Stella Dak, Ph.D., MBA, P.E.
Principal Consultant
Get to know the expert
Expert QuestionS & AnswerS
As Engineer of Record for more than 40 utility-scale solar facilities, what engineering issue do you most often see surface after a project is built and operating?
The issues I most often see surface after a solar facility becomes operational are structural foundation problems, tracker installation errors, and inadequate stormwater and erosion control. Pile embedment, alignment, tolerances, and tracker-component installation deficiencies may not be obvious during construction, but they lead to excessive movement, misalignment, premature component failure, and reduced system performance.
Battery energy storage systems are relatively new infrastructure. What risks do you think the industry still underappreciates?
The industry still underappreciates how thermal runaway, fire protection, structural design, drainage, and emergency response interact. Risks are often assessed separately, yet failures can cascade. For example, water intrusion may trigger an electrical event with environmental and operational consequences. Long-term degradation, installation quality, flooding, and extreme heat also deserve greater attention.
Your Ph.D. focused on machine learning applications in structural systems — how do you see ML changing infrastructure assessment and risk visibility?
Machine learning can shift infrastructure assessment from periodic, reactive inspections to continuous, risk- based monitoring. By combining design data, sensor readings, inspection records, weather, and performance history, ML can identify patterns and early warning signs that traditional review may miss. For owners, insurers, and engineers, this creates greater visibility into where failures are most likely, how risks are evolving, and where further investigations should be prioritized.
What is the most surprising finding you have encountered in a solar or energy storage investigation?
One of the most surprising findings was that a major system failure was not caused by a defective component, but by small installation and alignment errors that accumulated across the tracker system. Individually, each deviation appeared minor; together, they created unintended structural demands, accelerated fatigue, and led to visible failures. It showed that no matter how robust the engineering is, poor construction quality or inadequate operations and maintenance can cause a site to fail.
What is one thing developers consistently underestimate when planning large-scale energy projects?
Developers often underestimate how much site conditions and construction quality influence long-term performance. Drainage, soil variability, tracker/foundation installation deviations may seem manageable during design and construction, but across thousands of piles, trackers, or battery units, small issues can compound into major operational and financial risks. Investing early in site-specific engineering and strong QA/QC is less costly than correcting systematic failures after commissioning.
Background
Dr. Stella Dak is a licensed Professional Engineer with over 18 years of experience in civil and structural engineering, specializing in utility-scale solar and battery energy storage system (BESS) infrastructure. She has served as Engineer of Record (EOR) for more than 40 solar facilities across the United States and has led multidisciplinary engineering efforts spanning feasibility, design, construction, and technical due diligence.
Her expertise includes grading optimization, foundation systems, geotechnical evaluation, stormwater management, and infrastructure performance assessment. Dr. Dak has extensive experience aligning engineering design with constructability, cost, and schedule considerations, and has developed engineering frameworks to improve risk visibility and project reliability in large-scale energy developments.
She holds a Ph.D. in Civil Engineering with a focus on machine learning applications in structural systems and an MBA with distinction in management.
Education and Certifications
- Civil Engineering, Ph.D.: Southern Illinois University
- Business Administration, Master (with Distinction): Texas State University
- Civil Engineering, B.S.: Bogazici University
- Licensed Professional Engineer (P.E.): Arkansas, Arizona, California, Florida, Georgia, Indiana, Kansas, Kentucky, Louisiana, Maryland, Michigan, Minnesota, Nebraska, Nevada, New Mexico, North Carolina, Ohio, Oregon, Tennessee, Texas, Virginia, Washington, Wisconsin, and Wyoming
- Dissertation Research Award: SIUC, Fall-Spring (2014-2015)
- NSF CMMI Engineering Research and Innovation Conference Fellowship (July 2012)
Publications
- Co-author, “Bridge Deck Runoff: Water Quality Analysis and BMP Effectiveness,” Fairbanks, AK.
- Research on machine learning applications in structural engineering, “A Novel Approach for Modeling Nonstationary Ground Motions,” (Ph.D. research), 2011-2015.
- Tezcan, Y. Dak-Hazirbaba, and Q. Cheng. (2016). “A Kernel-Based Mixed Effect Regression Model for Earthquake Ground Motions,” Journal of Advances in Engineering Software, June 7, 2016.
- Dak-Hazirbaba, and J. Tezcan. (2015). “Image-Based Modeling and Prediction of Nonstationary Ground Motions,” Journal of Computers and Structures (CAS), 2015.
- Tezcan, Y. Dak Hazirbaba, and Q. Cheng. (2014). “A Semi-Parametric Tool for Seismic Risk Management Applications,” Proceedings of the Third International Conference on Urban Disaster Reduction Sustainable Disaster Recovery: Addressing Risks and Uncertainty.
- Y, Dak Hazirbaba, and J. Tezcan. (2014). “A Novel Approach for Modeling Nonstationary Ground Motions,” H.V. Topping, P. Iványi (Editors), Proceedings of the Twelfth International Conference on Computational Structures Technology, Civil-Comp Press, Stirlingshire, UK, Paper 56, 2014. doi:10.4203/ccp.106.56.
- Tezcan, Y. Dak Hazirbaba, and Q. Cheng. (2014). “Least-Squares-Kernel-Machine Regression for Earthquake Ground Motion Prediction,” B.H.V. Topping, P. Iványi (Editors), Proceedings of the Twelfth International Conference on Computational Structures Technology, Civil-Comp Press, Stirlingshire, UK, Paper 47, 2014. doi:10.4203/ccp.106.47.
- Tezcan, Y. Dak Hazirbaba, and Q. Cheng. (2014). “Semi-parametric Tool for Seismic Risk Management Applications,” International Conference on Urban Disaster Reduction, Boulder, Colorado, USA, September 23-October 1, 2014.
- Dak Hazirbaba, and J. Tezcan. (2014). “A Novel Approach for Modeling Nonstationary Ground Motions,” 12th International Conference on Computational Structural Technology, Naples, Italy, September 2-5, 2014.
- Dak Hazirbaba, J. Tezcan, and Q. Cheng. (2012). “Maximum Direction to Geometric Mean Spectral Response Ratios using the Relevance Vector Machine,” Proceedings of 15th World Conference on Earthquake Engineering.
- Dak Hazirbaba, J. Tezcan, and Q. Cheng. (2012). “Maximum Direction to Geometric Mean Spectral Response Ratios using the Relevance Vector Machine,” 15th World Conference on Earthquake Engineering, Lisbon, Portugal, September 24-28, 2012.
- Dak Hazirbaba, J. Tezcan, and Q. Cheng (2012). “Maximum Direction to Geometric Mean Spectral Response Ratios using the Relevance Vector Machine,” Poster Presentation. NSF CMMI Engineering Research and Innovation Conference, Northeastern University, Boston, MA, July 9-12, 2012.