Long-Term Pavement Performance Research Articles

Bayesian dynamic modelling for probabilistic prediction of pavement condition

Significant funds have been allocated to maintain road networks each year in developed countries. Performance prediction is crucial for pavement management systems to adjust working plans and budget allocation. As a Bayesian nonparametric method, Gaussian process regression (GPR) is powerful in predicting nonlinear time series and quantifying uncertainty. However, it remains computationally intensive and fails to adapt to the time-varying characteristics. To address such issues, a dynamic GPR model is proposed for probabilistic prediction of the International Roughness Index (IRI) for flexible pavements. A moving window strategy is developed to substantially shrink the size of training data, which effectively alleviates computational cost and thus leads to a dynamic GPR. A genetic algorithm is then adopted to determine the optimal window size by considering the trade-off between computational efficiency and accuracy. A dataset acquired from Long-Term Pavement Performance (LTPP) is used to demonstrate the feasibility of the dynamic GPR. Its performance is compared to traditional GPR as well as dynamic and static Bayesian linear regression (BLR) models. The comparison results indicate that the proposed dynamic GPR can increase the accuracy by 0.86, 1.52, and 2.27 times for dynamic BLR, static GPR, and static BLR, respectively. It exhibits the best results in terms of accuracy and uncertainty metrics due to its nonlinear modelling and time-varying ability.

Performance prediction for network-level flexible pavements using an optimised ANN approach

ABSTRACT Considering that the road network shares traffic flow as a system, this study aimed to predict the network-level flexible pavement performance, using an optimised artificial neural network (ANN) approach to predict the pavement performance of each individual road, and the Ford-Fulkerson algorithm to determine the weight coefficients of roads. ANNs were developed to predict distress conditions, functional conditions and structural performance using the inputs of the pavement age and structural, traffic and climatic conditions. ANNs were trained with Long-Term Pavement Performance (LTPP) program data, and almost all the coefficient of determination values between ANN outputs and measured results are larger than 0.9. This ANN approach was optimised by establishing grey models to provide predictive short-term performance data for training ANNs, and employing Kalman Filter to modify the long-term performance prediction. Applications in typical LTPP sections validated the effectiveness of the optimisation method. This study used the reduction in road network capacity due to the assumed unavailability of a certain road to quantify its role (weight coefficient), which was calculated using the Ford-Fulkerson algorithm. This network-level performance prediction approach was applied in a 25-expressways network in Shanghai, which validated the good generalisation capabilities of ANNs and feasibility of the Ford-Fulkerson algorithm.

Development of a Machine Learning (ML)-Based Computational Model to Estimate the Engineering Properties of Portland Cement Concrete (PCC)

AbstractPortland cement concrete (PCC) is the construction material most used worldwide. Hence, its proper characterization is fundamental for the daily-basis engineering practice. Nonetheless, the experimental measurements of the PCC’s engineering properties (i.e., Poisson’s Ratio -v-, Elastic Modulus -E-, Compressive Strength -ComS-, and Tensile Strength -TenS-) consume considerable amounts of time and financial resources. Therefore, the development of high-precision indirect methods is fundamental. Accordingly, this research proposes a computational model based on deep neural networks (DNNs) to simultaneously predict the v, E, ComS, and TenS. For this purpose, the Long-Term Pavement Performance database was employed as the data source. In this regard, the mix design parameters of the PCC are adopted as input variables. The performance of the DNN model was evaluated with 1:1 lines, goodness-of-fit parameters, Shapley additive explanations assessments, and running time analysis. The results demonstrated that the proposed DNN model exhibited an exactitude higher than 99.8%, with forecasting errors close to zero (0). Consequently, the machine learning-based computational model designed in this investigation is a helpful tool for estimating the PCC’s engineering properties when laboratory tests are not attainable. Thus, the main novelty of this study is creating a robust model to determine the v, E, ComS, and TenS by solely considering the mix design parameters. Likewise, the central contribution to the state-of-the-art achieved by the present research effort is the public launch of the developed computational tool through an open-access GitHub repository, which can be utilized by engineers, designers, agencies, and other stakeholders.

