Evaluation of void beneath the airport pavement slab corner based on the strain monitoring

ABSTRACT

Void beneath the slab corner as a significant problem of airport pavement maintenance is always paid more attention by airport operators. The objective of this paper is to establish neural network models to predict the void sizes and load position based on the monitored strain values. A finite element model was established and validated by the data collected from a full-scale test pavement. Pavement strain values of 3168 scenarios considering different pavement structure parameters, load positions and void sizes were calculated. The simulated bottom strains at concrete pavement slabs were analysed for different void sizes and loading conditions. A coefficient and its warning thresholds were proposed to predict the occurrence of voids. Then, two Genetic Algorithm Backpropagation (GA-BP) neural network models were developed to predict void sizes and load position using the numerical simulation results as the training dataset.

KEYWORDS:

• Airport pavement
• Finite element model
• Void beneath the slab corner
• Strain monitoring
• GA-BP neural network

Disclosure statement

No potential conflict of interest was reported by the authors().

Data availability statement

All data, models, and code generated or used during the study appear in the submitted article.

Nomenclature
$K_1$	=	the spring stiffness coefficient, MN/m;
$q$	=	the joint stiffness per unit length, MN/m2;
$\lambda$	=	the joint length, m;
$N_R,N_C$	=	the number of rows and columns of the node on the side of the slab;
$\mathrm{LT}{E}_{\delta }$	=	the joint load transfer coefficient, %;
$A_1$,$A_2$,$x_0$ and $p$	=	parameters of the Logistic function;
$K_x$	=	the horizontal spring stiffness coefficient of x direction, N/m;
$K_y$	=	the horizontal spring stiffness coefficient of y direction, N/m;
$K_z$	=	the vertical spring stiffness coefficient, MN/m;
$E$	=	the elastic modulus of base, MPa;
$S$	=	the contact pressure area, m2;
$h$	=	the contact pressure acting depth, m;
$n$	=	the number of spring nodes;
$R$	=	the length of the right-angled edge of void, m;
E11	=	transverse tensile strain value at the bottom of the pavement slab, μϵ;
x	=	the distance from the left edge of the tire to the left longitudinal joint of the slab, m;
${\epsilon }_{p,max}$	=	the maximum compressive strain value at all sensing position, μϵ;
${\epsilon }_{t,max}$	=	the maximum tensile strain value at all sensing position, μϵ;
Ratio	=	the absolute value of the ratio of maximum compressive strain value to maximum tensile strain value at the five sensing position;
$y_i$	=	the normalised value;
$x_i$	=	a certain value of each evaluation index;
$x_{imax}$	=	the maximum value of each evaluation index;
$x_{imin}$	=	the minimum value of each evaluation index;
$r_{\mathrm{xy}}$	=	the sensitivity coefficient;
$x_i$	=	the value of input variables;
$y_i$	=	the value of output variable;
$\overline{x}$	=	the average value of the input variables;
$\overline{y}$	=	the average value of the output variable;
${\sigma }_x$	=	the standard deviation of the input variables;
${\sigma }_y$	=	the standard deviation of the output variable;
$m$	=	the samples number of the datasets.

Additional information

Funding

This work was supported by National Natural Science Foundation of China [Grant Number 51978163]; Jiangsu Natural Science Foundation [Grant Number BK20200468].