Life-time prediction of high-performance concrete with respect to durability

Résultats exploitables

The hpc mixtures from the ConLife project were tested in their behaviour against acid attack. After the immersion testing of hpc specimens in diluted sulphuric acid a relative ranking of the mixtures in their behaviour under acid attack was established. In general it can be said that all high performance concrete mixtures showed a better resistance against acid attack than the reference ordinary type concrete that was tested in comparison. The main trends seen in the acid attack results include: - Using fly ash in combination with micro silica fume resulted in an improved acid resistance. The reason for the performance was a denser packing of the cement paste and aggregates due to the fly ash. The micro silica fume reacts pozzolanically and transforms the calcium hydroxide chemically and by that decreases its content in the binder matrix. This also increases the acid resistance. Without fly ash, the micro silica fume alone can result in micro cracks, which will increase the path for acid intrusion. The fly ash releases alumina ions and above a certain level of ion content the shrinkage is lower and therefore less cracks will occur. This was confirmed by the observation that using micro silica alone resulted in a lower acid resistance. - Using air entrainment resulted in an increased acid resistance because the entrained micro air voids improved the workability and the ability of the concrete to be packed denser. In other words, the concrete is more homogenous. Furthermore, the air voids block micro capillaries and prevent the acid from invading the concrete through these canals. - Using extra fine blast furnace slag did not provide any improvement to the acid resistance. It can be assumed that the combination of fly ash and silica fume is more effective than the use of extra fine blast furnace slag alone. The reason for this is probably that the slag does not change the amount of calcium hydroxide in the concrete. - To produce a dense concrete like in high performance concrete is favourable for the durability under acid attack. The dense matrix slows down the speed of the acid penetration and the dissolution of the matrix itself. The damage development is delayed. With that information it should be possible to select a good concrete test mixture which means to give recommendations or to make predictions to improve the mix design of hpc to resist acid attack prior to laboratory testing which has to be done before a special concrete mix design is used. From that scientific result a commercial result may off spring in the following manner. The acquired know how may be used to give advice to potential users that are trying to make acid resistant concrete. If this is possible a better acid resistant hpc will be available for general use. The main commercial value is additional knowledge about acid resistant high performance concrete combined with the existing information from former investigations documented in the literature. This knowledge can be used to give advice to potential users. A key innovative feature can be that an inherently acid resistant concrete can be produced that doesn't need any protective surfaces, which are quite often used. By that maintenance costs can be reduced. An acid resistant concrete may have a longer service life than a protective surface. This information or knowledge transfer to users in practice can be done by the project partners.

The results include laboratory and field studies of resistance of HPC to chloride attack. More than 22 mixtures of HPC, manufactured in Norway, Germany and Italy, have been investigated in the studies. The Nordic standardised rapid test method NT BUILD 492 was employed in the laboratory study for determination of chloride migration coefficient in different types of HPC cured at different ages (28 days, 6 months and 2 years). The specimens were also exposed in the seawater of different geographic zones from southern to northern European coasts. Chloride penetration profiles in the specimens were examined after one and two years field exposure. Overall, the results from the CTH method showed that the concrete blended with 7% silica fume reveals an excellent resistance to chloride ingress at both 28 days and 6 months ages. Blending with fly ash or slag results in a good resistance at the older age (e.g. 6 months), but not at the younger age (28 days), due to the slower hydration process. It should be noticed that the chloride diffusivity is just an indicator of chloride penetration and that chloride penetration itself does not make any damage of concrete, but induce corrosion of reinforcement steel, the chloride threshold value for corrosion must be taken into account when evaluating the resistance of concrete to chloride attack. Both the chloride migration coefficient from the laboratory rapid test and the chloride penetration profiles from the field exposure have supplied very useful information about the influence of the wide range of different constituents on the resistance of HPC to chloride attack. The chloride ingress data from such a wide range of different mixtures of HPC covering different geographic zones of chloride attack environments are of greatest interest for concrete consumers and suppliers. With these data it was possible to develop a feasible model for prediction of chloride ingress in HPC, so as to be able to assess the resistance of HPC to chloride attack in advance and to avoid therefore the corrosion of reinforcement steel in HPC structures in practice. The experimental results received by different partners have been already partly published and have been included within the final manual with recommendations for optimal use of HPC in different chloride attack environments (Deliverable D10 Report).

