Things that Have Gone Wrong and Why
We have learned these lessons the hard way: over the years we have had a number of “uncontrolled descents,” mostly of student-designed and -built aircraft (but not always). We try not to intervene too strongly into our student groups during their project work since we find they tend to learn more if allowed to make their own mistakes. What follows is a list of some the things that have gone wrong and why - it is a bit of a rogue’s gallery but we would not pretend that we only ever make perfect aircraft at the University of Southampton:
- • A failure to ensure a positive static margin led to an aircraft taking off and immediately stalling during a fast taxi test (this occurred when a test was being carried out without Paul being present).
- • A failure to ensure dynamic stability led one aircraft to exhibit uncontrollable Dutch roll. This was a flying boat and the very large area of hull forward of the aerodynamic center needed a far larger vertical tail volume than is normally required on conventional aircraft, see Figure 21.4 (note that an XFLR5 analysis would not have revealed this since XFLR5 does not model fuselage side forces).
- • A failure to ensure that a set of laser-cut plywood spar parts had been correctly placed in the clamping jig during manufacture led to a main spar failure during extreme maneuvers. A simple spar load or ground vibration test would have revealed this, but neither was carried out prior to flight.
- • An engine failure on a single-engine aircraft meant that the aircraft did not have sufficient height to glide back to our test runway, leading to a controlled ditching in the adjacent field. The low-speed screw on the carburetor was poorly set, and when opening the throttle rapidly from idle, the engine stalled.
- • The adoption of a radical, all-steerable and split elevator on a research aircraft had a highly nonlinear impact on controllability, causing the aircraft to suddenly nose-dive on a steep turn into toward final approach, see Figure 21.5. The aircraft had previously appeared to have benign characteristics. One should always fully explore the envelope with sufficient altitude to maximize the chances of recovery from such mishaps. The rebuilt aircraft has a fence at the elevator split, which mitigates this problem.
- • Engine vibrations caused an autopilot to fail. We now extensively test the mountings of our autopilots in the lab prior to operations, using signals measured on the target aircraft, see Figure 21.6.
Figure 21.4 Student-designed flying boat with large hull volume forward and insufficient vertical tail volume aft.
Figure 21.5 Aircraft with split all-moving elevator. (a) Without dividing fence. (b) With fence.
Engine vibrations caused a power isolation switch to momentarily open and thus initiated the engine ignition safety cut-out system, forcing the pilot to glide the unpowered aircraft back to the landing strip.
• A sizing error during concept design led to an aircraft being built with wings that were too small given the installed power; this was not revealed until its first catapult launch, Figure 21.7.
Figure 21.6 Autopilot on vibration test.
Figure 21.7 Student UAV with undersized wings. The open payload bay also added to stability issues.
Figure 21.8 2SEAS aircraft after failure of main wheel axle.
• Repeated autonomous landings on a concrete runway led to low-cycle fatigue of a main wheel axle due to the significant impact loads experienced. Subsequent use of laser height finding mitigated these issues on automated landings, Figure 21.8.
Despite this apparent litany of failures, our aircraft fly successfully much more often than not; so we would encourage anyone setting out to design and build his/her own unmanned air system (UAS) to go right ahead. It is not nearly as daunting as one might initially think (even given the size of this book) and it can be a great deal of fun. Certainly, our student teams enjoy the experience enormously and clearly get a great thrill when “their” aircraft takes wing for the first time. Even those whose aircraft do not perform as desired learn a great deal from actually being directly involved in the whole process from first ideas to final flight.