5 Ways A.I. and Software Will Change Structural Engineering

[@ing.baltazarteran]

📌Here are the most important lessons from the video: 1. Structural engineers will be able to do more with less: Software, such as Building Information Modeling (BIM) and finite element analysis software, is already being used in the structural engineering industry. These programs can quickly analyze structures and their structural behavior, allowing engineers to perform calculations much faster and analyze different structural options in less time. However, software and AI will still be tools of the structural engineer, who will still be the designer and make the important decisions. 2. There will be more complex structures: With the help of software, more complex and landmark structures can be designed. However, it's also important not to blindly rely on software for design, but to understand the basic behavior of structures. 3. Back to the basics of structural engineering: As equations become more complex, it's more important to understand the fundamentals of structural engineering, such as statics, free body diagrams, shear and moment diagrams, mechanics of materials, and the behavior of materials. 4. Structural engineers will need to adapt: With AI and software doing much of the automated design, structural engineers will be able to do more creative and important things. This could involve a larger role in the preliminary design phase, a greater emphasis on structural analysis, or even learning to program to improve automation. 5. Structural engineering could become a commodity: This will depend on how the structural engineering industry adapts to AI and software. If the industry simply markets the ability to do more with less without changing fees, it could devalue the profession. Instead, the industry should demonstrate to clients how value is added through the use of this type of software. In summary, although AI and software are changing the structural engineering industry, they will not replace the structural engineer in the short term. Engineering judgment, which is hard to replace, will continue to be a crucial component of the profession.

[@steevkeyes]

Structural engineers already make extensive use of classical (non-AI) software. I believe none of the 5 points you make are pertaining to artificial intelligent (AI) software. Popular examples (which I and my staff use) but which are classical programs include: Excel, Mathcad, Staadpro, Etabs, RISA, Sap, Autodesk suite, etc. In the late 1980's I was one of the programmers of a inhouse program (called SODA) that we developed and tried to market. I would classify it as being quasi-AI. It optimized the shape of a truss, within the constraints of the outside boundary. It did the optimization via taking a numerical derivative of the stiffness matrix for the truss. Then iterated changes to the panel locations by self-imposed changes to the stiffness matrix via changes of the panel locations. It worked; we did market it but with little commercial success at the time - I believe because the structural engineering marketplace was still at its infancy.

[@hdb-1999]

Great video Mat, I am glad you are talking about this. I don't think enough people in the industry think about this subject.

[@PedroLimeira]

Man, it's great to see Mola Models on your shelf, as a Brazilian I'm happy that Mola Models won the world.

[@emmanuelakintayo3977]

Wow. This is an eye opening video. It gives me an idea of what structural engineering would be in the future.

[@erichansen2860]

I am excited for computational and parametric design technology. Having gone from UBC in the 90s to the much more technical IBC to ASCE 7/ACI/AISC/NDS exploding in size and complexity, and all the advancement we've made in seismic design, we NEED a Jarvis to keep our work organized. It will also help us provide our clients with more fact-based information at the early stages of projects.

[@breakfastcrew]

i think companies like autodesk will start acquiring a bunch of design firms in various markets to train and sell AI models that can pretty much design and self check. That will eliminate most technical and advisory roles. The only roles left are client facing roles dealing with project management and business development.

[@erichansen2860]

I think it depends on the industry. Single family residence and strip malls in non seismic zones could definitely be designed by AI with no input from a structural engineer other than check and stamp, and Autodesk could provide all the tools needed. At scale, however, larger companies that do massive projects are building their own custom workflows as opposed to using canned products like Autodesk and Bentley. Those products are good but they have their limits. A custom workflow means that an engineer can jump in and provide the type of input only a human can when something awkward has to be navigated with human judgment.

[@MatPicardal]

I think that time, like with a lot of the industries, may eventually come, but who knows what will happen. Will the public be OK with AI completely designing their structures? Who then, will be legally responsible if the building collapses? If engineers with experience are the ones checking the AI software design, how will new engineers gain the experience to check if they don't do the design anymore? Let's wait and find out!

[@breakfastcrew]

@@MatPicardal the ai trainer will be legally responsible. In my example Autodesk would be legally responsible. I also think larger design firms with a sizable R&D budget might lease a more general ai model from say openai and train their own specific version for civil engineering.

[@michaelweyenberg6238]

Or architecture firms and contractors will just have structural engineering departments. Some big arch companies are already doing that.