The aim of reviewing in-service structures (Tasks 1.2 and 2.1, as reported in Deliverable D2 [2]) was to get information about damage behaviour of real structures as well as compatible and optimal mix design concepts. The investigated structures present a wide range of different exposure types and application forms (e.g. bridges, road pavements, pre-cast elements, harbour decks), age and strength categories. The influences of environmental aspects on the properties of HPC were grasped with help of extensive analysis of a high number of in-service structures under different climatic conditions and exposure classes. Therefore a description is given of the geographic location of every in-service structure and its exposure class in order to classify the environmental influences on the concrete. The laboratory examinations focused on the coverage of macro- and microstructure conditions of the concrete. The data collection was a basis for understanding the degradation mechanisms and the practical behaviour. 90 in-service structures and test elements were investigated under different environmental conditions. Compact descriptions of all structures and test results are inserted in a public open database in the section �Older structures�. The D2 Report contains the synopsis of the visual conditions as well as the macro- and micro-structural properties of the investigated constructions. 61 of the investigated structures and test elements were reported in D2. The following structural investigations and parameter studies were done: Quantitative investigations: -Compressive strength -Mercury Intrusion Porosimetry (MIP) -Air pore content, spacing factor, A300 content -Carbonation depth -Chloride profile -Capillary water uptake -Frost testing Qualitative investigations: -Visual analysis -Optical analysis by means of thin- sections -XRD -SEM and ESEM -EDX The volume of the past experiences with HPC differed between the participating countries. The most experiences exist in the Nordic region because of their extensive field test investigations in the past and many practical applications in different kinds of constructions. The survey of in-service structures covers a wide range of different concrete types and environmental and mechanical burdens, including: -Climatic zones from Northern to Southern Europe -12 types of structures -Ages of constructions between 0 and 30 years -7 attack classes according EN 206 -Concrete mixtures with equivalent w/b-ratios from 0.16 to 0.54 -Nominal compressive strengths from 30 to 158 Mpa

The goal of this database was to provide a tool for organizing the most relevant data that have been generated during the project. In agreement with all the partners of the CONLIFE project the code has been developed adopting Microsoft Access Language. A further relevant goal of the database was to allow the fast search of data for building of useful reports. A version can also be located on the CONLIFE web page. The final database version has been completed. It comprises two main sections: the first one is related to the mixes that have been cast and tested during the project; the second is related to coring of older HPC structures and their testing. The data can be edited, printed by reports and exported as Excel files for their elaboration. The section related to older structures has been completed with photos. Overall, the database contains information on more than 60 produced mixtures subjected to accelerated laboratory and real-time field tests and more than 50 in-service structures from which samples were tested. During the development of the project the database has proven to be a useful tool in aiding model building, furthermore it has helped to share the data and results in a clear form among the partners. In the future it will help concrete researcher and people from industry in designing HPC.

Cyclic temperature testing incorporates technology previously used for testing of other materials, such as marble facades. The brittle nature of HPC material may be altered by the thermal gradients within the mass, thus deterioration with the expansion and contraction of the actual matrix and not just the water (or ice, as seen in the freeze-thaw testing. The cyclic temperature results were used to assess the validity of laboratory test methods and theories supporting deterioration of concrete in harsh environments. Since there are no standardized tests for assessing cyclic temperature attack, past experience was applied. The cyclic temperature effects on HPC were tested in climatic chambers using two different methods: - A modified ASTM C666 method without moisture (sealed beams cycling daily +/-40C for over 100 cycles) - A weather exposure test of cyclic temperature and moisture attack (-25C to +60C including moisture period, for 8 cycles with each cycle lasting 6 weeks). Overall there were good results from the cyclic temperature attack tests, in that severe damage was not found. The best behaviour was generally found in concretes with a low w/c ratio. Some improvement was also seen when adding silica fume and normal blast furnace slag. In the weather simulation test, the mixtures made with special fine blast furnace slag had greater deterioration than ordinary slag. The improvement compared with normal OPC concretes requires appropriate curing and hardening time of the HPC prior to being subjected to climatic stresses. The main conclusions about all tests was that cyclic temperature variation is not a significant mode of failure in these HPC mixtures compared with the extent of damage in the frost tests or compared to OPC. The results will be useful to concrete suppliers and producers supplying materials to moderate and severe weather locations. Designers will have a better understanding of how the HPC behaves and deteriorates. The results will also be disseminated to other material users outside of the concrete field, such as the stone and tile industry.

