Under the dual regulation of government intervention and environmental constraints, user demand, low-carbon awareness, and the willingness of construction enterprise to reduce carbon emissions are crucial to the realization of low-carbon transportation infrastructure construction. Basing on prospect theory, this paper constructs an evolutionary game model for the government, construction enterprise, and user, explores the stabilization of the three-game players' strategy choices, and discusses the influence of key parameters on the behavior evolution of relevant players through simulation analysis. The results indicate that: the regulatory cost when the government adopts incentives or no incentives is an important factor affecting the strategy choice. The enhancement of incentives by the government can promote the construction of enterprise and the selection of low-carbon transportation infrastructure by user. However, when the benefit of government credibility outweighs the incremental cost of household subsidies, carbon subsidies, and regulation, the government changes its strategy and no longer adopts encouraging strategy. Improving benefits of build low-carbon transportation infrastructure reducing incremental costs, setting a ladder carbon tax rate, increasing carbon tax rate for traditional construction, and decreasing for low-carbon construction is an effective approach to stimulate enterprise to construct low-carbon transportation infrastructure. The user's choice of low-carbon transportation infrastructure strategy is positively correlated with the subsidy received and is less affected by environmental benefits. And the construction enterprise's strategy choice hasn't been influenced by user, which is related to the public service characteristics of transportation infrastructure.

The persistent carbon lock-in in transport infrastructure hinders low-carbon transition and sustainable urban development. This study, situated within the context of building energy and systems environments, examines the spatial and temporal dynamics of carbon lock-in across Chinese provinces over the past decade. A composite indicator system is constructed to quantify path dependence in transport-related carbon emissions. Additionally, a quantile regression model is applied to analyze the heterogeneous effects of influencing factors across varying levels of lock-in. The findings offer insights into developing region-specific decarbonization strategies and enhancing the sustainability of infrastructure systems. The key findings are as follows: (1) temporally, the composite path dependence index (PDI) increased from 0.13 to 0.24, with an average annual growth rate of 5.7%, although the growth trend exhibited noticeable fluctuations. (2) Spatially, a clear “East-High–West-Low, South-High–North-Low” pattern is observed, with the eastern region exhibiting a PDI approximately 29% higher than that of the northeastern region—mainly due to differences in economic vitality and the degree of policy intervention. (3) Threshold nonlinearity is evident: the influence of key factors on carbon path dependence varies across quantiles, and a critical threshold appears at the 0.9 quantile. Notably, the mitigating effect of low-carbon technological innovation on path dependence weakens beyond this threshold, suggesting diminishing marginal returns and the necessity for complementary governance strategies. This study deepens the understanding of transport carbon lock-in and provides guidance for region-specific strategies to enhance energy efficiency and sustainability in integrated building infrastructure systems.

To verify the feasibility of applying dry flue gas desulfurization ash (DFGDA) to asphalt pavement materials, the asphalt mastic (filler and asphalt composition) prepared by adding different proportions of DFGDA and LSP (limestone powder) into 70# matrix asphalt was studied experimentally. The asphalt mastics were subjected to the penetration test, the softening point test, and the ductility test. Moreover, the rheological properties of asphalt mastic were evaluated with dynamic shear rheometer (DSR) tests and bending beam rheometer (BBR) tests. An interaction ability index C-value based on the Palierne model was proposed to evaluate the interaction ability between DFGDA and asphalt. The influence of DFGDA asphalt on the interaction ability of matrix asphalt was observed and evaluated using Fourier transform infrared spectroscopy (FTIR) and scanning electron microscopy (SEM). The results showed that with the increasing proportion of DFGDA, the penetration of asphalt mastic gradually decreased, the softening point increased, and the ductility slightly decreased. At the same temperature, the dynamic shear modulus G* of the asphalt mortar significantly increased with increasing DFGDA content. The incorporation of DFGDA negatively affected the low-temperature plastic deformation resistance of asphalt, but the impact was weakened with the growing DFGDA amount and the powder mastic ratio. The combination mode of DFGDA and matrix asphalt depends on the physical blending, and their interaction ability mainly depends on the miscibility between DFGDA and matrix asphalt. In conclusion, DFGDA can be utilized as a novel filler in asphalt pavement materials.