Integrating Tensometer Measurements, Elastic Half-Space Modeling, and Long-Term Pavement Performance Data into a Mechanistic–Empirical Pavement Performance Model

Pavement performance models (PPMs) are utilized to predict pavement network conditions which is an essential part of any sustainable pavement management system (PMS). The reliability of a PMS and its outputs is proportional to the reliability of the PPM used. This article describes a mechanistic–empirical pavement performance model based on pavement response parameters—strains calculated in the pavement layers measured by tensometers embedded in the pavement surface and verified by calculations in the elastic half-space model and supplemented by empirical data from long-term pavement performance monitoring and accelerated pavement testing. Hence, the herein described PPM combines pavement serviceability evaluation, pavement bearing capacity, and the physico-mechanistic properties of paving materials. The analytical methods which were used to ascertain the physico-mechanistic characteristics, the material fatigue degradation model, and the surface degradation, unevenness in particular, are described. A comparison of the empirical PPM created in the last century used by the national road administrator to this day and the newly created PPM is presented. The comparison shows the difference in the calculated socio-economic benefits and subsequent cost–benefit analysis results. The comparison shows that the use of the old PPM may have produced false economic evaluation results that have led to poor decision making, partially explaining the unsustainable trend of road network management in our country.

Preventive Maintenance Decision-Making Optimization Method for Airport Runway Composite Pavements

Long-term preventive maintenance planning using finite annual budgets is vital for maintaining the service performance of airport runway composite pavements. Using the pavement condition index (PCI) to quantify composite pavement performance, this study investigated the PCI deterioration tendencies of middle runways, terminal runways, and taxiways and developed prediction models related to structural thickness and air traffic. Performance jump (PJ) and deterioration rate reduction (DRR) were used to measure maintenance benefits. Based on 112 composite pavement sections in the Long-term Pavement Performance Program, this study analyzed the influences of five typical preventive maintenance technologies on PJ, DRR, and PCI deterioration rates. The logarithmic regression relationship between PJ and PCI was obtained. For sections treated with crack sealing and crack filling, the DRR was nearly 0. For sections treated with fog seal, thin HMA overlay, and hot-mix recycled AC, the DRR was 0.2, 0.7, and 0.8, respectively. To solve the multi-objective maintenance problem, this study proposed a decision-making optimization method based on dynamic programming, and the solution algorithm was optimized, which was applied in a five-year maintenance plan. Considering different PCI deterioration tendencies of airport regions, as well as PJ, DRR, and costs of maintenance technologies, the preventive maintenance decision-making optimization method meets performance and financial requirements sufficiently.

Ensemble Learning Approach for Developing Performance Models of Flexible Pavement

This research utilizes the Long-Term Pavement Performance database, focusing on devel-oping a predictive model for flexible pavement performance in the Southern United States. Analyzing 367 pavement sections, this study investigates crucial factors influencing asphaltic concrete (AC) pavement deterioration, such as structural and material components, air voids, compaction density, temperature at laydown, traffic load, precipitation, and freeze–thaw cycles. The objective of this study is to develop a predictive machine learning model for AC pavement wheel path cracking (WpCrAr) and the age at which cracking initiates (WpCrAr) as performance indicators. This study thoroughly investigated three ensemble machine learning models, including random forest, extremely randomized trees (ETR), and extreme gradient boosting (XGBoost). It was observed that XGBoost, optimized using Bayesian methods, emerged as the most effective among the evaluated models, demonstrating good predictive accuracy, with an R2 of 0.79 for WpCrAr and 0.92 for AgeCrack and mean absolute errors of 1.07 and 0.74, respectively. The most important features influencing crack initiation and progression were identified, including equivalent single axle load (ESAL), pavement age, number of layers, precipitation, and freeze–thaw cycles. This paper also showed the impact of pavement material combinations for base and subgrade layers on the delay of crack initiation.