The main objective of this deliverable was the selection of compatible constituents for representative mixtures as basis for the development of representative mix designs. The mix design of mixtures for an European project have to cover on one hand general and on the other hand regional specific properties relating to very different climatic conditions within Europe. Therefore one typical cement was chosen for every climatic region. All partners decided to use the same aggregate, the same type of silica fume and fly ash, the same 2 types of blast furnace slag as well as the same type of air entrainment agent and superplasticizer to ensure the compatibility of the mixtures among each other. Extensive tests were conducted to assess the material properties. The following list gives an overview about the test which has been carry out: -cement: chemical analysis, density, XRD, Particle size distribution, compressive strength, flexural tensile strength, specific surface area, water demand, setting -aggregate: sieve curves, mineralogical description, water uptake -silica fume: chemical analysis, particle size distribution, activity index, -fly ash: chemical analysis, particle size distribution, setting activity index, water demand -blast furnace slags: chemical analysis, particle size distribution superplasticizer: effectiveness in different cements -air entrained agent: effectiveness in different cements A high number of test mixtures were performed to assess the compatibility of the materials and to establish consistent mixture guidelines for all 3-production partners. Totally, 40 at IBPM, 5 at HUT, 25 at NBI, 5 at ITC trial mixes have been carried out. Some rheological testing was done at IBPM with HWU from the SUPERPLAST project additionally to assess the fresh concrete properties with the CONLIFE admixtures. These tests are part of the Cluster activities. As result of this extensive examination of every single constituent as well as its behaviour in combination with the other components 22 mixtures for each climatic region were developed.

The aim of this part of the project was to relate fresh and hardened concrete properties to the mix design of the HPC. From the test results obtained it was possible to choose what concrete properties relate best to lifetime prediction. A very comprehensive laboratory programme has been carried out. A total of 22 mix-designs have been cast in Italy, Germany and Norway with different cements. Except for the three different cements used and the local water supply the rest of the mix design was identical in the 22 mix designs. The specimens were exposed both to laboratory and field conditions. The field exposure sites were located in Iceland, Sweden, Finland, Germany and Italy. Fresh concrete properties were tested in the three laboratories casting. The testing includes: Temperature, density, air-content, consistency, and loss of consistency and heat development. The test methods used were EN standard test methods. The mean target consistency of all mixes was in the range of 460 mm for flow and 160 mm for Slump, as typical for HPC. The air entrained mixes showed a mean air content around 5%. The testing of hardened concrete properties showed that the compressive strength (28 days) was in nearly all cases in the range of 70-130 MPa, as originally fixed. Testing of other hardened concrete properties included frost resistance under different conditions, shrinkage, chloride penetration, oxygen penetration, porosity, thin section analysis. The received results formed the basis for the further evaluation of the durability data and for the description of damage models.

