Abstract
This study presents a framework for evaluating the vulnerability of the electrical grid to storm outages, based on multi-year atmospheric reanalysis datasets. The underlying methodology encompasses the classification of outage event severity and machine learning-based outage prediction models (OPMs), toward identifying relevant weather events and quantifying the associated outages within the European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5 and ERA5-Land) records. The proposed framework is tested for the Eversource Energy distribution grid over the State of Connecticut, for a period from 1981 to 2020. Within this context, and using as benchmark outage data reported by the utility for the period from 2005 to 2020, the accuracy of the classification for events of high-impact and extreme-severity proved to be high (i.e., 0.84 and 0.95, respectively). Especially for the latter case, the OPMs exhibited acceptable mean absolute percentage errors and high coefficient of determination (R2) values. Further, an analysis based on the annual maxima of the total number of outages, as well as the number of events with outages above various thresholds, indicated an intensification of extreme events over the last decade. Within this context, and given its importance to long-term planning and investment, we ultimately assess the potential impact of climate change on the resilience of the distribution grid, by evaluating the non-exceedance probabilities of six historical hurricanes that impacted the Eversource Energy service territory in Connecticut through a parametric statistical approach.
Similar content being viewed by others
References
Amanulla B, Chakrabarti S, Singh S (2012) Reconfiguration of power distribution systems considering reliability and power loss. IEEE Trans Power Deliv 27(2):918–926
Arif A, Wang Z (2018) Distribution network outage data analysis and repair time prediction using deep learning. Paper presented at the 2018 IEEE international conference on probabilistic methods applied to power systems (PMAPS)
Balkema AA, de Haan L (1974) Residual lifetime at great age. Ann Probab 2:792–804. https://doi.org/10.1214/aop/1176996548
Benjamin BR, Cornell AC (1970) Probability, statistics, and decision for civil engineers. 1st edn. McGraw-Hill, Inc., New York, p 64. ISBN: 978–0070045491
Campbell RJ, Lowry S (2012) Weather-related power outages and electric system resiliency
Castillo A (2014) Risk analysis and management in power outage and restoration: a literature survey. Electr Power Syst Res 107:9–15
Caves DW, Herriges JA, Windle RJ (1990) Customer demand for service reliability in the electric power industry: a synthesis of the outage cost literature 1. Bull Econ Res 42(2):79–121
Cerrai D, Watson P, Anagnostou EN (2019) Assessing the effects of a vegetation management standard on distribution grid outage rates. Electr Power Syst Res 175:105909
Coles S (2001) An introduction to statistical modeling of extreme values, Springer London, p 209. ISBN: 978–1–84996–874–4. https://doi.org/10.1007/978-1-4471-3675-0
Coulston JW, Moisen GG, Wilson BT, Finco MV, Cohen WB, Brewer CK (2012) Modeling percent tree canopy cover: a pilot study. Photogramm Eng Remote Sens 78(7):715–727
Cox DR, Isham V (1980) Point processes. CRC Press, p 188. ISBN 978–0412219108
Davison AC, Smith RL (1990) Models for exceedances over high thresholds. J R Statist Soc B 52(3):393–442. https://doi.org/10.1111/j.2517-6161.1990.tb01796.x
Deng Z, Singh C (1992) A new approach to reliability evaluation of interconnected power systems including planned outages and frequency calculations. IEEE Trans Power Syst 7(2):734–743
Dobson I, Carreras BA (2010) Number and propagation of line outages in cascading events in electric power transmission systems. Paper presented at the 2010 48th annual allerton conference on communication, control, and computing (Allerton)
Duffey RB (2019) Power restoration prediction following extreme events and disasters. Int J Disaster Risk Sci 10(1):134–148
Duffey RB, Ha T (2009) Predicting electric power system restoration. Paper presented at the 2009 IEEE toronto international conference science and technology for humanity (TIC-STH)
Embrechts P, Klüppelberg C, Mikosch T (1997) Modelling extremal events for insurance and finance, p 167. ISBN: 9783540609315
Emmanouil S, Langousis A, Nikolopoulos EI, Anagnostou EN (2020) Quantitative assessment of annual maxima, peaks-over-threshold and multifractal parametric approaches in estimating intensity-duration-frequency curves from short rainfall records. J Hydrol 589:125151. https://doi.org/10.1016/j.jhydrol.2020.125151
Emmanouil S, Langousis A, Nikolopoulos EI, Anagnostou EN (2022) The spatiotemporal evolution of rainfall extremes in a changing climate: a conus-wide assessment based on multifractal scaling arguments. Earth’s Future 10(3):e2021EF002539. https://doi.org/10.1029/2021EF002539
ERA5-land hourly data from 1981 to present. https://doi.org/10.24381/cds.e2161bac
Eskandarpour R, Khodaei A (2016) Machine learning based power grid outage prediction in response to extreme events. IEEE Trans Power Syst 32(4):3315–3316
