Rut Depth Prediction Research Articles

Physics-guided neural network for predicting asphalt mixture rutting with balanced accuracy, stability and rationality

The prediction of rutting performance of asphalt materials poses a significant challenge due to the intricate relationships between the rutting performance and its influencing factors. Machine learning models have gained popularity to address this challenge by offering sophisticated model structures and algorithms. However, existing models often prioritize model accuracy over stability and rationality. The increasingly complicated model structure may lead to an imbalance between the data and the model, resulting in issues such as overfitting and reduced model applicability and interpretability. In this context, this study proposes a novel modeling framework to predict the rutting performance of asphalt mixture by utilizing autoencoder for feature selection and feedforward neural network for rut depth prediction. Notably, physics information of the selected variables is implemented into the neural network to achieve the appropriate balance of model accuracy, stability, and rationality. The results demonstrate that while maintaining high model accuracy, the implementation of physics information significantly enhances the model’s stability and rationality. This framework holds great potential for accurate and reliable predictions of pavement distress by leveraging the complementary strengths of data-driven machine learning and physics-based domain knowledge.

Prediction of asphalt concrete flexibility index and rut depth utilising deep learning and Monte Carlo Dropout simulation

ABSTRACT Asphalt concrete (AC) balanced mix design (BMD) relies on laboratory testing to meet both the Illinois Flexibility Index Test (I-FIT) and the Hamburg Wheel Tracking Test (HWTT) by varying aggregate gradations and asphalt cement contents. AC mix designers would benefit from estimating the impact of constituent materials’ properties on I-FIT and HWTT before conducting a BMD trial. This study focused on the development of two deep learning models to predict I-FIT flexibility index (FI) and HWTT rut depth. The models were created based on a I-FIT database of 19,138 datasets and a HWTT database of 7602 datasets (after data preprocessing was conducted). Two deep neural networks (DNNs) were then trained to predict FI and rut depth. Monte Carlo Dropout simulations were then used in the DNN models to compute a distribution of predicted FI and rut depth. The distribution of predicted FI and rut depth provides a best estimate and range of FI and rut depth. The developed models provide a distribution of predictions with a coefficient of variation (CoV) lower than 30% for both the I-FIT and HWTT models, respectively.

Creating Rutting Prediction Models through Machine Learning Techniques Utilizing the Long-Term Pavement Performance Database

Over time, roads undergo deterioration caused by various factors such as traffic loads, climate conditions, and material properties. Considering the substantial global investments in road construction, it is crucial to periodically assess and implement maintenance and rehabilitation (M and R) plans to ensure the network’s acceptable level of service. An integral component of the M and R plan involves utilizing performance prediction models, especially for rutting distress, a significant issue in asphalt pavement. This study aimed to develop rutting prediction models using data from the Long-Term Pavement Performance (LTPP) database, employing several machine learning techniques such as regression tree (RT), support vector machine (SVM), ensembles, Gaussian process regression (GPR), and Artificial Neural Network (ANN). These techniques are well-known for effectively handling extensive and complex datasets. To achieve the highest modeling accuracy, the parameters of each model were meticulously fine-tuned. Upon evaluation, the results revealed that the GPR models outperformed other techniques in various metrics, including Root Mean Square Error (RMSE), R-squared, Mean Absolute Error (MAE), and Mean Square Error (MSE). The best GPR model achieved an RMSE of 1.96, R-squared of 0.70, MAE of 1.32, and MSE of 109.33, indicating its superior predictive capabilities compared with the other machine learning methods tested in this study. Comparison Analysis was made for 10 randomly selected sections on our novel machine learning model that outperforms existing models, with an R2 of 0.989 compared with 0.303 and 0.3095 for other models. This demonstrates the potential of advanced machine learning in accurate rut depth prediction across diverse climates, aiding pavement management decisions.