A Machine Learning based approach to predict road rutting considering uncertainty

Roads as vital public assets are the backbone for transportation systems and support constant societal development. Recently, data-driven technologies such as digital twins and especially machine learning have shown great potential to maintain the service level of the existing road infrastructure by accurate future condition modelling and optimal maintenance treatment recommendations. However, the pavement community suffers from inadequate data and errors experienced in data collection, which unavoidably limits machine learning performance. In addition, focusing solely on data without considering the underlying physical behaviour remains as a challenge for the practical implementation of machine learning. To this end, this study provides a machine learning based approach to predict road rutting taking into account the machine learning uncertainties. The US Long-Term Pavement Performance public database has been used as the main data source while supplementary synthetic data was added using Finite Element simulations based on physics. The obtained results indicate that adding extra simulation data improved the model’s short-term prediction accuracy by 4.4% and reduced the long-term prediction uncertainty by 6.76%. The approach could potentially mitigate the issue of lack of data and the uncertainties around the data collected, by integrating existing understanding of pavement physical behaviour into the machine learning modelling pipeline.

Impacts of climate change on environmental and economic sustainability of flexible pavements across China

Pavements are susceptible to accelerated deterioration due to changing climate conditions, leading to increased maintenance and excess fuel consumption through pavement-vehicle interaction. China's diverse climates raise concerns about the environmental and economic sustainability of flexible pavements amid climate change and the effectiveness of preventative maintenance strategies. This study examines climate change's potential impacts on long-term pavement performance, greenhouse gas (GHG) emissions, and costs, employing the Mechanistic-Empirical Pavement Design Guide method. The calibrated World Bank's Highway Development and Management Model 4 assesses the impacts of surface characteristics on vehicle fuel consumption. Life cycle assessment and life cycle cost analysis quantify GHG emissions and costs. Results reveal significant pavement deterioration in Southeast and Central China. Preventative maintenance strategies reduce fuel consumption, with GHG emissions and cost savings from smoother driving conditions outweighing those from maintenance. These insights stress the importance of proactive maintenance strategies for mitigating climate-induced deterioration and enhancing sustainability.

Long-Term Performance of Flexible Pavements Containing Crumb Rubber Modified Asphalt in Louisiana

The use of crumb rubber modifiers (CRMs) in asphalt mixtures plays a significant role in creating a sustainable circular economy. This study aims to evaluate the field performance of CRM asphalt sections with service years greater than eight. Five field projects were evaluated. Each project had one or two CRM sections constructed alongside a companion control section. Five types of CRM products were considered. A 16-mesh proprietary CRM, a Rouse-80 powder (R-80), and a Neste Wright powder were used in the wet blending process. Further, a 16-mesh generic crumb rubber, a PlusRideTM shredded rubber, and R-80 were used in a dry blend. Field performance data, including rutting, roughness, and cracking, were collected and analyzed. Moreover, economic analysis was performed to ascertain the cost-effectiveness (CE) of constructing CRM sections. Furthermore, laboratory performance indicators measured during construction were analyzed to ascertain their ability to rank field performance. Generally, CRM sections exhibited significantly similar or better ride quality, cracking and rutting resistance, and pavement condition index than their corresponding control sections. In some cases, the styrene-butadiene-styrene polymer modifier showed better cracking resistance in the control sections than the CRMs. CRM asphalt sections generally exhibited higher CE than their corresponding control sections because of their high initial unit costs. The Rouse wet CRM asphalt section in field project 4, which had a relatively lower unit cost, was found to be cost-effective compared to its corresponding control section. The limited laboratory mechanical performance parameters considered were ineffective in consistently ranking field performance indicators.

Modelling Asphalt Overlay As-Built Roughness Based on Profile Transformation-Case for Paver Using Automatic Levelling System.

