An integrated damage modeling and assessment framework for overhead power distribution systems considering tree-failure risks

• Published in: Structure and infrastructure engineering Vol. 19; no. 12; pp. 1745 - 1760
• Main Authors: Lu, Qin, Zhang, Wei
• Format: Journal Article
• Language: English
• Published: Taylor & Francis 02.12.2023
• Subjects: CNN-based image classifier; Computer vision technique; fragility analysis; machine learning; physics-based damage modeling; power distribution systems; strong winds; tree failure risk
• ISSN: 1573-2479; 1744-8980
• DOI: 10.1080/15732479.2022.2053552

The overhead power distribution system (OPDS) is vulnerable to strong winds, such as hurricanes. Due to the challenges of including tree damage risks to the OPDS, tree failures are usually ignored in the risk assessment of the OPDS against strong winds. In the present study, an integrated damage modeling and assessment framework for the OPDS is proposed considering tree failure risks. The geographical information of trees surrounding the OPDS is extracted from satellite images using computer vision techniques, including CNN-based (convolutional neural network) image classifier and sliding window approach. The tree failure risk models are developed using tree geographical information in conjunction with tree height data, tree allometry and finite element analysis. With further integration of the conditional probability failure of poles under fallen tree impacts, the pole's failure probability considering the combined wind and fallen trees is obtained using series system reliability analysis. The failure probability of the pole is obtained using physics-based modeling facilitated by Bayesian regularisation neural network (BRNN) algorithm. The poles and wires are connected for system reliability assessment using connectivity-based theory. When the wind direction is counterclockwise from the east and the wind speed is 57 m/s, tree-failure can introduce 68.6% differences in OPDS' failure probabilities compared with that without consideration of fallen trees.