Development of Prediction Model for Rutting Depth Using Artificial Neural Network

One of the most common pavement distresses in flexible pavement is rutting, which is mainly caused by heavy wheel load and various other factors. The prediction of rutting depth is important for safe travel and the long-term performance of pavements. Factors that are considered in this paper for the prediction of rut depth are Temperature, Equivalent Single Axle Load, Resilient modulus, and Thickness of hot mixed asphalt. The input data for all factors are collected from the Long-Term Pavement Performance Information Management System for the state of Texas. Regression analysis is performed for dependent and independent variables to obtain the empirical relationship. In various fields of civil engineering, artificial neural networks have recently been utilized to model the qualities and behavior of materials and to determine the complicated relationship between various properties. An Artificial Neural Network is used to develop a predictive model to predict the rutting depth. A total number of 70 observations were considered for the predictive model. A mathematical relation is developed and verified between rut depth and variable input data.

Calibration of Inverted Asphalt Pavement Rut Prediction Model, Based on Full-Scale Accelerated Pavement Testing.

This study investigates the establishment and calibration method of the rut depth (RD) prediction model of inverted asphalt pavements (IAPs), based on full-scale accelerated pavement testing (APT), which facilitates the accurate and reliable design or maintenance of IAPs. A power function is adopted for the prediction model construction of the rut progression before the failure stage, based on the typical permanent deformation progression curve of flexible pavements. The APT loading history is divided into units, according to the difference in physical conditions, providing the basis for a cumulative RD analysis and model calibration. The nonlinear incremental recursive (IR) principle is applied in the RD analysis to consider the influence of the nonlinear material property, performance deterioration, and loading history on the RD development. Further, the rut shift function relating prediction models obtained from laboratory tests and full-scale APT is established to introduce the APT data in the calibration process. Accordingly, the mechanistic-empirical RD prediction model calibration method, based on APT and the IR principle, is proposed and applied in a case study of a IAP RD prediction model calibration. Four 3.5 m × 4 m IAP test sections S1-S4 are constructed and instrumented and 700,000- and 900,000-wheel loads are applied on test sections S1-S2 and S3-S4, respectively, using the heavy vehicle simulator. The test data from the different APT load units are utilized for the model calibration, and the resultant prediction errors range from -2.16 mm to 1.18 mm. The calibrated model can also be used for the RD prediction of IAPs with other design schemes, by updating the corresponding material-related coefficients and the finite element model, which is essential for the design and maintenance of IAPs. The proposed calibration method could be a useful reference for the establishment of flexible pavement performance prediction models.

Establishment of Rutting Model of Wheel-Tracking Test for Real-Time Prediction of Rut Depth of Asphalt Layers

The construction process control of asphalt layers directly affects the road service life and quality. The objective of this study was to establish a rutting model of the wheel-tracking test used for the real-time prediction of the rut depth of asphalt layers in the construction process. The gradation of asphalt mixture, asphalt content, and molding temperature were considered in the development of the new rutting model of the wheel-tracking test. The effects of these three factors on the high-temperature performance of asphalt mixture were analyzed. The order of importance of the factors affecting the high-temperature performance of asphalt mixture is the gradation of asphalt mixture, asphalt-aggregate ratio, and molding temperature. Overall, the predicted values of the rut depth of the wheel-tracking test are very close to the measured values. Furthermore, the difference between the rut depths of asphalt layers of the test group and the control group is small. These comparison results indicate that the new rutting model of the wheel-tracking test has high accuracy and good applicability for the test road.

Towards dynamic forest trafficability prediction using open spatial data, hydrological modelling and sensor technology

Abstract Forest harvesting operations with heavy machinery can lead to significant soil rutting. Risks of rutting depend on the soil bearing capacity which has considerable spatial and temporal variability. Trafficability prediction is required in the selection of suitable operation sites for a given time window and conditions, and for on-site route optimization during the operation. Integrative tools are necessary to plan and carry out forest operations with minimal negative ecological and economic impacts. This study demonstrates a trafficability prediction framework that utilizes a spatial hydrological model and a wide range of spatial data. Trafficability was approached by producing a rut depth prediction map at a 16 × 16 m grid resolution, based on the outputs of a general linear mixed model developed using field data from Southern Finland, modelled daily soil moisture, spatial forest inventory and topography data, along with field measured rolling resistance and information on the mass transported through the grid cells. Dynamic rut depth prediction maps were produced by accounting for changing weather conditions through hydrological modelling. We also demonstrated a generalization of the rolling resistance coefficient, measured with harvester CAN-bus channel data. Future steps towards a nationwide prediction framework based on continuous data flow, process-based modelling and machine learning are discussed.

