Determination of the effect levels of variables affecting the fiber distribution in steel fiber reinforced concrete

The variation in daily needs and sustainable production affect the desired concrete properties. The concrete industry is constantly improving itself for this purpose. Fiber concrete is a type of concrete that has been used extensively in the construction industry in recent years. For this reason, the predictability of fibrous concrete properties becomes important. This study, it was aimed to determine the effect levels of the variables affecting the fiber distribution in fibrous concretes on the fiber distribution. In the study, fiber volume ratio (%1, %2. and %3), mixing time(5 dk., 8.5 dk., and 12 dk.), and fiber length(35 mm, 50 mm, and 60 mm) were selected as the effect variables, and then fiber distribution was determined as the response variable. In order to determine the fiber distribution, a total of 45 square prism beam samples of 150150600 mm dimensions were produced at 15 trial points (concrete class C30/37 was chosen). Each specimen was divided into three cross-sections. The fiber distributions in the sections were calculated by dividing the fiber number determined by the eye and digital image processing (DIP) method by the total fiber number in the sample. The effect levels of the selected variables on the fiber distribution were determined by ANOVA. Moreover, the variations of the ultrasonic pulse velocity and bending strength according to fiber distribution was investigated. In addition, the effect of the direction of concrete casting into the molds (bottom area of the troweled surface) on the fiber distribution in the cross-section was also determined. The highest fiber distribution of the troweled surface lower region is 61.12% in the third cross-section and 55.92 in the first cross-section. Therefore, it can be said that for the concretes produced within the scope of the study, the direction of concrete casting does not affect the homogeneity of the fiber distribution in the cross-section. Only the fiber length has a significant effect level (p-value < 0.0001) on the fiber distributions. The R2 values of the models for the first and third cross-sectional surface fiber distribution were obtained as 0.9055 and 0.9604, respectively. The high R2 values indicate that the prediction accuracy of the models is high. The difference between the fiber distribution ratios obtained by the visual counting and the DIP method is a maximum of 0.58. Keywords: Fiber Concrete, Fiber Distribution, Digital Image Processing, Ultrasonic Pulse Velocity, Bending Strength