The as-built roughness, or smoothness obtained during pavement construction, plays an important role in road engineering since it serves as an indicator for both the level of service provided to users and the overall standard of construction quality. Being able to predict as-built roughness is therefore important for supporting pavement design and management decision making. An as-built IRI prediction model for asphalt overlays based on profile transformation was proposed in a previous study. The model, used as basis for this work, was developed for the case of wheeled pavers without automatic screed levelling. This study presents further development of the base prediction model, including the use of an automatic screed control system through a long-distance averaging mobile reference. Formulation of linear systems that constitute the model are presented for the case of a wheeled paver using contactless acoustic sensors set-up over a floating levelling beam attached to the paver. To calibrate the model, longitudinal profile data from the Long-Term Pavement Performance SPS-5 experiment was used, obtaining a mean error of 0.17 m/km for the predicted IRI. The results obtained demonstrate the potential of the proposed approach as a modelling alternative.

Novel Instance-Based Transfer Learning for Asphalt Pavement Performance Prediction

The deep learning method has been widely used in the engineering field. The availability of the training dataset is one of the most important limitations of the deep learning method. Accurate prediction of pavement performance plays a vital role in road preventive maintenance (PM) and decision-making. Pavement performance prediction based on deep learning has been widely used around the world for its accuracy, robustness, and automation. However, most of the countries in the world have not built their pavement performance historical database, which prevents preventive maintenance using the deep learning method. This study presents an innovative particle swarm optimization (PSO) algorithm-enhanced two-stage TrAdaBoost.R2 transfer learning algorithm, which could significantly increase the pavement performance prediction database. The Long-Term Pavement Performance (LTPP) database is used as the source domain data, and one of the highways in China is chosen as the target domain to predict pavement performance. The results show that the proposed PSO-Two-stage TrAdaBoost.R2 model has the highest accuracy compared with AdaBoost.R2 model and traditional regression decision tree model. The validation case study shows significant consistency between the predicted International Roughness Index (IRI) and the whole-year measurement data with an R2 of 0.7. This study demonstrates the great potential of the innovative instance-based transfer learning method in pavement performance prediction of a region’s lack of data. This study also contributes to other engineering fields that could greatly increase the universality of deep learning.

Machine Learning Modeling of Wheel and Non-Wheel Path Longitudinal Cracking

Roads degrade over time due to various factors such as traffic loads, environmental conditions, and the quality of materials used. Significant investments have been poured into road construction globally, necessitating regular evaluations and the implementation of maintenance and rehabilitation (M&R) strategies to keep the infrastructure performing at a satisfactory level. The development and refinement of performance prediction models are essential for forecasting the condition of pavements, especially to address longitudinal cracking distress, a major issue in thick asphalt pavements. This research leverages multiple machine learning methods to create models predicting non-wheel path (NWP) and wheel path (WP) longitudinal cracking using data from the Long-Term Pavement Performance (LTPP) program. This study highlights the marked differences in distress conditions between WP and NWP, underscoring the importance of precise models that cater to their unique features. Aging trends for both types of cracking were identified through correlation analysis, showing an increase in WP cracking with age and a higher initial International Roughness Index (IRI) linked to NWP cracking. Factors such as material characteristics, kinematic viscosity, pavement thickness, air voids, particle size distribution, temperature, KESAL, and asphalt properties were found to significantly influence both WP and NWP cracking. The Exponential Gaussian Process Regression (GPR) emerged as the best model for NWP cracking, showcasing exceptional accuracy with the lowest RMSE of 89.11, MSE of 7940.72, and an impressive R-Squared of 0.63. For WP cracking, the Squared Exponential GPR model was most effective, with the lowest RMSE of 12.00, MSE of 143.93, and a high R-Squared of 0.62. The GPR models, with specific kernels for each cracking type, proved their adaptability and efficiency in various pavement scenarios. A comparative analysis highlighted the superiority of our new machine learning model, which achieved an R2 of 0.767, outperforming previous empirical models, demonstrating the strength and precision of our machine learning approach in predicting longitudinal cracking.

