Dragons of the Air

The methodology of model construction and how we obtain our data is constantly changing, however, this post is here to cover some of basics and give any interested parties an idea of how we are progressing through.

Template of Coloborhynchus robustus (SMNK PAL 1133) in gliding flight

Building the models

The first stage of the project was involved with reconstructing the pterosaur bauplan and the various angles that were possible between elements of the fore arm (leading edge of the wing) and hind

limb so that an accurite model could later be constructed for any number of pterosaur taxa. Luckily there are a number of complete specimens in which it is simple enough to rearrage the limb elements back into their natural configurations and several more fossils, preserved in three dimensions, where it is possible to investigate the range of freedoms available to each joint of the fore and hind limb. This allowed us to build up several templates from which we could later use to reconstruct the physical models (see below). While most other studies have concentrate

d on model reconstruction of only one pterosaur taxon the work group selected 6 taxa for experimentation to give a broad cross section through the pterosaur lineage. These were Anurognathus, Rhamphorhynchus, Pterodactylus, Coloborhynchus, Auroazhdarcho [in press], and Sinopterus.

Left to right: The mounting strut fixed to the base; a headless Anurognathus sits in the prep lab; the finished Coloborhynchus mounted in the wind tunnel.

The starting models were constructed from a combination of wood, modelling clay, foam and PVC in such a way that they closely resembled the desired template but were immobile and so were designed only to investigate what pterosaurs would be like as fixed wing aircraft. The base was made of plywood that was sawn to match the body outline of the chosen pterosaur and the wing spar was bolted on to this and fixed with modelling putty to produce a stable frame. In order to mimic the complex shape of the pterosaur fore arm the individual elements making up the wing spar were carfully sawn so that they slotted together at the correct angle and immobilised by an aluminium strut through the center of the joint. Any gaps were patched up with modelling putty. The body was constructed of foam while the “membrane” wing was simulated by a thin sheet of cambered PVC that spanned the length of the wing spar and extended down to the ankle of the model. A mounting strut used to fix the model on to the wind tunnel balance was later bolted onto the base and all gaps were patched up with modelling putty to give the desired shape. The model was lastly sanded down and painted.

Testing the models.

Once the models are complete they are mounted on a specially made bracket in the wind tunnel at the University of Karlsruhe. The balance is located at the bottom of this bracket and is hooked up to a computer that streams the data as the experiments are run. To simulate the actual flight speeds that pterosaurs would have encountered the flow velocity is increased from 4 – 20ms and calibrated against a zero wind run. After each experiment the angle of attack of the mode is altered and the balance is recalibrated. This provides a relative quick and simple method to record values for forces (lift, drag, side force) and torques for each of the various models.

Once the data has streamed off the balance plotting it is a fairly simple matter even using a basic programme such as Excel and the forces are converted to non-dimensional coefficients so that all models can be compared irrespective of size. One method of comparing flight performance involves plotting a polar of the total drag coefficient (CD) against that of the lift coefficient (CL) which typically gives us a U-shaped curve. Calculating the quadratic curve of this polar allows the ratio of lift to drag to be found at any given point and thus, in turn, the sink speed of the animal at a given velocity. This is the typical Vsink plot used by studies to compare the gliding peformance of different aircraft. Those familiar with these types of plots might draw some conclusions from the example below but these will be discussed in later posts.

More to come.

This post covers the basic starting points of the project, however; as many more models are currently in production including those with flexible joints and membrane wings expect to see some methodology updates soon. There are also a few new ideas we want to try out in the wind tunnel including flow visualisation so keep watching this space.