DURABILITY INDEXES AND THEIR USE IN CONCRETE ENGINEERING

Transcription

1 DURABILITY INDEXES AND THEIR USE IN CONCRETE ENGINEERING Mark G Alexander Department of Civil Engineering, University of Cape Town, South Africa Abstract Durability of reinforced concrete structures is a pervasive and serious problem worldwide. To address this problem, an engineering approach based on sound scientific principles has been developed, using so-called Durability Indexes. These indexes relate to transport processes of permeation, water absorption, and chloride ion conduction. They are measured in conceptually simple and relatively easy tests, which do not require expensive equipment, thus rendering them suitable for construction use. They are based upon the philosophy that durability will only be substantially improved when it is possible to measure related parameters in-situ that govern durability performance. This approach has been developed to the point where it can be used for material characterisation, for drafting durability specifications on a performance basis, as a means of site quality control, and as a basis for long-term predictions. The paper reviews the development of the approach, and gives examples of the use of durability indexes in practice. The status in South African concrete practice is reviewed, and future developments are outlined, including the need for internationalisation of this and similar approaches. 1. Introduction For the past decade or so, a concerted effort has been underway in South Africa to attempt to improve the quality of reinforced concrete construction. This was born out of a realisation that the quality of new construction was poorer than desired, and from increasing occurrences of premature deterioration of structures. In addition, older infrastructure began to show signs of severe deterioration, requiring costly repair and rehabilitation. It was recognised from the outset that the majority of problems were overwhelmingly related to corrosion of reinforcing steel, although other issues such as ASR were still relevant. Thus, a programme of universitybased research was put in hand in the early 1990s, which had the following guiding principles: an engineering approach was to be taken, that is practical solutions that industry could implement were to be found; real materials, i.e. different concretes, were to be investigated; a comprehensive approach was to be followed, involving characterisation of materials, appropriate test methods, and models for predicting durability performance 1,2. At the same time, industry, mainly the cement and concrete producers, supported the programme by research grants and by being involved in an advisory capacity. The fruits of the research were regularly shared with practitioners and industry by means of seminars and monographs 3. This led in time to industry tentatively adopting the results of the research in the form of new test methods for charactering the potential durability of concrete, and to the drafting of

2 performance-based specifications for construction based on the new test methods. Simultaneously, both researchers and practitioners undertook work into appropriate repair and rehabilitation methods and strategies for aging structures. Consequently, an indigenous and somewhat innovative approach to addressing problems of the quality of reinforced concrete construction has been developed, which has made a noticeable impact on local practice. This paper outlines aspects of these developments, and in particular covers the so-called Durability Index Approach to achieving reinforced concrete durability. 1.1 Concrete Durability Durability may be defined as the ability of a material or structure to withstand the service conditions for which it is designed over a prolonged period without significant deterioration. Concrete has generally been regarded as being inherently durable, and is expected to be relatively maintenance-free during its service life. We now know that this is not always the case, with many modern concrete structures needing substantial repairs and maintenance during their service life with the resultant costs to the economy reaching 3 5% of GNP in some countries. Neville 4 suggests that reasons for the widespread lack of durability include poor understanding of deterioration processes by designers, inadequate acceptance criteria of concrete on site, and changes in cement properties and construction practices. Concrete is a complex composite material and environmental and service conditions vary widely, so that deterioration mechanisms interact dynamically with material and structural influences. Deterioration is often associated with ingress of aggressive agents from the exterior such that the near-surface concrete quality largely controls durability. The interaction between the various material and environmental elements influencing durability is shown in Figure 1. Figure 1: Schematic diagram of concrete protection of reinforcement