[@bradmesser4807]

Mission critical software yes needs to be thoroughly checked triple checked and tested so many ways from Sunday, it'll make your head spin and there'd be a responsibility to ensure an accurate set of testing has been made available to check what needs checked. Software cannot assume the end liability, but you can see the chain of responsibility I hope.

[@coryblake5623]

that "well the model showed this" is a HUGE problem already and we're seeing it more and more of it.. I asked a younger engineer the other day to check something via hand calc and was told "oh wow.. i haven't done a hand calc in long time.... ".... keep in mind, I'm only 38 and I go through engineering paper like CRAZY... how are we that different in our approach when we're only 8 years separated in age... kids these days.. haha

[@MatPicardal]

How do you think AI and software will change the structural engineering industry? Will the public be OK with AI completely designing their structures? Who then, will be legally responsible if the building collapses? If engineers need design experience to check the AI software design, how will new engineers even gain design experience if they don't do the design anymore?

[@erichansen2860]

I don't think we'll ever get to a point where AI-only structure design is accepted. Even today no engineer worth his seal just allows designs to go out of the office with out checking them, and philosophically no QC process should be trusted if there's only one engineer involved in it. All engineering software today carries a disclaimer that indemnifies the software developer and I don't expect that to change. There are also many owners throughout the US that prohibit use--or require use--of specific types of software, so kicking back and letting the robot do all the work just isn't going to happen if you want to work on these types of projects. AI is just an advancement of software capability so I don't think that desire for legal protection and owner requirements change. It's also just another tool in the tool box and may not be always useful--or capable--in all cases. But you make a good point about training and QC/QA for AI. But it's not that scary to me. FEM like SAP2000 and BIM/BrIM wasn't always around, and we figured out how to validate and check the data. AI workflow will just be another challenge to overcome. It will be a rocky road, but we'll get through it.

[@luci3042]

I think to rely fully on AI is like flying an aeroplane without a human pilot. I will not get on board. But, perhaps for simple low risk structures full AI could be enough. Like taking over the counter medicine for cold, no need to go to the doctors.

[@melborja5152]

The controllers of this planet should be removed FIRST before building an uncollapsable safe superstructure, and of course the entire system should be changed

[@hypell0]

With the advent of the Internet and smart phones, the travel industry was worried that travel agents and the other front of house staff would be out of work, and although the industry has now changed, there are still travel agents and other ground staff, people still want the human touch

[@colum9675]

​@Eric Hansen Very thought provoking response. I'm just graduating with an MEng and have been very worried so your viewpoint is reassuring!

[@AKu-ig6lr]

As a structural engineer, I found this video helpful and refreshing!

[@bengregory4729]

Great video Matt, I agree completely. Imagine a doctor purchases a new machine for the operating room that allows them to perform a task better and faster than they did before. It would likely folloe that they would request a higher fee for the value added. Additionally, that equipment takes nothing away from the skill, knowledge, and decision making ability of that person. Analogous to FEA/automation to the structural engineer.

[@solboozyk4014]

For me 5 ways that AI and software will change the life of structural engineers are: 1. Faster and More Accurate Analysis: With the help of AI and software, structural engineers will be able to analyze complex structures much more quickly and accurately than before. This will save them a lot of time and effort, and will allow them to focus more on design and optimization. 2. Automated Design Optimization: AI and software will enable structural engineers to automatically optimize their designs based on various criteria such as cost, weight, and strength. This will allow them to find the best possible design for a given set of constraints and requirements. 3. Improved Collaboration: With the use of collaborative software and AI tools, structural engineers will be able to work more effectively with other professionals such as architects, builders, and contractors. This will enable them to reduce errors and ensure that everyone is on the same page. 4. Enhanced Safety: AI and software will enable structural engineers to simulate the behavior of structures under various loads and conditions, allowing them to identify potential safety hazards before they occur. This will help to prevent accidents and ensure the safety of people and property. 5. Reduced Costs: By automating many of the tasks that were previously done manually, AI and software will help to reduce the costs associated with structural engineering. This will make it more affordable for businesses and individuals to undertake construction projects, and will also help to reduce the overall environmental impact of the construction industry.

[@dougzimbelman7641]

I think I have that text book from which the FBD were taken. Awesome.

[@temia7944]

Thanks bringing this up. I think it would still take a while before AI can be of optimal use because of the sizeable number of constraints involved on structural engineering projects , design coordination among multidisciplinary teams, risk of failure associated with structures, complexity of brownfield & retrofitting projects. The value proposition of AI in this field might be defeated it does not accelerate deliverable or reduce cost.