How much road deterioration do heavy vehicles cause on Australia’s sealed flexible pavements?

ABSTRACT An evidence-based estimate of the load-related road deterioration due to heavy vehicles on the road network is needed for underpinning heavy vehicle road user charges on Australia’s arterial roads. A network level model for estimating the percentage (%) load-related road deterioration was developed for typical sealed unbound pavements in Australia’s sealed road network. This model was based on a network level roughness deterioration model using observational data from long-term pavement performance (LTPP) and long-term pavement performance maintenance (LTPPM) sites and accelerated loading facility (ALF) experimental data. The model estimates that the % load-related road deterioration varied with changes to the traffic load, pavement strength and an environmental coefficient. This outcome also implies that the roughness deterioration model for sealed granular pavements is likely to be more appropriate for predicting Australia’s sealed granular pavement deterioration than past internationally based roughness deterioration models which were originally based on the deterioration of asphalt bound pavements.

Pure mode I fracture specification of molybdenum disulfide nanoparticles (MDS)-reinforced hot mix asphalt (HMA) under long-term conditions

Low and intermediate temperature cracking are two key factors for fracture and failure of hot mix asphalt (HMA), which impose huge costs on transportation departments and related agencies around the world. These two kinds of cracking may apear in the short- or long-term owing to states such as freezeing-thawing damage (FTD) and the aging mechanism. On the other hand, due to its unique features, nanomaterials are a suitable option for improving the mechanical features of bituminous mixes, which can affect the short-term performance and long-term performance. In this study, low and intermediate temperature fracture specification of Nano-molybdenum disulfide (MDS)-reinforced hot mix asphalt (HMA) under long-term conditions was evaluated. For this goal, two semi-circular bend (SCB) geometries comprising 90 degree and angled cracks, which were called symmetric SCB and classical-reinforced SCB-2, were made. Then, two various conditions, i.e. three freeze–thaw cycles and six days of aging operation were performed on various mixes in order to simulate the long-term performance of the bituminous mixes. Finally, the HMA samples with and without MDS were subjected to the three-point SCB test at the final temperatures of −15 °C and + 15 °C. The findings demonstrated that mixes comprising 0.3 % and 0.6 % MDS had more increased fracture energy (FE), fracture toughness (KIC), and stiffness under pure tension at low and medium temperatures (short-term and long-term fracture specification). In fact, MDS caused an increase in the fracture indices by increasing the maximum force (Pmax) and displacement at the time of fracture. At ± 15 °C and for the geometry comprising the 90 degree and 35 degree cracks, the HMA and Nano MDS-reinforced HMA mixes had more suitable flexibility under 3 FTC compared to 0 FTC; while, the flexibility of the HMA and Nano MDS-reinforced HMA mixes decreased under aging mechanism. In discussing the long-term performance of HMA mixture (with and without modifiers), the findings showed that applying three FTD's cycles to the mixes caused a decrease in KIC, FE, and Pmax. Also, reducing the stiffness of various mixes under 3 FTC, and increasing the flexibility of the mixes were other findings of this study. In addition, the simultaneous use of both SCB geometries comprising 90 degree and 35 degree cracks was proposed to study the fracture specification of bituminous mixes owing to the more inferior average FE and KIC. Finally, the application of 3FTC on bituminous mixes can more suitable simulate the long-term performance of asphalt pavement (compared to aging mechanism).

Research and comparison of pavement performance prediction based on neural networks and fusion transformer architecture

