I consider myself lucky to have been able to perform contract engineering research and consulting for my entire career. The work environment has been somewhere between industry research and academia. I’ve had opportunities to work with recognized experts at universities, national laboratories, in government and in industry. In addition, without changing my position, I get the opportunity to work on a new and interesting area of research every few years. When I started my career, my initial projects were on the dynamic buckling and collapse of structures. Five years later I was working on damage and fracture of welded steel structures from blast and impact loads. Five years after that, I was working on a project to reverse-engineer the structural design of an automobile and developing a detailed computer model for crash safety applications.
In addition to working on a range of interesting and challenging projects, I was in an environment where I was exposed to many more areas of research. The group I worked with was highly collaborative, and there were opportunities to support other projects. There were also regular seminars where colleagues would present the results of their research projects.
For anyone who has had the opportunity to follow a similar career path, there will be one or two projects that stand out above the others. The reasons could be the importance of the work, the technical challenges that needed to be overcome, and the quality of the work that was achieved. For me, the work we did as part of the National Institute of Standards and Technology (NIST) World Trade Center (WTC) investigation is one such project.
The WTC investigation work stands out, in part, due to the nature of the project. Living on the West Coast, I woke up on Sept. 11, 2001, to news that an aircraft had collided with the north tower of the WTC. I turned on the television and, like most of us, watched the events unfold for the remainder of the day. By the end of the day I knew that the world had been significantly changed by what happened.
When NIST announced it was looking for a partner to perform the detailed analyses of the aircraft impacts and structural damage to WTC towers 1 and 2, I felt that my prior research experience, and the team of researchers I was working with at Applied Research Associates Inc. (ARA), made us ideally suited to the project. We had experience making detailed models of vehicles (train cars and automobiles) for crash analyses; we had experience modeling high-speed impact and penetration; and we had extensive experience with the software system that we believed had the highest potential to successfully achieve a good answer for all of the complex damage processes that would occur in the impacts. We also worked with a range of experts on related issues such as video trajectory analyses that could help define the exact conditions of the aircraft at impact.
The primary objective of our work was to determine the condition of the two WTC towers immediately following the aircraft impacts. This included both the determination of the structural damage that degraded their strength and the condition and position of nonstructural contents such as partitions, workstations, aircraft fuel and other debris that influenced the behavior of the subsequent fires in the towers. I remember our team performing the work under a high-pressure environment, with a tight schedule, and subject to a high degree of public scrutiny. I didn’t have any other point in my career where we all were working such consistently long hours. However, I also remember the work as being collaborative and rewarding. Everyone involved in the NIST investigation was committed to the integrity of the work they were performing.
The team at NIST was supportive of our efforts and had collected all the information needed for the analysis of the tower structures. They performed characterization work on materials recovered from the site. NIST had also developed an extensive database of the photographs and videos that were necessary for evaluating the impact conditions and post-impact observable damage to the towers.
Our work was divided into two parallel primary efforts: developing a model of the aircraft, and developing the models of the towers. Along the way there were constant technical challenges that needed to be overcome, such as getting access to detailed structural design information on the aircraft, modifying the aircraft geometry data collected on a parked aircraft to account for deformations from the flight loads (e.g., upward wing curvature), and modeling the fuel being carried by the aircraft. Similarly, the building structure looked relatively simple from the exterior, but design details such as the construction of the exterior columns included a large amount of variation in material strengths and thickness based on their position both in height and around the tower. An automated approach needed to be developed to incorporate all the variation into the model.
In addition to the model development, there were parallel efforts to evaluate and confirm the accuracy of the modeling methodologies that we were using. This started at the smallest level, where we compared the material models against the laboratory testing that was performed to measure the strength of the tower steels. These were then used in detailed submodels of individual columns to investigate the impact response. These could be compared to the impact damage on recovered columns being held at the NIST facility (a somber reminder of the reality of the work). Increasingly larger and more complex models were developed, and at each step the changes in the model were compared back to the previous level of modeling.
As we finalized the models for the aircraft and towers and assembled the global impact models, I was anxious. I believed these analyses were at the limits of the software and computer capabilities we had. They were by far the largest and most complex analyses I had ever performed. There were no guarantees that they would run to completion without difficulties.
Each of the analyses required approximately six weeks of continuous run time on a high performance (for that time) computer cluster. I was relieved that they ran smoothly to completion. Subsequently, when we began to evaluate the results, I was impressed with how well the calculated damage correlated with the observable impact damage obtained from photographs.
It was a satisfying feeling to know that the hard work we had performed, and methodical approach we developed, resulted in a high-quality simulation of the impact. The good correlation of our results with all of the observed responses of the towers to the impact gave us and the NIST team confidence in our predictions revealing the extent of the damage to the interior of the building, which was critical to the analyses performed by other investigators working on the project. Looking back, I’m still proud of the work we did and our contributions to the NIST WTC investigation.
Read other blogs in this series:
20 Years Later: NIST's World Trade Center Investigation and Its Legacy
How 9/11 Changed Me and First Responder Communications
Putting Together the Big Picture for the World Trade Center Disaster Investigation
Reflections on Assisting With the 9/11 World Trade Center DNA Identifications
Inside the Towers on 9/11: My Story of Investigating the WTC Evacuation
NIST/ARA's impact orientation of the 2nd plane in the LS-DYNA model looks way different than what can be deduced from visual records. In reality, it impacted WTC2 at a horizontal angle nearly perpendicular to the south face. It becomes especially apparent in the Hezarkhani and Courchesne footages, where both engines start to enter the facade at nearly same time. Why does the impact orientation of the 2nd plane in the NIST/ARA model extensively differ from the recorded reality?
Doesn't this inconsistency invalidate every other analysis step that bases on this assumed (false) impact orientation, like the damage estimate to the structure and fireproofing; the dispersion of debris, jet fuel and other combustibles inside the impact zone; as well as the modeled fire progression and collapse initiation?
See NIST NCSTAR 1-6 (Sept. 2005), Figure E-1: Critical analysis and inter-dependencies.