[@VeronJuraj]

I think it needs to be 100% reliable to be efficient

[@bradmesser4807]

@@VeronJuraj Mission critical software is huge and it is an entirely different way to write code than any of the apps or websites that we're used to. Bonafide software engineering at its best is writing mission critical because if it goes down, there will be hell to pay

[@PeterRanieriII]

Interesting bc the clients today say why you charging more for revit than cad when it’s supposed to streamline coordination. We are far from AI designing foundations. And why should structural engineers learn how to code if the ai can do it? Instead work on your spatial reasoning and people skills

[@Theneutralaxisclub]

Great video, Mat! This is a good insight into how AI and other new technologies will play a role in optimizing much of the technical aspects of our jobs. The great thing is that the soft/client facing aspect of our job is still up for grabs, and it's something we will be able to control for decades. At the end of the day, we are in the service business, for people, by people. Keep up the good work!

[@sanyivizi8288]

My thoughts on 5. I think engineering will never be a commodity, espacialy structural engineering.If there will be a method to reduce design time and cost and to increase efficiency in my opinion, there will be an even higher pressure on performance and we will have to work with even thinner margins for error, which will put even more pressure to our level of confidency and will eat more of our free time relaxation and well being.

[@trevorstephenlewis7001]

Structural engineering has been more and more of a commodity for the past 15 years in the US. The money is in good design, not engineering

[@mrmadame28]

​@@trevorstephenlewis7001what that even mean?

[@Max-js7zr]

Really wondering if the audio is low or is it just the backhoe outside

[@jingness]

Hi mat, do you have any insight on what software or tools someone doing structural engineering would look into if they want to utilize AI at this period of time?

[@academiclearner9390]

Could you please make a dedicated video on communication skills?

[@dizcommunity4208]

I need a spreadsheet to assess demographic and productivity of labour and plants in a construction site. Help please!!

[@kerimthedream1]

In light of the massive destruction the earthquake in Turkey has caused, could you do a video on how to assess the structural strength and weaknesses of existing buildings and potential methods of improvement?

[@IGJTHOMAS]

I like the idea of showing how the new technology adds value. The new programs cost money and it takes time to train how to use the programs. The software provides added productivity in some ways but dose not translate to less time. I believe PE and the next levels should increase the value or their work.

[@muhammadilman7721]

Hi seniors. I'm working on final year project which is finding vulnerability of existing structures to earthquake, and suggesting low cost remedial measures to make the buildings safe. Which softwares should I use?, and what things should I keep in mind?

[@runcycleskixc]

would be great to know your extended opinion on the Turkey disaster.

[@isaiah7577]

Great Video! Thanks Mat! This is from Chat GPT, it's been pretty helpful thus far but I've had to correct it a number of times already. What are some ways that A.I. will be useful to structural engineers - Generating design ideas and concepts - Conducting structural analysis and simulations - Providing information and data on building materials and codes - Generating reports and project documentation - Assisting with project estimating and cost analysis - Suggesting solutions to structural problems and design challenges - Improving collaboration and communication between project team members - Streamlining the design process through automation and efficiencies - Offering 24/7 support for quick answers to questions.

[@asiagreen5658]

For architecture- people need to stop saying architect dream is engineer nightmares because y'all have to know that new midsized to large airports will have nursing room was passing new laws in 2018 . Freestanding full 24/7 emergency department is becoming common in my area without using main hospitals. Does your area have that, too? Does you know some hospitals start plan make emergency department for elderly people? Autism friendly places become common in public spaces, schools/clinics and home, too. Someone who started studying about Deafspaces for deaf/hard of hearing/hearing aid community. Chris Downey is architect who lost sight later in his life. Now, he is becoming consultant for blind community in architecture. Also, we had the blueprint for blind and visually impaired community. In China, architect have to know what the traditional chinese architecture in their communities. Health and wellness in building becomes common around worldwide. Healthy buildings is next generation of green building.

[@lordoftherain]

Complex is because of that we have certain ways to manufacture and ways to build. One out of thousands example would be if you can melt and cast steel onsite with a 3d printer ( imagine how you pour concrete). Do you still need so called "members" to form a structure? and do you still need weld or bolts as connections? probably not. At the end of the day only a small group of people will take the liability of the machine outcomes. It just AI is touching so many peoples cheeses, if the revaluation comes , the resistance would be huge. Myself is a Structural Engineer, to be honest I would trust an AI more over a human being on designing a safer structure.

[@mustafajarman1331]

Could you please tell me that how much graphic card is needed in a laptop to make a start in structural engineering career? Pleasee sir

[@merryparde]

Hi sir! Follow-up question: in your opinion, what does the education system, particularly in CE need to improve with all the advancements in technology? And for us (students) what do we need to be focus on?

