness. Growing installations of PMU devices that provide high resolution synchronized measurements combined with the advancement of artificial intelligence and big data analytics techniques have recently attracted the RD community interest. Some supervised learning techniques have been proposed using PMU measurements, however, they are facing challenges in 1) limited interpretability, 2) biased learning models/results, and 3) insufficient labeled data for learning. To address these issues, we propose a machine learning-based framework for physically-meaningful interpretability, hybrid-learning method with indexes, and a flexible data-preparation approach. Specifically, a thoroughly designed feature selection method is proposed for discovering event signatures. Then, a hybrid machine learning process is constructed to reduce biases of different machine learners due to their diversified working mechanisms. Finally, we propose to utilize unlabeled data via semi-supervised learning and add strategical event data via active learning, e.g., simulations. The goal is to significantly improve the supervised learning results via computational efficient techniques. Extensive simulations are conducted using a commercial power system dynamics simulator and synthetic realistic transmission grid models. Significant improvements are observed via hybrid supervised learning methods, semi-supervised learning, and active learning.

Original languageEnglish (US)
Title of host publication2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728116075
DOIs
StatePublished - May 1 2019
Event2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019 - College Station, United States
Duration: May 20 2019May 23 2019

Publication series

Name2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019

Conference

Conference2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019
CountryUnited States
CityCollege Station
Period5/20/195/23/19

Fingerprint

Hybrid Learning
machine learning
Supervised learning
learning
Learning systems
Labels
Identification (control systems)
Machine Learning
Supervised Learning
Semi-supervised Learning
Active Learning
Interpretability
Event Detection
Situational Awareness
Learning Process
System Dynamics
Power System
Feature Selection
Artificial intelligence
Biased

Keywords

  • Active learning
  • Event identification
  • Feature selection
  • Hybrid learning
  • Phasor measurement unit
  • Semi-supervised learning

ASJC Scopus subject areas

  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering
  • Safety, Risk, Reliability and Quality
  • Numerical Analysis
  • Instrumentation

Cite this

Li, H., Weng, Y., Farantatos, E., & Patel, M. (2019). A hybrid machine learning framework for enhancing PMU-based event identification with limited labels. In 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019 [8784550] (2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SGSMA.2019.8784550

A hybrid machine learning framework for enhancing PMU-based event identification with limited labels. / Li, Haoran; Weng, Yang; Farantatos, Evangelos; Patel, Mahendra.

2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019. Institute of Electrical and Electronics Engineers Inc., 2019. 8784550 (2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, H, Weng, Y, Farantatos, E & Patel, M 2019, A hybrid machine learning framework for enhancing PMU-based event identification with limited labels. in 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019., 8784550, 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019, Institute of Electrical and Electronics Engineers Inc., 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019, College Station, United States, 5/20/19. https://doi.org/10.1109/SGSMA.2019.8784550
Li H, Weng Y, Farantatos E, Patel M. A hybrid machine learning framework for enhancing PMU-based event identification with limited labels. In 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019. Institute of Electrical and Electronics Engineers Inc. 2019. 8784550. (2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019). https://doi.org/10.1109/SGSMA.2019.8784550
Li, Haoran ; Weng, Yang ; Farantatos, Evangelos ; Patel, Mahendra. / A hybrid machine learning framework for enhancing PMU-based event identification with limited labels. 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019. Institute of Electrical and Electronics Engineers Inc., 2019. (2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019).
@inproceedings{d16dfbe3d32c4f10a142dad41cd216c1,
title = "A hybrid machine learning framework for enhancing PMU-based event identification with limited labels",
abstract = "The energy industry is experiencing rapid and dramatic changes on both the generator side and the load side, necessitating faster, more accurate, and robust event detection methods for situational awareness. Growing installations of PMU devices that provide high resolution synchronized measurements combined with the advancement of artificial intelligence and big data analytics techniques have recently attracted the RD community interest. Some supervised learning techniques have been proposed using PMU measurements, however, they are facing challenges in 1) limited interpretability, 2) biased learning models/results, and 3) insufficient labeled data for learning. To address these issues, we propose a machine learning-based framework for physically-meaningful interpretability, hybrid-learning method with indexes, and a flexible data-preparation approach. Specifically, a thoroughly designed feature selection method is proposed for discovering event signatures. Then, a hybrid machine learning process is constructed to reduce biases of different machine learners due to their diversified working mechanisms. Finally, we propose to utilize unlabeled data via semi-supervised learning and add strategical event data via active learning, e.g., simulations. The goal is to significantly improve the supervised learning results via computational efficient techniques. Extensive simulations are conducted using a commercial power system dynamics simulator and synthetic realistic transmission grid models. Significant improvements are observed via hybrid supervised learning methods, semi-supervised learning, and active learning.",
keywords = "Active learning, Event identification, Feature selection, Hybrid learning, Phasor measurement unit, Semi-supervised learning",
author = "Haoran Li and Yang Weng and Evangelos Farantatos and Mahendra Patel",
year = "2019",
month = "5",
day = "1",
doi = "10.1109/SGSMA.2019.8784550",
language = "English (US)",
series = "2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019",

