Artificial Intelligence -Driven Sustainability and Human-Centric Optimization in Structural Engineering

58 | P a g e DOI: 10.5281/zenodo.17358873 “Artificial Intelligence -Driven Sustainability and Human-Centric Optimization in Structural Engineering” Mahadeva M1*, Yashaswini N K2 1Assistant Professor, 2Undergraduate Students, Department of Civil Engineering, RNS Institute of Technology, Channasandra, Bengaluru, India *Corresponding author: [email protected] Abstract The fields of structural and civil engineering are changing significantly due to the rise of new technologies, sustainability goals, and the need to manage uncertainties. Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) are transforming the industry. They provide useful, data-driven tools for optimizing designs, predicting material properties, and detecting structural damage. At the same time, there is a strong emphasis on sustainability. This pushes the focus toward a complete assessment of building technologies, considering environmental, social, economic, and technical effects. This matter is especially important since buildings account for a large part of global energy consumption and resource use. Building Information Modelling (BIM) serves as an important digital framework for managing this complex information throughout a building's entire lifecycle, although its implementation in existing structures can be tricky. Also, a systematic approach is needed to tackle the various uncertainties in structural engineering, such as variations in material properties, model limitations, and human error. By combining AI and BIM with sustainability and uncertainty management practices, the industry can build stronger, more efficient, and environmentally friendly infrastructure. This integration shows the potential of these fields and points to a future with smarter, more sustainable, and safer buildings. Keywords: AI, Machine Learning, Structural Engineering, Sustainable Technology, Building Information Modelling, Energy Efficiency, Structural Damage Detection, Uncertainty Principles. Introduction A key reason for this change is the fast growth of Artificial Intelligence (AI), especially its branches, Machine Learning (ML) and Deep Learning (DL). AI is becoming a better option than traditional modelling techniques for simulating human intelligence in solving complex problems. This includes tasks like automating design improvements, monitoring structural health in real time, and predicting material properties. AI can handle large, complex datasets and spot non-linear patterns. This helps engineers get past many of the limits of old methods, providing more precise, efficient, and cheaper solutions. Salehi and Burgueno (2018) [1] At the same time, the build field faces big stress to cut its harm to the earth. Homes use lots of stuff & energy and add much to air 59 | P a g e DOI: 10.5281/zenodo.17358873 harm. This led to a focus on safe tech & ways, seen as "the craft & good care of a clean built zone" Nelms et al. (2005) [2]. Yet, a top find from new study is that the work of this tech can't be seen all by itself. A real study on home homes found that new tech only gave about 42% to use less energy. The real drop in energy use leans much on how tech & folk who live their work, like heat set & tool use Zhao et al. (2017) [3]. This shows the need for a full, both-tech-&-folk view that looks at the mix of tech, folk, & the earth. So many things make this work hard. One big thing is the lack of sure things in making plans. As builders, we pick from choices with gaps or change in info. This unsure state has two types: random (aleatory) or from no know-how (epistemic) See Bulleit (2008) [4]. The roots of unsure state range from change in how strong stuff is, limits of math models, to the risk of us messing up. Some use codes & safe steps to fix this. But mistakes by us need strong checks like peer looks & checks. Building Info Modelling (BIM) is key to sort big pile of info & linked hard jobs. BIM keeps all project data in one place, from shape plans to what stuff is made of, aiding all to work well & pick well all through the building life. It’s big for new builds, but use for old builds, do-overs, & taking apart is still an open door. The mix of three fields AI for smart insights, green ways for Earth care, and BIM for keeping tabs on info makes a strong way to deal with big hard parts of new-age building work & to step toward a firmer & greener place to live and work. Figure 1. Integrating AI, Sustainability, and Human Behaviour in Structural Engineering Source: (Andrew P. McCoy research paper 2016) Literature Review The field of building tech has drawn parts of many types of study. This includes AI, green tech, and sure plans. AI in Building Work: AI holds ways like ML, PR, and DL. It's changing how we work by giving new smart tools. A main paper by Salehi and Burgueno (2018) [1] talks about these new tools. They say that AI is a good choice over old methods, more so when in doubt or when real tests can't be done. AI and ML help a lot in four