Embodied GHG emissions of high speed rail stations: Quantification, data-driven prediction and cost-benefit analysis

• Published in: Journal of cleaner production Vol. 366; p. 133006
• Main Authors: Jia, Siyi, Liu, Minghui
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.09.2022
• Subjects: carbon; China; Cost-benefit analysis; Embodied GHG emissions; High speed rail stations; infrastructure; inventories; multivariate analysis; Neural network; prediction; rail transportation; risk; China; Cost-benefit analysis; High speed rail stations; Neural network; Embodied GHG emissions
• DOI: 10.1016/j.jclepro.2022.133006
• ISSN: 0959-6526

The embodied green-house gas (GHG) emissions from the high speed rail (HSR) stations are non-negligible given the considerable scale of modern HSR stations that serve as multifunctional complex. However, the magnitude of the GHG emissions is often underestimated owing to the limited use of reliable inventory data, as well as the simplified modelling that fails to capture the versatile structural/architectural features. To fill the knowledge gap, this paper reviewed the design documents of 117 HSR stations in China to provide more reliable quantification and reveal the factors associated with the GHG emissions. The embodied GHG emissions of the HSR stations were quantified by use of the actual inventory data. Based on the quantification results, a data-driven prediction model was constructed by an optimized general regression neural network (GRNN), which identified a link between the seven design parameters of HSR stations and the embodied GHG emissions. In addition, a cost-benefit analysis was conducted considering the balance between the embodied GHG emissions and serviceability. and discussion The results revealed a non-negligible contribution of HSR stations to the embodied GHG emissions of rail infrastructure. For regional HSR lines, the embodied GHG emissions from the stations could account for over 26% of the whole line. The results also revealed the large stations to be the major source of emissions, for which the embodied emissions per area are 2.84 times of the smaller stations. A Pareto-type distribution exists as the ten largest stations (in station area) contribute nearly 60% of the total emissions of the 117 samples. A canonical correlation analysis (CCA) revealed the station area to be the most prominent factor to the embodied GHG emissions, and the optimized GRNN with CCA-processed input vector make more reliable prediction. The regional HSR lines are subject to higher risk of unsound carbon investment, owing to the low service level of some stations. [Display omitted] •Real inventory data of 117 sample HSR stations in China are used for quantification.•Large stations are responsible for more embodied emissions per station area.•The top 10 largest stations contribute about 60% of the samples' total embodied GHG emissions.•Station area is the most prominent factor affecting the embodied GHG emissions.•Stations of regional HSR lines are more vulnerable to unsound carbon investment due to low service level.