3 Durability problems frequently concern the corrosion of reinforcing steel rather than deterioration of the concrete fabric itself. The problem for engineers is how to provide adequate protection to the steel by the concrete cover layer, which is subjected to the action of aggressive agents such as chloride ions or acidification from the surrounding environment. What is therefore required is the ability to quantify the quality of the cover layer in terms of engineering parameters that can immediately be useful to designers and concrete practitioners. Modern design and construction practices, such as faster concrete casting and hardening times, have made concrete more sensitive to abuse, which in turn has contributed to premature deterioration of modern concrete structures. One response to this has been more stringent construction specifications, particularly in regard to durability provisions. However, this has not always led to a corresponding improvement in performance. This is due to a lack of understanding of what is required to ensure durability as well as inadequate means of enforcing or guaranteeing compliance with specifications during construction. Most codes and specifications are of the prescriptive or recipe type, setting limits on w/c ratios, cement contents, cover, etc., but without really addressing the issue of achieving adequate quality of the concrete cover. It is also difficult to ensure compliance with these specifications since they comprise difficult-to-measure aspects of construction. The one notable exception is, of course, checking concrete cover to steel. Enforcing this one simple expedient would probably cure the majority of current durability problems as they relate to steel corrosion! 1.2 Philosophy of the Durability Index Approach In response to the situation described above, research in South Africa has focussed on the development of a durability index approach that seeks to characterize the potential durability of new concrete. (In the context of this paper, the potential durability of concrete can be defined as the degree of resistance of the cover concrete to the conduction of chlorides, permeation of oxygen, and absorption of water, as indexed by the tests described in the paper.) The approach has been laid out in a number of publications 5,6,7, and was developed in response to the need for practical durability tests that could be site-applicable. The philosophy of durability index testing is given below. 1. Improved durability will not be achieved unless relevant durability parameters can be unambiguously measured, preferably in-situ. A means is required of characterising the quality of the cover layer, using parameters that reflect the deterioration processes acting on the concrete. The use of strength parameters is not sufficient, since these merely measure the overall bulk response of the material to stress. 2. The surface layer is most affected by curing and by external deterioration processes, which are governed by transport mechanisms such as gaseous and ionic diffusion, water absorption, etc. Thus, a series of index tests is needed to cover the broad range of durability problems, each index test being linked to a transport mechanism relevant to that particular process. 3. The material requires to be indexed in terms of its ability to resist the ingress of aggressive agents. This requires quantifiable physical or engineering parameters to characterise the concrete at early ages and provide reproducible measures of microstructure and durability properties. These index values can be matrixed with design and construction-related factors, such as mix materials and proportions, nature and extent of curing, thickness of the cover layer, and so on. 4. The usefulness of index tests must be assessed by reference to actual durability performance of structures built using the indexes for quality control purposes. This is a long-term undertaking. A framework for durability studies is therefore necessary,

4 incorporating early-age material indexing, direct durability testing, and observations of long-term durability performance 2. Correlations are required between indexes, durability test results, and actual structural performance, such that the index tests can be used as follows: As a means of controlling a particular property of concrete, usually the quality of the surface layer. This control would be reflected by a construction specification in which limits to index values at, say, 28 days would be specified As a means of assessing the quality of construction for compliance with a set of criteria As a basis for fair payment for the achievement of concrete quality As a means of predicting the performance of concrete in the design environment Index properties fulfil the requirements of a measurable property that can be specified. The criteria for index tests require that the tests: Be site- or laboratory-applicable. Site-applicable could involve retrieval of small core specimens from the structure for laboratory testing Be linked to important fluid and ionic transport mechanisms and have a theoretical basis Be quickly and easily performed without unreasonable demands on operator skill Have sufficiently low statistical variability Involve a minimum of specimen preparation, with uniform preconditioning to ensure standardized testing Be conducted at a relatively early-age (typically 28 days) The suite of three durability index tests developed in the research will be discussed in the next section. 2. The Durability Index Tests Three durability index tests have been developed 5,8,9,10, namely the oxygen permeability test, the water sorptivity test, and the chloride conductivity test. Each test measures a different transport property of fluids or ions through the concrete cover layer, typically covering the main mechanisms related to deterioration. The tests have been developed and proved in the laboratory, and increasingly are being applied on site in actual construction 6,7. They have progressed to the point of being in regular use, and specifications are being written around their site application. At the same time, the performance of structures built using the index approach is being monitored as far as possible to validate the approach and implement improvements. 2.1 Oxygen Permeability Test This involves a falling head permeameter in which oven-dried (50º C for 7 days) concrete samples, generally 68 mm diameter and 25 to 30 mm thick, are placed in rubber collars secured on top of a permeability cell 8. The cell is pressurised with oxygen to 100 kpa before being isolated, after which the pressure decay is monitored, from which the Darcy coefficient of permeability, k, may be determined. The oxygen permeability index (OPI) is defined as Oxygen permeability index = -Log (k) (1) Oxygen permeability indexes are logarithmic values and range generally from 8 to 11, i.e. three orders of magnitude; the higher the index, the less permeable the concrete. A diagram of the test apparatus is shown in Figure 2.

