This lecture was presented at the 3D Digital Documentation Summit held July 10-12, 2012 at the Presidio, San Francisco, CA
3D Technology and the H.I. Hunley: Beyond Documentation
Since raising the H.L. Hunley submarine (1863) from the seabed in 2000, the project has incorporated 3D technologies including laser scanning, patterned light scanning, Vulcan point system, and computer modeling. Archaeologists developed the use of 3D technologies primarily for site plan development and artifact documentation. As the project progressed, conservators found other applications for the technology in conservation treatment assessments and in the rotation of the H.L. Hunley in 2011.
After the recovery and throughout the excavation, 3D technology has been instrumental in documenting the Hunley and establishing the site plan. Before the excavation began, surface mapping with the Cyrax©3D laser scanning system generated a global 3D model of the submarine’s outer surface. With that documentation in place, the submarine was opened to allow for excavation of the interior. During the excavation, point mapping with the Vulcan 3D point system spatially recorded artifacts, human remains and submarine features. After these items were removed from the submarine, detailed 3D scans were taken with a Minolta © Vivid 910 laser scanner and a Breuckmann© OptoTOP-HE structured light scanner. Integrating this information into the master site plan has produced a powerful visualization of the site and archaeological record (Rennison et al. 2009).
The most recent documentation has utilized the Breuckmann© OptoTOP-HE structured light scanner, which provides high resolution 3D data capture integrated with color photography, to document the outer concretion of the hull. Since conservation treatment demands the removal of this concretion, which contains archaeological data regarding site formation, a comprehensive 3D documentation was undertaken. Conservators have also taken advantage of this sophisticated scanning technology to assess conservation treatments (Crette et al 2010). Organic materials such as wood and cork have been scanned before and after conservation, to digitally compare the shrinkage of the objects. Iron artifacts have also been scanned to assess experimental conservation treatments.
3D technology served an important role in the engineering studies of the hull, as well. To prepare for the rotation of the Hunley, successfully completed in 2011, a 3D geometric model was created to perform finite element analysis using nominal measurements from the submarine (Blouin et al 2011). This model was used to visualize potential stress distributions on the Hunley and aided in developing an overall approach to the rotation. Once decided that the rotation would proceed using the pull and release of 15 supporting slings to rotate the Hunley, more detailed spatial information was needed to compute the amount of pull and release for each sling. The simplified model created for the finite element analysis did not take into account the submarine’s relationship to its supporting slings or the outer concretion on its surface. To acquire this information, conservators and engineers collaborated with archaeologists to process existing data and collect new data. Using the site plan, virtual cross-sections were created for each sling location that accurately accounted for the concretion and overall shape. New data was collected using the Vulcan 3D point system to define the spatial relationships of the submarine to its supporting elements. With this information, engineers calculated the amount of pull and release for each sling using numerical modeling. Cross-sectional measurements were also used to create a full scale sectional mockup to test and refine the computer-based calculations.
Although some problems arose while using 3D data collected for archaeological documentation for engineering purposes, the outstanding contribution of 3D technology to the project remains undeniable. Having completed the comprehensive scanning of the hull and hundreds of artifacts, archaeologists have now laid the foundation for future collaborations.
Church: Chris joined the Warren Lasch Conservation Center in 2008. He works on the salinization treatments for the H.L. Hunley submarine and associated artifacts. Working on the June 2011 rotation of the Hunley, he helped apply engineering techniques to the conservation field. His current research aims at facilitating technology transfer for industries in conservation to develop viable technology for the stabilization of large archeological artifacts. Without further ado, here’s Chris.
Watters: Thank you Jason. That’s’ a tough one to follow. From Toledo Lombardi to the H.L. Hunley, who knew they had so much in common.
