محمد كاشاني نوين

Ground Penetrating Radar (GPR) has found various applications in civil engineering today, especially in investigating various parameters of cohesive an‎d non-cohesive layers of pavement, due to its many advantages such as non-destructiveness, high-speed data acquisition, providing a large volume of information, providing subsurface information, high safety, an‎d high accuracy. On the other han‎d, one of the important pavement design parameters in road construction an‎d maintenance projects is the modulus of resilient of the base layer, an‎d its measurement methods have many limitations, such as high cost, destructiveness, spot-specificity, an‎d the inability to test on expressways. Therefore, a method that can provide comprehensive information about the pavement, including the modulus of resilient of the base layer, at low cost, non-destructively an‎d continuously along the entire road, in a relatively short period of time, can, in addition to reducing the costs of studies an‎d their verification, assist experts in pavement design an‎d choosing the correct repair an‎d maintenance option. In this study, by first examining field data collected from the Qazvin-Saveh Road, a significant relationship was detected between dielectric parameters as the output of the GPR device an‎d the modulus of resilient. Then, by making laboratory samples with the base layer granulation an‎d changing the parameters of compaction, moisture, an‎d clay content, as variables that affect both the dielectric an‎d the resilience modulus, the resulting changes on the dielectric an‎d resilience modulus variables of the base layer were tested. An examination of the linear an‎d nonlinear correlation of the variables showed that the clay content variable in the range of 0 to 5% had very little effect on the dielectric an‎d resilience modulus an‎d was removed from the set of modeling variables. It was also shown that the modulus of elasticity in the range of 175 to 420 MPa had the highest correlation with the density variable (in the range of 78 to 100 percent) an‎d dielectric (in the range of 6.2 to 14.4), an‎d the humidity variable, despite a significant correlation, had a lower correlation with the modulus of elasticity in the range of 2.5 to 7.5 percent. Then, two linear regression models were developed to predict the modulus of elasticity, the first with the compaction an‎d moisture variables an‎d the second with the dielectric variable, an‎d both models were examined in terms of correlation statistics an‎d residual error. The correlation coefficient R2 was 0.893 in the first model an‎d 0.719 in the second model, which indicates that the models predict the resilience modulus with appropriate accuracy. Also, the adjusted correlation coefficient of 0.889 for the first model an‎d 0.714 for the second model an‎d their closeness to the R2 value indicate that the models performed optimally in terms of variable selec‎tion an‎d did not suffer from overfitting. The stan‎dard error of estimation in the first model was 20.7 an‎d in the second model was 33, which shows that these values are relatively appropriate errors compared to the scale of the data. Also, the calculation of F, VIF an‎d Tolerance statistics showed that the models were statistically highly significant an‎d there was no multicollinearity in the models. Analysis of the residual statistics also showed that the models did not have systematic errors an‎d biases.

دي الكتريك , رادار نفوذي زمين , مدول برجهندگي

GPR , Ground Penetrating Layer , Base Layer , Resilience Modulus , Dielectric

Mohammad Kashani Novin

Mahmoud Ameri