5 Figure 2: Schematic diagram of oxygen permeability apparatus Laboratory work on the OPI test has shown that the OPI increased (i.e. the quality improved) with increasing grade of concrete and extent of moist curing. Fly ash and slag concretes were less permeable than plain portland concretes when well cured, but more permeable when drycured. The test is sensitive to compaction, bleeding, and extent of moist curing, with high water:binder (w:b) ratio concretes being more adversely affected by poor curing than low w:b concretes 11,12. For example, the OPI of Grade 35 OPC concrete increased from 8.50 to when duration of wet curing increased from 1 to 28 days, while that of Grade 55 OPC concrete only increased from 9.43 to A particular index value can be obtained either by extending the duration of curing of lower strength concrete or by decreasing the w:b ratio in the event that curing is likely to be minimal or ineffective. This is illustrated in Figure 3. Figure 3: Iso-Permeability charts for OPI test OPC and Fly Ash concretes

6 2.2 Water Sorptivity Test Sorptivity is defined as the rate of movement of a wetting front through a porous material. The water sorptivity test involves the uni-directional absorption of water into one face of a preconditioned concrete disc sample 5,12. At predetermined time intervals, the sample is weighed to determine the mass of water absorbed, and the sorptivity is determined from the plot of mass of water absorbed versus square root of time. The lower the water sorptivity index, the better is the potential durability of the concrete. Sorptivity values typically vary from approximately 5 mm/ h, for well cured Grade concretes, to mm/ h for poorly cured Grade 20 concrete. A diagram of the test is shown in Figure 4. Figure 4: Schematic diagram of water sorptivity test Sorptivities measured on lab concretes showed that absorption rates reduced with increasing grade of concrete and duration of moist curing see Figure 5. Dry-cured concrete had significantly higher sorptivity values than wet or moist-cured concrete. The sorptivity test measures a near-surface property and should therefore be sensitive to early-age drying effects that influence the microstructural porosity gradients in the concrete. Differences in sorptivity values for wet and dry cured concrete are typically between 25 % and 70 %, indicating that the test method may be appropriate for assessing curing effectiveness on site 11,13. With 28 days of moist curing, the sorptivity of surface concrete becomes almost insensitive to changes in the normal range of water: binder ratios. Figure 5: Typical water sorptivity results for OPC concretes

7 2.3 Chloride Conductivity Test Streicher developed a rapid chloride conductivity test in which virtually all ionic flux occurs by conduction due to a 10 V potential difference between the two faces of a sample 9,10. The apparatus consists of a two-cell conduction rig, each cell containing a 5M NaCl solution so that there is no concentration gradient across the sample and chloride migration is the result of conduction from the applied potential difference see Figure 6. The concrete disc sample is pre-conditioned by vacuum saturation with a 5M NaCl solution. Figure 7 gives typical results for a range of concretes made with different binders. Figure 6: Schematic diagram of chloride conductivity apparatus Figure 7: Typical chloride conductivity results for different concretes

