Luke Muscutt, University of Southampton
Plesiosaurs were a unique reptile that swam in the seas during the time of the dinosaurs. Fossils show that they had four flippers (two at the front and two at the back), a design unlike anything else in the animal kingdom. I wanted to undertake physical experiments to identify how plesiosaurs swam, a question that has baffled scientists for generations.
I built a test rig that consisted of a robotic plesiosaur in a flume tank. I used CompactRIO to control the flippers, collect force data, and release dye into the water to visualise the fluid flow and wake.
Prof. Bharathram Ganapathisubramani
Dr. Gabriel Weymouth
Dr. Gareth Dyke
Dr. Colin Palmer
Plesiosaurs were apex predators of the oceans during the Mesozoic era (200–66 million years ago). These marine giants were an evolutionary relative of terrestrial reptiles that adapted to life in the water. As such, they had lungs instead of gills, which means they held their breath under water like modern cetaceans (such as whales) do today.
Plesiosaurs have four flippers, which is unique in the animal kingdom. What are the reasons behind this propulsion strategy? Did four flippers give plesiosaurs an advantage over having just two flippers of the same total area? These unanswered questions have been the subject of rigorous academic debate. As the hind flippers are behind the fore flippers, they would encounter the wake that the fore flippers create. Would this effect be noticeable and, if so, which motions would deliver the most efficient and effective propulsion strategies? The answers to these questions would determine whether the plesiosaur flapped its flippers in-phase (that is, all four up and down together), out of phase (that is, the fore pair moving up as the hind pair moves down), or with a completely different phase between the front and back pair.
Past studies reached no clear consensus on these matters. Most research has been based purely on morphological and anatomical factors, such as the shape of the flippers and the mobility of the hip and shoulder joints. Although illuminating, scientists cannot prove theories based on these factors conclusively. Another approach scientists have taken is to use human swimmers with plastic plesiosaur flipper shapes strapped to their arms. However, I decided that a more consistent and quantitative method would be to build a robotic plesiosaur.
It was important that the system I design would represent plesiosaur swimming to an acceptable degree of accuracy, so I started by looking at fossil evidence. I was mainly interested in flipper geometry and limits of motion.
The main fossil that I used to determine the flipper geometry was the Collard specimen, which was found recently on a beach in England, and has been carefully handled and preserved. This gives good data about the bone shape, but offers no information about the soft tissue like skin and tendons because these do not fossilise readily. To estimate the soft tissue shape, I compared the bone structure of the plesiosaur to that of other animals that use flippers as their primary propulsion method (such as penguins, turtles, and sea lions). This led me to propose the following flipper shape shown in Figure 3.
I determined the range of motion of the flippers through analysis of the fossilised limb joints. This information combined with the flipper shape suggested that the swimming style would have been an underwater flight stroke, rather than rowing. In other words, the plesiosaur moved its flippers up and down like a bird’s wings rather than moving them forward and back like an oar.
There are a variety of compelling reasons why we should care. For a start, knowing how an animal moved around can give palaeontologists an idea of how fast it could travel, and thus which animals it could hunt or be hunted by. Furthermore, in engineering, there is the longstanding trend of biomimicry, which takes advantage of the way that nature efficiently and creatively solves engineering problems through evolution, and uses this as design inspiration. One application of this project would be in autonomous underwater vehicles. Replacing a propeller with foils that flapped like a plesiosaur’s flippers could produce quieter and more manoeuvrable aquatic vehicles, which would be useful for applications that require stealth and agility. Scientists could use these to monitor animal populations or carry out underwater inspections.
When designing the test rig, I considered using low-cost hardware such as Arduino. I soon saw that this could not match up to my measurement and control requirements. We needed something more powerful and I determined that CompactRIO met my needs. The use of scan mode streamlined my code by removing the need for FPGA programming or compiling, and ensured that my measurements were synchronised. Also, the reconfigurable nature of the CompactRIO platform meant I could reuse some modules from a past project, saving considerable money.
For storing the data, the TDMS file format seemed the most suitable. I stored the readings of each of the three load cells (analogue input voltage levels) at 200 Hz for 100 flap cycles, which gave me up to 600,000 data points per case.
The final step was to complete the software design. I took advantage of the University of Southampton’s active Academic Site License by sitting training courses. First, I took LabVIEW Core 1 and 2 online, and then received hands-on classroom instruction for LabVIEW Real-Time 1 and 2. This helped me build and maintain my own code. Even with this training, however, the complexity of the system with its many interacting components and protocols produced many problems and challenges that were hard to overcome. Some of these problems were due to the fundamental code architecture, so I sought help from an NI Applications Engineer to restructure and debug the code. I settled on a producer-consumer architecture that utilises queues in which the producer loop checks the state of the controls on the front panel and controls the state of all the loops. There are five separate consumer loops for each axis and the load measurements, which work in parallel and allow for more systematic and reliable motion and acquisition. After rearchitecting the code to an established LabVIEW design pattern, the system worked well.
I conducted my experiments in the recirculating flume tank in Southampton University’s Experimental Fluid Mechanics Laboratories. I designed experiments to gather two types of data: quantitative force and moment data, and qualitative flow visualisation using dye injected into the water.
I found that the wake of the fore flipper greatly affected the forces on the hind flipper. When a flipper moves up and down, it creates vortices on its leading edge, which detach from the surface of the flipper and are shed into the wake. The way that the hind flipper interacts with these vortices determines its hydrodynamic performance, which is how much thrust and efficiency it produces. When it intercepts the incoming vortices, it has low performance, but when it avoids the vortices it has a high performance. In fact, the thrust on the hind flippers can be 50 percent higher than the fore flippers. This means that if the plesiosaur ensured that the hind flippers avoided the vortices, it would produce the thrust of five flippers instead of four. Whether the hind foil intercepts or avoids the vortices depends on the flapping frequency, flow speed, spacing, and phasing between the flippers. So, there is not one simple answer to the question ‘how did plesiosaurs move their flippers?’, as it depends on all of these kinematic parameters.
NI hardware and software has helped me deliver unique insights into how plesiosaurs propelled themselves through our oceans, and showed how the relative motion between the two pairs of flippers has a critical effect on thrust and efficiency. This is because the kinematics control how the hind flipper interacts with the wake of the fore flipper. If the hind flipper avoids the vortices its performance would increase and the plesiosaur would swim faster. Whether the hind flipper avoids the vortices depends on the swimming speed, flapping frequency, spacing, and the phasing between the flippers. There is not one motion that gives the best performance over all conditions. This means that each plesiosaur would have its own motions that it would use for different speeds, and it would learn these as it grew up. I am currently continuing the research as a post-doctoral fellow through funding from the EPSRC doctoral prize and am investigating how these findings apply to different species of plesiosaurs.
The Hydrodynamics of Plesiosaurs
Publications related to this project can be found on ResearchGate
Luke Muscutt
University of Southampton