1. CHANGES OVER TIME
Over the years, the changes in materials and reductions in safety factors make it more important to understand the behavior of reinforced concrete and provide more care. Rules of thumb and empirical methods may have been developed for different conditions and may not be applicable for today’s materials’ properties and design criteria need to be checked to determine whether they are still applicable. Another result of the development of and changes to material properties is that the ultimate limit state is often no longer critical, and a design now often depends on the serviceability limit states, apart from punching shear.
|Figure 1 Increase of concrete strength during 20th century.|
Concrete: There has been a continuous increase in the strength of concrete over the last hundred years; much of the increase has developed since 1980 (see Figure 1). Around that time, the value of blended cements and the use of admixtures was realized. Modern concretes have become complex with almost infinite variations available depending on the requirements.
|Figure 2 Change in stress block for high strength concrete.|
The understanding of how to change the properties of concrete and reinforcement is developing rapidly. It includes:
- The use of admixtures and blended cements. Admixtures are essential for modern concrete. Self-compacting concrete is one important example. Blended cements allow the control of the rate of strength gain and the amount of heat created.
- The use of stainless steel will increase for situations where durability is paramount.
- The use of higher strength concrete will become more popular for floor slabs, particularly flat slabs. This will result in thinner and longer span slabs.
- Serviceability limit states have already become critical to flat slab design and it will become more common to check vibration of floors.
- The use of fibres will increase; the use of steel fibres has already been proven for ground floor slabs.
All these developments add complexity and cost to concrete construction. Above a cylinder strength of 50 MPa, the stress–strain properties change with increases in strength. Figure 2 shows this change diagrammatically. The concrete itself becomes more brittle as the strength increases, but it should be noted that in flexural members (beams and slabs), the ductility and brittleness are dependent mostly on the properties of the reinforcement. The increase in concrete strength and reduction in overall factor of safety (see Figure 3) have meant that, for many structural elements, the design for the serviceability limit state is becoming more critical than that for the ultimate limit state.
|Figure 3 Reduction in concrete partial safety factor during 20th century.|
Reinforcement: A similar pattern of change has occurred for reinforcement both in strength and partial safety factors (see Figures 4 and 5).
|Figure 4 Increase in steel yield strength during 20th century.|
Figure 5 Reduction in reinforcement overall safety factor during 20th century.
In 1964, the construction of a 35 m high dust bunker for a coal-fired power station included an external concrete cantilever staircase to be built on t the face of the outside wall of the bunker. The construction of the wall meant that the reinforcement required for the stairway would be cast flush with the wall and then bent out after removal of the formwork. At that time, proprietary reinforcement systems for such a situation did not exist and the bars were bent before fixing within the shutter. The radius of bend would have been to a standard of three times the bar diameter.
After the formwork had been removed, the surface of the concrete was scabbled to expose these bars and the scaffold tubes were threaded over them. The scaffold tubes were then used to lever the bars out of the wall into their final positions. About 30% of all the bars bent out snapped off during the operation. The reason was a combination of factors:
- The bars should have been bent out with a special tool that ensured that the radius of bend was at least three times the bar diameter.
- The particular batch of reinforcement was found to be more brittle (less ductile) than specified.
- The work was carried out at a temperature just above freezing.
Comment: The bending out of reinforcement cast into walls is a common procedure and, all too often is done with scaffold tubes that are readily accessible on site. It is regrettable that a proper re-bending tool is not often used which is a reflection of poor understanding of the material’s physical and chemical properties. In the manufacture of reinforcement, special procedures are in place to check the re-bending of bars to ensure that the reinforcement is sufficiently ductile. It is unfortunate that some manufacturers have continually tried to eliminate such tests from the reinforcement standard.
3. TACK WELDING OF REINFORCEMENT
The design of a building with large columns required 32 mm diameter starter bars projecting from the pile caps. The temporary works for construction included tack welding some small diameter bars to the starter bars. When the time came to fix the column reinforcement to the starter bars, the contractor attempted to bend the starter bars to ensure that they would fit into the column shutter with sufficient cover to the concrete face. A large sledge hammer was used to effect this. During this operation, two of the 32 mm diameter bars snapped off.