This result included the study of all involved mixtures under pure frost as well as under combined deicing agent attack, simulated by accelerated test methods. For the simulation two reliable test procedures have been used - the CDF/CIF- and Slab-test. By using these test methods the reduction of the relative dynamic modulus of elasticity or length change for internal damage, the continuous water uptake as well as the scaling rate were determined. Therefore, the results could clearly show the influence of the wide range of different constituents on the frost resistance of HPC. Dependencies of the different mixture compositions on the internal damage parameters as well as surface scaling could be also clearly demonstrated. Great differences between pure frost resistance and resistance against deicing salt could be observed, so that these results formed the basis for the modelling of frost damage mechanisms. From the received results, it could be seen, that the moisture uptake - as most important parameter for the deterioration progress could be subdivided into three different sections, depending on the w/b-ratio and the type i.e. amount of additives. Each section could be characterised by different effects. Overall, the results have shown, that additives have a high impact on the deterioration process and the gradient of loss of RDM. In addition, the influence of ageing has been studied, which showed high influences on the frost resistance. A summary of the data can be found in the D6 report as well as an evaluation of the damage process within the D8 report. All deliverable reports can be downloaded from the Conlife web page. Basically, fundamental data of such a wide range of different HPC mixtures under frost attack are of greatest interest for concrete consumers and suppliers. With this knowledge it is possible to assess the frost resistance of HPC mixtures even in advance and to avoid therefore frost or deicing salt damages on HPC structures in practice. These experimental results received by different partners have been already partly published within different scientific journals or conference proceedings. A summary together with hints for optimal frost and de-icing salt suitable HPC mix designs are also given within the manual (Deliverable D10), which will be disseminated at the final CONLIFE workshop.

This result details how HPC performed in field tests with regards to various degradation attacks. 22 mixtures from each of the 3 regions (Nordic, Central Europe, Southern Europe) were placed in various field conditions for a period of 2 ½ years to assess how they perform when subjected to various environmental conditions. Field stations were established in various countries to account for the varying environments. The field stations were: - Mulheim Ruhr (N. Rhine), Germany; Moderate exposure, 130 m elevation; testing Freeze-thaw, with and without deicing salt - Kruft, Germany; Moderate exposure; testing Freeze-thaw, with and without deicing salt - Northern Italy; Moderate exposure; testing Freeze-thaw with deicing salt (scaling) - Southern Italy; Mild exposure; testing Seawater, chloride penetration - Kopavogur, Iceland; HAKOP floating seawater station; testing freeze-thaw with salt (scaling), seawater attack, chloride penetration - Reykjavik, Iceland; KORPA station; testing freeze-thaw without deicing salt - Boras, Sweden; Severe exposure; along Highway Rv40; testing freeze-thaw with deicing salt - Traslovslage Harbour, Sweden; Seawater station; testing seawater attack, chloride penetration - Sodankyla, Finland; Harsh exposure; testing freeze-thaw without deicing salt It was intended that the results from the field exposure would be used to improve upon the scientific understanding of HPC durability. The results would hopefully indicate which types of concrete could better withstand environmental attacks, such as frost or salt. Many of the samples at these field stations will continue to be monitored in the future, beyond the duration of this project. The main conclusions drawn after the field exposure testing are: - Nearly all of the 22 concretes in each region performed extremely well and did not show much damage. - Samples exposed to real-time field-testing did not have as severe of damage as accelerated laboratory tested samples. - The two-year (three winters) test period was not long enough to provide enough winter exposure for extensive correlations to laboratory tests or for service-life estimates. - The only location where a significant difference in freeze-thaw performance between the 22 concretes could be detected was at Field Station 9 (Finland). In this climate two mixtures showed severe damage after just 1 winter of freeze-thaw cycles. - In the case where there was damage (Field Station 9), there was a good correlation between the field and laboratory tests. - Chloride testing showed the field results (Field Station 8) were comparable with the results of chloride diffusivity measured in the laboratory using an accelerated method. - In the German region (Field Station 1), additional lab testing after field exposure showed that good correlation could be found by measuring the chloride content in different depths compared to the values found by accelerated lab testing. Compressive strength results showed on the other side high variations and could therefore not help to detect the start of deterioration. Additional frost and de-icing salt testing showed comparable results as found by testing of specimens at later ages. Here, the influence of ageing altered the frost resistances in some cases to a great extend. The exposure time and environment of these real-time tests was taken into account when modelling the life expectancy of the concretes. A wider range of concrete behaviour was expected from the 22 different mixtures in each of the 3 regions to provide more insights for modelling of HPC deterioration. The data obtained from the field stations was used in combination with field-cored structures (see Deliverable D2) and laboratory tests (see Deliverable D6) for modelling within Workpackage 4 (Deliverables D8 through D10). The findings from the field station results will be shared with the engineering community, to provide them with information for improved decision in future design of buildings and structures. Designers and engineers will have better tools for specifying concrete for high performance in the various environments across Europe and worldwide.

