Integrating the Finite-Element Method and Artificial Intelligence for Optimal Design of Transmission Tower Structures

Abstract

Overhead power lines are more economical to install than underground systems, and high voltages are frequently transmitted across the transmission network. This paper aims to introduce a hybrid optimization approach integrating the FEM and artificial neural networks (ANNs) to optimize the design of transmission tower structures. This research seeks to develop a practical ANN-based predictive model to predict the optimal weight of transmission line towers based on base width, wind span, and deviation angle. A method is developed in this paper for optimizing transmission line tower structures in two types of systems (self-supported suspension towers and self-supported tension towers) under limits of optimal weight. A numerical parametric study utilizing the FEM model was used to accomplish these goals. Seventy-two transmission line tower models were designed and developed using STAAD Pro V8i SS6 software. The results of the parametric research were utilized in the process of training and validating the ANN prediction model that was developed. In this study, the fit is quite good across all data sets, with R2-values of 0.979 for self-supported suspension towers and 0.976 for self-supported tension towers, according to the research findings. Additional machine learning models, including support vector machines (SVM) and random forest, were evaluated for comparative accuracy. The study also implemented Monte Carlo Simulation to assess uncertainty and propose future enhancements by incorporating deep learning models for structural optimization. Furthermore, a sensitivity analysis was conducted to examine the impact of material type, member cross section, and structural height on weight optimization. Results indicated that integrating artificial intelligence (AI)-based predictive modeling significantly enhances the efficiency of structural design processes.