The experimental and statistical analyses of soil stabilization with deep soil mixing (DSM)

Within the scope of the thesis study, soil improvement methods are mentioned and deep mixing method is discussed with an application. The data obtained from core samples taken from deep mixing columns were evaluated by statistical analysis. In the first part of the study, the significance relationship between the dependent variable "γ" and the independent variables "depth, length, diameter, weight and volume" was examined using the SPSS program. Normality test, Kolmogorov-Smirnov, Q-Q graph, kurtosis/skewness values were compared with the SPSS software. According to the normality test results, it was decided to use one-way ANOVA from parametric tests. In addition, regression equation for W (g)- γ (g/cm3) was calculated and estimated γ (g/cm3) was obtained by Regression and Correlation Analyses. The performance of the estimated γ was measured by calculating the Mean Absolute Relative Error (MAPE) value. Thus, the degree of accuracy of the proposed regression equation was determined. Strong-weak numerical relationships were determined and correlation graphs were created by looking at the range of the Coefficient of Determination (R2) calculated as a result of the regression analysis between the independent variables "depth, height, diameter, weight and volume" and the dependent variable "γ". In the second part of the study, uniaxial compression test was performed and the fracture load data obtained from the results were compared and evaluated by statistical analysis. As in the first part, Kolmogorov-Smirnov, Q-Q graph, kurtosis/skewness values were evaluated comparatively and the distribution of the data was determined and then the significance relationship was examined with one-way ANOVA and Homogeneity test. Keywords: Soil Improvement Methods, Deep Mixing Method, Uniaxial Compression Test, Statistical Analysis, Quality Control Tests