Eskandarpour R, Khodaei A (2017) Leveraging accuracy-uncertainty tradeoff in SVM to achieve highly accurate outage predictions. IEEE Trans Power Syst 33(1):1139–1141
Greenwood JA, Landwhr JM, Matalas NC, Wallis JR (1979) Probability weighted moments: definition and relation to parameters of several distributions expressable in inverse form. Water Resour Res 15(5):1049–1054. https://doi.org/10.1029/WR015i005p01049
Guha S, Moss A, Naor J, Schieber B (1999) Efficient recovery from power outage. Paper presented at the proceedings of the thirty-first annual ACM symposium on Theory of computing
He J, Cheng MX (2021) Machine learning methods for power line outage identification. Electr J 34(1):106885
He J, Cheng MX, Fang Y, Crow ML (2018) A machine learning approach for line outage identification in power systems. Paper presented at the international conference on machine learning, optimization, and data science
Hersbach H, Bell B, Berrisford P, Hirahara S, Horányi A, Muñoz-Sabater J, Schepers D (2020) The ERA5 global reanalysis. Q J R Meteorol Soc 146(730):1999–2049
Ho TK (1995) Random decision forests. Paper presented at the proceedings of 3rd international conference on document analysis and recognition
Homer C, Dewitz J, Jin S, Xian G, Costello C, Danielson P, Stehman S (2020) Conterminous United States land cover change patterns 2001–2016 from the 2016 national land cover database. ISPRS J Photogramm Remote Sens 162:184–199
Hosking JRM, Wallis JR (1987) Parameter and quantile estimation for the generalized pareto distribution. Technometrics 29(3):339–349. https://doi.org/10.2307/1269343
Hosking JRM, Wallis JR (1997) Regional frequency analysis: an approach based on L-moments. Cambridge University Press, UK. https://doi.org/10.1017/cbo9780511529443
Iešmantas T, Alzbutas R (2019) Bayesian spatial reliability model for power transmission network lines. Electr Power Syst Res 173:214–219
Kenward A, Raja U (2014) Blackout: extreme weather climate change and power outages. Clim Cent 10:1–23
Küfeoğlu S, Prittinen S, Lehtonen M (2014) A summary of the recent extreme weather events and their impacts on electricity. Int Rev Electr Eng (IREE) 9(4):821–828
LaCommare KH, Eto JH (2006) Cost of power interruptions to electricity consumers in the United States (US). Energy 31(12):1845–1855
Langousis A, Mamalakis A, Puliga M, Deidda R (2016) Threshold detection for the generalized Pareto distribution: review of representative methods and application to the NOAA NCDC daily rainfall database. Water Resour Res 52(4):2659–2681. https://doi.org/10.1002/2015WR018502
Leadbetter MR, Lindgren G, Rootzen H (1983) Extremes and related properties of random sequences and series, Springer series in statistics, 1st edn., Springer, NY, p 336. https://doi.org/10.1007/978-1-4612-5449-2
Liu H, Davidson RA, Apanasovich TV (2007) Statistical forecasting of electric power restoration times in hurricanes and ice storms. IEEE Trans Power Syst 22(4):2270–2279
Mills E (2012) Electric grid disruptions and extreme weather. Lawrence Berkeley National Laboratory
Muñoz-Sabater J, Dutra E, Agustí-Panareda A, Albergel C, Arduini G, Balsamo G, Hersbach H (2021) ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. Earth Syst Sci Data Discuss, pp 1–50
Nateghi R, Guikema SD, Quiring SM (2014) Forecasting hurricane-induced power outage durations. Nat Hazards 74(3):1795–1811
NCEI N (2018) US billion-dollar weather and climate disasters: Retrieved 22/10/18 from https://www. ncdc. noaa. gov/billions
Panteli M, Mancarella P (2015) Influence of extreme weather and climate change on the resilience of power systems: impacts and possible mitigation strategies. Electr Power Syst Res 127:259–270
Pickands J (1975) Statistical inference using extreme order statistics. Ann Statist 3(1):119–131. https://doi.org/10.1214/aos/1176343003
Sheridan SC, Zhang W, Deng X, Lin S (2021) The individual and synergistic impacts of windstorms and power outages on injury ED visits in New York State. Sci Total Environ. 797: 149199
Solutions CfCaE (2018) resilience strategies for power outages
Swaminathan S, Sen RK (1998) Review of power quality applications of energy storage systems: Sandia National Laboratories
Touzani S, Granderson J, Fernandes S (2018) Gradient boosting machine for modeling the energy consumption of commercial buildings. Energy Build 158:1533–1543
Walsh T, Layton T, Wanik D, Mellor J (2018) Agent based model to estimate time to restoration of storm-induced power outages. Infrastructures 3(3):33
Wanik D, Anagnostou E, Hartman B, Frediani M, Astitha M (2015) Storm outage modeling for an electric distribution network in northeastern USA. Nat Hazards 79(2):1359–1384
Wu J, Wang P (2021) Post-disruption performance recovery to enhance resilience of interconnected network systems. Sustain Resil Infrastruct 6(1–2):107–123
Xiao J, Zhang W, Huang M, Lu Y, Lawrence WR, Lin Z, Tan W (2021) Increased risk of multiple pregnancy complications following large-scale power outages during Hurricane Sandy in New York State. Sci Total Environ 770:145359
