weeks after the casting of the beams, with a record of the dead load deflection at that date, and readings were taken at frequent intervals for the following 36 weeks. The control beam (unloaded), with a deflection of 0.3 inches at commencement of readings, showed a gradually increasing deflection up to 0.85 inches afte7 1 36 weeks. The beam, with 30 lbs. centre ffiad, showed a deflection increasing from. 0.35 inches, at time of loading, to 1.02 inches in the same period. The other two beams showed similar increases. Extensometers were fitted to two of the beams (those loaded 30 lbs. and 60 lbs.), and readirgs were taken from the sixth to twelfth weeks of test, recording the top shortening and bottom lengthening of the beams.
The continued deflection, exceeding that calculated for in the usual way (taking Young’s modulus for concrete), after four weeks of testing, the author attributes to shrinkage and plastic yield; but further experiments were undertaken to determine, as far as possible, the relative responsibility of these two factors. For this purpose, two additional beams were constructed, in all respects like the others, save that they were symmetrically reinforced top and bottom. One of these carried dead load only, the other having a load of 30 lbs. applied mid-span. Shrinkage in these beams would be resisted equally by the steel top and bottom; but not only did they continue to deflect, but the deflections increased at first at practically the same rate as those of the corresponding beams in the first four. That it is not easy to distinguish between deflection due to shrinkage and that due to plastic yield, is admitted, but the inadequacy of the shrinkage explanation rests on the fact that, if attributable solely to shrinkage, the deflection of all four beams should have increased equally, whereas in the highly-stressed ones it increased more than in those moderately stressed; also, on the fact that the measured shortening of the top flange was greater than the shrinkage of unstressed concrete.
Proceeding from the extensometer readings, the deformation of the tops of all four beams due to yield and shrinkage was determined. The strain due to elastic compression was also calculated (from the known value of Young’s modulus) and the concrete stresses, and all the results were plotted in a graph. From this the author deduces that:—
(1) At a stress of 1,000 lbs. per square inch, the plastic yield in 36 weeks is approximately equal to the total shrinkage, and each is approximately twice the elastic strain.
(2) The plastic yield varies with the elastic strain, being roughly double its value, while the shrinkage remains constant.
(3) The total deformation in 36 weeks is roughly six times the elastic strain at 600 lbs. per square inch, about five times that at 1,000 lbs. per square inch, and about three-and-a-half times that at 1,400 lbs. per square inch.
The author did not expect to find that the plastic yield varied proportionately with the stress, but the tests indicated that this was so, and it is not difficult to believe or account for. The effect of shrinkage and plastic yield in increasing the respective stresses on compression and tensile steel are then calculated in regard to the sixth beam, on the basis of neglecting tension in concrete. From this the total stress on the compression steel was determined after yielding of the concrete at 15,253 lbs. per square inch, or an additional stress of 11,700 lbs. per square inch, which is clearly important.
The effects of plastic yield in various directions were discussed at considerable length. One point was the possibility that if plastic yield occurs in the concrete in compression, it might also occur in concrete in shear and adhesion, which would allow some
gradual slip and relief of stress in the rods. Though it is impossible to speak dogmatically about this, the author, for various reasons, is of opinion that this effect, if it occurs at all, was too slight in the experiments described to be important.
One of the beams was wetted and kept wetted, with the object of throwing further light on shrinkage and plastic yield. The idea was that if shrinkage was due merely to drying-out, wetting should remove it and part of the deflection on account of shrinkage should disappear. Actually, a slight reduction of
deflection was observed at first, but this was followed by a further increase in deflection. This would seem to indicate that shrinkage is attributable more to the irreversible changes associated with hardening than with drying-out, the subsequent increase of deflection resulting from the restarting of the hardening process by the wetting; and this is consistent with the fact that concrete only grows in strength when kept moist.
Removal of the load from one of the beams 30 weeks after casting, resulted in a reduction of deflection corresponding with the elastic deflection due to the load; all deflection due to breaking down of concrete in tension, shrinkage and plastic yield remaining as a permanem set.
The effect of shrinkage and plastic yield on stresses and design was then discussed, and the results, of calculation based on these factors are claimed to be entirely consistent with actual stresses and strains recorded on buildings given earlier in the paper. They are held to prove that rods stressed in compression to 21,500 lbs. per square inch require to be provided with close and efficient cross ties.
The author, in conclusion, recognises that the safety of reinforced concrete rests, not on laboratory tests, but on practical experience on a large scale. The laboratory tests help to explain phenomena encountered in practice and to avoid weak forms of construction. Columns with longitudinal rods and little binding have proved to be weak in practice. The tests, indicating very high stresses in the rods, supply the reason; but the weakness can be avoided by the use of close and ample bindings and hoopings. The question arises as to the effect of this newlyrecognised factor in structural members of large size as compared with the small specimens used in the tests described. Arguing that large members will have more volume for a given surface and will take longer to dry out, it is at least possible that they may remain in a (colloidal and) plastic condition so that the yield may be greater. The possibility _ of the steel stresses ultimately reaching the yield point may have to be considered. It is clear that they will not go beyond this point, sincp any further shortening of the concrete will be less than the shortening necessary to raise the stress in the steel after the yield point. If this is regarded as the final condition which may in time be reached, the stresses can easily be determined.
It remains to consider whether a concrete column with a low stress in concrete and the steel stressed to the yield point would be dangerous. A steel strut stressed to the yield point would be highly dangerous, and would buckle if it were unsupported; it is not clear that it is dangerous when encased in concrete and hooped laterally with efficient and close lateral bindings. Experience in actual buildings, many of which have carried their full load for much more than twelve years, is very strong evidence that the construction is entirely safe. Failure of concrete buildings of reputable design and construction are practically unknown.
