The study of the conformity criterion for compressive strength of concrete based on order statistics

Abstract

The test of concrete compressive strength conformity with current regulations PN EN 206-1:2003 is the sampling inspection based on a numerical evaluation. Satisfying the compound criteria, including the adoption of statistical quality control plan, confirms the conformity of the examined batch of concrete defined by the declared class of compressive strength defined by characteristic value of compressive strength. The conformity criteria recommended by EN 206-1 for the initial production are not without flaws and they are critically evaluated by many authors. This paper presents a new criterion of conformity based on order statistics. A preliminary evaluation of the criterion was made for the series with a small number of the test results with the use of probability of acceptance determined by means of the Monte Carlo method with the assumed 5% fraction defective. The analysis of the results showed that the presented criterion does not depend on the dispersion of results whereas the probability of acceptance is maintained at a constant level approached to the appropriate one at the stage of continuous production.

Proposition of a New Conformity Criterion for the Assessment of the Concrete Compressive Strength

“…with symbols according EN 206). As it was shown in [8], probability of acceptance in accordance with compound criteria (Method A) is increasing in the case of using test results, obtained on cube specimens, than cylinders. This is particularly noticeable for low strength concrete classes.…”

“…With increasing of the individual test results number, obtained during an assessment period, probability of acceptance P a is decreased. As it was shown in [7,8], for the fraction of defectives θ = 5% (fraction of the test results below specified characteristic concrete strength, ck f ) and standard deviation σ = 3 N/mm 2 , probability of acceptance in accordance with Method A EN 206:2013 (for group of test results n = 3) is equal to 0.706. However for the concrete the same quality level and overall number of tests results n = 6, probability of acceptance P a = 0.498 (for case of two non-overlapping group) and only Pa = 0.410 (for case of four overlapping group n = 3 from n = 6 results).…”

“…How it was pointed in [5], "the basic assumptions and guidelines which are mentioned in EN 206 are often vague and sometimes based on arbitrary suggestions without a coherent probabilistic background". Some basic aspects of conformity criteria in EN 206 were discussed in detail in publication [6,7] The main problem of the conformity assessment plan according EN 206 for stage of initial production is related with uncertainties of the statistical estimation based on a small group of the individual test result (n = 3) and limited prior information about statistical parameters of the whole population. Based on the results of their own researches [8] and practical application, three major paradoxes that appear in the case of practical utilizing of compound conformity criteria for initial production (Method A, EN 206), should be stated.…”

“…Method of statistical quality control of the construction materials and products based on the use of order statistics as estimators of qualities, recommended in ISO 12491:1997 [10] has not found wide practical application in the development of compliance (conformity) criteria in case of concrete strength assessment. However, own studies [7,8] has shown that the use of order statistics as estimators of low-order quantities can be effectively used in the case of limited (small) group of the test results. A new criterion is based on the known method of interval estimation of quantities (confidence interval quantities) using the properties of non-parametric statistics [11].…”

Economical aspects concerning quality control of concrete

“…The compound criteria for initial production and number of samples n = 3 can have unfavorable effects to both -producer and customer. The most important defects of the criteria are following [8,9,10,11]:…”