Page:The Building News and Engineering Journal, Volume 22, 1872.djvu/404

 382 THE BUILDING NEWS. may ees May 10, 1872. De SS) opportunity of seeing the structure when it is built. We have still to notice the works of one Academician (Mr. Barry), and of Mr. Burges, Mr. Pearson, Mr. Somers Clarke, Mr. Talbert, and many other architects, together with various studies of architecture, some of them too remarkable to be passed over slightly, and we therefore propose to return to the subject next week. (To be continued.) EE LOADS ON GIRDERS.* N designing a girder for moving loads, the strains to which the various parts of it will be sub- jected result from either the simultaneous or inde- pendent action of the dead weight and of the live weight which may at any moment be brought upon it. The former of these, or the fixed load, is pre- sumed to be uniformly distributed over the whole length of the span, while the latter, or moving load, obviously affects in succession each individual unit of it. The fixed or stationary load consists, first, of the weight of the entire girder itself ; and, secondly, of that of the platform or constant weight resting upon the girder. Itis proposed to point out in the present article that, although in cases where the span of the girders is small the assumption of a uniformly distributed dead load may be adopted without sensible error, it is very far from affording accurate results when the span is large. The form and character of the girder will also have some in- fluence upon the question. Let us first consider the weight of the girder itself, and the manner in which the flanges and web are separately affected by it. Suppose the girder to have parallel top and bottom flanges. Theoretically, the strain at the centre of each of these is a maximum, while at the extremi- ties, or immediately over the bearings, it is equal to zero. Consequently the sectional area of the flanges at the centre is likewise a maximum, and but for practical considerations would be also nil at the extremities. As a fact, itis considerable at these points. The width of the flange is usually main- tained constant, and the reduction in the sectional area effected by diminishing the thickness. This practical necessity for con- structing the ends of girders with parallel flanges a good deal heavier than actual theory demands, tends to favour the assumption of the uniformly distributed load, at any rate, so far as their own weight is concerned. The weight of a girder, and consequently that of the flanges, is obtained in the preliminary calculations either from a similar existing example, or by a rough estimate. In both cases the result is approximative only. From this the sectional area required at the centre is calculated, the weight having been regarded as constant for each unit of length in the span. But in reality this is incorrect, since the average weight is not identical with that at the centre, but equal to two-thirds of it. There is an accumulation of weight at the centre of the flanges in large girders which should be taken account of. A careful caleu- lation of the difference in the strains caused by the supposition that the weight of the flanges is con- stant, and the more accurate method which causes the central to exceed the average weight in the pro- portion of three to two, will demonstrate that in the latter case the sectional area at the centre would be increased by a quarter. The reason why, in small girders, the assumption of a uniformly distributed load may be adopted without error is that, in the first place, the total weights are small, and secondly, any disproportion in any one part is compensated for by a similar disproportion in another. This sometimes occurs in large girders, in which the great weight of the flanges at the centreis nearly counterbalanced by the extra amount of material in the web at the ends. Among the many advantages possessed by the bowstring type of girders is that the weight of the flanges is theoretically nearly uniform, and in prac- tice the sectional area is usually constant throughout each flange, When the span and depth of the girder are so proportioned that the parabola may be re- placed by the segment of the circle, the percentage of the increase of strain towards the ends is so small that it may be disregarded when the object is the determination of the weight of the flanges. Instead of the weight of the flanges only, if the total weight of the girder be taken into considera- tion, this excess towards the ends will be found to be compensated for by the corresponding increase in weight in the central part of the web. This is due partly to the greater length of the bars there, na and partly to the fact that the strains upon the web increase slightly towards the centre. The reverse occurs in girders with parallel flanges in which the strains on the web are greatest at the ends, and the bars therefore heavier than at the centre, thus tending to counterbalance the inerease in weight of the flanges at that point. It is rather singular that while the disproportion in the relative sectional areas of the central and extreme parts of a girder with parallel flanges, compared with the similar difference in the bowstring, is very large, the corresponding compensation in their respective webs follows the same law, both with regard to direction and amount. In addition to the extra weight at the ends of the web of a girder with parallel flanges required by the increase in the strains, there is also that due to the pillars and stiffening, which in deep girders is some- thing considerable. Not only is the shearing strain a maximum over the bearings, but there is at those points a tendency to deformation which must be practically guarded against. This was well exem- plified in the experiments conducted on the strength of different forms of tubes in connection with the Britannia Bridge. To preserve the shape of the tubes during the experiments, and enable them to exert their full power of resistance, without being prematurely crippled by deformation, blocks of wood were inserted at the extremities. In the experiments, also, upon the ‘‘ model tube,” the advantages of angle iron pillars was soon rendered manifest. The manner in which the permanent dead weight is ultimately brought upon the girder must determine whether it can be regarded as uniformly distributed over it or not. In other words, the distance between the cross-girders must practically decide the point. Formerly the cross girders were placed closer together than they generally are now, and the opinion of many engineers is that, where cireum- stances permit, they should be placed from 10ft. to 20ft. apart, rail-girders being employed to maintain the communication between them and carry therails, This plan has, undoubtedly, the merit of economy, but the localisation of the strains resulting upon the main girders is open to~ objection. One point is certain. Weights acting upon a girder, per saltum, at intervals of 10ft. or 20ft., cannot be considered uniformly distributed. It isclear that the duty of distributing the weight uniformly over the flanges falls upon the web. In this respect the continuous web has the first place, and the ‘‘ Warren” the last. When the interval between the cross girders does not exceed 4ft. or 5ft., thereis very little doubt but