skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Early Prediction of Power Outage Duration Through Hierarchical Spatiotemporal Multiplex Networks
Long power outages caused by weather can have a big impact on the economy, infrastructure, and quality of life in affected areas. It’s hard to provide early and accurate warnings for these disruptions because severe weather often leads to missing weather recordings, making it difficult to make learning-based predictions. To address this challenge, we have developed HMN-RTS, a hierarchical multiplex network that classifies disruption severity by temporal learning from integrated weather recordings and social media posts. This new framework’s multiplex network layers gather information about power outages, weather, lighting, land cover, transmission lines, and social media comments. Our study shows that this method effectively improves the accuracy of predicting the duration of weather-related outages. The HMN-RTS model improves 3 h ahead outage severity prediction, resulting in a 0.76 macro F1-score vs 0.51 for the best alternative for a five-class problem formulation. The HMN-RTS model provides useful predictions of outage duration 6 h ahead, enabling grid operators to implement outage mitigation strategies promptly. The results highlight the HMN-RTS’s ability to offer early, reliable, and efficient risk assessment.  more » « less
Award ID(s):
2125985
PAR ID:
10625012
Author(s) / Creator(s):
; ; ; ;
Editor(s):
Cherifi, H; Donduran, M; Rocha, LM; Cherifi, C; Varol, O
Publisher / Repository:
Springer Nature Switzerland
Date Published:
ISSN:
1860949X
ISBN:
978-3-031-82434-0
Page Range / eLocation ID:
320 to 334
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; (, Springer Verlag)
    Iliadis, L; Maglogiannis, I; Kyriacou, E; Jayne, C (Ed.)
    Weather-related power disruptions present significant challenges to public infrastructure, societal well-being, and the distribution grid. Predicting outage durations in distribution grids is another challenge compared to transmission line outage durations due to distribution networks’ complexity and finer granularity. While forecasting forced power outages is crucial, accurately estimating their duration is essential for timely response and mitigation measures. This study introduces the Spatiotemporal Multiplex Network (SMN-WVF), a methodology designed to predict power outage durations across varying lead times, tackling the difficulties posed by small, high-complexity spaces within distribution grids. SMN-WVF employs multiplex networks that incorporate multi-modal data across both time and space, including layers such as power outages, weather conditions, weather forecasts, vegetation, and distances between substations. We demonstrate the importance of incorporating additional layers of data sources as they are shown to help the model’s predictions through gradual improvement in the macro F1 score performance. 
    more » « less
  2. ; ; ; (, Natural Hazards)
    Large-scale damage to the power infrastructure from hurricanes and high-wind events can have devastating ripple effects on infrastructure, the broader economy, households, com- munities, and regions. Using Hurricane Irma’s impact on Florida as a case study, we examined: (1) differences in electric power outages and restoration rates between urban and rural counties; (2) the duration of electric power outages in counties exposed to tropical storm force winds versus hurricane Category 1 force winds; and (3) the rela- tionship between the duration of power outage and socioeconomic vulnerability. We used power outage data for the period September 9, 2017–September 29, 2017. At the peak of the power outages following Hurricane Irma, over 36% of all accounts in Florida were without electricity. We found that the rural counties, predominantly served by rural electric cooperatives and municipally owned utilities, experienced longer power outages and much slower and uneven restoration times. Results of three spatial lag models show that large percentages of customers served by rural electric cooperatives and municipally owned utilities were a strong predictor of the duration of extended power outages. There was also a strong positive association across all three models between power outage duration and urban/rural county designation. Finally, there is positive spatial dependence between power outages and several social vulnerability indicators. Three socioeconomic variables found to be statistically significant highlight three different aspects of vulnerability to power outages: minority groups, population with sensory, physical and mental disability, and economic vulnerability expressed as unemployment rate. The findings from our study have broader planning and policy relevance beyond our case study area, and highlight the need for additional research to deepen our understanding of how power restoration after hurricanes contributes to and is impacted by the socioeconomic vulnerabilities of communities. 
    more » « less
  3. (, The Hawaii International Conference on System Sciences – HICSS, Maui, Hawaii, January 2019)
    This paper describes the implementation of a prediction model for real-time assessment of weather related outages in the electric transmission system. The network data and historical outages are correlated with a variety of weather sources in order to construct the knowledge extraction platform for accurate outage probability prediction. An extension of the logistic regression prediction model that embeds the spatial configuration of the network was used for prediction. The results show that the developed model manifests high accuracy and is able to differentiate an outage area from the rest of the network in 1 to 3 hours before the outage. The prediction model is integrated inside a weather testbed for real-time mapping of network outage probabilities based on incoming weather forecast. 
    more » « less
  4. ; ; ; ; (, Nature energy)
    Climate-induced extreme weather events, as well as other natural and human-caused disasters, have the potential to increase the duration and frequency of large power outages. Resilience, in the form of supplying a small amount of power to homes and communities, can mitigate outage consequences by sustaining critical electricity-dependent services. Public decisions about investing in resilience depend, in part, on how much residential customers value those critical services. Here we develop a method to estimate residential willingness-to-pay for back-up electricity services in the event of a large 10-day blackout during very cold winter weather, and then survey a sample of 483 residential customers across northeast USA using that method. Respondents were willing to pay US$1.7–2.3/kWh to sustain private demands and US$19–29/day to support their communities. Previous experience with long-duration outages and the framing of the cause of the outage (natural or human-caused) did not affect willingness-to-pay. 
    more » « less
  5. ; (, Energies)
    Power outage prediction is important for planning electric power system response, restoration, and maintenance efforts. It is important for utility managers to understand the impact of outages on the local distribution infrastructure in order to develop appropriate maintenance and resilience measures. Power outage prediction models in literature are often limited in scope, typically tailored to model extreme weather related outage events. While these models are sufficient in predicting widespread outages from adverse weather events, they may fail to capture more frequent, non-weather related outages (NWO). In this study, we explore time series models of NWO by incorporating state-of-the-art techniques that leverage the Prophet model in Bayesian optimization and hierarchical forecasting. After defining a robust metric for NWO (non-weather outage count index, NWOCI), time series forecasting models that leverage advanced preprocessing and forecasting techniques in Kats and Prophet, respectively, were built and tested using six years of daily state- and county-level outage data in Massachusetts (MA). We develop a Prophet model with Bayesian True Parzen Estimator optimization (Prophet-TPE) using state-level outage data and a hierarchical Prophet-Bottom-Up model using county-level data. We find that these forecasting models outperform other Bayesian and hierarchical model combinations of Prophet and Seasonal Autoregressive Integrated Moving Average (SARIMA) models in predicting NWOCI at both county and state levels. Our time series trend decomposition reveals a concerning trend in the growth of NWO in MA. We conclude with a discussion of these observations and possible recommendations for mitigating NWO. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email