Civil Engineer | Numerical Modeling & Simulation | EERI WA State Board of Directors
Bridging advanced numerical modeling and real-world engineering together to build resilient infrastructure. Scroll through the page to see real-life applications of FEA simulations. See the disclaimer here.
Untitled video.mp4Using LS-DYNA, I developed a 3D seismic soil–structure interaction model to simulate the response of the Nihal Atakaş Mosque during the 2023 earthquakes. The simulation reproduced the combined effects of shaking-induced settlements, soil–structure interaction stresses, and partially drained soil behavior.
Click here to access the series of papers published on this study.
Following the earthquakes, the ground adjacent to the mosque settled by 230–300 mm, with clear evidence of liquefaction ejecta. These measured deformations provided a rare opportunity to compare field observations with advanced numerical modeling, showing strong agreement between computed and measured settlements.
3. Details of the Measurements at the Mosque
I observed significant ejecta material at the north-east corner of the mosque, and the ejecta material covered approximately 850 square-feet area. This indicated that liquefaction occurred at the site. I also observed and measured 230 mm to 300 mm settlement of the adjacent soil near the mosque structure. Mosque itself did not suffer significant damage and in fact, it was open to public shortly after the earthquake. However, the amount of settlement occurred adjacent to the structure was significant and worth investigating more.
4. Quick Glance at the Simulation Results
I modelled the 151-meter-deep soil profile, extending more than 100-meter at each horizontal direction, and placed the structure on top of the soil. The simulation modelled different strata of gravels, clays, sands, and silty and sandy gravels with advanced constitutive models. The simulation included structural components such as the mosque itself, its minarets, domes, and door/window spaces. Deep foundation system of the structure was modelled explicitly.
The finite element simulations (settlement contours shown on the left) computed more than 250 mm vertical displacements adjacent to the structure successfully reproducing what we measured at the field. This shows the practical and useful application of FEA to a real life problem.
2007 Niigata-ken Chuetsu-oki Earthquake and Kashiwazaki-Kariwa Nuclear Power Plant
The Niigata-ken Chuetsu-oki earthquake was a 6.6 magnitude earthquake. It impacted many coastal areas, including Kashiwazaki and Kariwa. The recorded peak ground accelerations (PGAs) reached values larger than gravitational acceleration (1 g). After this event, the Kashiwazaki-Kariwa Nuclear Power Plant was shut down. The critical structures performed well and were in good condition. However, a fire occurred near the critical structures and delayed the re-opening of the plant. The fire was under control after several hours and did not impact the stability of the NPP but disrupted the service. The fire likely occurred due to the differential settlement of backfill material placed at the site, and there are studies (e.g., Tokimatsu 2008) describing the differential settlements in detail. In 2011, Fukushima experienced a NPP disaster due to the Tohoku earthquake. Combined, the Niigata-ken Chuetsu-oki earthquake and the Tohoku earthquake demonstrated that, although generally founded on competent ground, these structures may still experience significant damage or safety issues due to extreme events. In response to these real-life occurrences, both Japan and USA took tangible actions for re-evaluating the seismic safety of NPPs.
Within the broader re-evaluation efforts, the estimation of earthquake-induced ground settlements under realistic, multi-directional earthquake loading conditions emerged as one of the critical technical challenges. Existing approaches to evaluate ground settlements rely on overly simplified empirical and semi-empirical tools that may not represent soil and system response of NPP sites under earthquake loading conditions, thus may not be enough for estimation of earthquake-induced settlements. My PhD thesis focused on developing an advanced constitutive model to better represent ground/soil conditions, enabling soil-structure interaction simulations under realistic, multi-directional, three-dimensional earthquake loading conditions, and better estimating earthquake-induced settlements of NPPs and their surroundings. For this purpose, I developed a constitutive model (I-soil), which represents the cyclic response of soils reasonably well. The model originally focused on representing the earthquake response of dense granular material, which can be found underlying nuclear power plant structure. We demonstrated the applicability of the model for simulation of soil-structure interaction at the nuclear power plant sites, and eventually submitted our report to Nuclear Regulatory Commission (NRC).
The idealized seismic conditions we solved for nuclear power plant containment structure is shown in the upper two figures shown on the left, where the grey colored hockey pack-like part represent nuclear power plan containment structure (Yes, it is simplified!), and the mustard-yellow (or brown in the first figure) color domain represents dense sand. In the third figure from top, the computed versus measured settlements are shown. The results indicate that the three-dimensional numerical simulations reasonably represent the settlement of the power plant containment structure modelled in a centrifuge test. Further engineering details of this study is published in several papers and I encourage civil engineers to read the articles listed below. The third article has won the 2024 Hogentogler Award from ASTM International. The award is given annually to the authors of an ASTM paper of outstanding merit on soil or rock.