Pavement Rut Depth Prediction for a Three-Span Suspension Steel Box Girder Bridge Based on Two-Year Temperature Monitoring Data

AbstractThe paper presents a study of temperature distribution and effect on the asphalt pavement of the Fourth Nanjing Yangtze River Bridge based on the 2-year continuous monitoring data. The temp...

Measurements and simulations of rut depth due to single and multiple passes of a military vehicle on different soil types

Accurate prediction of rut depth caused by military vehicles trafficking is crucial for the success of their mission. In addition to the operational aspect of trafficability and mobility, military vehicle traffic can lead to severe degradation of soil functions. The objective of this study was (i) to examine rut depth formation and bulk density changes after one and repeated passages of a military vehicle on different soils and land uses, and (ii) to assess the ability of an analytical soil compaction model to predict these changes. The wheeling tests were carried out with a 7 Mg military truck at seven sites in Estonia covering soil textures from sand to clay and a peat soil. Rut depths and bulk density at various depths were measured after one and ten vehicle passes. Undisturbed soil samples were collected next to the tracks for determination of initial soil compressive properties used for parameterisation of the model. The smallest rut depths were measured on arable soils, both after one and ten vehicle passes. The largest rut depth after one pass was obtained on the peat soil, while the largest rut depths after ten passes were measured on the grassland soils. The model overestimated rut depths on the peat soil, and underestimated the increase in rut depth from the first to the tenth pass on the grassland soils, while predictions for the arable soils were adequate. We identify a need for studies on the mechanical behaviour of peat soils during wheeling and for better quantitative description of multi pass effects, especially on soft soils.

Modelling asphalt pavement analyzer rut depth using different statistical techniques

Different computer-operated temperature-controlled equipments are used to predict asphalt mixture performance in laboratory at approximately similar conditions to actual in-service pavements. Asphalt pavement analyzer (APA) is one such multi-functional sophisticated equipment. Wearing course layers are primarily attributable to be deteriorated by rutting distress in flexible asphalt pavements. Therefore, 30 representative wearing course mixtures commonly used in road pavements were tested for rutting distress using APA. Mixtures were evaluated based on their rutting performance and found that consistent ranking has been observed at temperature conditions of 40°C and 50°C. It was observed that APA rut depth could be effectively modelled using statistical techniques of multiple non-linear regression and artificial neural network. Novel aggregate material indices designated as ‘aggregate source index’ and ‘aggregate gradation index’ were developed in this study. Based on past literature; temperature, aggregate source index (ASI), aggregate gradation index, bitumen penetration values, and number of APA loading cycles/strokes were included as independent variables in the model developed for selected asphalt mixtures. It has been concluded from the values of non-linear regression parameters that temperature has the highest significance for rut depth prediction, followed by the bitumen penetration value, ASI, aggregate gradation index and number of loading cycles. Statistical parameters i.e. coefficient of determination, root mean squared error, values account for, and mean absolute percentage error showed that the artificial neural network technique better predicts the APA rut depth as compared to multiple non-linear regression techniques.

Artificial neural networks approach to predicting rut depth of asphalt concrete by using of visco-elastic parameters

Performing comprehensive research on the functional behavior of asphalt pavements under the influence of various environmental and structural parameters can better assist the engineers in the design and maintenance of asphalt pavements. Using a solution that can reduce the cost and time of assessment is very important. Using artificial neural networks in many facilitates operations on data engineering sciences. It is necessary to ensure that a comprehensive study is performed through considering all or most of the parameters affecting the behavior. The aim of this study is to provide an experimental model to estimate the rut depth of asphalt concrete by using viscoelastic parameters and artificial neural networks. Accordingly the asphalt concrete specimens containing 3, 5 and 7 percent void with two types of limestone and siliceous aggregates and PG64-22 and PG58-28 bitumens were made and exposed to dynamic creep tests under 50–60 °C and the stress range of 100–300 kPa. Then the viscoelastic parameters of asphalt specimens were extracted from the creep diagrams and eventually the asphalt concrete’s rut depth prediction model was trained and provided by artificial neural network. Comparing the output results with the experimental test results show that by using this model it is possible to estimate the creep behavior and rut depth of asphalt concrete pavements based on the effective parameters without the need for costly and time-consuming tests.