key parts Building Health Check (SHM) & Hurt Spot AI can look at sensor data from a build to find and place hurt. Not like old ways that need much repeat work, AI leads, data-led ways learn from past sensor info. This makes them better and strong. Tools like stats PR, fake brain nets (ANN), & help vector tools (SVM) are much 60 | P a g e DOI: 10.5281/zenodo.17358873 in use for this Salehi and Burgueno (2018) [1]. The use of deep learn plans, like online convo brain nets (CNNs), lets for live, shake-led hurt find by learning from raw quick data Abdeljaber et al. (2017) [5]. Stuff Make and Guess Work. Things like rocks are not easy to guess how strong they are. AI tools help to make good guess models for stuff such as rock press strength and how bendy they are. This can cut the need and cost to test them in real life. Tools like net brains & help line tools work best here Zhang, (2023) [6]. Plan Make Better: AI that makes on its own, like fight-make nets, help to make plan draws & floor maps. This cuts the need for man work and gives many best plan picks that look at stuff use and how well it works Zhang (2023) [6]. Build Lead: AI has lots of good points. It makes cost guess better, ups how well we see risks, makes time maps flow smooth, & uses stuff well by looking at past data and seeing what comes next Ahmad Abdulla Aldashti (2025) [7] Green Tech & People-First Ways The push for green in the build world has made us look at new green tech. It has made us think on how we judge its use. Nelms et al. (2005) [2] find fault with the old, one-issue view (like just money) and give a plan to check green tech from all sides: earth, folks, cash and tech. By means of green roofs, they show the good goes past saving power. It fights city heat, cuts storm water, and gives a fun place. A key find in this study is how key human acts are. A look at home power use found that the true power use is from how tech & folks act. The study found four acts tied with power use: heat & cold knob sets, how much we use the washer/dryer, & smarts on home tech. They noted that bad habits can drop more than half of a green home's power-save touch. This flags that green must be seen as a link of both tech & how folks act, needing tech moves & a need for folks to shift Zhao et al. (2017) [3]. Figure 2. The AI integrated and sustainable into construction (Source: journal of emerging sustainable researchers) Research Gap Many big gaps & flaws are seen in the checked work. These stop us from a full mix of new ways in building. For AI, deep stuff needs lots of good, big, tagged data sets. But, to get this data is hard. For checking the health & hurt of builds, not much data exists. The same goes for green build stats. Data on how folks act & use tech is often not full AI and deep teach relies on trial runs in made-up spots, like SUMO for cars or fake sets for build checks. But these AI tools might not do well or be safe in real, changing spots. They need more checks before use in life. Mix and join tech Clear perks are there, but no smooth ways yet to join AI, BIM, and green tools. 61 | P a g e DOI: 10.5281/zenodo.17358873 Linking these different techs and data is hard.AI can make very good guesses. But it's "black box" ways can hide the true thought behind an answer. This can stop folks from making smart, safe choices. This is true in key jobs like building work. We also must deal with big worry about data safety, bias in code, and bad effects from AI on folks Work on AI in road control shows hard times are met when trying to grow AI across big and mixed roads. The ever change of such places may mess up teamwork &and slow gains. This cuts how well the AI works in big cities. Implementations Key gaps in current studies can be addressed by establishing a clear, integrated plan. This plan should combine the strengths of artificial intelligence (AI), green technologies, and data management. By doing this, we can create a strong yet flexible work framework that responds to changing conditions and needs. At the heart of this strategy is a multi-step flow chart that guides the entire process from data collection to decision-making and continuous improvement. The first stage focuses on data acquisition. This system must gather data from multiple sources to ensure accuracy and diversity. Sources include real-time data from IOT-based sensors, on-site mobile data collection during field inspections, and historical records from existing structures. To tackle the common problem of insufficient or incomplete datasets, techniques like data augmentation and synthetic data generation using AI-driven Generative Adversarial Networks (GANs) can be used. These methods create rich, artificial datasets that closely resemble real-world conditions, which helps strengthen model training and analysis. A standardized data structure is crucial. Without a common format, it is very difficult to integrate data from different systems. By creating a unified framework, interoperability between Building Information Modeling (BIM) systems, AI models, and other digital platforms