1. Simplified Three-dimensional Constitutive Model for Seismic Modeling of Dense Sands
2. A Multidirectional Cyclic Direct Simple Shear Device for Characterizing Dynamic Soil Behavior
3. Response of Sands to Multidirectional Dynamic Loading in Centrifuge Tests
Last couple years, while the AI frenzy was happening all over the place, virtual reality was slowly cooking. Meta poured billions of dollars to the virtual Reality Labs and developed two remarkable products, Meta Quest 2 and 3, with affordable purchase prices. Apple, following Meta, also invested billions of dollars to the virtual reality research and came up with Apple Vision Pro, which ended up being quite an obvious frustration for the consumers due to extremely expensive pricing. Furthermore, the rise of AI shattered the developments in the virtual reality arena (e.g., Meta fired 1,500 people from the Reality Labs in the Q1 of 2026 and switched gears to focus on AI) and VR stayed under the shadow of AI.
The consumers did not adopt the VR tools in a mass scale, and these tools ended up being a niche toy for specific audience mostly for gaming purposes. However, this does not undervalue the incredible potential of VR immersive technology to be used in general engineering and specifically civil engineering industry. These potentials include viewing products, structures and fields in three dimensions (3D) in detail, visualizing construction plans, material testing and simulation results in 3D, site surveying and conducting reconnaissance after an extreme event.
Boeing, the world's one of the largest aerospace companies designing commercial airplanes, military level aircrafts, and space systems, adopted use of VR and AR, and demonstrated the applicability of VR to design processes. Industry news commonly encountered in the websites claim that Boeing was able to reduce the design training time by 75% and processing / manufacturing time by 25%, by simply reducing the cognitive load of design details via use of VR/AR tools. Although I could not not trace the exact percentages, sentiment of reduction in training, processing and manufacturing is plausible and here is why.
Many millennials and Gen-Z workers are now replacing the existing workforce and, they are already not interested reading .pdf or CAD files where only the 2D drawings of the actually 3D components are shown. Such drawings are not intuitive at all and require heavy cognitive price to pay, which make them ineffective and inefficient. Whereas, a 3D VR model, if accurately built, requires less significant cognitive load, and is naturally easier to understand.
The same applies to site visits for civil engineers. Many civil / geotechnical engineers working at an office space / or in work-from-home condition are getting themselves familiar with work sites through photos of the sites taken by others. Similarly, civil engineers generally rely on Google Earth to understand the site conditions. And, those who use these tools already know that the output you get are very crude and often are not enough to fully understand the topography or other features of the site. These approaches are already outdated and can be readily replaced by VR / AR as well as GoPro Max 360 Action cameras (no I am not paid to advertise for GoPro). Similarly, post-disaster site reconnaissance can be conducted using VR and advanced video-capture tools and real-time satellite data.
In 2023, soon after the February Earthquakes, I deployed to the southern-east parts of Turkiye and conducted site reconnaissance. I collected perishable geotechnical / civil engineering data to study the lessons offered from the damages caused by these earthquakes. One of the sites I collected data from and then studied was Nihal Atakas Mosque. As I described in one of my previous articles, this mosque performed well and did not suffer significant damage, but it experienced liquefaction. I collected reasonable amount of data and photos from the site, but once I arrived at USA and started to study, I wanted to further inspect the surrounding areas of the mosque to see if I missed anything. Well, that meant I either re-deploy to the site, which costs 2,000 USD plane ticket, 1,000 USD accommodation and likely another 1,000 USD food / site expenses. Orrrr! Maybe I could use Meta Quest 2 and virtually visit the site to see if the evidence of liquefaction is still there or not. For this purpose, firstly I set up my virtual office space using Steam VR as shown in the video at the bottom of this article to the left. And then used Meta Quest 2 and Google Earth VR to fly back to the earthquake site as shown in the second video to the left. To my luck, there was a satellite data available in the Google VR that corresponded to the timeline consistent with my visit of the site. And, in fact, the ejecta material I was looking for was still right there (see the second video to the right where the right joysticks points out to multiple spots for liquefaction evidence towards the end of the video). It is obviously not the best quality visual data but the fact that it conceptually worked and in fact provided me with what I needed for (the further visuals for the ejecta material) was more than enough. Considering that this technology is only at "floppy disk-stage" and many more improvements will be made within relatively short time of period, the VR technology has a potential. Now, imagine inspecting the design of a high-rise, or a bridge, dam, or a levee using this technology. Would we still need 2D .pdf plans or CAD drawings? Can't we place dozens of GoPro 360s around a construction site, connect them to wireless, and 7/24 have an access to the 3D data through VR headsets? Can't we visit earthquake sites virtually, and collect evidence for engineering analysis ?
Media1.mp4Virtual Office Space Using Steam VR
Media2.mp4VR Assisted Locating of Liquefaction Evidence Spots