by clicking the arrows at the side of the page, or by using the toolbar.
by clicking anywhere on the page.
by dragging the page around when zoomed in.
by clicking anywhere on the page when zoomed in.
web sites or send emails by clicking on hyperlinks.
Email this page to a friend
Search this issue
Index - jump to page or section
Archive - view past issues
Concrete In Australia : June 2008
TECHNICAL Table 5. Statistical data for “TEL immersion diffusion test” from 4 concretes. Type GB cement is 65% BFS. Status Trial 1 Trial 2 Binder (kg/m3 ) 375 GP + 30 CSF 0.32 425 GB Trial 3 390 GB + 30 CSF Delete 1 outlier Production Quaternary blend 0.38 0.38 w/c Samples D (x10-12 8 8 8 7 ~0.36 9 preparation, but refer to other documents, which may have kept pace better with durability testing requirements. This issue will be discussed further later in this paper. Compressive strength as a statistical benchmark One of the most important parameters in concrete is the characteristic compressive strength, f´c . It is not only used in specifi cation of concrete for purchase as discussed above, but it also forms the basis of structural design in accordance with the relevant standards. Based on normal distribution statistics, f´c is a 95% passing value, and the actual target mean strength of the concrete is equal to f´c +1.65s, where s is the standard deviation of the test results. For a given batch plant and grade of concrete, monitoring of s indicates the overall precision of the production process, affecting both mix design in terms of how close the target mean strength will be to f´c , and costs in terms of cement usage to ensure that the requirements of AS 1379 are met. Since compressive strength monitoring has a long track history and is routinely performed for all grades of concrete regardless of other parameters that may be specifi ed for a given project, it is sensible to use these statistics as a benchmark for assessment of the discrimination ability of durability related tests for assessment of compliance. Typically, modern computer controlled batch plants will achieve coeffi cients of variation (CoVs) in the order of 4-6% in strength values taken over a signifi cant number of production batches. In assessment of potential durability tests, this level of consistency may be used as a statistical benchmark, but may not be achievable due to the inherently greater complexity of many potential durability related test methods. Chloride diffusion testing As indicated in Table 4, the Nordtest NT Build 443 appears currently to be the default method used in Australia for determination of chloride diffusion coeffi cients at early age. Clause 6.6 of the standard provides typical coefficients of variation for the estimated surface chloride concentration (CS of 15%. It is not stated whether these values reflect single batch variation (repeatability), or include batch variations (reproducibility). In 1993 the Perth laboratory of Taywood Engineering Limited undertook some repeatability testing of the “TEL Immersion Diffusion Test”, the mechanics of which test closely parallel NT Build 443. The statistical results for the diffusion coeffi cients obtained are presented in Table 5. Standard ) of 20%, and the estimated diffusion coefficient (De 54 Concrete in Australia Vol 34 No 2 ) ) m2 /s 4.0-4.7 (4.4) 0.4-0.6 (0.53) 0.43-0.80 (0.51) 0.43-0.58 (0.46) 0.40-0.90 (0.63) (x10-12 ) m2 0.24 0.08 0.12 0.05 0.17 /s CoV (%) 5.4 15.7 23.5 10.7 27.4 test cylinders were prepared by an east coast laboratory in accordance with AS 1012 and shipped to TEL for testing. Each of the “trial” sets are understood to be from single batches of laboratory prepared concrete, the “production” data set represents single cylinders from nine production batches of concrete, so the CoV includes a contribution from the innate variability of the batch plant involved. All four data sets were produced using a common operator and equipment for the profi le grinding; and a common operator, equipment and analytical method for analysis and calculation. The data presented in Table 5 raises several interesting points for further discussion and consideration: • Under favourable conditions, a CoV of under 6% can be obtained (viz Trial 1), which is comparable with the benchmark of compressive strength, and contrary to the common perception of such tests as lacking repeatability. • There is a possibility that the CoV increases as the value of D decreases, implying that the CoV has high dependency on parameters in the testing process that are not related to the magnitude of the analytical quanta (c.f. Trials 1 v 2/3). • The CoV is also likely to be sensitive to consistency of sample preparation (c.f. Trials 2 v 3). The mean CoV for the three Trials is 14.9%, comparable with 15% published in NT Build 443, reducing to 10.7% if a single outlier value is deleted. The slightly better reproducibility is attributed to use of constant weight pre-saturation. The CoV for production concrete is comparable in magnitude with the mean CoV plus the batchwise variation for the batch plant as demonstrated by typical CoV values in compressive strength testing. This suggests that both the inherent variability in the test method and the inherent variability in the production plant should be accounted for in establishing production compliance criteria for inclusion in specifications. ASTM C1202, the rapid chloride permeability test (RCPT) First published in February 1992 as C1202-91, the current issue of this ASTM test method is C1202-05. Other versions were published in 1994 and 1997. As indicated in Table 6, this method is an ASTM adaptation of an earlier method developed under the Strategic Highways Research Program as AASHTO T277. A summary of the precision data quoted in the various versions is presented in Table 6, including the maximum allowable differences between results. The information collected in Table 6 is somewhat