8 Diffusion and conduction are related by Einstein s equation, allowing the conductivity test to be used as an index of concrete diffusivity. The test is sensitive to changes in the pore structure and cement chemistry (mainly binder type), which might appear to be insignificant when using the permeation process 11. Typical chloride conductivity index values range from > 3 ms/cm for Grade OPC concretes, to < 0.75 ms/cm for Grades slag or fly ash concretes. The lower the index, the better is the potential durability of the concrete. 3. Use And Applications Of The Durability Index Approach 3.1 Material Characterisation For the durability index tests to be useful to industry, they must be able to characterise the potential performance of concrete. Some examples of this have been given above. The tests have been shown to characterise the important variables governing durability performance, that is, mix materials and proportions, compaction and bleeding effects, and type and extent of curing. Figure 8 illustrates the influence on OPI of bleeding measured in a 3 meter high Grade 30 concrete column 14, while Figure 9 shows the influence of early site curing practices on conductivity 7. OPI (Log) Low bleed Medium Bleed High Bleed Top Middle Bottom Position of core in column Figure 8: Characterisation of effects of bleeding in 3 m high concrete column, in terms of OPI A valid concern for any new test method is its repeatability and reproducibility. Typical results for variability of durability indexes measured on samples from a range of concretes used in actual construction including ready-mix concrete, and in the laboratory, are shown in Table 1. In the study quoted 6, the highest variability occurs for indexes of actual structures, as expected. The coefficients of variation for chloride conductivity and OPI of the actual structures were approximately double those of wet-cured site concrete, ascribed to variations in curing effectiveness and degree of compaction. The variability of indexes for ready-mixed concrete was low, and of the same order as indexes for the concrete made in the laboratory. A series of round robin tests is currently underway to further assess the variability of the tests in a number of commercial laboratories in South Africa. Preliminary findings indicate that the OPI and sorptivity tests give acceptable results in most laboratories, but the chloride conductivity test is more difficult to control 15. The suite of index tests is also currently being used in a RILEM round robin in order to evaluate the validity of a wide range of test methods for characterising covercrete properties 16.

9 1.6 Chloride Conductivity (ms/cm) days CSF 120 days CSF 28 days FA 120 days FA 0.2 Wet Compound Hessian Curing Method Sand Air cure 28 days GGBS 120 days GGBS Figure 9: Influence of early site-curing practices on chloride conductivity results for slabs at 28 and 120-days 7 Table 1: Estimates of the single operator coefficients of variation (1s%) 6 Sorptivity OPI Chloride Concrete source Conductivity Actual structures 13 % 3 % 14 % Wet-cured, site mixed concrete 12 % 2 % 7 % Wet-cured, ready mixed concrete 7 % 1 % 5 % Laboratory A 5 % 1 % 4 % Laboratory B 6 % 1 % 6 % 3.2 Quality Control and Assessment of Construction Quality The suite of DI tests is being used to monitor the quality of actual construction. The usual procedure involves retrieving samples of the concrete mix at the point of discharge (usually a ready-mix truck), as well as retrieving small surface core samples from the as-built structure approximately 28 days after casting. The former samples are cured in a standard way and are used to assess concrete mix quality, while the latter samples are used to assess construction quality. Appropriate limiting values for the different indexes are still being actively debated. However, one example of a typical scheme is given in Table 2, taken from an actual durability specification currently in use in South Africa 17, which has found acceptance with authorities, consulting engineers, and constructors.

10 Table 2: Acceptance limits for durability indexes 17 Acceptance Criterion OPI (log scale) Sorptivity (mm/ h) Conductivity (ms/cm) Laboratory concrete > 10 < 6 < 0.75 Full acceptance > 9.4 < 9 < 1.00 Conditional acceptance 9.0 to to to 1.50 Remedial measures 8.75 to to to 2.50 As-built Structures Rejection < 8.75 > 15 > 2.50 A limitation of the above approach is that it does not recognise the matrixing effect of binder type and exposure environment, particularly in relation to chloride environments. This is further considered below. 3.3 Performance-Based Specifications 18 Most current specifications contain prescriptive requirements, that is they require limits on factors such as w:b ratio, cover, period and type of curing, minimum binder contents, etc. If written appropriately, they can have merit by providing information to assist constructors in achieving the performance requirements demanded of the structure, by giving guidance on best practice ; and by covering particular requirements necessitated by local conditions, e.g. materials and environment. However, they do not directly address the quality of the concrete cover layer likely to be achieved in practice. Durability index values can be used to overcome this deficiency by being incorporated in appropriate performance-based specifications. In keeping with typical structural design codes, two possible approaches to specifications can be identified Deemed to Satisfy Approach This approach would be adequate for the bulk of construction. It has an analogy in structural design with the deemed to satisfy rules associated with, for example, span/depth rules for deflection checks. The approach involves requiring as-built structures (and laboratory trial specimens) to conform to limiting criteria for durability indexes, such as those given in Table 2. If conformance is achieved, the structure is deemed to satisfy the durability requirements. This approach can be coupled with penalty measures for cases of non-conformance. Incentives for excellent performance can also be introduced. To illustrate the limitation of this approach in terms of single values, particularly for chloride conductivity, consider Table 3. The exposure classes in the table are suitable for South African marine conditions, and the various binders are all in regular use. It can be seen that limiting chloride conductivity values depend on both exposure conditions and binder type. The values in Table 3 for any horizontal row can be regarded as giving approximately equal protection against chloride ingress, but a single nominal value is an oversimplification.