That the effects of gradually increasing deflections have not been more noticed in practice, the author
(Continued on p. 798.)
The continued deflection, exceeding that calculated for in the usual way (taking Young’s modulus for concrete), after four weeks of testing, the author attributes to shrinkage and plastic yield; but further experiments were undertaken to determine, as far as possible, the relative responsibility of these two factors. For this purpose, two additional beams were constructed, in all respects like the others, save that they were symmetrically reinforced top and bottom. One of these carried dead load only, the other having a load of 30 lbs. applied mid-span. Shrinkage in these beams would be resisted equally by the steel top and bottom; but not only did they continue to deflect, but the deflections increased at first at practically the same rate as those of the corresponding beams in the first four. That it is not easy to distinguish between deflection due to shrinkage and that due to plastic yield, is admitted, but the inadequacy of the shrinkage explanation rests on the fact that, if attributable solely to shrinkage, the deflection of all four beams should have increased equally, whereas in the highly-stressed ones it increased more than in those moderately stressed; also, on the fact that the measured shortening of the top flange was greater than the shrinkage of unstressed concrete.
Proceeding from the extensometer readings, the deformation of the tops of all four beams due to yield and shrinkage was determined. The strain due to elastic compression was also calculated (from the known value of Young’s modulus) and the concrete stresses, and all the results were plotted in a graph. From this the author deduces that:—
(1) At a stress of 1,000 lbs. per square inch, the plastic yield in 36 weeks is approximately equal to the total shrinkage, and each is approximately twice the elastic strain.
(2) The plastic yield varies with the elastic strain, being roughly double its value, while the shrinkage remains constant.
(3) The total deformation in 36 weeks is roughly six times the elastic strain at 600 lbs. per square inch, about five times that at 1,000 lbs. per square inch, and about three-and-a-half times that at 1,400 lbs. per square inch.
The author did not expect to find that the plastic yield varied proportionately with the stress, but the tests indicated that this was so, and it is not difficult to believe or account for. The effect of shrinkage and plastic yield in increasing the respective stresses on compression and tensile steel are then calculated in regard to the sixth beam, on the basis of neglecting tension in concrete. From this the total stress on the compression steel was determined after yielding of the concrete at 15,253 lbs. per square inch, or an additional stress of 11,700 lbs. per square inch, which is clearly important.
The effects of plastic yield in various directions were discussed at considerable length. One point was the possibility that if plastic yield occurs in the concrete in compression, it might also occur in concrete in shear and adhesion, which would allow some
gradual slip and relief of stress in the rods. Though it is impossible to speak dogmatically about this, the author, for various reasons, is of opinion that this effect, if it occurs at all, was too slight in the experiments described to be important.
One of the beams was wetted and kept wetted, with the object of throwing further light on shrinkage and plastic yield. The idea was that if shrinkage was due merely to drying-out, wetting should remove it and part of the deflection on account of shrinkage should disappear. Actually, a slight reduction of
deflection was observed at first, but this was followed by a further increase in deflection. This would seem to indicate that shrinkage is attributable more to the irreversible changes associated with hardening than with drying-out, the subsequent increase of deflection resulting from the restarting of the hardening process by the wetting; and this is consistent with the fact that concrete only grows in strength when kept moist.
Removal of the load from one of the beams 30 weeks after casting, resulted in a reduction of deflection corresponding with the elastic deflection due to the load; all deflection due to breaking down of concrete in tension, shrinkage and plastic yield remaining as a permanem set.
The effect of shrinkage and plastic yield on stresses and design was then discussed, and the results, of calculation based on these factors are claimed to be entirely consistent with actual stresses and strains recorded on buildings given earlier in the paper. They are held to prove that rods stressed in compression to 21,500 lbs. per square inch require to be provided with close and efficient cross ties.
The author, in conclusion, recognises that the safety of reinforced concrete rests, not on laboratory tests, but on practical experience on a large scale. The laboratory tests help to explain phenomena encountered in practice and to avoid weak forms of construction. Columns with longitudinal rods and little binding have proved to be weak in practice. The tests, indicating very high stresses in the rods, supply the reason; but the weakness can be avoided by the use of close and ample bindings and hoopings. The question arises as to the effect of this newlyrecognised factor in structural members of large size as compared with the small specimens used in the tests described. Arguing that large members will have more volume for a given surface and will take longer to dry out, it is at least possible that they may remain in a (colloidal and) plastic condition so that the yield may be greater. The possibility _ of the steel stresses ultimately reaching the yield point may have to be considered. It is clear that they will not go beyond this point, sincp any further shortening of the concrete will be less than the shortening necessary to raise the stress in the steel after the yield point. If this is regarded as the final condition which may in time be reached, the stresses can easily be determined.
It remains to consider whether a concrete column with a low stress in concrete and the steel stressed to the yield point would be dangerous. A steel strut stressed to the yield point would be highly dangerous, and would buckle if it were unsupported; it is not clear that it is dangerous when encased in concrete and hooped laterally with efficient and close lateral bindings. Experience in actual buildings, many of which have carried their full load for much more than twelve years, is very strong evidence that the construction is entirely safe. Failure of concrete buildings of reputable design and construction are practically unknown.
That the effects of gradually increasing deflections have not been more noticed in practice, the author
(Continued on p. 798.)