that, with acontinuous web, the permanent dead weight does fulfil the conditions which are erroneously sup- posed to obtain in other cases. On the other hand, there are many existing examples of Warren girders, in which the points of junction of the web with the flanges are placed a great deal too far apart to admit of a uniform distribution of the load. In these girders the assumption that the total load is evenly divided upon the apices of the triangles must be re- ceived with caution. It may, however, be accepted in the case of lattice girders, which have almost entirely superseded the Warren type. It is scarcely necessary toremark that the latter type, as originally designed, has long smce become obsolete. The result of designing a girder with parallel flanges on the in- correct basis we have alluded to, would be that it would be too weak at the centre and too strong at the ends. It sometimes happens that a method of calcula- tion which is correct when applied to a certain class of structures is retained and employed in instances to which it is only partially, or perhaps not at all, adapted. It has been admitted that a girder or beam will practically bear, uniformly distributed over it, twice the weight that it will carry concentrated at its centre. This conclusion, originally deduced from experiments on solid beams of timber, and true both theoretically and empirically, became in course of time received as holding good for flanged girders, without any suspicion that it was only valid for certain parts of them. So far as the flanges are concerned, the rule may be considered sufliciently correct, but it is not so for the web, as a little con- sideration will at once indicate. If a given weight be placedon the centre of a flanged girder, the shearing strain upon each half of the web of the girder is equal to half the weight, and moreover is constant throughout the whole of the web. Thus, in a lattice girder, the strains upon each bar of the web resulting from a weight placed at the centre would be equal, and their vertical component would be equal to half of the said weight, supposing there were only one system or series of triangulations. Now, imagine twice the above weight to be uni- formly distributed over the girder. The vertical reaction will evidently be twice as great as before. But this reaction is equal to the vertical component of the strain upon the last bars of the web, or those nearest the points of support. Manifestly, therefore, the strain upon each of these bars respectively will be twice as great as before. Again, at the centre, the strain upon the web with a uniform load is theoretically nothing, and practically very slight indeed. A girder designed on the datum described would, so far as its web was concerned, be too strong at the centre and too weak at the ends. The former anomaly would, unless in large examples, not be of much importance, as the central bars of the web are, for practical reasons, not reduced below a certain scantling. But it is otherwise with the web at the extremities of the girder. A deficiency of 50 per cent. in strength in a girder at a point where the shearing strain is a maximum is too serious a ques- tion to leave unnoticed. Our object in drawing at- tention to this subject is, that it is a common practice to test flanged girders of moderate dimen- sions by a weight supported at the centre, and when they have undergone the trial in a satisfactory manner their strength and power to resist twice the testing weight distributed uniformly over them is at once taken for granted. Such a method of caleu- lation is obviously so fallacious that, were it not that the webs of small girders are very frequently excessively strong, the truth must have become ap- parent when the girder was subjected to its actual working load. The fact is that, until a certain span is reached, it is not possible to make those scientific adjustments of strain and material which result in true economy. A certain amount of scope is indis- pensable to the advantageous application of theory to practice. When reduced to insignificant examples science becomes degraded to mere pedantry. eee THE LIVERPOOL ARCHITECTURAL 5 SOCIETY. N the 1st inst. the concluding meeting of the present session of this Society was held in the Royal Institution, Liverpool, Mr. Henry H. Vale. F.R.1.B. A., the president, in the chair.—The Secretary stated that the number of members was about the same as at the beginning of the session, 130.—The President said by exercising economy their funds were now in a flourishing state——Mr. T. D. Barry, was elected President for the ensuing ses- sion. Mr. Barry gave a criticism upon the drawings of the students in the competition for the best design for three semi-detached model cottages. 7 The Present then delivered the closing address of the session. He gave a comprehensive sketch of architectural progress in Liverpool during the session, singling out certain new buildings for praise. He congratulated the society upon the good it continued to do for local architecture and as a training school for the students, who had now advantages which were not available at any former period. The stu- dents’ competitions for prizes offered by the society, the President said. had been well responded to, and some highly meritorious designs submitted which would do credit to any architect’s office, and he spoke highly of the good drawing and good taste of the students. The questions of architectural competitions and others affecting the profession were handled by the President in a clear, practical manner ; he depre- cated the practice of making pretty sketches and coloured perspectives, merely to catch the eye of the unwary and mislead clients, and which got honest professors of the art a bad name, and gave rise to the popular saying, “Oh, yes, very pretty on paper,” whereas every true architect made his buildings very much handsomer than they could possibly be made on paper, by the selectionof suitable materials whose colours harmonised, and by first-class workmanship. Buildings should be made far too beautiful to be faithfully portrayed in a drawing, even by the first artists of the day.- He stigmatised the practice of some architects sending in designs for public buildings accompanied by fictitious esti- mates, and, as a check and safeguard to committees, recommended the Scotch method of calling in the professional “ measurer” to estimate the cost of exe- cuting each design before the premiums were adju- dicated. The President spoke in terms of praise of the press as a great help towards the progress of architecture, and instanced the spirit of candour and fair play in which the designs for the Law Courts had been discussed in all the leading journals. He con- sidered that the fitness of architects to enter the lists as competitors for important works should be ascertained from their existing buildings, and not from their reputation either as critics or draughts- men, which were often unreliable tests of their powers as practical architects. The progress of church architecture occupied a considerable portion of the President's address. He gave a lucid sketch of the present phases of thought as evidenced in our ecclesiastical buildings, and said that the Gothic revivalists had gone too far in their enthusiasm
 * From the Engineer.