Revolutionizing Structural Engineering with ARTIFICIAL INTELLIGENCE

Key Concepts

AI in structural engineering, process automation, forensic engineering, generative design, machine learning, digital workflows, interoperability, design optimization, data processing, large language models (LLMs), return on investment (ROI), legal risks, automation scripting, data security, intellectual property protection, search and retrieval, cost minimization, talent acquisition, incremental implementation, continuous learning.

Background and Journey to SML Consulting

Jesse Light, President/Senior Structural Engineer at SML Consulting Structural and Forensic Engineers, shares his journey. He initially pursued physics before obtaining a BS in Mechanical Engineering from Cal Poly Pomona in 2001. After 9/11 impacted aerospace opportunities, he worked in track home construction and earned his PE in 2006. He joined SML in 2007 after a stint at Orbital Sciences. He became an associate in 2012 (requiring an SE, obtained in 2014), a partner in 2000, VP in 2003, and President in 2024. Becoming president allowed him to allocate resources and implement technology-driven ideas.

Involvement with NCSEA AI Grant Team and NSF IUCRC Visor Center

Jesse is involved with the NCSEA AI grant team and will present at the upcoming summit in New York (October 14th-17th). He also joined the NSF IUCRC Visor Center, focused on AI in structural engineering.

• NSF IUCRC Visor Center: Visor is the name of the structural engineering artificial intelligence center. IUCRC is the designation given by the National Science Foundation for these industry-university cooperative research centers. The University of Houston is teaming up with Purdue University and NSF. The goal is to get 6-10 industry leaders to invest in the center.
• Funding Model: Industry partners invest (e.g., $75,000), universities contribute (e.g., $200,000), and NSF adds funding (e.g., $300,000), creating a pool of ~$1.3 million for R&D. Investors get access to research outcomes and influence the research direction.

Jesse's involvement stemmed from a Reddit post that led to a summer internship with a PhD student from the University of Houston's structural engineering and artificial intelligence lab (SALE).

The Role of AI in Structural Engineering and Project Delivery

Jesse believes AI and process automation are the future of engineering. The key is not whether AI will replace structural engineers, but that structural engineers who leverage AI will replace those who don't.

Examples of AI and Automation in Practice

1. Forensic Report Parsing: The intern created a tool to search through old forensic reports (Word documents) by converting them into JSON files and vectorizing them for keyword searches. This addressed the challenge of accessing information after the original owners of SML retired.
2. Automated Workflow for Repeat Projects: For a client with a repetitive product, Jesse automated the design process using RISA 3D, Excel, and Mathcad.
   • RISA 3D models are exported to Excel.
   • Mathcad updates loads in Excel and creates a new model.
   • Mathcad reads RISA 3D list reports (matrix form) for connection checks.
   • Hilty is used for FA analysis of column welds, base plates, stiffener plates, and anchor bolts.
   • Results: Reduced design time from 8-10 hours to a fraction of that, optimized designs, reduced material usage (e.g., base plate thickness reduced by 1/4 inch), and lowered shipping costs.
3. Automated Letter Generation: Using Microsoft Power Automate and ChatGPT, Jesse created a system that automatically populates site addresses on client letters. A form is filled out, the system finds the correct PDF, adds the address, flattens the PDF, digitally signs it, and emails it to the client. This process takes two minutes.

Generative Design, Machine Learning, and Emerging Technologies

Jesse sees value in generative design and machine learning for early-stage design development and feasibility studies. These technologies can help optimize structural designs and accommodate organic architectural designs.

• Leaders should encourage teams to use AI for automation scripting: AI can help connect the dots in a process, but the core calculations should remain under engineer control.
• Example: Using ChatGPT to troubleshoot scripting for the automated letter generation process.

Data Security and Intellectual Property Protection

Jesse emphasizes the importance of protecting intellectual property, especially with the ease of modifying PDFs. He digitally signs documents to prevent unauthorized alterations. He has experienced instances of stolen seals, rewritten letters, and copied plans being resubmitted to different jurisdictions. Blockchain is mentioned as a potential solution for signing and sealing.

AI in Forensic Engineering

Jesse believes AI can influence investigations and post-event analysis.

• Data Processing: High-powered tools are needed to process data from laser cloud point data collection and drone technology (LiDAR).
• Large Language Models (LLMs): LLMs can be trained to identify trouble areas in structures (e.g., cracked concrete, corroded steel) more quickly and accurately than humans.
• Consortium for Data Sharing: The structural engineering community could form a consortium to share anonymized data for training AI models.
• Example: Scanning a concrete parking structure and using LLMs to identify cracked areas.

Overcoming Hurdles to AI Adoption

The biggest hurdles are cost, ROI concerns, and legal risks.

• Cost Minimization:
  • Hire interns with data science and Python coding skills.
  • Negotiate with clients to fund AI-related projects.
  • Use international talent platforms like Top Tall for cheaper programming resources.
  • Hire engineers with both engineering and programming skills (e.g., mechanical engineers).
• Return on Investment (ROI):
  • Start with small steps and build incrementally.
  • Use systems you already know (e.g., Excel, Mathcad, Intercalc, RISA 3D, AutoCAD).
  • Connect existing tools and templates.
  • Leverage cloud-based tools with API access (e.g., BQE Core).
• Legal Risks:
  • Focus AI on process automation and scripting, while maintaining control over core calculations.

Final Advice

Start now, even with small steps. Don't expect miracles and be prepared for failures along the way. Continuous learning and incremental progress are key.

Resources Mentioned

• NCSEA AI Grant Team
• NSF IUCRC Visor Center
• University of Houston's Structural Engineering and Artificial Intelligence Lab (SALE)
• RISA 3D
• Mathcad
• Hilty
• Microsoft Power Automate
• ChatGPT/Claude
• Top Tall
• BQE Core
• AIC (American Institute of Steel Construction) AI assistant
• NCSEA's Structure magazine archive search
• LinkedIn (for connecting with Jesse Light)
• SML Consulting Structural and Forensic Engineers website

Synthesis/Conclusion

The interview with Jesse Light highlights the transformative potential of AI in structural and forensic engineering. While challenges like cost, legal risks, and data security exist, Jesse provides actionable strategies for firms of all sizes to begin integrating AI into their workflows. Key takeaways include starting small, leveraging existing tools, focusing on process automation, and continuously learning. By embracing AI and process automation, structural engineers can enhance efficiency, optimize designs, and deliver greater value to their clients.