Optimum location belt truss system and pattern recognition with neural network based on belt truss system

Belt truss systems have gained significant traction in the past two decades for their effectiveness in mitigating lateral loads caused by wind and earthquakes in tall buildings. This research investigates the structural behavior of belt trusses in high-rise buildings, focusing on their potential to reduce lateral displacement through movement and energy dissipation. The study complements the existing body of research by employing a neural network (NN) for optimal placement of the belt truss structure. The investigation analyzes two steel frame structures, a ۳۰-story and a ۴۰-story building, incorporating belt truss systems. Through displacement and weight balance analysis, the optimal location for the belt truss in each structure is identified. Furthermore, a pattern recognition-based neural network is developed using input and target functions derived from structural analysis software. This NN offers a computationally efficient alternative for analyzing the vast number of potential configurations (approximately ۱۴۰۰ states) for these high-dimensional structures with belt trusses. Compared to traditional software, the proposed NN approach has the potential to deliver more accurate results, including optimal truss location, weight, and movement targets, within a significantly reduced timeframe. The research is structured in two sections: the first section explores the belt truss system and its impact on structural behavior, while the second section delves into the development and application of the neural network system. These sections work synergistically to provide a comprehensive approach to optimizing belt truss placement in high-rise buildings.

کلیدواژه ها:

Belt truss system ، Target function ، Input functions ، Neural network ، Optimum weight ، Displacement