First, a little background on the project. This is the H.L. Hunley. It was built in 1863 out of cast and wrought iron in Mobile, Alabama and shortly moved to Charleston, South Carolina in the effort to break the naval blockade. Eight men manned the crew. It took the forty foot vessel over three miles off shore powered by a hand crank and detonated an explosive charge into the 205 foot USS Housatonic, becoming the world’s first successful combat submarine. The crew in the vessel didn’t make it back home until the year 2000, when it was recovered from the ocean and brought to the Warren Lasch Conservation Center, now part of the Clemson University Restoration Institute for further study and conservation.
Here we see the H.L. Hunley supported by its slings, and the slings are attached to a truss and that sits on the tank floor and it remained in that configuration through the excavation until 2011, just last year, where we rotated the submarine, putting it flat on the tank floor. This was a big project. It required the collaboration of engineers, ship builders, industrial measurement specialists, professional riggers and as a conservator; I was kind of a liaison between all those different groups.
I’m not an archeologist. I don’t actually take 3D capture and I’m definitely not an engineer, but I just wanted today, to kind of give you a recap of the 3D technology that the archeologists have employed over the last decade and what we used for the engineering study for the rotation and the overlap therein.
So, a quick recap. These are our archeologists. The Hunley is an artifact but it’s also a self-contained archeological site. To investigate the site, we had to modify the artifact. Before that, we wanted to do a pre-disturbance survey and 3D scanning was chosen as the technique to do it. This was way back in 2000 and laser scanning was kind of a new technology. We used Cyrax 3D laser scanning and this kind of gave us a very good indication of the overall structure and formed the foundation for future 3D site plan.
Before we opened it up, we had no idea what was going to be inside. The ensuing excavation uncovered more than 1500 artifacts and the remains of all eight crew members. To document the position of all those objects during that excavation, a Vulcan 3D measurement system was used and that gave an x, y, & z point on different points on the objects. Points on the objects were selected and given an x, y, & z value and this particular system was used because the measurement one has an offset laser triangulation and that was really handy for getting in tight spaces and working around the truss and the slinks. Often, line of sight was impossible.
The Cyrax scan and the network of points developed with the Vulcan system were integrated using Rhinoceros software and this gave a flexible platform that allowed for CAD model elements. You can see some of the architectural features of the submarine to be integrated with the points and the laser scans or the direct scanning of the artifacts. For example, in 2004 we used Minolta Vivid laser scanner to scan all the bones before they were reinterred. It was a fairly high resolution at the time so this resolution can be compressed back down and then reinserted into the 3D site plan according to those Vulcan point sites. So we have to have three points of course, and that gives an exact archeological placement of that bone and that was done for all the bones. These are each different color which represents a different crew member. It provides just an incredible visualization tool and really sets the bar for archeological documentation I think.
This continues, this campaign continues. You can see on the right, they’re a little transparent but those are the crew benches that are modeled with CAD. Here you see a direct capture data that was done using the Breuckmann pattern light scanner, the up to top HE and this system was acquired in 2008 and has really been the workhorse of the project. It was chosen largely because you can use a wide range of field of views. You can do small artifacts, you can do big artifacts and you can also collect color data with it. The accuracy allowed us to do assessments of conservation treatment. This is a cork, a water-logged cork that was dried and was scanned before it was dried and after it was dried and then the shrinkage was measured here. Yellow and green is relatively little shrinkage, the blue is a little bit more, so we had a little bit around the edges. The red would indicate expansion. We didn’t have much of that. This work was presented in the last WOAM Conference. I think the proceedings just went online at lulu.com and this equipment was also a great answer to a problem we faced.
As conservators, we’re usually chiseling, breaking, generally destroying concretion to get to the original surface of the artifact. On the H.L. Hunley, this concretion had very valuable archeological data in it. It was a record of everything that’s happened to the sub since it sank, possibly holding the clues to how and why it sank. So it needed to be preserved and what ensued was a campaign to document the concretion of the hull. The Cryrax scan that was good for general distance only had about 30 million points in it, so it wasn’t really high resolution enough to be able to define the morphological features of the concretion. So the Breuckmann was used, and these are all the scans that have been stitched together to create an overall model. This is over 1500 color calibrated scans but this model couldn’t be completed until after the rotation. That was because of the limitations of working in the tank and the standoff distance of the equipment. This wasn’t the only consideration in rotating. We also worried about the long term structural stability of the sub and we worried about the slings inhibiting the desalinization treatment.