[@MatPicardal]

I think there needs to be more practical design work and emphasis on detailing and load paths.

[@michaelweyenberg6238]

They should teach Bluebeam skills. Marking up PDFs is half the job.

[@kerolsmosad4159]

I’m currently studying civil engineering but I’m a bit confused about being a structural engineer or make a career shift to backend web development so if someone whose currently working in structural engineering field can give me some informations about the work or the salaries or if it’s a good field of work or not ?

[@mohdfarhan5697]

Hi Mat, In the the Etabs model which you have shown at 0:42 in your video you pinned the shear wall at base. Can you please tell the reason why?

[@PeterRanieriII]

Because “pinning” it is cheaper than “fixing” it.

[@lordoftherain]

why? have you assigned out of plane bending stiffness to your shear wall? hahah

[@tylerjohnson5233]

Because you can pin shear walls. This is common in residential. Fixing isn't necessary in every circumstance

[@chillyman1459]

What is that earthquake software?

[@nikhilpatil3087]

hey, do you know how can I clear my basics in the best way possible? so that i can get to a n advance level

[@bradmesser4807]

I don't want my structural engineers working in code, especially for mission critical software. Learning a new skill step by step is good. Really properly designing mission critical software that people won't die from is a whole different journey.

[@andrositee]

NOOOOO!!!! MY LIVELIHOOD!!!

[@brisingr12]

cool, time to become a structual enginner

[@aneetta_D]

Hi...can you please reply to my question.iam a civil graduate in 2015 and haven't worked since .if i did my masters in structural will i be able to grab a job or will i be rejected due to my year gap.hoping for a reply.i got admission in one of the UK university but I'm super confused 😕

[@Toxicflu]

well said. AI is a tool. With quicker edits, expect more edits. Ai will just give engineering so much more capacities for cooler designs.

[@tomoki-v6o]

i think ai will be useful for buiding codes and reguelations

[@hilasela4444]

Which book of codes and regulations do you recommend to purchase for residential in CA? Also what engineering calculations program is most used for residential in CA?

[@MatPicardal]

I'm guessing the Wood NDS code books and some residential design books. I'm assuming wood single family homes? Forteweb is a common software for wood design.

[@hilasela4444]

Thank you so much!

[@wsapalas]

I think that in less than 5 years there will already be AI specialists in the great areas of knowledge such as medicine and law, engineering I think it should take a little longer for the issue of analysis but not much like that, I say accessible AIs, because today we already have AIs "doctors" and lawyers, I myself have a Python script that gives me the percentage of having pneumonia by analyzing lung x-ray images. He has more accuracy than a doctor with decades of experience because he works with massive data, he doesn't rest, he doesn't ask for a raise...

[@MatPicardal]

The Architecture/Engineering/Construction is also infamously slow for implementing new technology as well, but we'll see. Yes, I think that's what AI and software should be used for, it's better at calculations and analysis than you, but the doctor should still be the one taking ownership of that data and verifying the results to make the final call. But who knows, there may be a time when an AI makes better health judgement calls than a doctor.

[@bradmesser4807]

This one's a slippery slope as there's still statistics at play. If the AI gets it wrong and the engineer misses it, then we're screwed. That can't be allowed to happen.

[@JKS323]

The pressure on engineers will be increased because we have to verify things done by AI

[@trevorstephenlewis7001]

Consider AI as being fast at presenting 100s of options to select the top three that need to be verified through full code checks, documentation and stamping

[@mohammadarslanali3943]

Hi Matt I'm Truly your ancient follower would you help me find a Job CA

[@abq901]

Ha Ha Ha ... I will tell you a story .. A building was designed completely by Computer.. one fine day a Steam water pipe burst ... and a joint failed leading to the collapse of the Building ... Confidence of the Structural Engineer is what makes a Building stand .. if AI can do that ... then Hat's off ... also remove the disclaimer at the beginning of every Code that Engineer is responsible ... A b Quadri---India

[@RoqueMatusIII]

Hmm Construction 5.0 may be here before you know it

[@leonardobattezini1476]

I sent you an email. Hope you have the time to take a look. Thank you so much in advance.

[@TheTruthisGold]

I don’t think this guy really understands what AI is capable of at all. I think an AI generated avatar would speak more intelligently than he is.

