Issue 78

M. S ł owik et alii, Fracture and Structural Integrity, 78 (2026) 393-408; DOI: 10.3221/IGF-ESIS.78.23

0.73 f  



(6)

f

0.78

ct

ct

,15/30

,15/15

in which the constant term is in MPa, f ct,15/15 – splitting tensile strength of concrete for 150 mm long cylinders in MPa, f ct,15/30 – splitting tensile strength of concrete for 300 mm long cylinders in MPa.

The proposed regression equations are empirical conversion functions established from the experimental data to provide the best agreement within the investigated range of splitting tensile strengths. Therefore, the intercepts result from the statistical regression analysis and should not be interpreted as having direct mechanical significance or extrapolated beyond the investigated range. To assess the fit between the proposed models and the experimental results, the Mean Absolute Percentage Error (MAPE) was determined using Eqn. 7:

Y Y 

T

1

 

i

ip



MAPE

(7)

T n

Y

i T n  

i

where: T – calculation and forecast periods total number, n – forecast periods number, Y i –the actual value of the variable in period i, Y ip – the predicted value of the variable in period i. The MAPE between the generated function and the test data was: • 1.8% for concrete with gravel aggregate, • 2.6% for concrete with granite aggregate.

The obtained values of MAPE did not exceed 5%, and it was concluded that the generated functions were well-matched to the laboratory results. Consequently, the regression models (Eqns. 5 and 6) were applied to transform the results obtained from short specimens into equivalent values for standard-length specimens. The re-calculated values of the tensile splitting strength for short specimens were then taken together with the test results for splitting tensile strength for long specimens in the final analysis of the experimental results, which focused on assessing the effect of cement matrix strength and the type of aggregate on concrete’s splitting tensile strength.

T HE ANALYSIS OF TEST RESULTS

W

The impact of aggregate properties on the splitting tensile strength of concrete

hen analysing the influence of aggregate type and cement matrix strength on the splitting tensile strength of concrete, the experimental results obtained on standard cylinders, together with the corresponding values normalised from short cylinders (as it was presented earlier in “The conversion of test results – from the length of 150 mm to 300 mm”), were considered within one unified dataset (f ct,sp ). The strength of cement mortar is a basic parameter that influences the strength properties of hardened concrete, and in concrete mixture design, it is expressed as the water-cement ratio (w/c). It was observed that higher tensile splitting strength was noticed for concretes with w/c = 0.38 (stronger concretes) compared to concretes with w/c = 0.55 (weaker concretes). Therefore, in order to isolate and evaluate the effect of aggregate type, the test results for the two w/c ratios (0.38 and 0.55) were analysed independently. Four statistical samples were built: • for weaker concrete w/c = 0.55 with gravel aggregate, • for weaker concrete w/c = 0.55 with granite aggregate • for stronger concrete w/c = 0.38 with gravel aggregate, • for stronger concrete w/c = 0.38 with granite aggregate, and then examined according to the aggregate type. The comparison of line plots and regression functions is shown in Fig. 7 (for weaker concrete, Fig. 7.a., and for stronger concrete, Fig. 7.b). Concrete containing crushed granite aggregate exhibited higher splitting tensile strength than concrete made with natural gravel. However, this increase was less pronounced for concretes with lower strength in comparison to those with higher strength. To enable a quantitative comparison of the results, the Mean Percentage Error (MPE) was evaluated based on Eqn. 8.

401