The reason was that tack welding the small bars onto the larger diameter starter bars changed the molecular structure of the latter. Unlike structural welding, tack welding heats just the local spot, and the heat sink of the main bar cools it very rapidly. The result was that the starter bars became brittle and required only a sharp blow to fail. In the past, tack welding on site was forbidden. Today it is sometimes permitted if carried out by a skilled specialist. Unfortunately once permitted, it is all too easy for a nonspecialist to do this work, believing that it will do no harm.
Comment: Too many people are unaware that tack welding can have significant structural effects. This is another case where a material’s chemical and physical behavior was not properly understood.
4. HIGH ALUMINA CEMENT
High alumina cement concrete has achieved a certain notoriety following the collapse of several buildings in the 1970s. By the end of 1974, up to 50,000 buildings had been reported as suspect and a major effort was made to check their safety. Fortunately, many of the affected beams stood in dry conditions and the chemical deterioration had not reached an advanced stage. The worst affected elements were positioned in damp environments.
Description: High alumina cement is manufactured from limestone or chalk and bauxite (the ore from which aluminium is obtained). The two materials are crushed and fired together using pulverised coal as a fuel. The materials fuse together, and after cooling are crushed and ground into a dark grey powder.
The predominant compounds are calcium aluminates; calcium silicates account for no more than a few percent. The calcium aluminates react with water and the primary product is calcium aluminate decahydrate (CAH10).
One of its main characteristics is that the concrete made with it achieves its full strength after 24 hours compared with 28 days for a concrete with Portland cement. However, its crystal structure is unstable and changes to tricalcium aluminate hexahydrate (C3AH6) spontaneously (albeit slowly). This process occurs at room temperature and is accelerated by an increase in temperature. The crystal structure transforms itself to a more compact form, with the result that the cement matrix of the concrete becomes porous and weaker. The extent to which this conversion, as it is known, occurs is largely a function of the:
- Original water/cement ratio of the concrete
- Temperature rise in the concrete during hardening
- Temperature and moisture to which the hardened concrete is subsequently exposed
Degree of conversion: It was found that at the time of the collapses most HAC concrete used in buildings was 90% or more converted. A concrete from a wet mix exposed subsequently to the sun was found to have its strength reduced from 40 MPa at 24 hours to an average of about 10 MPa after less than 10 years. In contrast, concrete from prestressed precast beams with a low water-to-cement ratio and hence a 24 hour strength of 65 MPa from the same building but in a dry environment was found to have retained a strength of about 35 MPa.
5. CALCIUM CHLORIDE
Many reinforced concrete structures have suffered from too much chloride in the concrete mix. This causes the breakdown of the high level of alkalinity. When moisture and oxygen are present, carbonation occurs. This allows the reinforcement to rust and leads to spalling of the concrete surface.
Before 1980, calcium chloride was used extensively for in situ concrete works, frequently without adequate supervision. It was used principally for frost protection and to facilitate the rapid stripping of shutters. However, all too often, too much was added. In the 1980s, the codes of practice and concrete specifications were tightened to ensure that the rusting and spalling should not happen again. The following three examples describe where too much chloride in concrete caused structural failures.
Example 1: A primary school (built in 1952) was shut in 1973 due to extensive corrosion of the reinforcement of factory-made precast concrete beams. This was due to the presence of too much calcium chloride added during the manufacture of the beams to hasten the hardening of the cement. The condensation under the beams accelerated the corrosion by combining with the calcium chloride to produce hydrochloric acid.
Example 2: In 1974, the concrete roof of a school collapsed. The reason was found to be too much calcium chloride in the concrete, causing the reinforcement to deteriorate and eventually fail.
Example 3: An independent investigation of the collapse of a 100 m long pedestrian bridge found the cause to be high levels of calcium chloride in the grout used in the ducts for the prestressed tendons. This led to corrosion and failure of the prestressing tendons.
6. ALKALI–SILICA REACTION
The alkali–silica reaction (ASR) is a heterogeneous chemical reaction that takes place in aggregate particles between the alkaline pore solution of cement paste and silica in the aggregate particles. Hydroxyl ions penetrate the surface regions of the aggregate and break the silicon–oxygen bonds. Positive sodium, potassium, and calcium ions in the pore liquid follow the hydroxyl ions so that electro-neutrality is maintained. Water is imbibed into the reaction sites and eventually an alkali–calcium–silica gel is formed.
