INFLUENCE OF DISTRIBUTION TYPE ON THE PROBABILITY OF ACCEPTANCE OF CONCRETE STRENGTH

摘要

The subject of this paper are criteria and methods of evaluations of ready-mixed concrete. These methods may influence the rational decisions about the quality of concrete and may be favorable to the producer or contractor. The Monte Carlo simulation method has been applied to obtain operating-characteristic curves (OC-curves) for different compound conformity criteria and different distribution types of concrete strength. The distribution type of concrete strength significantly influences the risk that the suitable lot of concrete can be rejected (producer's risk) as well as the risk that the defective lot of concrete can be accepted (contractor's risk), especially for small sample sizes. The most favorable for a producer and unfavorable for a contractor is lognormal distribution. The normal distribution protects mostly the contractor's profit, whereas the Weibull distribution gives balanced results. The analyses presented in this paper show that the compound criteria in cases of small sample sizes n < 15 can have unfavorable effects to the both producer and contractor, for example groundless increase of production costs and high level of contractor's risk.