Effective temperature for predicting permanent deformation of asphalt pavement

The determination of a single temperature to stimulate permanent deformation is quite significant in pavement design and maintenance. The practical use of this single temperature to replace the seasonal temperature fluctuation throughout the whole year can apparently simplify the prediction of permanent deformation of an asphalt pavement. This single temperature is termed effective temperature (Teff) and is defined as a simulated temperature to achieve a permanent deformation, which is accumulated due to the traffic loading and pavement temperature within an entire year. However, most previous studies on effective temperature focus on investigating the performance temperature for permanent deformation instead of simplifying the prediction of this distress. For this reason, the primary goal of this study was to develop a Teff model and simplify the prediction of permanent deformation of asphalt pavement. The predicted distress and the Teff results used to build the models and find the optimized coefficients were obtained by the rut depth prediction model, which was based on shear stress and widely applied in China. The proposed Teff models incorporated degree-days and the earth temperature, which was represented by mean annual air temperature (MAAT). By incorporating those parameters, these temperature models predominantly considered the effect of high temperature on the permanent deformation and can be utilized to improve their prediction in various regions. In addition, the asphalt mixture properties and vehicle speeds were also incorporated in the models. By comparing with different models in previous studies, the proposed models were testified to be reasonable.

Soil rut depth prediction based on soil strength measurements on typical Estonian soils

Deep ruts in soil hinder the movement of off-road vehicles and harm drastically the soil environment. The aim of this study was to determine the most suitable cone and remoulding index based methods for rut depth estimation for single and repeated military vehicle passage in typical Estonian soils. Cone index based sinkage models – US Army ERDC, Maclaurin (1990) and Willoughby and Turnage sinkage models applicable for military vehicles were tested, using different critical layers (0–15 cm, 7.5–23 cm or 15–30 cm). The test vehicles were a 7 Mg truck and a 23 Mg armoured personnel carrier. At eight experimental sites covering mechanically very variable soil conditions (from peat to clay soil), rut depth was measured after one and ten vehicle passes, cone penetration tests were conducted in situ, and samples were collected for determination of bulk density, organic carbon content, texture, gravimetric water content, plastic and liquid limits of the topsoil and subsoil. According to average RMSD values for military truck the Willoughby and Turnage model was the most accurate prediction method. For one pass and 10 passes the rating cone index values from the 7.5–23 cm and 15–30 cm soil layer, respectively, produced the lowest RMSD values.

RETRACTED: Effect of segregation on rutting resistance of asphalt pavement

Mix segregation is generally described as localized non-uniform zones of mix that do not conform to the original Job Mix Formula in gradation or asphalt content. Segregation may cause field performance issue such as rutting but no good trend between segregation and field/laboratory rut depth is observed. In this paper, a field road with segregation was monitored and those reported parameters used to detect segregation were collected. The segregated areas and the control non-segregated areas were identified by visual assessment. Rut depth were measured within one year after the road was opened to traffic. Rut depth from segregated areas and non-segregated areas are compared to evaluate the effect of segregation on rutting performance. The possibility of using different parameters such as pavement texture and construction temperature to detect segregation is also analyzed. Acknowledging that the rut depth may be affected by multiple parameters, a statistical Partial Least Squares (PLS) regression method is applied to select predictor variables that have the most significant effect on rut depth, and to develop rut depth prediction model. Results indicate that rut depth from segregated areas (visual) are statistically higher than rut depth from the non-segregated areas (visual). It is also found that pavement texture and construction temperature are the potential factors to detect segregation since they are correlated well with both visual assessment and rut depth. The developed rut depth prediction model shows good predictability and is well validated by considering segregation (mean texture depth), climatic effect (aging days), material property (HMA layer moduli), and pavement structure (entire HMA thickness).

