Xinzheng Lu

Keynote Lecture at the 18th World Conference on Earthquake Engineering (WCEE2024): From "Simulation-driven" urban disaster mitigation to "Generative AI-powered" seismic design. July 3, 2024 youtu.be/UdqIeSUVV_c?si… via @YouTube

Milestone 3,500 real projects! AI-structure Copilot now: • One-click CAD recognition (DL-powered) • Export PKPM/YJK models in 1 step • Smarter shear wall (incl. fused AI models) • Stronger frame design (half/secondary beams) Faster. More compliant. Engineer-approved.

New feature of AIstructure-Copilot v0.4.0 (ai-structure.com): One-click, deep-learning-powered extraction of architectural CAD drawing information. Box-select the architectural floor plan to be designed, then click once to have the AI recognize components (doors, windows,…

Excited to share our new paper: an AI framework that auto-designs fire-sprinkler layouts for any floorplan. We've extended Generative AI from structural to building services design. Our GAN-based model automates fire sprinkler layouts, boosting efficiency by 76% & cutting costs…

The latest version of the generative AI design software for building structures, AIstructure-Copilot V0.3.8 (ai-structure.com), has enabled better design of inclined beam members.

On July 20, 2025, the Structural Engineering Committee of the Chinese Society of Theoretical and Applied Mechanics organized the session "Intelligent Structural Design and Optimization" (S14) at the Chinese Congress of Theoretical and Applied Mechanics-2025.

City-scale disaster simulation and resilience assessment: From physics-based method to AI method youtu.be/P1AIJxVFHQQ Presentation at the 2nd SIP International Seminar, Earthquake Research Institute, the University of Tokyo, Japan.

Generative-AI Design and Intelligent Optimization of Building Structures youtu.be/9DuFKkFt2CM?si… Semi-plenary lecture at COMPSAFE2025, Kobe, Japan.

AIstructure-Copilot Version V0.3.7 (ai-structure.com) enhances its frame structure design module with AI design functionality for partial frame beams and secondary beams, resulting in more consistent design results that align with practical engineering requirements.…

Our latest research addresses a critical bottleneck for generative AI in structural design: low-quality training data. We propose a data enhancement method using structural optimization to refine shear wall layout datasets, powering diffusion models. This results in a 67%…

Earthquake-induced falling debris from building envelopes poses a remarkable threat to public safety. Our new paper reveals that building collisions during earthquakes significantly worsen infill wall damage, thereby increasing the risk of falling debris. A testing methodology…

New review paper: Leveraging data-driven artificial intelligence in optimization design for building structures. Data-driven AI opens new horizons in structural design optimization by efficiently managing complex constraints, accelerating performance evaluations, and generating…

Last week, the number of engineering design projects completed by ai-structure.com surpassed 3000. Since the official launch of AIstructure-Copilot in July 2023, the number of actual online engineering design projects reached 2000 by December 2024. Six months later, the…

AIstructure-Copilot-V0.3.6: Added intelligent design function for concrete frame structures youtu.be/F7XpFg9MKJ0?si… Since the launch of AIstructure-Copilot's intelligent design function for shear wall structures in July 2023, after nearly two years of development, the design of…

This work demonstrates that generative AI design is applicable not only to buildings but also to infrastructure. Utilizing image-parameter fusion GANs for intelligent zoning design of concrete-faced rockfill dams boosts efficiency 10×. This method reduces design time from hours…

AI-driven deep learning can predict safe egress times in complex building fires with 95% accuracy. The developed new framework analyzes occupant density and exit usage in seconds, eliminating hours of manual simulation. Journal of Building Engineering doi.org/10.1016/j.jobe…

Generative AI is transforming the way we analyze fire scenarios in complex building layouts by accurately predicting smoke flow and temperature distribution with high accuracy. By reducing simulation time from hours to minutes, these models significantly alleviate the challenges…

On Mar/31, the updated Chinese national standard "Code for Residential Projects" officially required that the story height of new residential buildings shall not be less than 3 m. To address this important update, we quickly responded and made targeted upgrades to…

This is the most interesting case we have seen recently, and also the one that best reflects the value of ai-structure.com. This is the 2777th engineering project designed by —a residential building that is 42.9 m tall with a seismic design…

Version upgrade, AIstructure-Copilot-V0.3.4: Integration of Multiple Algorithms, Offering a New Algorithm Trial Feature With the progress of our research, the algorithm library for AI design has become increasingly rich. We have integrated various algorithms, complementing each…