Dynamic model for predicting the strength of road surfaces based on georadar data

Abstract. Problem. The current state of Ukrainian automobile roads is characterized by significant wear level of road pavements caused by intensive growth of transport loads, increase in heavy transport proportion and climatic factors influence. Premature destruction leads to significant economic losses, traffic safety deterioration and road capacity reduction. Subgrade soil moisture is a key factor determining road pavement durability. Research proven that on Category III-IV roads with "weak" strength structures, subgrade soil moisture most significantly affects bearing capacity. The greatest contribution to total strength is made by subgrade soil strength (50.2%) for "weak" structures on light silty clay subgrade. Increasing soil moisture from 0.55 Wт to 0.80 Wт leads to 83.9% bearing capacity loss. Traditional soil moisture determination methods are labor-intensive, require pavement integrity disruption and lack necessary operational control efficiency. Goal. To develop a dynamic model for predicting road pavement strength that uses ground-penetrating radar sounding results to account for subgrade soil moisture influence on road structure degradation processes. Methodology. Development of dynamic methods for accounting subgrade soil moisture influence using Taylor series expansion for predicting deviations from baseline models. The methodology involves decomposition of deviation into random component and component related to subgrade soil moisture and deformation characteristics changes over time. For "weak" pavement structures most sensitive to moisture changes, a modified dynamic modeling algorithm is proposed with correction accounting for subgrade soil moisture influence. A baseline model characterizing safety factor coefficient change over time and residual service life prediction algorithm are developed. Results. A dynamic adaptive model for predicting road pavement strength has been developed integrating georadar sounding data. The model accounts for correction between model and georadar-determined safety factor coefficient. A baseline model for safety factor coefficient based on shear resistance in subgrade soils and residual service life prediction algorithm were created. Forecasting methodology using dynamic modeling with parabolic interpolation and adaptive correction over three years was implemented. Originality. The research combines empirical and mechanical approaches within unified practical model for assessing and predicting road pavement structure strength based on georadar sounding results, developing and improving existing diagnostics and forecasting methods for flexible pavements. Practical Value. The approach reduces initial data set and increases forecasting efficiency. Sufficient predictive accuracy is achieved using three years of observations. Ground-penetrating radar enables operational subgrade soil condition assessment and moisture determination based on electrophysical properties, allowing continuous linear monitoring at relatively low costs.