Statistical evaluation of factors affecting the laboratory rutting susceptibility of asphalt mixtures

Rutting or plastic deformation is one of most frequently observed distress types in the wheel paths of flexible road pavements. This distress may cause the pavement structure to be attributable to early failure by expediting the deterioration due to load and environmental detrimental agents. Therefore, it is ideal to simulate the rutting performance of different asphalt mixtures in laboratory-controlled conditions before exposing them to aggressive field conditions. Different rutting simulation computer-controlled equipments are available to predict the mixture performance in laboratory at nearly similar conditions to actual in-service pavements. Cooper wheel tracking equipment is one such precise equipment. Mostly, the wearing course is likely to be damaged by this distress type in flexible asphalt pavements. Therefore, in this research study, thirty representative wearing course mixtures commonly used in road pavements were tested for rutting distress using Cooper wheel tracker. Three aggregate sources, three binders from two sources, five mixture gradations and two temperature conditions were included in the testing scheme. Mixtures were evaluated based on the terminal value of rut depth. Consistent ranking of mixtures has been observed for variable combinations of aggregates, binders, temperatures and gradations. Flakiness index (FI) of aggregates, binder penetration values, per cent coarse material retained on 4.75 mm sieve and temperature values are considered as independent variables based on past researches. The significance of these independent variables on the dependent variable of rut depth has been analysed by analytical tools of general full factorial design analysis. It has been concluded that temperature has highest significance for the rut depth prediction, followed by FI of aggregates, bitumen penetration value and percentage of coarse fraction.

Airfield Flexible Pavement Responses under Heavy Aircraft and High Tire Pressure Loading

This paper investigates airfield flexible pavement responses under heavy aircraft and high tire pressure loading considering the realistic aircraft tire–pavement interaction. An advanced three-dimensional finite element model was developed; it characterized the hot-mix asphalt layer as a viscoelastic material and used implicit dynamic analysis to predict time-and temperature-dependent pavement responses under various loading conditions. The tire loadings were simulated as moving loads having uniform and nonuniform contact stress distributions. To illustrate the effect of moving load, stationary loading analysis was conducted with the equivalent modulus determined from the middepth pulse time under moving loading. Two temperature profiles were considered in the analysis, as compared with the constant temperature profile using the average temperature. The pavement responses in the asphalt layer (tensile, compressive, and shear strains) under different loading conditions were calculated and analyzed in terms of pavement failure mechanisms. The results emphasized the importance of considering nonuniform contact stresses and moving load in airfield pavement analysis. On the contrary, applying the average temperature profile in summer was a conservative approach for predicting fatigue cracking potential but underestimated rutting or near-surface cracking potential in the airfield pavement. The rut depths in the asphalt layer were predicted by using mechanistic–empirical performance models and compared with the measurements from the full-scale test. The results suggested that specific calibration parameters should be developed to provide accurate prediction of rut depth for airfield pavements.

Influence of Air Voids of Hot Mix Asphalt on Rutting within the Framework of Mechanistic-empirical Pavement Design