11 Table 3: Allowable maximum chloride conductivity values (ms/cm) at 28 days (Marine Exposure) Moist Cured (3-7 d) Concrete Type (Binder) Marine Environment 100% PC 10% CSF 30% FA 50% Slag Extreme Very Severe Severe Marine Exposure Zones are those for SA conditions as follows: Extreme: Structure exposed directly to seawater with heavy wave action and/or abrasion Very Severe: Structure exposed directly to seawater under sheltered conditions, little wave action Severe: Structure located near shore in an exposed marine location Service Life Approach The essence of this approach is to matrix the key elements of concrete and binder type, likely on-site curing, environmental exposure conditions, concrete cover to reinforcement, notional design life or Service Life of the structure, and optimisation for best economy. This represents a sophisticated approach, likely to be adopted only in critical or important cases. It will rely on the ability to characterise concrete properties at the construction stage to give an assurance of long-term durability. Concrete cover to reinforcement has an important influence on the economics of construction. Larger covers require less durable concrete to provide protection to the steel, but at the same time may increase the cost of construction. In critical cases, it will be necessary to optimise concrete type and steel cover, subject to the exposure conditions and the economics of construction. An example of this approach, which also introduces a design life allowance, is given in reference 18. In practical terms, applying this approach indicates that, for PC and CSF concretes, adequate durability in marine conditions usually requires concrete grades in excess of 60 MPa. These mixes often given rise to other problems, such as early age autogenous shrinkage and excess hydration temperatures, which may induce internal microcracking. Furthermore, PC matrixes are not highly resistant to chloride ingress. For FA and slag concretes, the use of larger covers and less onerous exposure conditions results in concrete grades less than 30 MPa and/or water:binder > For reasons of conservativeness, it is probably wise not to permit such mixes in marine or chloride environments. Thus, only a small range of mixes is both acceptable and practical, and usually requires use of a cement extender (typically FA or slag). Such mixes can be used with confidence over a wide range of cover and exposure conditions. 3.4 Prediction Models Presently, the OPI is being used as a predictive tool for estimating carbonation depths in structures, while the chloride conductivity index is being used similarly for chloride ingress estimations. The prediction models are based on empirical correlations between early age index values and subsequent rates of ingress of aggressive agents. An example of the use of the chloride conductivity index is given in ref. 19, while the basis of the carbonation model is shown in Figure 10 (a). The experimental results in Figure 10 (a) were reworked to give the nomogram in Figure 10 (b) 20. These preliminary prediction models are relatively crude, but have been shown to perform with acceptable accuracy in practical situations. Current work is

12 aimed at more fundamental modelling based on hydration and microstructure development coupled with transport models and including environmental influences (a) Carbonation depth (mm) PC-60% FA-60% SL-60% PC-80% FA-80% SL-80% Oxygen permeability index Carbonation depth (mm) (b) Oxygen permeability index R.H. 60% 80% 90% Figure 10: Basis of construction from experimental results (a), of carbonation prediction nomogram (b), using OPI Closure The paper lays out the background and philosophy to the durability index approach for achieving durable reinforced concrete structures in cases where the main process of deterioration is reinforcing steel corrosion. The durability index approach is based on the premise that suitable transport-related material parameters must be measured on in-situ concrete in order to characterise the multiple effects of materials, construction, and environment. Three durability index tests have been developed which are sensitive to these important effects, and which have sufficiently low statistical variability to make them useful in