The cracks are empty (not gel-filled) when formed. Small or large amounts of gel may subsequently exude into the cracks. Small particles may undergo complete reaction without cracking. Formation of the alkali–silica gel does not cause expansion of the aggregate. Observation of gel in concrete is therefore no indication that the aggregate or concrete will crack. ASR is diagnosed primarily by four main features
- Presence of alkali–silica reactive aggregates
- Crack pattern (often appearing as three-pointed star cracks)
- Presence of alkali–silica gel in cracks and/or voids
- Ca(OH)2 depleted paste
In mainly unidirectional reinforced members, the cracks become linear and parallel to the reinforcement. The degree of cracking depends on the amount of confining reinforcement, i.e., links, etc. One major concern was that ASR caused cracking that led bits of concrete to fall off structural elements and hit people below. This led to demolition of the structures in some cases. Examples of ASR effects are given in Figure 6.
7. LIGHTWEIGHT AGGREGATE CONCRETE
During the 1960s, a medium-size civil engineering contractor wanted to join the housing drive, then at its peak. At the time, an Austrian construction firm used crushed brick rubble as aggregate in un-reinforced concrete walls for six- and seven-storey blocks of flats. Inspired by this, it was decided to try to develop a similar form of load-bearing wall with adequate thermal insulation, made of lean-mix plain concrete with light expanded clay aggregate (LECA). A 12-storey block was constructed as a pilot project.
The strength of the wall concrete was reduced in stages—about 2000 psi (14 MPa) at 28 days for the four bottom storeys, 1600 psi (15 MPa) for the next four, and 1200 psi (8 MPa) for the top storeys. The floor slabs were of traditional reinforced concrete, but the roof slab was reinforced LECA concrete with a strength of 3000 psi (21 MPa).
There was no significant adverse feedback from the tenants nor the building authority. The block remained standing and in use for over 40 years. Encouraged by the apparent success, the contractor started promoting the ‘system.’ About the same time, lightweight aggregate concrete was included in the code of practice and a minimum strength of 3000 psi (21 MPa) was
stipulated. This required a richer mix than that used for the walls of the earlier block. The resulting effects of this on the thermal insulation and shrinkage properties of the LECA concrete appear to have been overlooked by the design team.
A few blocks were built for local authorities outside the London County Council area. These were higher than the first block utilising the higher concrete strength required by the code in the walls. Many of the flats were allocated to tenants in poor financial circumstances, who could not afford the charges for the underfloor heating and used paraffin heaters instead.
This, combined with the reduced thermal insulation of the external walls, led to severe condensation. Structurally more important, however, were the diagonal cracks that developed on the top floor of one of the blocks within a short time after hand-over. From their geometry, they appeared to be due to lower shrinkage and greater thermal expansion of the roof slab relative to the wall concrete.
Definitely alarming was the occurrence of horizontal cracks in one of the 200 mm thick internal cross walls connected to the 300 mm thick external wall at right angles. One of these cracks on the 13th floor of a 16-storey block opened suddenly with a noise like a gun shot. The wall was 200mm thick and, according to the design assumptions, carried the floor slabs that spanned about 3.5 m on either side. This meant that the building contained three storeys of unreinforced concrete cross wall, with the load from approximately 3.5 m width of floors plus the roof hanging or cantilevering off the external wall!
Structurally, the only explanation for these cracks seemed to be that the internal wall was drying out, and therefore shrinking and shortening, while the external wall with very little load to carry (at least initially) and exposed to the British weather was not shortening at the same rate.
Discussion: The porous LECA pellets were soaked just before the mixing of the concrete, to prevent them from absorbing water from the fresh mix and thus making it too stiff. They therefore constituted a reservoir of water, over and above that required for the hydration of the cement. This extra water meant that the LECA concrete needed more time to dry out and the 300 mm external walls would have a slower rate of drying out than the 250 mm internal walls even if they had the same environment on both faces.
Comment: These cracks were due to the changed properties of the wall concrete. A proper study of the properties of the materials along with a review of the design would have shown that the two-stage extrapolation from medium rise to high rise and from lean mix, brick rubble, unreinforced concrete to dense, albeit lightweight aggregate, concrete could not be sustained.