With the rapid expansion of metro tunnel construction, safety risks such as collapse, water inrush, and gas explosion have become increasingly critical. Existing warning models often lack fine-grained disaster type identification and dynamic risk assessment capabilities. This paper proposes a two-stage intelligent warning framework based on multi-source data fusion. First, a dual-autoencoder structure (MLP-AE and LSTM-AE) extracts deep features from static geological parameters and dynamic monitoring sequences. Then, a multilayer perceptron (MLP) classifier identifies four typical states: normal, collapse, water/mud inrush, and gas explosion. Subsequently, a regression model predicts a continuous risk score, mapped to three risk levels: Safe, Moderate Risk, and Significant Risk. Experimental results demonstrate that, compared with Decision Tree (DT), Gradient Boosting Decision Tree (GBDT), and Bayesian Network (BN), the proposed framework achieves superior performance in risk level identification, with an accuracy of 91% and an F1-score of 0.87. Notably, it exhibits particularly strong recall for severe (Level III) risks, which is crucial for practical engineering applications. The proposed framework provides a practical and intelligent approach for safety warning in metro tunnel construction.

Plant‐mixed hot recycling technology enables efficient resource utilization of reclaimed asphalt pavement (RAP); However, the resulting mixtures often exhibit inferior anti‐cracking stability compared to virgin mixtures, and the cracking evolution mechanisms remain insufficiently understood. This study investigated the effect of six core production parameters: RAP preheating temperature, new aggregate preheating temperature, asphalt preheating temperature, mixing temperature, mixing regime, and RAP content. A mixed‐level orthogonal experiment was designed to conduct multi‐factor coupling tests. The Semi‐Circular Bending (SCB) test combined with Digital Image Correlation (DIC) technology was employed to precisely capture the entire cracking evolution process, including crack initiation, stable propagation, and unstable failure. Three evaluation metrics—Flexibility Index (FI), Crack Initiation Energy (CIE), and Energy Consumption Rate (ECR)—were adopted to systematically analyze the primary and secondary effects of each process parameter on anti‐cracking performance. Furthermore, meso‐structural uniformity analysis was incorporated to explore the relationship among process parameters, meso‐structural characteristics, and macro‐scale anti‐cracking performance. The results indicated that production process parameters affect the comprehensive performance of recycled mixture in the following descending order: mixing time regime > RAP content > RAP preheating temperature > new aggregate preheating temperature > asphalt preheating temperature > mixing temperature. Notably, mixing regime appeared to be the sole parameter capable of simultaneously controlling both meso‐structural uniformity and macro‐scale anti‐cracking performance, with unequal‐step multi‐stage mixing proving superior to equal‐step mixing. The findings suggested a differentiated control mechanism for crack initiation and propagation: the former was primarily associated with aggregate‐asphalt interfacial bonding, while the latter was linked to mastic integrity and aggregate skeleton stability. The optimal process combination balancing both crack initiation and propagation resistance was established as follows: RAP content 40%, RAP preheating temperature 150°C, new aggregate preheating temperature 200°C, asphalt preheating temperature 180°C, mixing temperature 180°C, and 4‐min unequal‐step staged mixing. These findings offer both theoretical foundations and technical guidance for optimizing production processes and controlling cracking in plant‐mixed hot recycling asphalt mixtures.

Mountain tunnel construction involves complex geological conditions and frequent cross‐operations, making safety accident prevention particularly challenging. To enhance the foresight and specificity of construction safety management, this paper develops a feature element system for safety accidents in mountain tunnel construction from the “object–cause–consequence” three‐dimensional perspective, and establishes a safety accident early warning model that integrates association rule mining and text vectorization techniques. First, based on 145 real‐world mountain tunnel construction safety accident cases, the TF‐IDF algorithm and k‐means clustering are used to extract and condense 6 types of object elements, 5 categories (20 specific items) of cause elements, and 10 types of consequence elements. Second, the Apriori algorithm is applied to mine strong association rules among “object–cause–consequence” triplets, resulting in a corpus of 41 accident rule chains. The Word2vec word embedding model and Pearson similarity calculation are then combined to match target cases against these rule chains. Finally, a simple majority principle and a fluctuation principle are introduced to define warning thresholds, establishing a three‐tier early warning mechanism. Two empirical studies are carried out: one on the XCS tunnel, where an accident actually occurred, and the other on the DPS3 tunnel of the GN railway, which has not experienced accidents. The results indicate that the XCS tunnel triggered a red warning in the excavation zone, predicting a possible collapse or roof fall accident, which aligns well with the actual collapse that happened. The DPS3 tunnel, in its support and lining zone, also triggered a red warning, signaling the risk of collapse and roof fall. These findings provide methodological support for dynamic safety control and precise early warning in mountain tunnel construction.