Concrete pavement prediction life model based on electrical response of concrete-CNTs sensors under fatigue loading

Structural Health Monitoring (SHM) of important infrastructures such as Portland cement concrete pavements plays a key role in pioneer societies, to guarantee the optimum usability and performance of the infrastructure system. As an example, this is imperative to appraise the destruction level in Portland cement concrete pavements along time as to plan their maintenance with proper actions at the right time. Based on this premise, the goal of this study is to extend an extensive feasibility study to set up a novel approach for SHM of Portland cement concrete pavements and remaining life estimation, based on self-sensing concrete-CNTs sensors added with multi-walled carbon nanotubes (CNTs). The concrete-CNTs sensors are applied as piezoresistive sensors (i.e. to collect information such as weight and species of passing vehicles), or as destruction identification sensors (i.e. to detect crack propagation in the Portland cement concrete pavement). For casting the concrete-CNTs sensors, some parameters are important, such as the number of CNTs, species of surfactant, dispersion quality etc. In this research, the dispersion quality of multi-walled carbon nanotubes in the aqueous phase and cement matrix was experiment examined using Ultraviolet-Visible Spectrophotometry (UV-VIS), Scanning Electron Microscopy (SEM) and mechanical experiments for compressive and flexural strength. The outcomes demonstrated that a new specific surfactant composition (sodiumdodecylbenzene sulfonate (SDS) and superplasticizer carboxylate base (SP-C) with the ratio of 1:9 respectively) could disperse MWCNTs around 64% more than other surfactant combinations indicated in the previous studies while this has enough compatibility with the concrete to omit antifoam in the mixing process and maintain the concrete mechanical specifications pretty constant. To appraise the influences of the main parameters affecting concrete-CNTs sensors function, diverse criteria in static and dynamic loading patterns were defined such as sensitivity of the sensor (Se), the standard deviation of the prediction error, repeatability (Re), cross-correlation (CC) and hysteresis (SSE). The dynamic criteria such as sensitivity, internal repeatability, cross-correlation and hysteresis declared that the dispersion energy levels for the dispersion of MWCNTs have a major impact on enhancing the function of the sensor rather than the number of CNTs. The repeatability criterion, contrariwise, showed that the number of MWCNTs has a major impact on the function of the sensor compared to the dispersion quality (dispersion energy level) of MWCNTs. Consequently, both parameters have to be regarded as relevant. The overall outcomes showed that the sensors fabricated with 0.15 wt% CNTs, superplasticizer and SDS as a surfactant using the maximum dispersion energy level (ultrasonic bath for 2 hours and 90 minutes of ultrasonic probing) have the best function in both static and dynamic load mode. To explore the influences of traffic loads on the pavement, concrete-CNTs sensors were experiment examined under various value of dynamic loads. The outcomes demonstrated that the maximum exterior load applied on the sensor (Fmax) is linearly correlated to the maximum response of the sensor (Smax) via a constant coefficient tagα/tagβ, in which tagα is defined as the slope of the Force vs. Time graph and tagβ is defined as the slope of the Sensor’s response vs. Time curve. So, this can be concluded that the application of the indicated concrete-CNTs sensors for piezoresistive applications is feasible and the sensors can appraise the load with of high-goodness of fit (R2adj.> 0.99). In addition, to study the response of the concrete-CNTs sensor under the traffic loading, fatigue experiments were run. An alternative data processing log(G)-log(N) was applied instead of traditional S-log(N) fatigue graphs (Goodman curves), based on the electrical sensor response with a linear regression approach and the outcomes were verified by statistical tests. In this research, the concept of G has been defined for the first time as the slope of a sensor’s response that reflects the destruction created in a pavement because of one pass of vehicle load. Based on these findings, two various types of remaining life models for Portland cement concrete pavements were proposed.