Structural frames are usually designed to satisfy deflexion requirements by trial and error. Recent optimization methods are restricted to small or very specialized structures. In the general method proposed herein an efficient analysis program is used to determine linear approximations to the rates of change of deflexions with the section properties of the frame members. A simple linear programming procedure then determines the most economical changes of section required to satisfy the deflexion constraints. Cycles of analysis and optimization continue until a satisfactory design is obtained. Comparisons with other methods show that this design is likely to be close to the optimum. Haunched frames can be designed and the reduction in frame stiffness due to compressive axial loads can also be allowed for. Examples of the application of the method to plane steel frames are given. D. Anderson and J. Salter
The President: Before l propose a vote of thanks to our authors, I would like to make a few remarks about the profession. We, in fact, hear many remarks about our profession when collapses or other troubles occur. This is a paper about an extremely difficult job-one which posed ever-changing problems. The outcome was successful-after all, Ove Arup Et Partners are the leading structural consultants-and yet, what does the Press say? Ove Arup were getting advice; they were using outside experts. I think they ddeserve considerable praise.
The paper examines the behaviour of multi-storey structures employing the concept of belt truss system for wind bracing. Optimum locations of belt trusses which minimize wind sway are discussed. A simple method of analysis based on flexibility approach is presented and an example problem is worked out using the method. B.S. Taranath
The Institution's Basic Features of Design Panel at the invitation of the Structural Codes Advisory Committee has examined and reported on the values for and ways in which partial safety factors are being introduced into Codes of Practice for building structures following the adoption of limit state design in the Code of Practice for Structural Concrete, CP110:1972.
Situations are described in which it is advantageous to use composite steel-concrete beams with fewer shear connectors than the number required for full interaction. From a study of the results of tests and computations, simple rules are derived for estimating the ultimate flexural strength of such beams, and for checking deflexions in service, both by calculation and by the use of limiting span-depth ratios. R.P. Johnson and I.M. May
A new ultimate load method of design is presented which can be used for predicting the failure load of conventional plate girders having webs reinforced by both longitudinal and transverse stiffeners, hybrid girders and unsymmetrical girders. This new design method provides identical lower and upper bound solutions, and, furthermore, it is shown that many of the existing solutions are contained as special cases of the present, more general solution. D.M. Porter, K.C. Rockey and H.R. Evans
Queries from several members convince us that there is a prevalent misunderstanding about written comments on papers published in the Journal; many seem to believe that these are not acceptable, but this is not so. The miscomeption probably stems from the decision not to give notice of a closing date for written comment as was at one time done for each paper. The reason for this was that quoting a terminal date inhibited and suppressed worthwhile observations, many of which were not formulated in good time and consequently not submitted. Comment which makes a useful contribution is welcome at any time; obviously the sooner after the publication of the paper concerned the better, since it will be fresher in the reader's mind and probably more readily accessible. Nevertheless, good comment is always useful no matter how far removed in time from its stimulus, and it must often be that grounds for comment do not become apparent until some application reveals a flaw, an inconsistency or, for that matter, a confirmation. Apart from the normal submission of comment to the Editor, there is lways this column in which we will be glad, as we always have been, to air the observations of members, anonymously if they so wish. Verulam