becomes seamless. This ensures engineers, contractors, and policymakers can work on the same platform with consistent data sets. The next step involves using AI and machine learning models for predictive and optimization tasks. In the design phase, these models can forecast material behaviour, structural performance, and potential risks while suggesting ways to optimize layouts, construction processes, and resource use. In the operational phase, deep learning models can automatically detect structural damage through image analysis, drone surveys, or vibration data from Structural Health Monitoring (SHM) systems. This greatly reduces the need for costly and time-consuming manual inspections while increasing the speed and precision of analysis. A long side the technical enhancements, the plan also includes an evaluation of green technology in a comprehensive way. Instead of focusing only on direct energy consumption, this approach considers environmental, social, economic, and technological aspects. This includes full lifecycle assessments of materials, evaluation of community and social acceptance, long-term economic costs and benefits, and overall efficiency of resource use. By incorporating this data into BIM models, engineers and planners can simulate and test various sustainable options before actual implementation, ensuring that choices are both practical and environmentally responsible. All these outcome’s predictions, sustainability tests, and risk evaluations contribute to a decision-support system. This system uses probabilistic methods to handle uncertainty, applying load resistance factors and other design codes while incorporating real-time review data such as peer checks or field validations. This significantly reduces the risk of human error and supports smarter, evidence-based decision-making. Stakeholders are empowered to make well-informed judgments by balancing 62 | P a g e DOI: 10.5281/zenodo.17358873 performance, cost, safety, and sustainability. Finally, the framework encourages a continuous feedback loop. Data from the performance of completed projects is cycled back into the AI models and design processes. This ensures ongoing improvements in technical performance and a better understanding of the interactions between technology, human behaviour, and environmental impacts. Such a self-improving cycle turns the system into a living, adaptive framework capable of becoming more accurate, sustainable, and efficient over time. Smart building setup with AI running the show. First up, they grab loads of random data sensors pinging away, cameras snapping pics, tracking what people do around the place, even checking out which materials went into the walls. Basically, they’re snooping everywhere. Goes straight to the nerd squad think brainy AI models like ANNs, CNNs, GANs, and a few Bayesian whatnots. These models eat up the data, trying to spot weird patterns, like why everyone cranks the AC at noon, or which floors are energy hogs. Basically, these digital brains figure out what’s normal...and what’s just wasting cash and electricity. It gets interesting they don’t just stare at the numbers. They pull in the sustainability wizards. We’re talking Life-Cycle Assessments and green tech, all that jazz, just to see how to stop trashing the planet and start sneaking in better materials and smarter systems. Once these computer geeks and ecowarriors hash things out, they dump their findings into a dashboard for the building boss. That’s where the magic happens: real, simple advice. More stuff like, “Hey, swap out your lights,” or “Maybe, stop running heating and cooling at the same time, genius. “And then, know, they do it. This isn’t just pie-in-the-sky theory the changes go live, and the AI keeps nosing around, learning from what happens next. If it works, awesome. If not, back to the drawing board. They just copy-paste the system onto a bunch of other buildings like it’s no big deal. That’s its AI, smart moves, greener buildings, rinse and repeat . Figure 3: Flow chart of above integrated AI sustainability in civil engineering 63 | P a g e DOI: 10.5281/zenodo.17358873 Future Scope AI, green tech and BIM in building join up. This makes many new paths for work growth. Mix Models for More Toughness: We should make new mix models. They use the best bits of each way. This has blending path-based guess with flow ways for road events. It also links AI guess with old tech to make it clear and trusted. AI That Knows Risk We need AI that can tell how sure it is. These "risk-know" sets would give safe ranges with their guess. This is key for safe choices in big areas like building. Grow Multi-Agent Sets For city-wide use, we need more work on big multi-agent learning ways. This could look at tech like shared learning event calls to cut data work and boost ties among many AI agents but keep their own rule. Acts-Linked Design We need to learn how tech in buildings and human acts mix. Work should make tools that learn from and talk to folks to push less power use. This could link AI watch with smart home