Home Engineering Small Unmanned Fixed-Wing Aircraft Design. A Practical Approach
When embarking on a concept design (sometimes termed preliminary concept design), the designers must first decide which aspects of the product must be considered from the outset, and which tools will be used to make a start. Often, these are based on considerations of
Table 1.1 Design system maturity.
previous designs: it is rare to design from a blank sheet of paper - even Whittle knew that attempts had been made to design and build jet engines before he filed his first patent; the Wright brothers were by no means the first to attempt powered, heavier-than-air flight. Given a few previous designs, it is usually possible to construct some basic design rules that will allow a designer to make outline predictions on a “back-of-the-envelope” basis, see, for example, the many approximate formulae and design curves in the books by Torenbeek  or Stinton . If a great deal of data is available or the product very complex, such design rules may well be encoded into statistically based concept tools, such as those developed by Airbus . Even today, however, it is not possible to provide first-principles analysis for all of the topics given in the list. In a traditional design process, the best that can be hoped for, particularly during the early concept stages, are empirical relationships and simple constraint analyses. The initial framework used to design our airframes is based on a linked series of highly structured Excel spreadsheet pages. Each page deals with a separate aspect of the design and is set up in a highly formalized way with input, calculation, and output sections, definitions, units, notes, and so on, all set in fixed locations. These are supported by static tables of key information for items such as propellers or engines.
The designs produced during concept work can usually be characterized by relatively few quantities. A typical aircraft design may be summarized at this stage by less than 100 numbers. Such designs are commonly produced fairly rapidly, in timescales measured in days and weeks. In the work described by Keane , a wing is characterized by just 11 variables, and in that by Walsh et al. at NASA Langley  a little under 30. The designs produced during concept design will normally be used to decide whether to proceed further in the design process. As Torenbeek  notes, the “aim is to obtain the information required in order to decide whether the concept will be technically feasible and possess satisfactory economic possibilities”. Thus traditional concept design may be seen as part of forming a business case, as opposed to an engineering process. The designs considered following the decision to proceed may well differ radically from those produced by the concept team. The concept designers should ideally allow for this dislocation, so that performance specifications can be delivered in practice by those charged with preliminary and detailed design. The merging of the concept and preliminary design phases within DECODE eliminates these difficulties to a large extent (though it requires a greater effort to set up and develop).
It is in the area of concept design where formal optimization has perhaps had its greatest effect so far. Even so, the decision-making process, where many competing design topologies are traded off against each other, is still normally carried out manually. This aspect of design is beginning to change with the advent of multiobjective optimization and the use of game theory and search methods in decision making. The researchers at the Computational Engineering and Design Research Group at the University of Southampton have built a series of state-of-the-art tools for use in this phase of design.
|< Prev||CONTENTS||Next >|