Alasdair’s Engineering Pages © A. N. Beal 2024 www.anbeal.co.uk
www.anbeal.co.uk
The Structural Engineer 87 (10)19 May 2009
Viewpoint
Concrete strength testing -
Alasdair Beal proffers his views on concrete strength testing and the Eurocodes
Paul Toplis’s ‘Viewpoint’ [1] is timely in drawing attention to the problems specifications for concrete strength testing can create for contractors and suppliers. It is true that some of these are caused by inexperienced engineers (‘designer/specifiers’ in Eurospeak) specifying unrealistic sampling rates and misunderstanding BS EN 206/ BS 8500 compliance rules. However the underlying problems have been created not by these engineers but by code writers and the rules they have written for the engineers.
Thus I agree that a concrete slab specification which calls for a sampling rate of 1/2 and two 28-
Given that a single cube test to measure the strength of a batch of concrete has been regarded as acceptable for a very long time, it is not clear why the BS EN 206 committee decided that we now need double this number of tests. However until BS EN 206 is changed, engineers are stuck with it -
Thus the problem of engineers specifying excessive numbers of cube tests could be solved by a simple amendment to BS EN 206. However Mr Toplis’s comment that results from the BS EN 206/ BS 8500 compliance rules ‘may look a little counter-
BS 8500 requires tests to be assessed in non-
In the table below, the BS EN 206/ BS 8500 rules are applied to a set of 18 results (specified characteristic cube strength 40N/mm²). The table also shows how changing the starting point in the data can affect the analysis.
The rows all contain the same set of test results but each starts from a different point in the sequence. As can be seen, in five cases all of the concrete would pass but in one case two groups of results would fail, leading to rejection of two thirds of the concrete. In the latter case, a batch of concrete with a strength of 37N/mm² would be accepted, yet batches with strengths as high as 46 and 48N/mm² would be rejected.
Thus the concrete supplier has a problem: depending the order he delivers the batches, he may find that the concrete is all accepted, or else two thirds of it may be rejected, even though he has supplied identical concrete in every case.
The customer has a problem too. Instead of no more than 5% of test results failing, 22% (4 out of 18) are below the specified characteristic strength. The characteristic strength calculated from the mean and standard deviation of the results is only 35.5N/mm², compared with 40N/mm² specified. The concrete has well below the specified characteristic strength, yet in five scenarios out of six the test results would have passed the BS EN 206 compliance criteria.
In this case, the BS EN 206 compliance rules mean that:
(i) the supplier faces uncertainty, with acceptance or rejection of the concrete depending on what order the batches are delivered in;
(ii) the engineer is forced to make decisions which appear completely nonsensical: he may end up accepting a batch of concrete which has less than the specified strength, while at the same time other batches which are well above the specified value are rejected;
(iii) although the concrete characteristic strength is only 89% of the specified value, it has a 5/6 chance of being accepted.
BS EN 206 states that its criteria give only a 1% chance of conforming concrete being rejected, which may well be true. However no figures are given for the risk of substandard concrete being accepted, which appears to be much higher. Concrete which passes the BS EN 206 test criteria may not comply with the specified characteristic strength.
This type of problems with strength test compliance rules is not new. BS EN 206 brings back ideas which were fashionable in the UK in the 1960s and 1970s, when CP 114 and CP 110 included similar statistically-
The CP 114 and CP110 compliance rules were replaced by BS 5328, which introduced simpler rules: the mean of four results should be at least 3N/mm² above the characteristic strength and individual results should be no more than 3N/mm² below it. However in practice the ‘mean of 4’ criterion was normally treated as a ‘warning light’ and only individual batches which fell below the ‘-
The current BS EN 206 rules are impractical, illogical, give unpredictable results and put too much concrete at risk on single decision. They also fail to ensure that the concrete supplied actually complies with the specified ‘characteristic strength’. (The inadequacy of the compliance rules undermines the calculations of partial factors and failure probabilities in Eurocode 2.)
The best way of checking concrete strength is also the simplest: a straightforward ‘minimum strength’ requirement for test results is logical, practical and avoids all the problems associated with criteria based on groups of tests. However if this idea would be a bit too radical for the codewriters, they could reduce the problems in the present BS EN 206 rules to a tolerable level by amending them as follows:
* balance risks more fairly between suppliers and customers by changing the criteria from ‘group mean +2N/mm²/ individual batch -
* reduce the volume of concrete at risk on a decision by changing from groups of six to groups of four and treat the ‘group mean’ limit as a warning rather grounds for rejecting the entire volume of concrete;
* when a test falls below the limit on individual results, adjacent batches should be investigated but only rejected if they fall below the minimum strength;
* encourage the use of 7 day cube tests to give early warning of quality problems and allow early action to minimise consequences.
References
1. Toplis, R: ‘Concrete compressive strength testing -
2. Beal, A.N.: ‘Concrete Cube Strengths -
The original copy of this paper is available from
No. |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
11 |
12 |
13 |
14 |
15 |
16 |
17 |
18 |
Strength |
41 |
52 |
37 |
49 |
47 |
51 |
53 |
36 |
37 |
42 |
44 |
48 |
43 |
41 |
40 |
46 |
44 |
39 |
Mean/pass |
|
[ |
46 |
pa |
ss |
] |
[ |
43 |
pa |
ss |
] |
|
[ |
42 |
pa |
ss |
] |
|
Strength |
52 |
37 |
49 |
47 |
51 |
53 |
36 |
37 |
42 |
44 |
48 |
43 |
41 |
40 |
46 |
44 |
39 |
41 |
Mean/pass |
|
[ |
48 |
pa |
ss |
] |
|
[ |
42 |
fail |
|
] |
|
[ |
42 |
fail |
] |
|
Strength |
37 |
49 |
47 |
51 |
53 |
36 |
37 |
42 |
44 |
48 |
43 |
41 |
40 |
46 |
44 |
39 |
41 |
52 |
Mean/pass |
|
[ |
45 |
pa |
ss |
] |
|
[ |
42 |
pa |
ss |
] |
|
[ |
44 |
pa |
ss |
] |
Strength |
49 |
47 |
51 |
53 |
36 |
37 |
42 |
44 |
48 |
43 |
41 |
40 |
46 |
44 |
39 |
41 |
52 |
37 |
Mean/pass |
|
[ |
45 |
pa |
ss |
] |
|
[ |
43 |
pa |
ss |
] |
|
[ |
43 |
pa |
ss |
] |
Strength |
47 |
51 |
53 |
36 |
37 |
42 |
44 |
48 |
43 |
41 |
40 |
46 |
44 |
39 |
41 |
52 |
37 |
49 |
Mean/pass |
|
[ |
44 |
pa |
ss |
] |
|
[ |
44 |
pa |
ss |
] |
|
[ |
44 |
pa |
ss |
] |
Strength |
51 |
53 |
36 |
37 |
42 |
44 |
48 |
43 |
41 |
40 |
46 |
44 |
39 |
41 |
52 |
37 |
49 |
47 |
Mean/pass |
|
[ |
44 |
pa |
ss |
} |
|
[ |
44 |
pa |
ss |
] |
|
[ |
44 |
pa |
ss |
] |
Table 1: BS EN 206/ BS 8500 rules are applied to a set of 18 results