}

TY - GEN

T1 - A hybrid machine learning framework for enhancing PMU-based event identification with limited labels

AU - Li, Haoran

AU - Weng, Yang

AU - Farantatos, Evangelos

AU - Patel, Mahendra

PY - 2019/5/1

Y1 - 2019/5/1

N2 - The energy industry is experiencing rapid and dramatic changes on both the generator side and the load side, necessitating faster, more accurate, and robust event detection methods for situational awareness. Growing installations of PMU devices that provide high resolution synchronized measurements combined with the advancement of artificial intelligence and big data analytics techniques have recently attracted the RD community interest. Some supervised learning techniques have been proposed using PMU measurements, however, they are facing challenges in 1) limited interpretability, 2) biased learning models/results, and 3) insufficient labeled data for learning. To address these issues, we propose a machine learning-based framework for physically-meaningful interpretability, hybrid-learning method with indexes, and a flexible data-preparation approach. Specifically, a thoroughly designed feature selection method is proposed for discovering event signatures. Then, a hybrid machine learning process is constructed to reduce biases of different machine learners due to their diversified working mechanisms. Finally, we propose to utilize unlabeled data via semi-supervised learning and add strategical event data via active learning, e.g., simulations. The goal is to significantly improve the supervised learning results via computational efficient techniques. Extensive simulations are conducted using a commercial power system dynamics simulator and synthetic realistic transmission grid models. Significant improvements are observed via hybrid supervised learning methods, semi-supervised learning, and active learning.

AB - The energy industry is experiencing rapid and dramatic changes on both the generator side and the load side, necessitating faster, more accurate, and robust event detection methods for situational awareness. Growing installations of PMU devices that provide high resolution synchronized measurements combined with the advancement of artificial intelligence and big data analytics techniques have recently attracted the RD community interest. Some supervised learning techniques have been proposed using PMU measurements, however, they are facing challenges in 1) limited interpretability, 2) biased learning models/results, and 3) insufficient labeled data for learning. To address these issues, we propose a machine learning-based framework for physically-meaningful interpretability, hybrid-learning method with indexes, and a flexible data-preparation approach. Specifically, a thoroughly designed feature selection method is proposed for discovering event signatures. Then, a hybrid machine learning process is constructed to reduce biases of different machine learners due to their diversified working mechanisms. Finally, we propose to utilize unlabeled data via semi-supervised learning and add strategical event data via active learning, e.g., simulations. The goal is to significantly improve the supervised learning results via computational efficient techniques. Extensive simulations are conducted using a commercial power system dynamics simulator and synthetic realistic transmission grid models. Significant improvements are observed via hybrid supervised learning methods, semi-supervised learning, and active learning.

KW - Active learning

KW - Event identification

KW - Feature selection

KW - Hybrid learning

KW - Phasor measurement unit

KW - Semi-supervised learning

UR - http://www.scopus.com/inward/record.url?scp=85071380183&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071380183&partnerID=8YFLogxK

U2 - 10.1109/SGSMA.2019.8784550

DO - 10.1109/SGSMA.2019.8784550

M3 - Conference contribution

AN - SCOPUS:85071380183

T3 - 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019

BT - 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, SGSMA 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -