From Corrosion of Existing to Durability of New Concrete Structures

Transcription

1 IABSE Symposium - Rio de Janeiro - August 25-27, 1999 From Corrosion of Existing to Durability of New Concrete Structures Eugen BRÜHWILER Professor, Civil Engineer Swiss Fed. Inst. of Techn. Lausanne, Switzerland Eugen Brühwiler received his civil engineering and doctoral degrees from the Swiss Federal Institutes of Technology (ETH) in Zurich and Lausanne. He is currently the professor for the Maintenance, Construction and Safety of Structures Institute at ETH Lausanne and the chairman of the IABSE Publications Committee. Pierre MIVELAZ Dr. Civil Engineer Swiss Fed. Inst. of Techn. Lausanne, Switzerland Pierre Mivelaz received his civil engineering and doctoral degrees from the Swiss Federal Institute of Technology (ETH) in Lausanne. He is a part time research associate at ETH Lausanne in the domain of durability of concrete structures and co-director of ESM Consulting Engineers. Summary The problem of reinforcement corrosion in new concrete structures can be avoided if measures are taken to increase the corrosion initiation time. These measures include providing adequate concrete cover thickness, ensuring a low permeability of the cover concrete, and limiting early age concrete cracking. As adequate concrete cover is normally not a problem, this paper highlights the findings of two studies that focus on the latter two measures. The first study investigated chloride ingress under given climatic conditions and evaluated the feasibility of measuring the permeability of concrete cover in situ. The second study used numerical models to investigate the effects of early age cracking and determine measures to be taken to limit their development during construction. The principal findings of both studies are given herein. It is concluded that permeability measurements should be conducted for quality control of the cover concrete of new structures and that early age cracking due to the hydration of young concrete may be limited by reducing the difference between concrete and ambient temperature..h\zrugvrebar corrosion, initiation phase, permeability, cover concrete, early age cracking, numerical simulations, concrete structures, durability. 1. Introduction Reinforced concrete structures have generally shown satisfactory performance in terms of strength and durability. Under certain climatic conditions, however, reinforced concrete structures have suffered from corrosion damage of steel reinforcement (rebar) resulting in premature and expensive repair work. Damage to concrete structures is the result of the interaction between the material and the environment. Gas, liquid and ionic transport mechanisms play a major role in the concrete deterioration. The nature of the damage can be chemical, such as rebar corrosion or sulfate and acid attacks on cement paste, and physical, such as abrasion or freeze-thaw cycles (Fig.1). Both the quality of cover concrete and the presence of cracks (especially those which allow water seepage) have a predominant influence on the durability of a concrete structure. Concrete not only has a load bearing function, but also acts as a protective coating for the steel reinforcement. 208

2 Structures for the Future - The Search for Quality Fig. 1 Vulnerability of reinforced concrete The majority of research on the deterioration phenomena affecting reinforced concrete has concentrated on material behaviour. The ultimate limit states for structural safety and serviceability, however, are based on structural considerations. These aspects must be combined. With this combination it is important to analyse the deterioration mechanisms considering both the observable steady state and a time evolutionary point of view. This paper investigates measures to reduce the risk of occurrence of rebar corrosion that may be used in the construction of durable and low maintenance concrete structures for the future. After a review of the relevant aspects of rebar corrosion, this paper investigates the roles of the permeability of cover concrete and early age crack formation in delaying the onset of corrosion. 2. Corrosion of steel in concrete During cement hydration, a highly alkaline pore solution (ph ) is formed in the concrete. In this alkaline environment, ordinary reinforcing steel forms a very thin oxide film (the passive film) that protects the steel from corrosion. This passive film remains stable as long as the pore water composition remains constant. The protective film is destroyed when sufficient chloride ions (from deicing salts or from seawater) have penetrated to the reinforcement or when the ph of the pore solution drops to values below 9 due to carbonation (a chemical reaction between the pore solution and carbon dioxide). The reinforcing steel is depassivated and thus vulnerable to corrosion attack. Once the reinforcing steel is depassivated and simultaneously in contact with both oxygen and water (humidity) metal dissolution (corrosion in the form of rust formation, loss in cross section) occurs (Fig. 2). oxygen electrolyte (humidity) corrosion damage carbonation rebar corrosion depassivation of steel chlorides limit state propagation corrosion rate time Fig. 2 Conditions for steel reinforcement corrosion Fig. 3 Development of corrosion of steel reinforcement in concrete with time 209

3 IABSE Symposium - Rio de Janeiro - August 25-27, 1999 The deterioration of concrete structures due to rebar corrosion can be divided into two phases (Fig. 3): (1) During the initiation phase, carbonation occurs and/or chlorides penetrate from the surface of the concrete to the reinforcement. The duration of the initiation phase is from the construction until depassivation of the reinforcing steel has occurred. (2) During the propagation phase, the reinforcement actively corrodes. The duration of the propagation phase is dependent on the rate of corrosion. The end of the propagation phase is determined when a specified amount of section loss has occurred. A durable concrete structure obviously has both a long initiation phase and a slow corrosion rate. (Usually the conditions that provide a long initiation phase also provide a slow corrosion rate.) The ideal situation in the design of new structures is to have an initiation phase that is longer than the design service life of the structure. This paper deals specifically with the measures affecting the initiation phase. Chloride penetration, carbonation and corrosion are highly complex and seemingly random phenomena. It is however well known that they depend on parameters such as the concrete cover thickness, the porosity and permeability of surface and cover concrete, the concentration of chlorides and carbon dioxide, and the microclimatic conditions (wetting, drying) at the concrete surface of a given structural element. In order to reduce the risk of rebar corrosion of new structures it is necessary to alter these parameters to achieve a long initiation phase. Three specific measures that should be taken are: (1) provide sufficient concrete cover thickness, (2) provide a dense surface and cover concrete of low permeability, and (3) limit or avoid early age concrete cracking. In today s design codes only the concrete cover thickness is defined. Sufficient concrete cover alone, however, does not guarantee durable reinforced concrete. The issues of the permeability of the cover concrete and the formation of early age cracks must be given more attention. The highly random nature of the phenomena involved in rebar corrosion means that in the design phase of reinforced concrete structures - it makes little sense to calculate the duration of initiation and propagation phase using numerical models. Durability of reinforced concrete structures should not be calculated! Numerical models are, however, valuable tools to investigate for given structural elements subjected to certain climatic conditions - time evolutionary properties of predominant parameters or the effectiveness of preventive measures. Probabilistic concepts can be introduced to account for the random nature of phenomena [1]. 3. Permeability of cover concrete 3.1 Transport processes Chloride ions penetrate by different transport mechanisms into concrete: capillary suction into dry or partially dry concrete (movement of chloride ions with the water), diffusion of ions into water saturated concrete (e.g. in submerged marine structures), or ion migration (under electric field). In most concrete structures the capillary suction is the most important and most rapid process. The rate of carbonation depends on the CO 2 ingress. Carbon dioxide can penetrate into concrete only in the gas phase, thus in (partially) dry concrete. The chemical reaction of CO 2 with the alkalis, in particular with Ca(OH) 2, requires the presence of water. (Thus the carbonation rate in very dry concrete is low.) In most cases, at least two transport mechanisms are superimposed. The transport processes depend on the exposure and permeability of concrete. This is demonstrated in the following example. 210

4 Structures for the Future - The Search for Quality 3.2 Permeability and chloride ingress In the simulations, a concrete face was exposed to chlorides for a 90 day period in three exposure zones corresponding to the different climatic conditions experienced by bridge components [1]. The mist zone was used for components such as the underside of a bridge cantilever slab. The direct zone was used for components such as piers immersed in seawater. The spray zone was used to simulate the curbs of bridge slabs, or the splash zones of bridge piers. The relative humidity in the concrete was varied with respect to exposure zone(figure 4). The water content of the concrete was assumed to be in equilibrium in an 80 percent relative humidity environment. The depth of chloride penetration into the concrete was calculated numerically using a finite element model as a function of exposure zone (Figure 4). 2,0 1,5 1,0 0,5 MIST: Constant 80 % relative humidity Concrete Surface Rebar 10 years 30 years 100 years DIRECT: Constant 100 % relative humidity 2,0 1,5 1,0 0, 'HSWKÃ>PP@ SPRAY: Alternated 80 % and 100 % relative humidity 2, 'HSWKÃ>PP@ Probability of rebar depassivation vs. total chloride content 1,5 1,0 0,5 Probability of rebar depassivation high low 0.5 % Total Chloride 1.0 % Content (%) 'HSWKÃ>PP@ Fig. 4 Influence of the microclimate on chloride penetration and probability of rebar depassivation with respect to chloride content. All components were assumed to have a design service life of 100 years, concrete cover thickness of 30 mm and average quality concrete (kt = mm 2 the in situ permeability coefficient, see section 3.3). A general assumption of probability of rebar depassivation was assumed with correlation to total chloride content in the concrete (Figure 4). In the mist zone, total chloride content did not significantly exceed 0.5 % at a depth of 30 mm until nearly 100 years of exposure: a low probability of depassivation. In the direct zone, the total chloride content at a depth of 30 mm was 0.5 % after 30 years of exposure and 1.0 % after approximately 100 years: a moderate probability of depassivation. In the spray zone, the total chloride content was 0.5 % after only 10 years of exposure: a high probability of depassivation. The analysis showed that the severest deterioration occurred in the spray zone, where there was alternate wetting and drying of the concrete. Variable humidity is the most unfavourable condition 211

5 IABSE Symposium - Rio de Janeiro - August 25-27, 1999 because it increases capillary suction. Under such severe exposure concrete should be protected by a coating. Numerical simulations also show that by decreasing the in situ permeability coefficient by a factor of 100, representing low quality cover concrete, the initiation phase is reduced by a factor of approximately 3.4. Low permeability coefficient values (good quality concrete) can be ensured by dense surface and cover concrete. These are obtained by using suitable formwork surfaces (with a certain capacity of water absorption) and with appropriate compaction. 3.3 Quality control by permeability measurements In order to verify the quality of a new concrete structure in terms of the permeability of surface and cover concrete, in situ permeability measurements can be conducted. Figure 5 shows a commercial permeability measuring device which was developed for the quality control of early age concrete [2]. The device has two chambers, an internal chamber (diameter 32 mm) and an external chamber (diameter 86 mm). It measures air permeability by creating a near vacuum on the concrete surface and measuring the increase in pressure in the internal chamber over a period of time. From the in situ air permeability coefficient, kt, corrected with respect to electrical resistivity measurements, - the water permeability of surface and cover concrete can be determined. 3XPS 3UHVVXUHÃ5HFRUGHU 3L 3UHVVRVWDW 3R 3R &KDPEHUV &RQFUHWH Fig.5 Schematic of the air flow permeability measuring device [2] This device may also be used to evaluate existing bridges [3]. Figure 6 shows results as obtained on one bridge at two different days. Measurements at six of the nine separate locations resulted in the determination of the same concrete class (based on the kt measurements), from average to good concrete and one of the nine locations showed decreased concrete quality from very good to good. The readings at the highly variable location, indicated as Location 8 in Figure 6, showed variation between very bad and good concrete (4 levels). This variation was due to variable humidity. Location 8 was the only location where humidity was detectable in the concrete. The readings were taken in slightly different locations. In general, the permeability measurements were repeatable when taken at exactly the same location and when the concrete had the same internal water content. Air and water permeability of concrete depends on the water content in the pores. Further work in this area should investigate adjusting the in situ measurements with respect to humidity. Measurements are not representative in the case where cracks were present at the surface. 212

6 Structures for the Future - The Search for Quality ÃP V H OX D Ã9 7 N , /12/97 20/03/98 5 very bad 4 bad 3 avg. 2 good &RQFUHWH&ODVV 0,01 1 very good Fig.6 0, /RFDWLRQÃRIÃ0HDVXUHPHQWV In situ permeability coefficients (kt) obtained at different locations on a bridge (dry environment) Cracking of early age concrete 4.1 Cracking due to cement hydration Crack formation due to cement hydration can lead to accelerated water penetration into concrete. The most severe concrete deterioration is observed when water regularly seeps through cracks. Early age cracking occurs when there is a large difference between the maximum temperature of the hydrating concrete and the ambient temperature. Normally measures are taken to limit this difference, such as reducing the cement content, adopting an appropriate curing process or reducing the temperature of the fresh concrete temperature by using cooling pipes or liquid nitrogen. The effectiveness of these measures can be evaluated using numerical models and experimentally determined concrete properties. To study the influence of cracks on transport mechanisms in concrete, a distinction between microcracks and macrocracks is made. - Microcracks are observed on the materials level, where hardened cement paste (hcp), aggregates and interface zones are distinguished [4]. They are shorter than the maximum aggregate size and usually evenly distributed throughout the cement paste. Microcracks are predominately caused by differential deformations between cement matrix and granulates due to endogenous and drying shrinkage of hcp, and differential dilatation between hcp and aggregates. Other causes of microcracks are chemical swelling, shrinkage of hcp and loads. The influence of microcracks on transport mechanisms is generally included in global parameters such as permeability. - Macrocracks are observed on a structural level. They are cracks visible to the naked eye at the concrete surface and represent linear discontinuities in concrete elements. Macrocracks result from applied loads on the structure, endogenous and drying shrinkage as well as thermal dilatation. The influence of macrocracks on transport mechanisms is investigated using transport models in discrete cracks [5]. 213

7 IABSE Symposium - Rio de Janeiro - August 25-27, Numerical Simulation This example deals with the superstructure of a 720m long continuous box girder. The bridge was constructed using the launching method [6]. The sectional geometry, in particular the box height and slab width, is variable. The complex geometry led the contractor to cast the box girder in two stages; the bottom slab and the webs first; and the top slab one week later. This casting sequence produces significant tensile stresses in the top slab, especially in cold weather, due to temperature differences between the young and the hardened concrete (Fig. 7). 2.5 MPa 2.5 MPa Tensile stress distribution - pouring and curing with temperatures maintained at > 5 C Tensile stress distribution considering the suggested precautions Fig.7 Analysis of early age tensile stress in a deck slab after two weeks, taken from [7] To limit the risk of traverse crack formation in the top slab before and during the launching of the deck, special precautions were taken to ensure appropriate curing. The primary difficulty was to ensure that the launching and prestressing of the bridge could occur four days after casting throughout the entire year regardless of climatic conditions. In addition to this the casting area was to be accessible from above (i.e. no overall encasing) throughout the duration of the project. Stage Stage 3ODQH YLHZ RI SRXULQJ DQG KHDWLQJ]RQHV Stage Launching direction External heating 6(&7,21Ã%% Light Internal heating roof Water pipes heating to maintained formwork temperature = 15 ± 5 C Insulation Fresh concrete 1 week old concrete ÃÃ6(&7,21Ã$$ Figure 8. Precautions to ensure adequate curing [6]. The precautions followed to ensure adequate curing were the result of thermal and mechanical analyses of early age deck behaviour using a finite element model. The analyses showed that local enclosure and formwork heating made it possible to maintain the optimum temperature difference between the freshly poured top slab and the already hardened bottom slab and webs [7]. The 214

8 Structures for the Future - The Search for Quality specific heating methods used are shown in Figure 8. The optimum temperature differences were determined by limiting the concrete tensile stresses to 1.5 MPa (without considering the stresses caused by prestressing). In situ thermal measurements were conducted with each concrete pour to confirm the maturity evolution of the young concrete, especially during cold weather. These precautions made it possible to follow a regular pouring schedule throughout the entire year regardless of climatic conditions. Early age cracks (traverse) in the top slab were avoided. 4. Conclusions 1. The objective of measures to reduce the risk of rebar corrosion should be to increase the corrosion initiation time until it is greater than the service life of the structure. Three methods to reduce this risk are (1) to provide sufficient concrete cover thickness, (2) to provide dense surface and cover concrete of low permeability, and (3) to limit early age cracking of concrete. These measures depend on the climatic exposure of the structural element. 2. Numerical models allow approximate prediction of the initiation phase for rebar corrosion and early-age cracking of concrete elements. They are reliable tools to evaluate time evolutionary issues of concrete structures and therefore investigate the effectiveness of measures to ensure durable new concrete structures. 3. Permeability measurements should be conducted for quality control of new concrete structures. They are reliable with adequate adjustments for moisture. 4. Cracking due to hydration of young concrete increases concrete permeability and thus its vulnerability to rebar corrosion. Early age cracking may be limited or even avoided by reducing the difference between the temperature of new concrete and ambient temperatures. 5. References [1] Roelfstra, G., B. Adey, R. Hajdin and E. Brühwiler, Condition evolution of concrete bridges based on a segmental approach, non-destructive testing and deterioration models, Proceedings, 8 th International Bridge Management Conference, Transportation Research Board U.S. Federal Highway Administration, Denver, April 26-28, [2] Torrent, R.J., A two-chamber vacuum cell for measuring the coefficient of permeability to air of the concrete cover on site, Materials and Structures, Vol. 25, No. 150, July 1992, pp [3] Adey B., G. Roelfstra, R, Hajdin, E. Brühwiler, "Permeability of existing concrete bridges",2 nd Int. PhD Symposium in Civil Engineering, Budapest, 1998, pp [4] Wittmann, F.H., "Structure of concrete with respect to crack formation", Developments in civil engineering n 7, Fracture mechanics of concrete, Elservier, 1983, pp [5] Mivelaz P, "Etanchéité des structures en béton armé - Fuites au travers d'un élément fissuré", Thèse EPFL n 1539, Lausanne, 1996, 213 pp. [6] Badoux, M., Favre, R. and Laurencet, P. Design of a curved incrementally launched bridge, Structural Engineering International, Vol. 9, no 2, IABSE Zurich, and Favre R., P. Laurencet, "Poussage cadencé d'un grand pont à géométrie variable - Viaduc Ile Falcon", Publication EPFL-IBAP n 146, 1998, 36 pp. [7] Mivelaz P., Charif H., "Viaduc Ile Falcon - Bétonnage du caisson", Rapport d expertise, ESM - Charif & Mivelaz S.A., St-Sulpice, 1997, 46 pp. 215