So how did we rotate it? The first step was to elevate the truss and the submarine and insert the supports. Half the slings supporting the submarine were dropped to make room for those supports and the foam was replaced with a more compliant, we actually used large planks of ethafoam instead of a hard polyethylene or polyurethane foam. Half of the slings were dropped and then we used chain hoists all along the port side to release it simultaneously. We lengthened one sling and just kind of let it drop down with minimal adjustments on the starboard side using turn buckles. That was really the culmination of years of study and research.
Why were we so cautious? Well in the words of the head conservator, Paul Mardikian, you should always know what you don’t know and as a conservator you kind of have an intuition of how to hold things, how to support things, when something is going to break, but when it reaches a certain complexity, you can’t really rely on just your intuition and with over fifty major architectural pieces riveted together, 140 years on the seafloor, unknown degrees of corrosion, and partially disassembled, that’s sufficient complexity to ask for help. We turned to Dr. Vincent Blouin, who’s a professor at Clemson. He’s a naval architect and a mechanical engineer and we asked him a couple of questions. They were simple questions but the answer was a little bit more difficult. Is it better to rotate the sub? Is it better for the sub to be upright or on its side or can we drop fourteen slings, would that be adequate support for the submarine? And can we rotate it without structural failure? To answer these questions, he came up with a finite element model and as we just heard finite element models take a very complex shape and take it down to more, simpler forms. Into those forms you can insert physical properties or ranges of physical properties, such as density, strength, and thickness. Then you can visualize that on a global level. We can look at the theoretical stresses that are being applied to the submarine. The answers we got were that yes, the submarine would be better supported in the upright position. Yes, it could be done without breaking it and yes, we could break it in the process. The way that we did that was to look at the theoretical internal forces and compare that to known yield strengths of historical iron. The reason that the buildup of stresses could cause it to break is the non-uniform profile of the bottom. So, in the worst case scenario, all the same length of the slings would all be released at the same time and that would induce distortional forces that you see in the bottom right. So, this is what we were trying to reduce.
Armed with some justifications for the rotation and a word of caution, we proceeded, and as we developed and as we got more into the planning, the finite element model actually became a very practical tool for planning things like how much chain do we need, how much of the release is going to happen, where is the sub going to land, all these kinds of things were answered by Vincent. It’s good to have the theory and it’s good to have what you think is going to happen but to double check that, we developed a full scale two-sling model of the sub, and we integrated a load cell system which measured the tension on these slings. The tension can be related to the vertical and horizontal forces that would be applied to each profile, so we’re kind of breaking the sub down into fifteen different profiles and looking at what’s going to happen when we rotate.
This was actually a very useful model. We figured out some assumptions that were wrong on the finite element model and that allowed us to refine the model. One of the assumptions was that the axis of rotation would stay fixed. Well it didn’t. It was more of a dynamic axis, so it would move up and down during the rotation. So that showed us that it was feasible to do a release only rotation. Before, we were thinking about more of a pull and release, and we actually performed it as basically a release only but with a little bit of adjusting on the starboard side with turn buckles. This was very practical information to have, you know, how much do we need chain hoists on both sides, what kind of materials do we need to use and how it worked out is Vincent was watching this load cell system that was applied to every sling on both sides, and he could visualize the load profiles and he could visualize the direction the sub was going and so he would call out, position 1 through 5 and 7 through 14 go a half turn, position 6 through 10 go a quarter turn or something like that. So that’s how we rotated it, millimeter, by millimeter, half turn by half turn.
