According to the World of Petroleum and Bitumen Journal, a new research framework has brought asphalt pavement surface texture design one step closer to a fully digital environment. The approach digitally reconstructs the three-dimensional texture of an asphalt surface by using the actual shape of aggregate particles and simulating how they are arranged and interact during compaction. Developed by researchers from Chang’an University and their collaborators in France, the framework combines high-resolution three-dimensional laser scanning with the Contact Dynamics method to create digital asphalt specimens and reconstruct their surface texture before physical production. The study was publicly highlighted on September 25, 2026.
The research addresses a practical challenge in pavement engineering. Pavement surface texture is closely linked to how a road interacts with vehicle tires and water, affecting properties such as skid resistance, drainage, and surface performance. However, the final texture of an asphalt pavement remains highly dependent on aggregate gradation, particle shape, particle arrangement, and construction conditions. Engineers can measure texture after constructing a pavement or laboratory specimen, but predicting the final three-dimensional surface from a mix design is more difficult. The new framework seeks to narrow this gap by digitally reconstructing the surface based on the internal structure of the material particles.
The basic concept is relatively straightforward. Instead of treating pavement texture as an independent surface property, the researchers model it as the geometric result of how individual aggregate particles are positioned near the surface of the asphalt layer. If particle shape, size distribution, and arrangement are realistically represented, the resulting surface can be simulated before producing a physical specimen or constructing the pavement. The framework therefore moves toward a predictive approach to pavement design that is not solely dependent on repeated physical testing and iterative mix adjustments.
To establish a realistic digital foundation, the researchers used a high-resolution three-dimensional laser scanner to capture the actual geometry of aggregate particles. Approximately 500 aggregate particles ranging from 4.75 to 26.5 mm were scanned and converted into three-dimensional models. Their elongation, flatness, and sphericity were then measured to create a statistical library of irregular particles rather than relying on simplified geometric shapes. This is important because aggregate shape affects how particles fit together and ultimately influences the structure observed at the pavement surface.
The study also examined the computational cost of using detailed particle geometry. The initial aggregate models were simplified to reduce computational requirements while preserving their key geometric characteristics. The results showed that reducing the particle models to approximately 60–65 facets kept deviations in the main shape indices within about 4% while reducing the computational cost of the simulations.
For numerical simulation, the researchers selected the Contact Dynamics method instead of conventional elastic contact models commonly used in the Discrete Element Method. This approach enables interactions between irregular particles to be modeled through nonsmooth contact mechanics and is suitable for simulating angular aggregates. The virtual mixtures were compacted under controlled conditions to reproduce the formation of the aggregate skeleton and the resulting surface. Particles smaller than 2.36 mm were also represented through an implicit asphalt mastic phase to reduce computational costs.
Two types of asphalt mixtures were investigated: AC-13, representing a dense-graded mixture, and SMA-13, representing a stone mastic asphalt mixture. The measured surfaces showed clear differences between the two mixtures. The Mean Profile Depth of SMA-13 was approximately 47% higher than that of AC-13, indicating a more pronounced macrotexture. The researchers also observed differences in the distribution of surface peaks and valleys, reflecting the different aggregate structures produced by the two gradation types.
One of the key technical findings was that direct reconstruction based solely on aggregate geometry produced an excessively rough surface. In an actual pavement, asphalt mastic fills part of the spaces between aggregate particles and smooths portions of the surface. A digital model that considers only aggregate geometry does not automatically reproduce this effect.
To address this issue, the researchers introduced a surface reconstruction method based on bridging. The method adjusts the elevations of the digital surface to simulate the continuity and smoothing effect produced by asphalt mastic. Following calibration, the method significantly improved the agreement between the simulated texture and the measured surface. For AC-13, the deviation in Mean Profile Depth decreased from 160.9% to 27.4%, while for SMA-13 it decreased from 119.1% to 18.2%. Deviations in the surface roughness parameters Sa and Sq were also reduced by more than 80%.
The study also provides a more detailed explanation of how aggregate size influences the final surface. Larger particles played the primary role in determining aggregate presence at the surface, while medium and smaller particles made a significant contribution to the formation of surface peaks. In AC-13, the 2.36–4.75 mm and 4.75–9.5 mm size ranges together accounted for approximately 98% of the particles associated with surface peaks. In SMA-13, particles in the 9.5–13.2 mm range had the greatest presence at the surface, while particles between 4.75 and 9.5 mm accounted for approximately 56% of the surface peaks.
These findings are relevant to pavement design because they directly connect mixture gradation and aggregate shape with measurable surface characteristics. Rather than simply selecting a gradation and determining the resulting texture after construction, engineers could eventually use a digital model to compare different mixture structures before committing materials to physical testing.
The practical implications of the technology also extend to pavement safety and durability. Surface texture plays an important role in tire–pavement interaction, particularly in wet conditions. A better understanding of how aggregate structure generates macrotexture could eventually help engineers evaluate different mixtures more predictively in terms of properties such as skid resistance and drainage. The current study is not a commercial pavement design system capable of directly replacing physical testing, but it provides a framework for quantitatively evaluating surface geometry in a virtual environment.
The research is also connected to the broader development of digital twins in road engineering. A full digital twin typically goes beyond a static three-dimensional model by connecting a physical asset with measurements, simulations, and performance changes over time. The present research should therefore be viewed as one building block within this broader concept rather than as a complete digital twin platform. Its main contribution is the ability to reconstruct surface texture based on the actual geometry of materials and simulate how that texture develops through particle arrangement and compaction.
Artificial intelligence is not the primary computational method used in this study. The main tools include three-dimensional laser scanning, aggregate shape analysis, and numerical simulation based on the Contact Dynamics method. This distinction is important because the research belongs to the broader field of data- and physics-based pavement engineering rather than representing an artificial intelligence product that is already ready for field deployment.
The approach also has limitations. The researchers noted that the bridging parameters used to represent the smoothing effect associated with the binder are numerical parameters rather than intrinsic and universal material properties. They may therefore require recalibration for different aggregate types, binders, or contact conditions. In addition, the computational cost of simulating realistic polyhedral particles currently limits the method primarily to digital reconstruction and small-scale investigations, and it is not yet ready for routine use at full project scale.
For asphalt producers and pavement contractors, the potential value of this technology will become clearer as the method develops. If these approaches become faster and can be applied to different materials with sufficient accuracy, digital screening could reduce the number of physical trial mixtures required during the early stages of development. It could also establish a more precise connection between aggregate selection, gradation, and expected surface performance.
For the broader bitumen and asphalt market, the main significance of this research is not the replacement of conventional materials but the improvement of how they are designed and evaluated. The study does not introduce a new bitumen grade or indicate that conventional asphalt binders are becoming obsolete. Instead, it adds a digital layer to the engineering process and demonstrates that surface texture can be reconstructed based on the geometry and organization of the materials that create it.
As of September 27, 2026, the study should be regarded as a developing research tool rather than a ready replacement for laboratory testing or field validation. Its most important technical contribution is demonstrating the relationship between actual aggregate shape, virtual compaction, the effects of asphalt mastic, and measured three-dimensional surface texture.
If further developed, this approach could allow future pavement design to incorporate virtual testing alongside conventional laboratory testing. Engineers could compare different aggregate structures, simulate compaction, examine surface texture, and identify potential performance differences before constructing a full physical pavement section.
For the asphalt industry, this development could represent an important shift: from measuring road texture after construction toward predicting and optimizing it before construction. The technology has not yet reached its final commercial stage, but the September 2026 study represents a technically significant step toward predictive, simulation-based pavement design.
By WPB
Asphalt, Bitumen, Technology, Innovation
