AIR PERMEABILITY ASSESSMENT IN A REINFORCED CONCRETE VIADUCT

Transcription

1 AIR PERMEABILITY ASSESSMENT IN A REINFORCED CONCRETE VIADUCT Rui D. Neves (1) and João Vinagre Santos (1) (1) EST Barreiro - Instituto Politécnico de Setúbal, Portugal Abstract When the deterioration key factor is steel corrosion, the lifetime of reinforced concrete structures highly depends on the permeability and thickness of concrete covering reinforcing bars. Adequate permeability is usually achieved by means of assuring a minimum cement content, maximum water to cement ratio and minimum compressive strength. The new European Standard for concrete (EN 206-1) allows performance design methods with respect to durability, for instance by the specification of values for some relevant concrete properties. However, these properties are to be determined on standard specimens and therefore only allow distinguishing the potential performance of different concrete compositions. As there are other factors, such as placing, compacting and curing, which influence concrete permeability results, it is extremely important to evaluate the real permeability of concrete in structures. This work presents an assessment of a viaduct s concrete air permeability using the Torrent Permeability Tester (TPT). The tests were performed in different structural elements made of concrete of two different strength classes. The obtained results from structural elements were compared with those obtained from cast specimens, cured, preconditioned and tested under laboratory conditions. This comparison shows significant differences between both situations, pointing out that a reliable durability approach should also include specifications for site durability related properties and rules for its conformity control. Keywords Concrete; durability; air permeability; NDT 299

2 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS INTRODUCTION The settlement of proper rules for reinforced concrete structural design and its correct use by the technical community led to constructions with suitable structural behaviour. However, the deterioration due to environmental action remains one of the most common related causes of its faults. This deterioration generates a reduction of the structural elements resistance or to constructions appearance. Besides mechanical actions, reinforced concrete structures are subjected to several physical and chemical attacks during its service life. In general, deterioration is caused by reinforcement corrosion. Steel is corroded in contact with humidity and oxygen, but when placed within concrete an oxide film appears and sets up a passivity condition, which protects the steel against corrosion. This film is stable in an alkaline environment, such as young concrete, although it can be destroyed if the ph lowers or if the chloride content reaches a critical threshold. Knowing that the decrease of ph can be caused by concrete carbonation, the key agents of the process that leads to reinforcement corrosion are carbon dioxide and chloride ions. So, it is important to assure that the separation board between reinforcement and these agents, known as cover, has a suitable quality and thickness to ensure steel protection against corrosion during the structure s service life. The concrete resistance against carbonation, or chloride ingress, increases with increasing cement content or compressive strength and with the decrease of the water-cement ratio, but the cement type or the presence of additions also has an important role [1]. The specification of mix characteristics based on the environmental aggressiveness is named as a prescriptive durability approach. However, in terms of technical and economical point of view, this kind of approach is questionable. In the document that goes more deeply into this subject, the European Standard for concrete [2], there has been an evolution towards the possibility of a more rational approach. The EN already allows the use of the equivalent performance concept, the use of methods based exclusively on performance and also of prediction models. That performance is evaluated by properties considered to be relevant for the deterioration mechanism(s). As a result of already developed research, nowadays there are some properties considered to be suitable to evaluate concrete durability [3], as well as some methods to determine those properties [4][5][6], but they are not yet standardized. Durability related properties specification values has been already made in important engineering structures such as the Vasco da Gama Bridge [7] or the Millau Viaduct [8], where the intention was to have service lifetimes of over 100 years. Nevertheless, values specification was based on concrete laboratory s performance knowledge, not on their real behaviour. Usually, the concrete is not delivered on site as a finished product. On the other hand, it will be submitted to a set of procedures (placing, compaction and curing) which may strongly influence its future performance [9]. Therefore, it is important to be able to evaluate durability related properties on concrete applied in structures, which are surely more related to its real performance. Nowadays, some information about the ability of different site methods to evaluate cover concrete durability related properties is available as a result of a study developed by a RILEM technical committee [10]. 300

3 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS 2008 The Torrent Air Permeability Test (TPT), considered to be a suitable method for that evaluation, was used in the present study. TPT basics, equipment and procedures details can be found in [11]. 2. TESTING PROGRAM The testing programme s main objective was to evaluate the influence on the concrete air permeability, and therefore on its durability, of placing, compaction and curing. That evaluation was performed by applying TPT to: - control specimens, cured in water for 7 days and then kept in laboratory (22 ºC e 60% HR) until testing age; - structural elements where the concrete corresponding to the control specimens was applied. tests were always performed over moulded surfaces. Slabs of cm 3, cast together with compressive strength control specimens (15cm cubes), were also used as control specimens. Air permeability tests were performed after 28 days or as soon as possible. Tests were performed on three different concrete compositions, indicated in Table 1. Table 1: Concrete composition Component C40/50.S4.D25 C40/50.S3.D25 C30/37.S3.D25 Fine sand (kg/m 3 ) Sand (kg/m 3 ) Coarse agg. 1 (kg/m 3 ) Coarse agg. 2 (kg/m 3 ) Cement I 42.5 R (kg/m 3 ) Fly ash (kg/m 3 ) Water (l/m 3 ) Admixtures (l/m 3 ) As for structural elements, tests were performed in box girder walls and upper slab, piers, prefabricated struts and also in abutments. For those surfaces there were no special protection procedures beside the formwork itself. The moulds were removed from the elements as indicated in Table 2. For TPT purposes the tested surfaces may be considered dry as they hadn t been in contact with water for more than 15 days. Table 2: Demoulding time Element Time Element Time Girder At least 3 days Struts 12 hours Piers 24 hours Abutments 24 hours 301

