Predicting single freestanding transmission tower time history response during complex wind input through a convolutional neural network based surrogate model. (15th April 2021)

• Record Type: Journal Article
• Authors: Xue, Jiayue; Xiang, Zhongming; Ou, Ge
• Is Part Of: Engineering structures. Volume 233(2021)
• Publication Date: 2021-04-15
• Subjects: Dynamic analysis -- Time history prediction -- Surrogate model -- Complex wind input -- Transmission tower; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2021.111859
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: A CNN-based surrogate model is used to predict transmission tower's time history. Complex wind is considered as the inducement of transmission tower. Configuration, window size and training data are discussed to optimize CNN network. The trained CNN can achieve RMSE around 0.01 for time history prediction. Abstract: As the steel towers in the power system are vulnerable to intensive wind loads, it is essential to understand their dynamics response to estimate its potential failure. Conventional structural analysis methods like the finite element analysis or the field test are either computational heavy or cost expensive. Thus, this paper proposes a machine learning approach based on convolutional neural network (CNN) to predict the time history response of the transmission tower during the complex wind input. By preprocessing the time history of wind load and the tower's dynamic response, a well-developed CNN can capture the time and spatial correlation of the wind load successfully and provide high accuracy results. CNN configuration, window size selection, and training data scale are carefully discussed to optimize the CNN design to maximize the prediction accuracy as well as minimize its computational time. Finally, to evaluate the performance of the surrogate model, the accuracy of the optimal CNN is tested in predicting the time history response of the transmission tower under 15 m/s to 70 m/s wind speed. The effectiveness of the CNN surrogate model is validated through a fragility model development, and its robustness is investigated using two wind inputs generated from a random wind profile and a random wind spectrum.