This result includes the information obtained when measuring shrinkage of laboratory produced mixtures. Three batches of 22 mixtures each were produced in the different European regions and tested in various labs. In the Nordic group, length change of 10 x 10 x 50 cm (length) samples was measured and was reported as a percentage of length change. Drying shrinkage measurements began after 1 month of water curing, followed by drying at 20C degrees and 40% RH. In Germany, the drying shrinkage was measured on beams 10 x 10 x 40 cm (length) at 20C degrees and 65% RH. In the Nordic group the autogenous shrinkage was measured from day 1, at the time of demoulding and while samples were stored at 100% RH (water soaked). Autogenous shrinkage was only measured on mixtures with the lower w/c ratios (< 0.40), since it has been shown that only in these cases is there a lack of water for hydration and thus autogenous deformation. In the German region during the first 24 hours up to time of demoulding the early age shrinkage was measured. After measuring the early age shrinkage the same specimens were used for measuring the autogenous shrinkage. The drying shrinkage results of the HPC mixtures ranged from 0.1 to 0.4 mm/m after 1 year of drying. This range was below the cracking risk level and lower than typically seen on OPC mixtures. By decreasing the w/b ratio or the paste content, the drying shrinkage was reduced while at the same time the autogenous shrinkage increased. Adding more mineral additives, such as fly ash, resulted in a higher paste content and thus greater drying shrinkage. The same trend was seen with additions of air entrainment or other pozzolans, such as silica fume or blast furnace slag. The autogenous shrinkage results of the HPC mixtures under sealed conditions ranged from 0.2 to 0.4 mm/m after 1 year. The magnitude of these results was equivalent to the drying shrinkage, indicating that autogenous shrinkage is of greater risk with HPC. The autogenous shrinkage was much higher than typically seen for OPC mixtures, which is as expected since the autogenous shrinkage only exists below w/b of approximately 0.45. In the tests, HPC mixtures with greatest autogenous shrinkage had low w/b ratios and dense microstructures (such as the silica fume mixtures). The autogenous shrinkage results for water-cured specimens ranged from -0.1 (swelling) to 0.1 mm/m, indicating that even though these small specimens were very dense they could still absorb water and lower the autogenous shrinkage cracking risk. When designing HPC mixtures, it is important to keep the w/b ratio and paste content to relatively low values to prevent excessive drying shrinkage. But these parameters should not be lowered too much (i.e. less than w/b of 0.40) or else the autogenous shrinkage can become significant. Both autogenous and drying shrinkage can be minimized by providing sufficient water curing immediately after placement, which is also critical during the early ages on a construction site to prevent plastic shrinkage cracking.

The test results from WP3 and the results of observations on in-service structures (WP2) were compared, enabling partners to project and assess the behaviour of the different HPC mixtures in the field. A correlation of these results with the laboratory studies was performed, leading to a summary of the durability behaviour of HPC mixtures against different exposure classes. On the basis of this summary, degradation models were created and/or analysed. Differences between the HPC mixtures exposed to different climatic conditions were expected which supported the creation of models for describing the degradation processes. Deterioration models included different damage mechanisms in dependence of the HPC mixtures and the exposure classes of EN 206, as follows: - Freeze-thaw attack (-20°C) - Freeze-thaw attack (-40°C) - De-icing agent attack - Acid attack - Sulfate/seawater attack - Cyclic temperature attack These models can understand the degradation mechanisms of different HPC design concepts with the results being presented by figures and diagrams. The estimation of mixtures with an improved resistance against different climatic conditions leads to the achievement of recommendations for the production of HPC with fixed limits for constituents and acceptance criteria due to exposure classes according to EN 206. These recommendations helped form the basis for lifetime prediction of HPC with respect to durability. From the modelling results, guidelines were also provided regarding service life evaluation, classification to EN206 service class, and finally recommendations to industry for application of HPC.

Recherche de données OpenAIRE...

Résultats exploitables

Partager cette page

Télécharger