Xie Y, Li C, Lv Y, Yu C (2019) Predicting lightning outages of transmission lines using generalized regression neural network. Appl Soft Comput 78:438–446
Yang F, Wanik DW, Cerrai D, Bhuiyan MAE, Anagnostou EN (2020a) Quantifying uncertainty in machine learning-based power outage prediction model training: a tool for sustainable storm restoration. Sustainability 12(4):1525
Yang F, Watson P, Koukoula M, Anagnostou EN (2020b) Enhancing weather-related power outage prediction by event severity classification. IEEE Access 8:60029–60042
Yang F, Cerrai D, Anagnostou EN (2021) The effect of lead-time weather forecast uncertainty on outage prediction modeling. Forecasting 3(3):501–516
Yodo N, Arfin T (2021) A resilience assessment of an interdependent multi-energy system with microgrids. Sustain Resil Infrastruct 6(1–2):42–55
Yue M, Toto T, Jensen MP, Giangrande SE, Lofaro R (2017) A Bayesian approach-based outage prediction in electric utility systems using radar measurement data. IEEE Trans Smart Grid 9(6):6149–6159
Zorn CR, Shamseldin AY (2015) Post-disaster infrastructure restoration: A comparison of events for future planning. Int J Disaster Risk Reduct 13:158–166
Acknowledgements
The Authors would like to thank Eversource Energy for their kind support, as well as Prof. Andreas Langousis for his valuable insights. Reported outage data for the State of Connecticut were obtained from Eversource Energy and can be available with the necessary permission from the company. We would also like to thank the Editor and two anonymous Reviewers for their constructive feedback and recommendations, which enhanced the quality of the presented work.
Funding
This work was supported by the Electric Power Research Institute and Eversource Energy.
Author information
Authors and Affiliations
Contributions
FY designed the study’s integrated framework, developed the proposed machine learning and return period scheme, performed the analysis and assessed the results, as well as wrote the main manuscript. MK contributed to the weather data analysis and review of the manuscript. SE developed the extreme value model presented herein, conducted the respective analysis and assessed the results, and wrote the manuscript. DC co-designed the study’s framework and contributed to the writing and review of the manuscript. EN Anagnostou advised on the project, conceptualized the research framework, acquired the funding, and contributed to the review of the manuscript. All authors have read and agreed upon this version of the study.
Corresponding author
Ethics declarations
Conflict of interest
The Authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this study.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
1.1 Variables used in the QWD classification and OPM
See Table
6.
1.2 Evaluation metrics for the storm event classification
For a given class label X (i.e., high or extreme), terms are defined as follows:
-
True Positive (TP): Observation is label X and is predicted as label X.
-
False Positive (FP): Observation is not label X but is predicted as label X.
-
False Negative (FN): Observation is label X but is not predicted as label X.
-
True Negative (TN): Observation is not label X and is not predicted as label X.
Precision is expressed as the proportion of events correctly predicted as label X, over all events predicted as label X:
For a given class label X, Recall is the proportion of events correctly predicted as label X, over the observed events of label X:
Then, the F1 score is the harmonic mean of Precision and Recall, given equal weights:
1.3 Evaluation metrics for the OPM validation
We use the Absolute Error (AE) to measure the difference between the actual (\({\text{o}}_{\text{i}}\)) and predicted (\({\text{p}}_{\text{i}}\)) outage totals for the service territory from each event (i). The first, second, and third quantiles of the sorted Absolute Errors are represented as AE Q25, AE Q50, and AE Q75, respectively. AE can be estimated as follows:
The first, second, and third quantiles of the sorted Absolute Percentage Errors (APE) are denoted as APE q25, APE q50, and APE q75, respectively. APE is estimated as follows:
We, also, employ the Mean Absolute Percentage Error (MAPE), which is estimated as follows:
Moreover, we use the Centered Root-Mean-Squared Error (CRMSE) to quantify both systematic and random error components:
Further, we utilize the coefficient of determination (R2) to assess the goodness-of-fit of the acquired model predictions to the actual outages:
Finally, we employ the Nash–Sutcliffe model efficiency coefficient (NSE), which ranges between negative infinite and 1:
1.4 Variable importance for the GBM-based and RF-based OPMs
See Fig.
10.
1.5 Model performance of the OPMs for the high-impact and and extreme-severity events
See Table
7.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Yang, F., Koukoula, M., Emmanouil, S. et al. Assessing the power grid vulnerability to extreme weather events based on long-term atmospheric reanalysis. Stoch Environ Res Risk Assess 37, 4291–4306 (2023). https://doi.org/10.1007/s00477-023-02508-y
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00477-023-02508-y