Environment, Coasts, Oceans, & Infrastructure

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

Hello Mark,

Thank you for sharing your story and your research idea-it's inspiring to see such dedication to improving infrastructure safety for your home community. Your AI-based triage concept using Random Forest is very promising. Machine learning models can indeed be used to prioritize inspections and identify high-risk bridges quickly, especially when data is limited. In practice, agencies like FEMA or local DOTs often use risk-based prioritization, so your approach aligns well with current strategies. However, real-world deployment would require sufficient quality and quantity of input data, integration with GIS and inspection databases, and validation against historical bridge failures.

Regarding bridge designs for coastal resilience, simplified precast or modular concrete structures can often be cost-effective for smaller municipalities. Elevated spans, flexible connections, and optimized pier designs help resist wave forces while keeping costs manageable. Collaboration with local engineering offices can help tailor designs to both budget and hazard level.

For your project, I recommend:

1. Start by building a small dataset of local bridges and their damage history.

2. Test your Random Forest model on this dataset to see if predictions correlate with observed damage.

3. Explore partnerships with professors or local engineers for guidance and validation.

4. Document your findings and lessons learned, as this can form part of your portfolio and demonstrate practical applicability.

Your personal connection to Puerto Rico and understanding of the local challenges will add a valuable perspective to your research. Keep iterating and testing-your idea could make a real impact.

Best regards,
Darya Stanskova

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

Dear Mark, your interest in addressing the structural vulnerability of bridges in Puerto Rico reflects both strong technical curiosity and a commendable commitment to public welfare.

The approach you describe-using machine learning techniques such as Random Forest models to support rapid screening or triage of infrastructure assets-is not only conceptually sound, but also aligned with current research trends in structural health monitoring and risk-informed decision-making. While such methods are not a replacement for detailed physics-based analyses, they are increasingly viewed as valuable tools for prioritization, especially in contexts where time, funding, and manpower are limited.

In this regard, I would like to point you to the work currently being developed by my colleague Dr. Carlos Arteta (<maskemail>carteta@...</maskemail>), who has been actively implementing and evaluating data-driven methodologies for structural damage identification and assessment. His experience may be particularly useful for you, both in terms of methodological choices and in understanding how these approaches can be realistically integrated with traditional engineering frameworks. I strongly encourage you to reach out to him to discuss possible alternatives and limitations relevant to your specific case.

The key challenge now is validation, transparency, and demonstrating that these tools can reliably inform-not replace-engineering decisions.

I wish you every success in your academic path and commend you for pursuing engineering solutions with real societal impact.

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

------------------------------
Mark Gibbs S.M.ASCE
Student
Massillon OH
------------------------------

Your background parallels mine ( Westchester CC and U of Arizona ). A Long bridge background and might spent money to stop traffic instead of tragic loss of life. Wish I was in warm weather ! Replying from cool NYC .

------------------------------

Original Message:
Sent: 01-15-2026 11:21 PM
From: Mark Gibbs
Subject: Student question: Using AI to help with bridge recovery for the people of Puerto Rico?

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

------------------------------
Mark Gibbs S.M.ASCE
Student
Massillon OH
------------------------------

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges

Mark – amid some great feedback and suggestions to your research question – here are something more I like to add.

As you know – Random Forest Model is a statistical tool based on ensemble analysis of large datasets incorporating individual Decision Trees. If not seen already – have a look at the 2024 NASA-NESC Technical Note – Statistical Analysis Using Random Forest Algorithm . . . It's a good source for clarity of understanding and approach.

On the question of training AI machine to learn the random forest model, and applying it – a NAP publication, the 2024 NAP 27902 Implementing and Leveraging Machine Learning at State DOTs, especially its Chapter 5 – will give you ample insights on AI applications and machine learning catering to DOT activities.

You may also look at my essays for more and to further your research: Artificial Intelligence – the Tool of No Limit and The World of Numbers and Chances.

--------------------

Dr. Dilip K Barua, PhD

Website Links and Profile

Hi everyone!

I'm a student member finishing my Associate's and transitioning to the University of Akron later this year. I'm a Phi Theta Kappa member, and in my free time, I enjoy playing chess and doing research to help me grow as a future engineer.

I've been looking into a major problem facing the people of Puerto Rico. The island's bridges have a D+ grade from ASCE. Because hurricanes happen every year, there is a constant cycle of damage. There is often not enough money or manpower to inspect every single bridge using complex simulations before the next storm hits. A 2024 U.S. GAO report shows the recovery is still struggling and costs are much higher than expected.

I'm trying to find a more strategic way to handle this. I've been reading research that suggests Random Forest (a type of machine learning) can be used to "triage" bridges in minutes instead of hours.

Here is the idea I'm looking at:

AI Triage: Using a Random Forest model to quickly flag which bridges are at high risk of washing away. This is much faster than doing a full physics simulation (FSI) for every single bridge.

Smart Construction: Since money is tight for many municipalities, faster and cheaper construction methods could be used for the bridges the AI identifies as high-risk so that local communities don't get hit as hard financially.

I'm just a student, so I'd really appreciate some help from the engineers here:

Does this actually sound doable in the real world, or is it too "out there" for agencies like FEMA or a local DOT?

Are there cheaper bridge designs you've seen work well against coastal waves that a small town could actually afford?

This hits very close to home for me. I am Puerto Rican and lived on the island from the age of five until 9th grade. I have seen firsthand what happens every year during hurricane season; one of those storms actually destroyed my house. I really want to find a way to help the communities back home stay safe.

Any advice or direction you can give a future Akron Zip would be awesome!

Supporting Research & Data:

https://2021.infrastructurereportcard.org/state-item/puerto-rico/ - 2019 Puerto Rico Infrastructure Report Card

https://www.sciencedirect.com/science/article/abs/pii/S0141029615004356?via%3Dihub -  Fragility surrogate models for coastal bridges in hurricane prone zones

https://www.researchgate.net/publication/383691493_Structural_analysis_of_the_economy_of_Puerto_Rico_after_the_elimination_of_Section_936_A_network-based_approach  -   Structural analysis of the economy of Puerto Rico after the elimination of Section 936: A network-based approach

Puerto Rico Disasters: Progress Made, but the Recovery Continues to Face Challenges