Traditional pavement design process involves limiting the strain on top of the subgrade layer to ensure that the pavement will not fail by rutting. The recently released Mechanistic-Empirical Pavement Design Guide (M-E PDG) and the associated „AASHTOWare Pavement ME Design” software of the United States considers permanent deformations in all the rut- susceptible layers (asphalt concrete and unbound granular material layers) on the overall cumulative rutting in the pavement. However, the rutting models incorporated in the design do not consider the influence of reduction of air voids during design life and the effect of confinement pressure.This investigation explores the influence of air voids reduction and confinement pressure on the rut depth prediction using AASHTOWare for a typical Indian pavement cross-section. An asphalt concrete mixture, Bituminous Concrete (Grade-1) confirming to the Ministry of Road Transport and Highways (MoRTH) specifications was cast at three air voids contents of 7, 4 and 2%. The dynamic moduli of these samples were determined at 5, 15, 20, 40 and 55°C for frequencies ranging from 0.01 to 25Hz at 0 and 200kPa confinement pressure. In addition, the dynamic modulus values of Dense Bituminous Macadam (Grade 2) mix at these temperature and frequencies for an air voids content of 4% were determined in the unconfined condition. A typical pavement cross-section as given in the Indian Roads Congress guidelines corresponding to 10% CBR and 150 msa traffic was chosen. Simulations were carried out for this pavement cross-section using the material properties measured for asphalt layers. Air voids and confinement pressure exhibited considerable influence on the rutting predicted for the asphalt layers of the pavement.While the AASHTOWare pavement design software may not be directly applicable to Indian conditions due to the use of considerable amount of data pertaining to USA and Canada, such exercise as carried out here, clearly showcases the limitations of the cross-sections used in India and provides guidelines on the directions India should take when it comes to material characterisation.

A torsional hollow cylinder research apparatus for studying the permanent deformation characteristics of asphalt

A large, innovative torsional hollow cylinder apparatus has been designed and commissioned for research investigations on asphalt. It can apply the complex shear and normal stress paths experienced in a pavement under a moving wheel load. The resulting high-quality data on permanent shear and volumetric strains will assist with the development of theoretical methods for rut depth prediction and act as an accurate reference to assess the performance of simpler test equipment. A wide range of repeated load stress paths can be applied under constant temperature conditions. The 350 mm long test specimens have a wall thickness of 50 mm. They are produced by accurate coring from a roller compacted test slab of asphalt. A detailed description of the apparatus is presented along with some preliminary results.

Rutting prediction for stabilized soils based on the cyclic CBR test

Rutting mechanism is complex and still being clarified. In order to predict rut depth on the basis of empirical tests and to establish the mechanism of its creation, sophisticated experiments, such as accelerate pavement testing (APT), are being carried out. To find the proper way of predicting rut depth, mathematic models are also created. It should also be noted, in order to obtain results in line with reality, that the nature of the soil affects the depth of rutting. This paper present results of the cyclic CBR test carried out on chemically stabilized sandy clay with the addition of the 8% hydrated lime prepared in CBR mold with respect to Proctor compaction rules. Sandy clay was subjected to cCBR test after either non-stabilized, 1 hour, 14 or 30 days from stabilization with lime. The aim of the cCBR test was to predict rutting depth, calculated with Plastic Displacement (P-D) model. Thus obtained data were compared with test results and some modification was proposed in order to obtain more accurate prediction of rut depth. Also, statistical analysis of correlation between the modeled permanent displacement and the one obtained during the test was made.

Analysis of Rutting Performance Impacting Factors in HMA Overlay Mixture Design

It is reasonable to assume that hot-mix asphalt (HMA) overlay rutting occurs only in HMA overlay(s) because of rutting from the existing pavement, if any, most of which had already occurred before the HMA overlay(s) was placed. So, this paper focuses only on the HMA overlay itself. Many factors—such as the asphalt binder grade, the aggregate gradation, the types of aggregates, the volumetric parameters of the mixtures, etc.—potentially influence the HMA rutting performance when designing an HMA overlay mix. In this paper, a Hamburg Wheel Tracking Test (HWTT) device is employed in order to evaluate the rutting performance of HMA overlay mixes in terms of the rut depth at specified wheel passes. An experimental design is developed to cover the different combinations of aggregates, gradations, and asphalt binder types. Through extensive laboratory tests and statistical analyses, the most significant impacting factors were identified. The corresponding HWTT rut depth prediction models were developed with high R2 values (>0.80). With these models and established rut depth criteria, the maximum allowable asphalt content that can be used without causing rutting problems was estimated for different overlay mixes. Furthermore, the models and estimated maximum allowable asphalt content were preliminarily verified by performance data of the National Center for Asphalt Technology test track sections and WesTrack test sections.