bottom surfaces of the model under the action of a point load applied, in turn, to each junction of the longitu- dinal and transverse stiffeners.
The results to be obtained from this model are the influence surfaces for deflection and moments at the various points. The torsional properties of the model will be measured by apply- ing a known torque and measuring the resulting displacements.
The occurrence, or otherwise, of uplift at the supports will be inferred from the displacements of the sup- ports when the model is loaded with a scaled distribution of load. This will be calculated from the experi- mentally determined influence sur- faces.
It was considered necessary to check the possibility of the prestressing force causing buckling phenomena in the central suspended span. Theore- tically buckling cannot occur provided the line of action of the force remains fixed with respect to the centroid of the section: which is the case in most prestressing applications including this one. The model of the suspended span has been provided with prestress- ing wires in the appropriate locations. After all the elastic tests have been completed the central span will be prestressed to several times the scaled prestressing force to demonstrate that buckling effects do not occur.
Concrete model
The concrete model of the central suspended span was built to a scale of 1/6th full size, geometrically similar to the prototype except that super- elevation and the vertical curvature The dimensions of the were ignored. model are:-
Length along centre-line 13 ft.-0 in. Width in radial direction 4 ft.-8 in. Depth at mid-span
Depth near supports
Thickness of flanges
Thickness of stiffeners Radius of curvature along
centre-line
61⁄2 in. 911⁄2 in. 11⁄2 in. 2 in.
26 ft.-6 in.
T
Prestressing force in the model was provided by 17 Nos. 0.276 in. high tensile steel wires having a steel area equivalent to the 17 Nos. 7 0.7 in. strand cables in the prototype. Shear and torsional rein- forcement was scaled down on the basis of steel percentage per unit volume of concrete, and comprised a wide mesh of 8 s.w.g. high tensile steel wire. in the stif- feners.
The concrete mix used was designed to give the same strength and modulus of elastici- ty as those assumed in the design. In order to increase the self weight of the model, the rea- son being explained later, magnetite_aggre- gates from the iron-ore mine in Ma On Shan were employed.
employed. These aggregates contain 28% 30% of iron and weigh 120 lb. cu.
ft. Fig. 4. when loosely packed.
Formwork and reinforcement of concrete model
The aggregates as delivered were not suitably graded for concrete. Coarse aggregate, 38 in, nominal size. was screened to remove the under- sized and the over-sized particles. The fine aggregate. which passed through the 3/16 in, sieve and lay in zone 2 of B.S. 882, was washed to reduce the silt content.
The concrete mix adopted was I 1.8 2.7 by weight with a water/ - cement ratio of 0.45. Rapid-hardening Portland cement was used in order to yield a high strength at an early age. Sealocrete double-strength premix was added as a plasticizer.
Batching and mixing were under strict control and the workability was constantly checked during concreting. The average slump recorded was 21⁄2 in.. which was the minimum to en-
sure full compaction of the concrete with mechanical vibration in such a complicated form with congested rein- forcement. To further increase the self weight of the model, the voids between the stiffeners were filled dur- ing concreting with loosely packed "iron-concentrate" which contains 56% iron.
The in-filling did not contribute any rigidity to the model whereas it increased the overall unit weight of the model. Prestressing was perform- ed at seven days by jacking at both ends with a pair of CCL MK1 hand- operated hydraulic jacks.
According to the principles of similitude the unit weight of the mo- del should be six times that of the prototype in order to induce the same dead load stresses in the model. By using heavy aggregates in the con- crete and by filling the voids in the model with heavy material a density factor of 112 was achieved which was still far from the required value of 6. In order to compensate for the low dead weight, the prestressing force ap- plied to the model was reduced so that the maximum compressive stress at mid-span at transfer was the same as calculated in the design of the pro- totype. This ensured that the stress condition at mid-span was faithfully reproduced in the model. The initial prestress in each wire was 8500 lb. and the total prestressing force for 17 wires was approximately 65 tons. Figure 4 shows details of the rein- forcement before concreting and figure 5 shows the completed model set up in the testing frame. The con- crete model will be tested according to the following schedule:-
(1) Individual point load tests:
A point load supplied by a hydrau-
Fig. 5. Concrete model in its testing frame
Far East BUILDER, May 1968.
49