Shape of things to come

For most engineering designers, the most useful tool is their instinct. When designing a particular component, they will often go for the shape that feels right. But according to Atul Bhaskar, an aeronautics specialist at

, very often instinct isn't good enough.

Bhaskar is leading a team which is developing a computer tool for searching out the most efficient shapes for aerospace components. This will help engineers understand and predict how a change in shape will affect the weight and strength of a particular component, and how it might perform in service.

'When we design mechanical components, it is important to make the best use of material by choosing an appropriate shape and material connectivity,' said Bhaskar. 'Designers often make this choice intuitively.'

The key factor for the design is the topology of the components — the level of 'connectivity' between the various load-bearing sections of the shape, which is related to the amount of material needed to make the component.

If great stresses need to be carried from one section of the component to another, then those sections need a high level of connectivity and a lot of material; if a section is not load-bearing, the material can be thinned down, or even removed altogether. But working intuitively, it's difficult to grasp which areas need connectivity and which don't.

'Given that human intuition can miss out on interesting design options that are mechanically efficient, this approach is never satisfactory in situations that are demanding in terms of the designs being light and strong, as in the aerospace industry,' said Bhaskar.

Currently, engineers 'search' designs by changing their parameters and assessing the mechanical performance of the new shape by calculating the stresses on it when given forces act upon it.

But aerospace components are very often complicated shapes, and a large number of calculations is often needed to model the performance after a relatively simple shape change. 'The search is very time-consuming,' said Bhaskar, 'and given the finite resources of the computer, it may be restrictive.'

However, there might be a simpler way. What interests the engineer, said Bhaskar, is what happens to the mechanical performance when a shape is altered in a relatively simple way, such as stretching it in one direction. This could lead to a short-cut. 'The nearness of shape will be compared with the nearness of mechanical response when similar — but not identical — objects are stressed,' he said. 'This correlation will be used to find efficient designs while avoiding unaffordable amounts of calculations.'

Bhaskar's team is carrying out a range of studies on the effects of changing both the topology of a component and its overall shape, and how this alters the performance of the component.

'The relative importance of connectivity and shape on mechanical efficiency of designs will be explored,' he said. 'The information about the change in connectivity and its relationship with with the performance of a design will be used for finding optimal designs.'