4 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS RESULTS For the present study were considered the casting of 9 elements. The obtained results are presented in the next paragraphs, grouped by structural element. Beside the air permeability values (kt), compressive strengths at 28 days (f c,28 ) are also presented. Two compressive strength values are presented for each casting: one corresponding to the average of all the samples taken during that casting, and the other to the specimens taken together with the air permeability sample. This last one is presented in the same row of the laboratory air permeability results. 3.1 Box girder In this study, three of the box girder castings were monitored. The site tests were performed in two walls (corresponding to two different concrete mixtures) and in one upper slab of the box girder. All tests were performed in the inner side of the box girder. The results are presented in Table 3. Table 3: Box girder test results Region Concrete kt (10-16 m 2 ) f c,28 (MPa) Wall E1P1 C40/50.S4.D25 Wall P1P2 C40/50.S3.D25 Upper slab C40/50.S3.D Pier The casting of one pier was monitored for air permeability purposes. Tests took place in 3 of the 4 pier faces belonging to the same region. The results are presented in Table

5 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS 2008 Table 4: Pier test results Region Concrete kt (10-16 m 2 ) f c,28 (MPa) Pier C40/50.S3.D Prefabricated struts Two sets of prefabricated struts were also monitored. Tests took place with the struts in its final positions. Different faces and strut positions were chosen for each set. The results are presented in Table 5. Table 5: Prefabricated struts Region Concrete kt (10-16 m 2 ) f c,28 (MPa) Set 35 C40/50.S3.D25 Set 37 C40/50.S3.D

6 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS Abutments Three elements from one of the abutments were monitored for this study. The site tests were performed in two walls and in one corbel. These are the only study regions where the lower strength class concrete was applied. The results are presented in Table 6. Table 6: Abutment test results Region Concrete kt (10-16 m 2 ) f c,28 (MPa) Wall 5/4 C30/37.S3.D25 Wall 4/3 C30/37.S3.D25 Corbel C30/37.S3.D DISCUSSION First, it is important to discuss which value is representative of a set of measurements. Although a coefficient of variation of 6.6% for TPT is indicated in the Swiss Standard for determining air permeability in structures [12], the variability associated to air permeability tests is well known. For instance, in TC NEC Round Robin Test for laboratory testing, the mean coefficient of variation was around 25% [10]. When working on site, higher coefficients of variation should be expected. In fact, single pores or microcracks, which are not relevant for concrete durability, cause air permeability to vary several orders of magnitude. Despite of this scatter, some works use the arithmetic mean to represent a set of values [10] while others [13][14] use the geometric mean for that purpose, which corresponds to the 50% fractile of a lognormal distribution which Denárie et al. [13] found to be the distribution of TPT values. 304

7 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS 2008 However, when there is a small set of samples, a single reading, which may be influenced by an isolated pore or a microcrack, may strongly affect any of the above means. Therefore authors believe that is preferable to use the median as the representative value for a set of measurements, as is standardized for rebound hammer testing [15]. The median values are presented in Table 7, as well as a comparison between site and laboratory results. Table 7: Summary of TPT results Region Concrete kt site (10-16 m 2 ) kt lab ktsite (10-16 m 2 ) ktlab Wall E1P1 (box girder) C40/50.S4.D Wall P1P2 (box girder) C40/50.S3.D Upper slab (box girder) C40/50.S3.D Pier C40/50.S3.D Set 35 (pf struts) C40/50.S3.D Set 37 (pf struts) C40/50.S3.D Wall 5/4 (abutment) C30/37.S3.D Wall 4/3 (abutment) C30/37.S3.D Corbel (abutment) C30/37.S3.D When comparing all values presented in Table 7, it can be noticed that the laboratory result from the corbel is abnormally high. This difference is believed to be caused by a defective manufacturing of the specimen, therefore this result will be disregarded in what follows. The laboratory results show some variability between samples from the same mix design. However, this variability is also present in compressive strength results, although with a lower magnitude. Air permeability assessed on site was generally higher than in laboratory specimens, increasing the difference in structural elements where casting is associated with large concrete volume (box girder). Actually, large volume castings, the high rates of pouring concrete, the existence of several work fronts and the large number of labor hours, usually lead to a lower quality in placing and compacting concrete when compared to smaller castings. As no special products or procedures were used, the curing was based only on concrete protection. The curing seems to have no influence on the observed permeability differences between laboratory and site measurements, since those differences are bigger in the elements where the formwork remained longer (box girder). Moreover the box girder tested surfaces were in the inner side, so after the formwork removal they were still protected from direct sunlight and wind. However, the large concrete volume applied in each box girder casting, associated to high cement content and environmental temperature, may have developed high hydration heat that may have caused concrete microcracking, even though it justifies the larger differences. 305

