After all the devastation and destruction wrought by Hurricane Katrina, you'd hardly think anyone would be up for a repeat performance. But that's exactly what's happening in university labs, government agencies and even at some private consulting companies. In the two years since the storm blew through, engineers and scientists have been poring over computer-based simulations and full-scale models, trying to recreate the conditions of the biggest storm to hit the U.S. in more than a century. By mimicking Katrina's effect on everything from the New Orleans levee system to the off-shore oil rigs in the Gulf of Mexico, engineers are uncovering invaluable insights that are aiding in the rebuilding effort while facilitating designs for new protection systems that will weather more effectively in the event of another catastrophic storm.
The U.S. Army Corps of Engineers, for example, tapped computer-aided engineering (CAE) software, including a wide range of Finite Element Analysis (FEA) simulation and modeling tools, to recreate Katrina's storm surge and wind speeds. This enabled them to evaluate the affect on New Orleans' series of flood gates and levee walls while providing a benchmark for the repairs and subsequent new designs. An interagency task force charged with evaluating what went wrong with the levees from an engineering standpoint employed similar simulation and modeling capabilities to do their forensic analysis. Across the Gulf, private companies consulting to the oil industry are engaging in simulation studies to evaluate Katrina's waves and wind speeds as part of their strategy to build more stability into their off-shore rigs and mobile drilling units.
“As we were designing and constructing, panels were conducting analysis as to why things failed,” says Col. Lewis F. Setliff III, of the Army Corps of Engineers and the commander of Task Force Guardian, the team responsible for restoring New Orleans' flood and hurricane protection system to its pre-storm levels. “We used computer-aided modeling … and other types of technology to find out why the original hurricane protection systems failed. We took the lessons from the analysis of the failures and spiraled it into the designs as we were going.”
While such simulation has always been part of the process for building and infrastructure design, the Katrina experience upped the ante, putting additional scrutiny on the need to fully analyze failure modes. In addition, much of the analysis and simulation effort prior to a storm like Katrina was informed guesswork. “In the past record, you could look at a spectrum of events, categorized in a broad way as to what a 10-year, 50-year or 100-year storm might look like,” says Dale Berry, director of technical marketing for SIMULIA, the Dassault Systèmes brand that markets the Abaqus FEA software. “When a storm of this magnitude hits you, you can go back and really know what it looks like. All of that information can be used in simulations that will aid in the rebuilding effort.”
Project of Unprecedented Proportions
In the post-Katrina era, the Army Corps of Engineers faced a rebuilding effort of unprecedented proportions. The Corps faced three major challenges simultaneously back in late summer 2005: Reestablishing New Orleans' emergency response system, trying to “unwater” the city and taking a long-range look as to how to best repair the hurricane protection system (a 350-mile stretch of flood walls, levees and gates) before the next hurricane season hit. That essentially left nine months to analyze the failures and make the requisite repairs to the more than 220 miles of system that was swept away and damaged by the storm. There was also intense media scrutiny and close supervision by the government. “It was an imperative to not only work fast to get the system complete by June 1, 2006, but quality was a key component of our work,” Setliff says. “We had to work concurrently in all aspects of the program and several key technologies were critical to our ability to assess and design the systems in an expeditious manner.”
Setliff's team was forced to design the new hurricane protection system as it was being built — a practice that presented novel challenges to the Corps' engineers and contracting partners. The Corps' standard building process is far more sequential, Setliff says, with detailed assessments leading to studies, which ultimately culminate in completed designs, that are then handed off to contractors prior to actual construction.
In the case of Katrina, contractors got started on the rebuilding effort with preliminary designs and as analysis from the simulation studies and full-scale models filtered in, they were redirected with the changes. During the planning and preparation phase, the Corps employed LIDAR (Light Detection and Ranging) and GPS technologies to evaluate the levees, many of which were still submerged at the time. The technology combination helped calculate the elevation of the levees and provided coordinates so the Corps could see what was happening with the infrastructure even though in many cases, it wasn't physically visible, Setliff says. Bentley Systems Inc.'s InRoads civil engineering software was also deployed to translate the LIDAR data into a digital terrain map that was then leveraged to assess the levees in terms of the types of damage that occurred.
