Tag Archives: Test method

Test method with electrolyte solution representative of actual service DIBt test

Various failures of prestressed concrete structures caused by hydrogen-induced stress corrosion cracking at the beginning of the 1980s led to the development of a constant load test based on laboratory as well as on-site investigations.20,60 In  this test, the prestressing steel specimens as delivered were stressed in a stressing frame under constant strain to 0.8 Rm, at 50ëC. Based on the results from chemical analyses of on-site samples of water taken from prestressing ducts, the electrolyte solution for the test contained 0.014 mol/l chloride, 0.052 mol/l sulfate and 0.017 mol/l thiocyanate (pH ˆ 7.0). The SCNÿ ions promote hydrogen uptake and therefore made conditions more severe than under normal conditions. As noted previously, thiocyanate ions have been found in some cases where damage had occurred,6 and were shown to have originated from concrete admixtures (in which SCN± is no longer permitted).

During a test period of 2000 h, a measurable amount of hydrogen was absorbed by the specimens and led to embrittlement. Specimens made of susceptible materials failed before 2000 h was reached. Compared to tests in concentrated thiocyanate or nitrate solutions, this test has the advantage that the decisive criterion is the occurrence or non-occurrence of a fracture within a certain time period, rather than within a relatively short failure time. Further advantages arise from the fact that the test solution closely simulates corrosion conditions around the prestressing steels within the ducts under construction conditions, and that the results obtained with this test for different prestressing steels are reproducible and correspond to practical experience.

In comparison with the FIP test, slightly higher initial and then nearly constant low hydrogen activity at the prestressing steel surface is required to ensure reproducible results. This has been proved by permeation measure- ments.20,60,61 The higher the pH of the solution, the lower the general corrosion rate. The more favourable conditions for surface layer formation result in a longer testing time. However, with this test the results can be better extrapolated to the behaviour of prestressing steels in practice. Experiments using what is now known as the DIBt test have been carried out successfully for a wide range of steels.

This method has been described by Grimme et al.20 It was found (Fig. 6.15) that all approved prestressing steels on the market passed the time limit of 2000 h without fracture, while steels known to be susceptible to SCC, from cases of reported damage or known not to be suitable for construction practice (St 1080/ 1320, bainitic microstructure3), failed within 2000 h due to hydrogen-induced SCC. A criterion based on time to failure could thus be defined that appeared valid for all types of prestressing steels tested.

Applying the DIBt test to prestressing steels that had failed due to SCC after long-term use in actual structures, showed that this test would have classified them as susceptible and hence not approved, supporting the suitability of this method. Results23,64 showed that annealed prestressing steels of the `old type’ (i.e. without Cr-alloying and manufactured prior to 1965) and also comparable steels from the former GDR do not pass the test and should be classified as `not approvable’ according to today’s standards.

Test method with electrolyte solution not representative of actual service FIP test

The demand for a test to assess the susceptibility of prestressing steels to stress corrosion cracking dates back to when prestressing steels were introduced. Federation International de la Precontrainte published the first reports,58,59 which described the evaluation of SCC for various prestressing steels. The data were obtained from research and round-robin tests with different electrolyte solutions that did not necessarily reproduce actual conditions, with and without electrochemical control (polarisation) and at prestressing levels between 0.5 and 0.95 Rm.

The following electrolyte solutions were evaluated:

· boiling Ca(NO3)2 solution
· saturated aqueous H2S solution with and without electrochemical control
· 20% NH4SCN solution with and without electrochemical control
· 5M H2SO4 with cathodic polarisation
· distilled water with cathodic polarisation and
· saturated Ca(OH)2 solution with anodic and cathodic polarisation.

As expected, the corrosion behaviour differed considerably. The uncertainty in hydrogen activity during the test contributed to a large scatter of time-to-failure values, so the stress corrosion cracking behaviour could not be assessed accurately or ranked. Evaluating the different methods, it was concluded that only the tests in the concentrated ammonium thiocyanate solution were promising. This method, known as the FIP test,36 used the following test conditions:

Corrosive medium: 200 g NH4SCN in 800 g H2O temperature: 50ëC  1ëC, preferably  0.2ëC

· Test cell: controlled heating, no circulation of the corrosive medium, plastic materials recommended
· Specimen: specimen is to be cleaned only with acetone (in view of health and safety concerns, trichlorethylene is presumably no longer permitted)
· Loading: constant load of 80% of the actual ultimate tensile strength, obtained on a reference, kept constant during the experiment
· Result: time to fracture of the specimen will be determined under these test conditions. Because of the relatively large spread of the results (which is normal with corrosion tests), it is recommended that 3 to 12 specimens are tested to obtain a representative value, evaluated by Gaussian distribution on a logarithmic time scale. A test is stopped if the specimen does not break within 500 hours.

More details can be obtained from the literature.36 The test conditions, under which specimens are at open circuit potential, demonstrate that a variety of parameters can overlap in an uncontrolled manner during the test. Values for the time to failure were therefore spread widely and the only classification possible was into prestressing steel classes. In fact, no results relevant to construction and service conditions could be achieved. Tables 6.6 and 6.736 show the effect of specimen characteristics and test conditions on the time to failure, but a quantitative correlation of parameters influencing time to failure is not possible. Lifetimes relevant to construction and service cannot be estimated from these results. The type of electrolyte does not allow these results to be extrapolated to behaviour in practice.

Thresholds for the time to failure can only be defined for single classes of prestressing steels. They depend on diameter, microstructure and type of delivery. The FIP test provides insufficient information to distinguish the particular susceptibilities of different types of prestressing steel. Nevertheless, for all types within one category (hot-rolled, quenched and tempered or cold- drawn) the susceptibility increases with increasing strength and higher prestressing level.22 To date, no reliable values have been proposed. In the draft for EN-standardisation of this test (prEN 10138-3 Prestressing steels ± Part 3: strand), the proposed time to failure is so short that almost any material would pass the test.

NuÈrnberger6 has used the FIP test to evaluate susceptibility to hydrogen- induced SCC, and determined that prestressing steels can be classified as extremely susceptible if time-to-failure remains below the following values:

· cold-drawn wire 2±3 h
· quenched and tempered wire 10±45 h
· hot-rolled bar 30±50 h.

Again it must be emphasised that FIP tests are not suitable for comparing different types of prestressing steel (hot-rolled, cold-drawn, quenched and tem- pered). Therefore, results must not be used as a ranking scale for susceptibility. More recent experiments59 show that the ranking achieved with FIP tests can also be obtained if hydrogen and mechanical loads are not applied at the outset. Susceptibility of steel can obviously be determined more easily from deforma- tion values (elongation after fracture, reduction of area after fracture) under slow-strain rate tests directly after hydrogen charging.