This paper investigates the application of artificial neural networks for predicting road traffic accident counts in Poland and Kosovo. The study uses annual accident data collected for the period 2010–2023 from official statistical records provided by the Polish Police and the Kosovo Police. Forecasting procedures were carried out with multilayer perceptron (MLP) models developed in the Statistica environment. To assess the robustness of the models, two alternative dataset division schemes were employed, namely, 70%–15%–15% and 80%–10%–10% for the training, testing, and validation subsets. Forecast accuracy was measured using Mean Absolute Error (MAE) together with Mean Absolute Percentage Error (MAPE). The obtained forecasts suggest that the number of road traffic accidents in both analysed countries is likely to remain relatively stable over the 2024–2030 period. The analysis further demonstrates that a larger share of training observations contributes to improved predictive performance and lower estimation errors. Nevertheless, the relatively short time series constitutes a methodological limitation and requires careful interpretation of the forecasting results.
In view of the shortcomings of large dead-weight and limited spanning capacity for traditional composite girders and the fatigue failure and pavement deteriorations for steel girders, composite girders with lightweight composite deck emerged as good alternatives. Thus, trial design schemes of long-span cable-stayed bridges with steel-NC (normal concrete) and steel-UHPC (ultra-high performance concrete) composite decks were proposed under different span conditions. The static performance and technical feasibility of the designed bridges and the influence of structural parameters of main girder on different static effects were analysed, based on which, the determinative static effects and their variation with span length were investigated. The results show that the designed NC-slab schemes with main spans no more than 900 m can meet the requirements of static strength, stiffness and stability, while the 1000 m NC-slab scheme cannot meet the compressive strength requirements of the concrete slab. The UHPC-slab schemes meet the static requirements even when the main span reaches 1100 m, and the most unfavourable static effect turns to be the maximum compressive stress of steel girders. The main structural parameters of the main girder that most affect the unfavourable static effects include the main girder height and the thickness of the lower flange steel plate. Using UHPC-slab instead of NC-slab improves the spanning capacity of cable-stayed bridges by approximately 25%, close to the reduction rate of the dead-weight of main girder.
A localized landslide that occurred on a specific section of the road to the Uluyazı Campus of Çankırı Karatekin University seriously threatened road safety and disrupted transportation continuity. In this study, the ground conditions of the landslide area were examined in detail. Soil parameters were determined through field observations, drilling, Standard Penetration Test (SPT) data, laboratory experiments, liquefaction potential, bearing capacity, and consolidation behaviour were analysed. The analyses showed that the soil was safe in terms of liquefaction and that the bearing capacity was at a level suitable for road construction and use. However, consolidation settlements and dissolutions in gypsum levels were identified as critical factors that could trigger landslides. To prevent landslides and ensure road safety, double rows of bored piles with a diameter of ∅120 cm, a length of 35 m, and spaced at 1.4 m intervals were connected with a cap beam. This engineering measure cut the landslide surface, limited road deformations, and ensured transportation safety. The study demonstrated the effectiveness of bored pile applications in formations where soil-water interaction was significant and offered an applicable solution for critical areas in highway engineering.
This study investigates how traffic heterogeneity affects both traffic-flow efficiency and bridge structural demand using one year of road-based Weigh-in-Motion (WIM) data collected on a 60 m two-span simply supported prestressed-concrete girder bridge in Nanjing, China. The objective was to develop an integrated framework for quantifying the joint influence of vehicle composition on traffic performance and bridge load effects, and to evaluate whether operational mitigation measures can improve both mobility and structural performance. The methodology combined data preprocessing, vehicle classification, macroscopic traffic-flow modelling, statistical analysis, influence-line-based structural load-effect estimation, machine-learning prediction, and simulation-based intervention evaluation. Traffic-flow relationships were analysed using the Greenshields’s, Greenberg’s, and Underwood’s models, while XGBoost and SHAP were applied to predict and interpret traffic and structural indicators derived from the processed WIM dataset. The results showed that increasing heavy-vehicle proportion reduced traffic efficiency by lowering flow and speed and increasing density, while simultaneously increasing predicted bending moment, shear force, and fatigue-related demand. Among the calibrated traffic-flow models, the Underwood’s model achieved the lowest RMSE, while the Greenshields’s model remained highly competitive and was retained because of its simpler and more interpretable formulation. The scenario analysis further indicated that both lane management and speed harmonization reduced structural demand relative to the baseline observed traffic condition, with lane management providing the greater overall benefit. The proposed framework is intended as an analytical and scenario-based decision-support methodology for evaluating bridge traffic-management strategies under heterogeneous traffic conditions.
This study established the ratio of the base layer's elastic modulus to that of the surface layer Rm as a key variable to systematically investigate how the base layer's elastic modulus influences the mechanical response and service life of flexible pavement. A three-dimensional finite element model of a three-layer pavement system was analysed, using EverStressFE, with a constant surface layer modulus and a variable base layer modulus. The analysis included two important interfacial bonding conditions: full bonding and full slip. It also considered that the actual wheel loads are not uniform but rather follow concave and convex distribution patterns. The mechanical responses, including deflection at the top of the asphalt layer, tensile strain at the bottom of the asphalt layer epsilon xx, and vertical compressive strain at the top of the subgrade epsilon zz, were quantified to predict fatigue life (for cracking) and rutting life (for permanent deformation). The results indicated that the maximum deflection reached 0.53 mm under a full slip condition with a convex load distribution at Rm = 0.75. Critical tensile strains at the bottom of the asphalt layer were most severe under full slip with a convex load, reaching 348 & times; 10-6, while the fully bonded, concave case resulted in a much lower value of 83 & times; 10-6. Similarly, the maximum vertical compressive strain on the subgrade was 365 & times; 10-6 for the fully slipped, convex case compared to 250 & times; 10-6 for the fully bonded, concave case. A lower Rm value under full slip with a concave load distribution significantly reduced the pavement lifespan, with predicted fatigue life decreasing by over 60% and rutting life by nearly 45% compared to the fully bonded case. On the other hand, a convex load distribution greatly increased the pavement's bearing capacity by raising the critical strain thresholds and lengthening the expected service life. These findings underscore the paramount importance of interface bonding and load distribution patterns, suggesting that they can outweigh the influence of the base layer's modulus alone on pavement design.
