As high‐performance Portland cement (PC), fly ash (FA) and metakaolin (MK) concrete have been developed in wide applications, it has growing interest in optimizing and predicting consistency of fresh PC‐FA‐MK concrete for efficient and practical design and construction. This paper presents statistical models for predicting the consistency of concrete incorporating PC, FA and MK from the experimental results of standard consistency tests. They reflect the effect of variations of pozzolanic replacement materials including FA and MK at graduated replacement levels of up to 40% and 15%, respectively. The predictions produced are compared with the experimental results of consistency of concrete blends. Models show that they can be used to predict the consistency parameters including slump, compacting factor and Vebe time with a good degree of accuracy in a wide range of FA‐MK blends. Design guidelines for evaluating consistency parameters are tentatively recommended along with their confidence intervals for prediction limits at 5% significance levels.
Santrauka
Straipsnyje aprašyti cementbetonio mišinio su lakiaisiais pelenais ir metakaolinu konsistencijos (slankumo, sutankinamumo, Vebe rodiklio) tyrimai. Parenkant betono mišinių sudėtis buvo naudojami lakieji pelenai, kurie pakeisdavo iki 40 % portlandcemenčio ir metakaolinas, kurio buvo dedama iki 15 % cemento masės. Atitinkamai buvo keičiami ir portlandcemenčio kiekiai. Remiantis tyrimų rezultatais, pasiūlyti statistiniai modeliai įvairių sudėčių betono mišinio konsistencijai prognozuoti. Palyginus prognozuojamus ir eksperimentinių tyrimų betono mišinio konsistencijos rodiklius nustatyta, kad jie labai gerai koreliuoja. Todėl pasiūlytus statistinius prognozavimo modelius galima taikyti betonų technologijos praktikoje.
Visualization Strategies for Regression Estimates with Randomization Inference