<abstract> <p>The decision-making process for pavement maintenance from a scientific perspective is based on accurate predictions of pavement performance. To improve the rationality of pavement performance indicators, comprehensive consideration of various influencing factors is necessary. To this end, four typical pavement performance indicators (i.e., Rutting Depth, International Roughness Index, Longitudinal Cracking, and Alligator Cracking) were predicted using the Long Term Pavement Performance (LTPP) database. Two types of data, i.e., local input variables and global input variables, were selected, and S-ANN and L-ANN models were constructed using a fully connected neural network. A comparative analysis of the predictive outcomes reveals the superior optimization of the L-ANN model. Subsequently, by incorporating structures such as self-attention mechanism, a novel predictive approach based on the Transformer architecture was proposed. The objective is to devise a more accurate predictive methodology for pavement performance indices, with the goal of guiding pavement maintenance and management efforts. Experimental results indicate that, through comparative analysis of three quantitative evaluation metrics (root mean square error, mean absolute error, coefficient of determination), along with visual scatter plots, the predictive model employing the fused Transformer architecture demonstrates higher robustness and accuracy within the domain of pavement performance prediction when compared to the L-ANN model. This outcome substantiates the efficacy and superiority of the model in terms of predictive performance, establishing it as a reliable tool for accurately reflecting the evolution of asphalt pavement performance. Furthermore, it furnishes a theoretical reference for determining optimal preventive maintenance timing for pavements.</p> </abstract>

Predicting flexural-creep stiffness in bending beam rheometer (BBR) experiments using advanced super learner machine learning techniques

BBR test is commonly used to assess the low-temperature performance grade (PG) of asphalt binders, with the flexural-creep stiffness being a critical parameter calculated through this test. However, it has notable limitations that demand attention. The significant amount of asphalt binder needed for test specimens increases costs and resource consumption. Additionally, the complex and time-consuming specimen preparation process hinders testing efficiency and introduces result variability, affecting the accuracy and reliability of PG determinations. In recent years, machine learning (ML) has emerged as a promising substitute for predicting various engineering values. In this study, the primary focus was on harnessing super learner (SL) techniques to predict the creep stiffness of asphalt binders. The SL approach combines multiple ML algorithms to enhance predictive accuracy and reduce individual model biases. Bagging, boosting, and stacking algorithms were employed in the construction of these prediction models. To conduct the investigation, data from 1350 samples sourced from the Long-Term Pavement Performance (LTPP) website were utilized to explore the influence of six crucial variables on the creep stiffness of asphalt binders. The proposed method demonstrated high accuracy, nearing 90% in the coefficient of determination. The Stacking model achieved a low Mean Absolute Percentage Error of 2.86% and robust Prediction Accuracy of 97.14% for randomly selected data points. Furthermore, the sensitivity analysis highlighted the significance of distinct input variables in influencing the creep stiffness of asphalt binders. Notably, the test temperature emerged as the most influential factor affecting creep stiffness, according to the conducted study

Modelling the effects of flexible pavement distresses in the long-term pavement performance database on performance

Evaluating flexible pavement performance is mandatory for managing transport infrastructure. This study focuses on modeling the relationships between international roughness index (IRI) and a total of 10 types of pavement distress, including alligator, block, wheel path length, wheel path longitudinal, non-wheel path longitudinal, transverse crackings, patches, bleeding, raveling areas, and pumping. The data recorded under the Long-Term Pavement Performance was used to develop the models. Data sets covering General Pavement Studies from seven states of the United States were used in modeling. The study used modeling approaches, including nonlinear regression analysis, multivariate adaptive regression splines, and artificial neural networks (ANN), in which IRI was the dependent variable and pavement distress was the independent variable. In the developed models, 0.516, 0.623, and 0.684 regression coefficients values were obtained for nonlinear regression analysis, multivariate adaptive regression splines, and artificial neural networks approaches, respectively. The analysis results have determined that the artificial neural networks technique performs more successfully than the other techniques. The statistical error analyses of the root mean square error, Nash-Sutcliffe coefficient of efficiency, mean absolute error, and normalized root mean square error also showed that the same modeling approach performs more successfully. With these data generated from a universally used database, it has been determined once again that ANN is the most efficient mathematical approach in modeling the relationships between surface distresses and IRI.

