Machine learning techniques for diagrid building design - Architectural–Structural correlations with feature selection and data augmentation

Aug 5, 2026

Upload by Hazel

Artificial intelligence and machine learning methods are reshaping building engineering by linking design inputs to structural performance. This study targets tall diagrid buildings and investigates how machine learning can strengthen early design decisions. A small collected dataset is expanded via data augmentation, enabling classification of diagrid design feasibility. Key architectural and structural parameters are selected using multiple filter and wrapper methods. Results show effective generation of synthetic data, stable learning accuracies, and reliable relationships between architectural parameters and structural responses, supporting more effective high-rise diagrid design processes.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages27

Reading Time1 h 8 min

AccessFree to download

Introduction

AI and ML for early-stage tall building design

Parametric modeling and data generation

Machine learning roles in structural assessment

Focus on outer diagrid buildings

Article structure (from provided text)

Feature selection and data augmentation approach

Dataset enhancement and feasibility classification

Results: synthetic data quality and correlation robustness

Journal of Building Engineering 86 (2024) 108766 | Full length article
Machine learning techniques for diagrid building design: Architectural–Structural correlations with feature selection and data augmentation
Pooyan Kazemi ∗, Alireza Entezami, Aldo Ghisi
Department of Civil and Environmental Engineering, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy | | | | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | Keywords:
AI in building design Architectural feature selection Advanced data augmentation High-rise buildings Generative architectural forms
Design informed by structural insights | | Artificial intelligence (AI) and machine learning (ML) techniques are transforming building engineering. This work goes through the critical role of architectural parameters in influencing the structural responses of tall buildings, with a special focus on diagrid structures. The main aim of this study is to demonstrate how ML can improve the early design phase of diagrid buildings. Using a small, initially collected data set, enhanced through data augmentation, the classification of diagrid buildings in terms of design feasibility is investigated. This study identifies key architectural and structural parameters, employing various filter and wrapper methods for feature selection. The results show that our methods are effective in producing high-quality synthetic data, maintaining stable learning accuracies, and establishing accurate and robust relationships between architectural parameters and structural responses in diagrid buildings. These insights are crucial for facilitating more effective design processes in the realm of high-rise diagrid building design. | 1. Introduction The rapid rise of AI in recent years has opened new possibilities for architectural and structural designs of tall buildings. In the early phase of building design, conflicts can arise because visually appealing forms for architects may not align with structurally effective design choices. The key challenge lies in accurately estimating the structural behavior. This difficulty is particularly pronounced in the case of seismic loading, even moderate, and represents a significant challenge in the design process, which is generally unfavorable for all parties involved, including architects, engineers and clients. With the current availability of computing power, integrated design, and advanced tools capable of analyzing large data sets [1], it is feasible to propose an improved design approach [2,3]. Tools like parametric design software [4] allow specialists to create numerous geometries for high-rise building using computeraided design (CAD). This information can then be passed onto structural codes, which can generate extensive data to describe the response, even under complex loading conditions. AI tools, based on ML [5,6] or even deep learning (DL) [7–9], can process these results, identifying correlations between input and output variables with an efficiency that surpasses human capabilities. While this AI and ML-driven approach is promising, it is important to recognize that numerous practical details, both in the design process and in training the ML models, still necessitate human intervention for optimal results. Only through the correct application and interpretation of these procedures it is really possible to gain an advantage. ∗ Corresponding author. E-mail addresses: [seyedpooyan.kazemi@polimi.it](seyedpooyan.kazemi@polimi.it) (P. Kazemi), [alireza.entezami@polimi.it](alireza.entezami@polimi.it) (A. Entezami), [aldo.ghisi@polimi.it](aldo.ghisi@polimi.it) (A. Ghisi). [https://doi.org/10.1016/j.jobe.2024.108766](https://doi.org/10.1016/j.jobe.2024.108766) Received 14 August 2023; Received in revised form 4 January 2024; Accepted 6 February 2024 Available online 15 February 2024 2352-7102/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC B

FAQs

What is the main goal of the study on diagrid building design?

To demonstrate how machine learning can improve the early design phase of diagrid tall buildings by relating architectural parameters to structural behavior and assessing design feasibility.

How is the limited dataset handled in the proposed machine learning workflow?

The study enhances an initially small collected dataset using data augmentation to support training and analysis for feasibility classification.

Which techniques are used to identify influential parameters?

The work uses multiple filter and wrapper methods for feature selection to determine key architectural and structural parameters that drive the structural responses.

Aug 5, 2026

Upload by Hazel