13 practical reinforced concrete construction. The use of the tests has progressed to the point where performance-based specifications are being drafted and tentatively used in construction in South Africa. Present work involves gaining greater confidence in the use of the tests by round robin testing, and selecting appropriate limiting values that can be used without ambiguity in performance specifications. 5. Acknowledgements The author acknowledges the important contributions made to the development of the ideas and practice presented in this paper by researchers and students active in the research programme, in particular Prof Y Ballim, Dr J Mackechnie, Mr B Raath, and Dr P Streicher. 6. References 1. Ballim, Y. and Alexander, M.G., Research in concrete deterioration and durability, in A review and some thoughts on future developments, Proceedings of a Colloquium on South African research needs in cement and concrete, Pretoria, 2000 (National Research Foundation) 44 pp. 2. Alexander, M.G. and Ballim, Y., Experiences with durability testing of concrete: a suggested framework incorporating index parameters and results from accelerated durability tests. Proceedings of 3rd Canadian Symp. on Cement and Concrete, Ottawa, August, 1993, (Nat. Res. Council, Ottawa, Canada, 1993) Series of Monographs published by Department of Civil Engineering, University of Cape Town; 6 No., Neville, A.M., Why we have concrete durability problems, ACI SP-100, Katherine and Bryant Mather International Conference on Concrete Durability, (American Concrete Institute, Detroit, 1987) Alexander, M.G.. Mackechnie, J.R. and Ballim, Y., Use of durability indexes to achieve durable cover concrete in reinforced concrete structures, Chapter, Materials Science of Concrete, Vol. VI, Ed. J. P. Skalny and S. Mindess (American Ceramic Society, Westerville, 2001) Gouws, S.M., Alexander, M.G. and Maritz, G., Use of durability index tests for the assessment and control of concrete quality on site, Concrete Beton, 98 (2001) Du Preez, A.A. and Alexander, M.G., A site based study of durability indexes for concrete in marine conditions. To be published in Materials and Structures. 8. Ballim, Y., A low cost falling head permeameter for measuring concrete gas permeability, Concrete Beton, 61 (1991) Streicher, P.E. and Alexander, M.G., A chloride conduction test for concrete, Cement and Concrete Research, 25 (6) (1995) Streicher, P.E. and Alexander, M.G., Towards standardisation of a rapid chloride conduction test for concrete, Cement, Concrete and Aggregates, 21 (1) (1999) Mackechnie, J.R., Predictions of reinforced concrete durability in the marine environment, PhD Thesis, University of Cape Town, Ballim, Y., Curing and the durability of OPC, fly ash and blast-furnace slag concretes, Materials and Structures, 26 (158) (1993)

14 13. Ballim Y., Taylor P.C. and Lampacher, B.J., Assessment and control of concrete durability. Proceedings CSSA National Convention: Concrete Meets the Challenge, 1994 (Concrete Society of Southern Africa, Midrand, 1994) 12pp. 14. Dixon, S., The effects of bleeding on the durability of concrete, Final Year Thesis (Unpublished), Department of Civil Engineering, University of Natal, Grieve, G.R.H., Alexander, M., Ballim, Y. and Amtsbuchler, R., Evaluation of the interlaboratory precision for three South African developed durability index test methods. Proceedings of 11 th International Congress on the Chemistry of Cement, Durban, May, 2003 (CD ROM, Cement & Concrete Institute, Midrand, 2003) 7 pp. 16. RILEM TC 189-NEC Non-destructive Evaluation of the Concrete Cover. 17. Raath, B., Notes from a Workshop on Durability: Reinforced Concrete for the Year 2100, (Concrete Society of Southern Africa, Midrand, 2001). 18. Alexander, M.G., Towards a performance specification for reinforced concrete, based on durability indexes. Proceedings of International Conference on Performance of Construction Materials, Cairo, February, Ed. A El-Dieb, S. Lissel, M.M. Reda Taha (Elmaarefa Printers, Cairo, 2003) Alexander, M.G. and Mackechnie, J.R., Concrete mixes for durable marine structures, Journal of the South African Institution of Civil Engineering, 45 (2) (2003) Mackechnie, J.R and Alexander, M.G., Durability predictions using early-age durability index testing. Proceedings of 9th Durability and Building Materials Conference, Brisbane, 2002 (Australian Corrosion Association, Brisbane, 2002) 11 pp. 21. Griesel, E. And Alexander, M.G., Modelling the durability properties of concrete cover. Proceedings of 11 th International Congress on the Chemistry of Cement, Durban, May, 2003 (CD ROM, Cement & Concrete Institute, Midrand, 2003) 10 pp.