Machine learning evaluation of storm-related transmission outage factors and risk

Taylor, William O.; Nyame, Sita; Hughes, William; Koukoula, Marika; Yang, Feifei; Cerrai, Diego; Anagnostou, Emmanouil N.

doi:10.1016/j.segan.2023.101016

Cited by 6 publications

(2 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most machine learning algorithms are unable to perform well with highly imbalanced data and require additional algorithms to solve these problems [36]. To address this issue while following the literature [37], in which similar circumstances are faced, we converted the target variable into a binary response variable ("0 = 27,411", "1 = 22,071") and predicted the probability that a given observation/outage occurrence belonged to one of the two classes, which are any outage presence or absence.…”

Section: Machine Learning Datasetmentioning

confidence: 99%

Non-Parametric Machine Learning Modeling of Tree-Caused Power Outage Risk to Overhead Distribution Powerlines

Wedagedara,

Witharana,

Fahey

et al. 2024

Applied Sciences

Self Cite

View full text Add to dashboard Cite

Trees in proximity to power lines can cause significant damage to utility infrastructure during storms, leading to substantial economic and societal costs. This study investigated the effectiveness of non-parametric machine learning algorithms in modeling tree-related outage risks to distribution power lines at a finer spatial scale. We used a vegetation risk model (VRM) comprising 15 predictor variables derived from roadside tree data, landscape information, vegetation management records, and utility infrastructure data. We evaluated the VRM’s performance using decision tree (DT), random forest (RF), k-Nearest Neighbor (k-NN), extreme gradient boosting (XGBoost), and support vector machine (SVM) techniques. The RF algorithm demonstrated the highest performance with an accuracy of 0.753, an AUC-ROC of 0.746, precision of 0.671, and an F1-score of 0.693. The SVM achieved the highest recall value of 0.727. Based on the overall performance, the RF emerged as the best machine learning algorithm, whereas the DT was the least suitable. The DT reported the lowest run times for both hyperparameter optimization (3.93 s) and model evaluation (0.41 s). XGBoost and the SVM exhibited the highest run times for hyperparameter tuning (9438.54 s) and model evaluation (112 s), respectively. The findings of this study are valuable for enhancing the resilience and reliability of the electric grid.

show abstract

Section: Machine Learning Datasetmentioning

confidence: 99%

Non-Parametric Machine Learning Modeling of Tree-Caused Power Outage Risk to Overhead Distribution Powerlines

Wedagedara,

Witharana,

Fahey

et al. 2024

Applied Sciences

Self Cite

View full text Add to dashboard Cite

show abstract

“…Different from traditional analysis methods, machine learning algorithm digs and dissect multidimensional data in a data-driven way, and builds a power grid fault prediction and evaluation model based on multishadow factors. William O. Taylor et al [2] developed a random forest model to assess the significance and correlation between vegetation, weather, infrastructure, physical environment, and storm-induced transmission line outages, and demonstrated that machine learning can be used to improve transmission line outage risk prediction. However, the above machine learning algorithm lacks the ability of temporal feature mining, so it needs to rely on artificial prior knowledge for feature extraction.…”

Section: Introductionmentioning

confidence: 99%

Substation flood control risk prediction based on improved neural network

Chen,

Xia,

2024

Eighth International Conference on Energy System, Electricity, and Power (ESEP 2023)

View full text Add to dashboard Cite

With the frequent extreme rainstorm and flood events at present, it is necessary to carry out flood control transformation of substation in operation to avoid its damage by flood. By identifying and evaluating risks and taking corresponding pre-control measures, the risk of substation suffering from flooding can be reduced to the greatest extent. This paper presents a method for predicting substation flood control risk. Firstly, the finite element method (FEM) is used to preprocess the historical data of substation flood control. Secondly, the improved neural network model is constructed, and part of the data is used to complete the model training. Finally, the actual substation data verification shows that the accuracy of the model is higher than that of the traditional neural network model, which is a simple and effective method.

show abstract

Assessing grid hardening strategies to improve power system performance during storms using a hybrid mechanistic-machine learning outage prediction model

Hughes,

Watson,

Cerrai

et al. 2024

Reliability Engineering & System Safety

View full text Add to dashboard Cite

Machine learning evaluation of storm-related transmission outage factors and risk

Cited by 6 publications

References 32 publications

Non-Parametric Machine Learning Modeling of Tree-Caused Power Outage Risk to Overhead Distribution Powerlines

Non-Parametric Machine Learning Modeling of Tree-Caused Power Outage Risk to Overhead Distribution Powerlines

Substation flood control risk prediction based on improved neural network

Assessing grid hardening strategies to improve power system performance during storms using a hybrid mechanistic-machine learning outage prediction model

Contact Info

Product

Resources

About