Design of Flush Endplate Connections in Composite Beams

And the wind still blows ... Several more readers have joined the debate on wind. Dr Nick Cook, who worked extensively on the preparation of BS 6399: Part 2, has responded to various points raised previously in this column and writes from St Albans: First, Mr M. G. Searle has found the typographical error in eqn C.l. It should, of course, read ...

The Chairman So many of us have been brought up on the assumption that the only way you can find out what a structure is going to do is by analysing it before you start the building and then you predict certain movements and you measure them. If the measurements don’t come up to your predictions, the structure is not behaving properly or the measurements are inaccurate - there is nothing wrong with your calculations! The paper outlines a very interesting way of coming at the job from both ends. We get both the feedback and the means of conveying accurately what we have to do. We have to make decisions before we build. We can’t build, monitor and then decide we have built the wrong thing.

Dr J. W. Smith (F)/Mr S. M. Williams (University of Bristol) The author poses the question, ‘is there anything in neural computing for the structural engineer?’ When the finite element method was first proposed its potential in structural engineering was immediately recognised. In contrast, neural computing has certain fundamental attractions and yet its application in structural engineering is not immediately obvious. The author concentrates on the principles of the method but does not provide an example of clear benefit over other methods. The writers would like to suggest reasons why neural computing has been slow to establish itself in structural engineering, and will suggest an application that may be tackled much more effectively by neural computing than by other methods.