Experimental Testing and Validation
Although computational methods have improved enormously over the last 20 years, it remains the case that the results obtained from computational fluid dynamics (CFD) and finite element analysis (FEA) cannot be relied upon to be completely accurate in a number of key areas. In particular, it is still very hard to predict the stall speed of an aircraft in landing configuration, to establish the ultimate failure loads for key aircraft structures, or to accurately compute the natural vibrational frequencies. These limitations mean that experimental testing retains an import role in the development of any aircraft design. Such tests can be used to validate computational results or to assess designs in the absence of computational modeling. Experimental testing is not, however, without its own problems. First and foremost is the cost involved, both in terms of the facilities needed and also in designing and building suitable aircraft parts for testing. Large facilities are expensive to build and maintain, and so it is relatively rare for prolonged experimental programs to be carried out for low-cost unmanned air vehicle (UAV) designs. Moreover, the parts used in testing must have sufficient structural integrity to survive in the wind tunnel or be realistic for structural load or vibration tests. This means that a good deal of design effort has to be expended in getting from the simplified geometries considered earlier to those that can be used in test part construction. We are fortunate in having routine access to the UK’s largest academic wind tunnel which can house full-sized UAVs up to around 15 kg maximum take-off weight(MTOW) and entire built-up wings of larger aircraft. We also have a number of facilities to help carry out static and dynamic load tests on key structural elements. Therefore we usually do carry out quite detailed experiments on our UAVs. In all cases, the aims are to check that preliminary design calculations agree with reality and to update and adjust any empirical factors used during design, for both the current and subsequent aircraft projects.
The design of parts for experimental testing leads to a key choice in the design process. Should a short program of experimental validation be carried out on simplified parts to establish that preliminary design calculations have been adequately good predictors of performance before the costs of full detailing are entered into; or should a separate program of design
Small Unmanned Fixed-wing Aircraft Design: A Practical Approach, First Edition. Andrew J. Keane, Andras Sobester and James P. Scanlan.
©2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.
be set up to make fully detailed test parts? Of course, given suitable facilities, experiments can be carried out on the completed airframe before the first flight to finally establish airframe properties; but if any shortcomings are revealed at that stage, it can be very difficult and costly to correct them if hundreds of hours of detailed design effort has already been committed.
Our approach to this problem is based around the rapid manufacturing processes we adopt combined with a heavy focus on as much parametric CAD modeling during detail design as possible. We then carry out experimental tests on full-sized preproduction-quality parts that can lack some details, and iterate the detailing effort where necessary. The biggest weakness of this approach is that one tends to be somewhat cautious in terms of the selective laser sintered (SLS) nylon structures, which represents the largest commitment of manual detail design effort, and this can make such parts rather heavier than one would wish. Much effort can often be devoted to controlling weight and carrying out detailed stress calculations to ensure that such parts are as optimized as possible. In our experience, even quite poor design concepts can be made to perform adequately with good detail design, while poor detail design will wreck even the best concepts.