8 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS 2008 The air permeability and compressive strength results are plotted in Figure 1. Both show a considerable scatter, due mainly to the above mentioned variability associated to air permeability tests, and also a weak correlation. Despite that weak correlation, also reported in other works [14][16], one can distinguish permeability ranges associated with each compressive strength class. Also the median for all measurements made for concrete C30/37 (kt= m 2 ) its more than twice the median for all measurements made for concrete C40/50 (kt= m 2 ). Actually, Torrent [11] suggests the use of classes also for air permeability. 0,30 Air Permeability - kt (x10-16 m 2 ) 0,25 0,20 0,15 0,10 0,05 site lab 0, Compressive Strength (MPa) Figure 1: Tests results Nevertheless some works [9][17][18] have already shown the strong correlation between air permeability and carbonation resistance or chloride diffusion, and some prediction models use it to forecast concrete performance in what concerns durability. 5. CONCLUSIONS The developed work has shown that there may be relevant differences between air permeability results from laboratory and site testing. As this property is often used as input in prediction models, it is of the outmost importance to achieve an accurate prediction to introduce in the model a value which represents the permeability of the concrete that will be exposed to the environmental actions. Despite the scatter of the results, TPT seems to be suitable to assess site air permeability of concrete cover, using the median as the representative value of a set, since the results have consistently reflected an expected different permeability between site and control specimen concrete, due to different placing, compaction and curing procedures. ACKNOWLEDGEMENTS The authors would like to acknowledge the Portuguese Roads Institute - EP, for all its support during the development of this work. 306

9 on Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS 2008 REFERENCES [1] Concrete Society, Developments in Durability Design & Performance-Based Specification of Concrete, [2] EN Concrete part 1: specification, performance, production and conformity,2005. [3] RILEM TC 116-PCD, Permeability of concrete as a criterion of its durability. Final Report: Concrete durability An approach towards performance testing, Materials & Structures 32 RILEM (1999) [4] RILEM TC 116-PCD, Preconditioning of concrete test specimens for the measurement of gas permeability and capillary absorption of water, Materials & Structures 32 RILEM (1999) [5] RILEM TC 116-PCD, Measurement of the gas permeability of concrete by the RILEM - CEMBUREAU method, Materials & Structures 32 RILEM (1999) [6] RILEM TC 116-PCD, Determination of the capillary absorption of water of hardened concrete, Materials & Structures 32 RILEM (1999) [7] Portaria n.º 366-A/93, Diário da República 76 INCM (1993). (in Portuguese) [8] Guerinet, M., La durabilité des bétons, Travaux 816 FNTP (2005) [9] Torrent, R., Towards a performance-based specification and conformity control of durability, In Structural Concrete and Time, Proceedings of an International Fib Symposium, La Plata, 2005, [10] RILEM TC 189-NEC, Comparative test - Part I - Comparative test of penetrability methods ; Materials & Structures 38 RILEM (2005) [11] Torrent, R., A two-chamber vacuum cell for measuring the coefficient of permeability to air of the concrete cover on site, Materials & Structures 25 RILEM (1992) [12] SN /1, Norme Suisse: Construction en béton Spécifications complémentaires, Annexe E: Perméabilité à l'air dans les Structures, (2003) [13] Denarié, E. et al., Air permeability measurements for the assessment of the in situ permeability of cover concrete, In Concrete Repair, Rehabilitation and Retrofitting, Proceedings of an International Conference, Cape Town (2005) [14] Jacobs, F., Hunkeler, F., Non destructive testing of the concrete cover Evaluation of permeability test data, In Performance Based Evaluation and Indicators for Concrete Durability, Proceedings of an International RILEM Workshop, Madrid (2006) [15] EN , Testing concrete in structures. Non-destructive testing. Determination of rebound number, CEN (2001). [16] Leeman et al., Assessment of destructive test methods to determine the covercrete quality of structures, In Performance Based Evaluation and Indicators for Concrete Durability, Proceedings of an International RILEM Workshop, Madrid (2006) [17] Ribeiro, B. et al., Durabilidade do betão. Contribuição para a especificação de requisitos de desempenho, In Betão Estrutural 2002, Actas do Encontro Nacional, Lisboa (2002) (in Portuguese). [18] Gonçalves, A. et al., Caracterização da durabilidade do betão, In Betão Estrutural 2000, Actas do Encontro Nacional, Porto (2000) (in Portuguese). 307