Simultaneously, an interagency performance task force was charged with doing a forensic analysis of the failures to determine the cause of the levee failures and to ensure subsequent designs addressed their conclusions. According to Dr. Ed Link, a researcher with the University of Maryland and the leader of the Interagency Performance Evaluation Task Force (IPET), the group was handed five directives: Evaluation of the original hurricane protection system's design; an accounting of Katrina's conditions and its impact on the structures; an analysis of the performance of those structures in light of those forces and their history; a study of the consequences of that behavior and identification of any risks going forward.
As part of its work, IPET employed sophisticated modeling and computing technology, including two of the world's largest 200g centrifuges, as well as supercomputer horsepower from the Dept. of Defense to replicate the storm's effects. The Advanced Circulation Model (ADCIRC), a finite element hydrodynamic model for coastal oceans, inlets, rivers and floodplains, was also employed to simulate the storm surge, while the Wave Predication Model (WAM) came into play to mimic the wave conditions at the time of the storm.
“A vast majority of the measuring devices in New Orleans were totally blown away by the storm so there was very little measurable information on the conditions,” Link says. “We had to go backwards and 'hinecast' the storm, and modeling and simulation were the only tools available to help understand in a spatial and time context what happened.”
Using the simulation data, IPET and the Corps were able to uncover a number of things that affected their new design. For example, IPET discovered a deflection in some of the flood walls that was not anticipated in the original design and which gave the Corps a fundamental principle for reconstruction. Specifically, computer modeling determined the levees breached from the backside, a result of overtopping and erosion, which was a departure from the general way of thinking they were washed away from the frontside due to strong currents and waves resulting from Katrina. “That information told us that we needed to have a greater degree of resistance on the backside of the levees as we were rebuilding to prevent erosion,” he says. “In the rebuilding effort, we used all clay materials so the levees would be more resistant.”
Along with the government effort, the private sector also ramped up its use of simulation software to recover after Katrina. The oil industry in the Gulf region was particularly affected and work is underway to determine if structures are stable enough to withstand storm damage going forward. Houston Off Shore Engineering, for one, is using Ansys Inc.'s simulation software to build models that will better predict the impact of storms on the firm's specialty deep water floating systems used to house oil and gas production facilities. Using actual data from Katrina, the firm is now doing more sophisticated analysis on its existing structures and new structure designs to account for the higher stress loads going forward. “What Katrina taught us is that we have to design for higher loads,” says Philip Poll, manager of projects for House Off Shore Engineering. “We're now looking at things in more detail, trying to understand things about the structures we didn't have to before.”
For its part, Delmar Systems Inc. has been working with Ansys’ AQWA simulation solution for offshore floating structures to make its mooring systems for the oil and gas industry more reliable in the wake of the Katrina lesson. Using the software, Delmar builds a hydrodynamic model of a particular rig and mooring system as it was installed before Katrina, then runs a Katrina simulation on that location to evaluate its response. With rig tracking data and GPS information, Delmar is able to detect when exactly a mooring system failed, which is the first step in creating new designs to make the units more reliable. “We have to first understand where we currently are,” says Evan H. Zimmerman, engineering manager for the firm. “It would be difficult to have confidence in proving reliability without having a yardstick and Katrina presented an opportunity to really figure out where we were.”
Based on their analysis, many of oil and gas companies have invested millions in the last two years to upgrade their mooring systems. Many have increased the number of mooring lines or increased the strength of their lines, Zimmerman says, while others are introducing fresh materials like polyester ropes into their new designs to provide more elasticity. Dr. William Chirdon, assistant professor of chemical engineering at the University of Louisiana, in Lafayette, has also done some simulation work in the area of concrete analysis, that he believes can play a role in New Orleans' rebuilding effort. Chirdon is using SIMULIA's Abaqus to interpret the thermal properties of concrete before it hardens, which can be an important factor in modeling the strength and microcracking when rebuilding with this material. “If the numbers you're putting into an equation are not accurate, than your model isn't accurate,” Chirdon says. “My research will ensure that more accurate thermal properties be used in those models. As a result, engineers will be able to predict failures more accurately.