[@sandygrungerson1177]

you guys have been saying this since the 1980s

[@gerritbearda154]

AI in engineering will only work if the data is online. How many engineers will publish their work online? Who is responsible for the AI outcomes? Who signs the AI generated design, structural engineering, etc. off and takes any claims for failure? All grandiose but like many engineers, designers, architects, civil engineers, etc. we wont be publishing anything online for an AI algorithm to plunder our work. Customers pay good money for our integrity, our experience, our designs, our engineering competence, our quality control, etc. Not going to give it away for free so that some lazy AI using "engineer" to certify a building using our calculations. Dream on.

[@bradmesser4807]

I'm sorry it makes you feel that way. AI is incredibly helpful, but there is a large trust and responsibility component to it to really allow for us to do the mission critical software. Feel free to reach out if desired.

[@gerritbearda154]

@@bradmesser4807 "trust" and "responsibility" are foreign words when using AI. For there to be trust and responsibility there has to be accountability. AI does not provide that. Without accountability nothing engineering wise will be certified. Without certification nothing will be build. Bit like building a submersible without certification. It will simply implode on its way to see the remnants of Titanic due to no third party certification or oversight. OceanGate is an example where ignored sound reasoning by third parties and trust by one person (or AI) without overview or accountability led to an engineering failure.

[@drticktock4011]

Management and coordination...ugh...boring! Just go into Construction Management (Not an engineer)

[@trevorstephenlewis7001]

Lolz more managers managing engineers where the manager is not an engineer, has never been a licensed engineer telling engineers how long something should take, at what cost, because I worked on this project 10 years ago. Yet said manager is not the one stamping the calcs and draiwngs. There is a reason why the SE is a difficult exam.

[@drticktock4011]

@@trevorstephenlewis7001 exactly. I didn't spend all that time in school and doing real engineering for nothing. Technical Lead is the way to go!

[@Nikesh977]

Bro, i think you have no concept of AI. AI has potential to completely replace humans in structural designing work. Structural engineers might need to shift to other thinfs like site management etc.

[@SAA-hr9gr]

I think you are also over estimating AI too...AI comes up with solutions it finds in their database. I have been working over 5 years and never once had the same building built twice. Everything is so specifically designed to accomodate the customer, then practical experience plays a big role in choosing what route you go down. The actual building is still done by hand and the workers often make mistakes. AI can't find an answer to all these mistakes because it's based off experience. Also you can program it to look at norms but you don't ALWAYS follow the norms, they also at time contradict themselves or you can combine them...honestly there are still SOOO many variables that I find it hard to imagine AI can replace a structural engineer full stop. Also it would easily overdimension everything, which isn't cost efficient.

[@Nikesh977]

@@SAA-hr9gr General AI, something that can learn on itself i.e. you do not need to program it to do specific tasks. You just need to program it to be intellegent. It will learn all th knowledge that a human can learn in a span of 30 years career in a matter of hours, and if you feed it with Big Data, it will learn the wisdom of all structural engineers, all buildings. and will be far more efficient than any humans.

[@lordoftherain]

100% agree. plus huge amount of data already existed. I am pretty sure CSI has already collected huge amount of data. if they want, they don't even need to " design" a certain building. because every new building is a modular of a lot of existing ones which already sit at certain corners of the world. Client : I need a 10 story office building in 777 Queen street. Project manager from CSI: ok just give me a second. (5 seconds later) with that site configuration and soil info from GIS, here are few options for you ( put the VR on clients head). client, OK this one. Project manager from CSI: Perfect, let me do a 3D scan of the site tomorrow and fit our model into there and then we are ready to submit to Siri for a final approval. Based on this true story, where the hell did you see the Structural Engineer?

[@melborja5152]

Great pyramid in Egypt was built by AI, high tech w/ anti gravity supreme quantum computing

[@FloridaZoe305]

I am a REAL engineer (Licensed) not the ones that went to a boot camp and became an "engineer" or whatever they call themselves these days. Professional engineers AREN'T going to be replaced whatsoever. As long as humans exist on this planet and population continues to grow Doctors and Engineers (Real) will ALWAYS be needed! Mat I agree with you that we will need to know programming, I had to learn different programming languages to assist me with my design work!

[@bradmesser4807]

Coming into this as a software engineer who went through undergrad and did a grad self-study that I benchmarked regularly against my IBM colleagues, I've found that there are multiple ways to success. That being said, you have to have the skills to do what needs done period and that self-study gave me ways to reinforce my knowledge and think differently over time, especially when directed in a consistent effort over 2000 hrs over multiple years. I think as long as you know what you're missing, get it, and prove it, then you're ok, but you had better be able to back it up when the rubber meets the road.