From the very beginning, I wanted to get the archeologists and their great reserve of incredible shape data together with the engineers because of course, the finite element model is only as good as the data you put into it. The problem with this of course was there was just too much data. One of Vincent’s jobs was to pare down the data and figure out only what he needed for the model and this was just too much information, but we used the Vulcan point system which was the same system that we used to place the artifacts in the 3D site plan to get a relation between the tank floor, the slings, the sub, and the truss. So that allowed us to refine the geometry of his model.
It was a very important to have a good understanding of exactly how far this sub was going to drop because all of our supports were prefabbed and so it would be very difficult in the middle of the rotation to change the design. It really came down to a matter of inches, a matter of millimeters, and it was very handy to have the archeologists scan data on hand. I could just run up to them and say, “Hey, what’s the offset for the keel blocks, how much of a give and take factor do we have there?” For the construction of the test rig, they gave me sections from the data, and I went next door to the shipyard and they put it on their dual plasma CNC milling machine and cut out perfect profiles and that’s how we got the test rig.
Now, during the rotation we knew the sub was going to move. We just didn’t know how much it was going to move and whether that moving was going to be a good thing or a bad thing because we didn’t really have an idea if it moves, is it going to be relieving tension or is it going to be creating tension. So we kind of had some boundaries established and we had some monitoring tools in place to watch that. Of course, visual inspection, we were constantly looking for cracks. What I would think is that probably the crudest piece of 3D equipment we had was a laser pointer mounted to the stern with a target on the bow. That gave a very quick and dirty kind of visualization of any kind of movement and of course, we had the load cell system but we also wanted to use the expertise of the archeologists and incorporate a photogrammetric survey.
When I first came to the archeologists, I said, “Hey, you guys have got this great laser scanner. Can you measure the deflection between the bow and the stern?” Looking back and knowing the technology now, I know how foolish that was. They came up with a photogrammetric survey. So we had points on the truss. We had points on the sub and we had points on the tank wall. You can see the photo positions and the positions on the sub and the truss. This is the truss actually rotating around the sub. You can see that. They kind of normalized the points on the sub with the points on the wall and the truss. Out of 77 points, we had an average deflection of about a half a millimeter and a max deflection of about 2.7 millimeters and that kind of roughly corresponded to what we were seeing with the crude laser pointer and target system. According to the finite element model, we kind of had an ideal max below 5 millimeter deflection and an upper range of 10 millimeters. Like I said, is that a good thing or a bad thing, we’re not sure, but we have plans to continue the photogrammetric survey and watch for any kind of movement of the sub and we also have a load cell system in the supports that it’s now sitting on. So we measure that weekly if not daily sometimes.
In conclusion, at the Warren Lasch Center, the archeologists, the conservators, and the engineers continue to work together to integrate the direct capture into the engineering field more. The rotation was really just the first step, and I really see the digital documentation as the first step that will lead to more analytical applications that are really going to alter how conservators go about their daily work. That’s only going to happen if you have a good understanding of the technology and we incorporate into our planning, we make room into our planning for scanning and rescanning and I think this is just the beginning.
Christopher Watters joined the Warren Lasch Conservation Center in 2008. He works on the stabilization treatments for the H.L Hunley submarine and its associated artifacts. Working on the June 2011 rotation of the Hunley, he helped apply engineering techniques to the conservation field. He was also on the organizing committee for the interim meeting of the ICOM-CC Metal Working Group that was held in Charleston SC, in October 2010 and co-editor of the proceedings.
Starting his career in the studio of a paintings restorer, during graduate school Chris switched his focus to objects conservation. Interning at the National Gallery of Art in Washington, DC, he gained experience in the conservation and technical studies of sculpture dating from the Renaissance to the present. His specialization in archaeological conservation began while working on-site with the Middle Eastern Cultural Center of Japan at the Kaman- Kalehöyük excavation in Turkey and in the objects lab of the Brooklyn Museum of Art treating and preparing artifacts to travel in the exhibition To Live Forever: Egyptian Treasures From the Brooklyn Museum. His current research aims to facilitate technology transfer from industry to conservation and develop viable technology for the stabilization of large archaeological artifacts.