Performance-oriented road structure and material design method based on enhanced XGBoost algorithm

ABSTRACT A well-designed road structure is critical for reducing road operational period diseases and saving construction management and maintenance funds throughout the life cycle of roads. Excessive design may cause waste in materials or pavement distresses. This study proposes an intelligent optimisation design method for road structure and materials based on pavement performance and cost. The Long-Term Pavement Performance (LTPP) database were used to training the prediction model using the particle swarm optimisation (PSO) algorithm and the extreme gradient boosting algorithm (XGBoost) algorithm. Subsequently, an automated design model encompassing pavement thickness, pavement material, and subgrade material was established based on the predictive model derived from the PSO algorithms. The results indicate that the R2 between the actual values and the predicted values of the PSO-XGBoost model is above 0.935, and the Pareto optimal solution is obtained using the multi-objective Particle Swarm Optimisation algorithm. This investigation showcases the potential of data-driven methodologies in furnishing valuable guidance for the initial stages of road design and construction.

Optimizing Faulting Prediction for Rigid Pavements Using a Hybrid SHAP-TPE-CatBoost Model

Faulting refers to the common and significant distress in Jointed Plain Concrete Pavement (JPCP), which has an adverse impact on the pavement roughness. Nevertheless, the existing fault prediction models continue to heavily rely on conventional linear regression techniques or basic machine learning approaches, which leaves room for improvement in training efficiency and interpretability. To enhance training efficiency and accuracy, this study developed five novel faulting prediction models. These models are based on five basic machine learning algorithms: Random Forest (RF), Additive Boosting (AdaBoost), Gradient Boosting Decision Tree (GBDT), Light Gradient Boosting Machine (LightGBM), and Categorical Boost (CatBoost), combined with the tree-structured Parzen estimator (TPE). The five models are TPE-RF, TPE-AdaBoost, TPE-GBDT, TPE-LightGBM, and TPE-CatBoost. In addition to selecting the best-performing model, this study incorporated the Shapley Additive Explanation (SHAP) technique and developed TPE-SHAP-CatBoost to improve the interpretability of the model’s predictions. The process involved extracting historical data on pavement performance, including 17 variables, from the Long-Term Pavement Performance (LTPP) database for 160 instances of observation. Firstly, the Boruta method was used to identify the final set of input variables. Secondly, the TPE technique, which is a Bayesian optimization method, was applied to automatically select the optimal hyperparameters for the base models. Finally, SHAP was used to provide both global and local explanations of the model’s outputs. The results indicate that the TPE-CatBoost model achieves the highest accuracy with an R2 value of 0.906. Furthermore, the TPE-SHAP-CatBoost model identified the primary factors influencing faulting by incorporating SHAP and provided explanations of the model’s results at both the global and local levels. These research findings highlight the ability of the proposed model to accurately predict faulting, providing precise and interpretable guidance for pavement maintenance while reducing workload for pavement engineers in data collection and management.

Holistic multiphysics simulation of climatic responses of cold region pavements

In cold regions, the environment dynamics lead to variations of soil temperature, water content, and deformation, which are characterized by highly coupled physical interplay. The hydraulic and thermal properties of unsaturated soils are highly nonlinear, which is further complicated when subjected to freezing. This paper presents a comprehensive multiphysics coupling model to evaluate these complex processes. The model considers the behaviors of unsaturated frozen soils. It accounts for the influences of meteorological, geothermal, and hydrological factors. The model is validated through two pavement case studies using Long-Term Pavement Performance (LTPP) road section data. The first case analysis is performed for a pavement section in Vermont, and the simulation lasted for 30 days during a non-freezing season on an hourly basis. The results validated the performance of the model considering unsaturated soil behaviors. The second case study is based on a daily analysis of a pavement section in South Dakota over a freezing–thawing cycle over 194 days. The results validated the model in considering the frozen unsaturated soil behaviors. Both case studies demonstrate the performance of this comprehensive model in quantifying the spatial and temporal variations of soil temperature and water content in response to environmental stressors. The capability of the model in accurately predicting the responses of pavement to the meteorological factors unleashes the potential of this model to assess the effects of climate and climate change on cold region pavement, as well as other types of geo-structures.