sets. It would use less power in the now based on how folks live. Sim-to-Real Move It is key to link sim and the real world. We must aim to set up and test AI tools in small city tests. This helps fix rules with real data and grow trust in these techs for wide use. As AI grows, we need strong ethics rules plans fit for the build civil world. We must make clear data rules, fair AI steps human checks. This makes sure AI tools are made used in a good way. Next BIM for Old Builds: We should work on new BIM tools that auto-make "as-built" models from scans photos. This makes using BIM on old works easy, opens big chances for site care and makeovers. Summary The document discusses how new technologies like machine learning are changing civil and structural engineering by helping to optimize designs, detect damage, and predict material properties. It highlights the importance of sustainability and managing uncertainties in construction, noting that buildings use a lot of energy and resources. The paper emphasizes that for real energy savings, you must consider both new technologies and the behaviour of the people who use the building. Building Information Modelling (BIM) is presented to manage all this complex information. In conclusion, the paper suggests that combining these three areas technology, sustainability, and information management can lead to stronger, more efficient, and environmentally friendly buildings. Conclusion The files show that tech, care for the earth, solid risk plans will shape the build trade's path. AI and its parts, ML and DL, give strong tools that go past old ways. They help guess material traits, how strong builds are, and the way they look. At the same time, a full take on care for the earth is key. It looks at how we act, which is big for real green and cost gains. Building Info Mod (BIM) links these bits. It acts as the core net hub for all stages of a project. The trade still faces tough tests. These include poor data, how clear AI is, and how hard it is to scale up new ways. The mix of these studies shows we must work hard at setting data norms, making joint plans, and testing new tech. By taking this path, build and hard build fields can start a new time. They will be smart, tough, and green. In the end, we will have a built space that is good for us and the earth. 64 | P a g e DOI: 10.5281/zenodo.17358873 References [1] Salehi, H., and Burgueño, R. (2018). Emerging artificial intelligence methods in structural engineering. Engineering structures, 171, 170-189. [2] Nelms, C., Russell, A. D., and Lence, B. J. (2005). Assessing the performance of sustainable technologies for building projects. Canadian Journal of Civil Engineering, 32(1), 114-128. [3] Zhao, D., McCoy, A. P., Du, J., Agee, P., and Lu, Y. (2017). Interaction effects of building technology and resident behaviour on energy consumption in residential buildings. Energy and Buildings, 134, 223-233. [4] Bulleit, W. M. (2008). Uncertainty in Structural Engineering. Practice Periodical on Structural Design and Construction, 13(1), 24-28. [5] Abdeljaber O, Avci O, Kiranyaz S, Gibbous M, Inman DJ. Real-time vibration-based structural damage detection using one-dimensional convolutional neural networks. Journal of Sound and Vibration 2017; 388:154-70. [6] Zhang, S. (2023). Artificial Intelligence and Applications in Structural and Material Engineering. Highlights in Science, Engineering and Technology, Vol. 75 (ACEGE 2023). [7] AL Dashti, A. A. (2025). What an Artificial Intelligence (AI) is Being Utilized in Structural Engineering. International Journal of Novel Research in Engineering and Science, 12(1), 14-19. [8] Volk, R., Stengel, J., and Schultmann, F. (2014). Building Information Modelling (BIM) for existing buildings - literature review and future needs. Automation in Construction, 38, 109-127. [9] Raman, R., Patnaik, D., Lathabai, H. H., Kumar, C., Govindan, K. Nedungadi, P. (2024). Green and sustainable AI research: an integrated thematic and topic modelling analysis. Journal of Big Data, 11(55). [10] Liu, D., Lei, L., and Jie, L. (2025). A congestion prediction method based on trajectory mining algorithm. Computational Urban Science, 5(1). [11] Zai, W., and Yang, D. (2023). Improved deep reinforcement learning for intelligent traffic signal control using ECA_LSTM network. Sustainability, 15(13668). [12] Fathi, H., and Brilakis, I. (2019). "Automated Vision-Based Crack Detection Using Deep Learning in Civil Structures." Automation in Construction, 91, 317–329. [13]. Wei, H., and Chen, Y. (2017). "Adaptive Structural Health Monitoring Using Artificial Intelligence." Journal of Advanced Engineering Research. [14] Baduge, S., (2022). "AI Applications in Green Building Design and Sustainability Assessment." Journal of Cleaner Production. [15] Wang, S., Yan, C., and Xiao, F. (2012). "Quantitative Energy Performance Analysis with Occupant Behaviour Modelling." Energy and Buildings, 43(2), 614–624. [16] Kusiak, A., and Xu, G. (2012). "Data-Driven Optimization of Building Energy Systems with AI." Energy, 42(1), 27–34.