数据驱动的城市高架车站立面光伏设计及改造潜力研究OACHSSCD

Given the in-depth advancement of the"carbon emission peak and carbon neutrality"strategy,the green and low-carbon transformation of urban rail transit systems has become a significant development direction for the industry.By 2025,the total operational mileage of urban rail transit in China exceeded 10,978 km.The elevated stations have significant potential for integration with photovoltaic(PV)power generation due to their large roof areas and favorable solar radiation conditions.However,existing research and applications primarily focus on rooftop PV systems,with insufficient attention paid to PV on building facades.In fact,elevated stations are often located in suburban areas with minimal surrounding building shading,allowing annual energy outputs several times those of rooftop systems,indicating substantial development potential.However,the introduction of facade PV systems will alter the thermal performance of building envelopes and the indoor lighting environment,presenting a challenge in simultaneously enhancing renewable energy output while ensuring thermal comfort and natural lighting for passengers.To address this issue,this research aims to reveal the renovation application potential of PV systems on elevated station facades,establishing a multi-objective collaborative optimization design method that integrates PV power generation,indoor thermal comfort,and natural lighting performance,thus providing a scientific reference for the green and low-carbon transformation of urban rail transit.A case study based on field research and morphological classification of elevated stations in Shanghai was conducted,and four common benchmark models(two-story and three-story side platforms,with and without skylights)were constructed.Initially,Ladybug Tools and PVsyst software were used to simulate and compare solar radiation and power generation performance for different facade and roof orientations,demonstrating the feasibility of facade PV applications.Subsequently,typical stations were selected to build simplified performance models,with 12 design variables including orientation,window-to-wall ratio,skylight roof area ratio,and shading device angle and width.A total of 1,020 samples were generated using the Monte Carlo sampling approach,and performance simulations were conducted using EnergyPlus.To address the time-consuming nature of physical simulation calculations,a LightGBM regression model was trained as a surrogate,achieving R2 values exceeding 0.93,accompanied by mean absolute errors below 0.2.Ultimately,a multi-objective optimization focusing on the goals of thermal comfort time percentage(TCP),useful daylight illuminance(UDI),and annual PV power generation was performed on the Wallacei Platform using the NSGA-II algorithm.It generated 2,000 design options and extracted the Pareto optimal solution set.The study finds that the efficiency of facade PV systems could reach over 80%,only 6%to 8%lower than that of rooftop systems,and that the annual energy outputs of east-west-oriented facades could exceed 1.5 times those of rooftop systems.The maximum annual output of the largest application area facade model exceeded that of the skylight roof model by 150%.After optimization,TCP reached a maximum of 11.84%,UDI reached 71.70%,and annual PV power generation reached 963,669 kWh,an improvement of 122%over the baseline scenario.Based on the Pareto front solution set,the study further summarized engineering practice-oriented design renovation methods:the greatest PV potential lies in east-west orientations,followed by north-south orientations;shading device angles are recommended to be between 40 ° and 50°,with horizontal positioning for east-west orientations and enhanced shading for west-facing sides;the window-to-wall ratio is recommended to be 0.2～0.4,along with a skylight roof area ratio of 0.2～0.3;and the angle between PV panels and the facade is suggested to be 0 °～30°,with integrated prefabricated PV shading modules proposed for low-intervention installation.An integrated technical process of"parametric modeling-performance simulation-machine learning-multi-objective optimization"was established,enhancing computational efficiency by nearly 50%through the use of a LightGBM surrogate model and overcoming the time-consuming limitations of traditional physical simulation calculations.It achieved accurate,fast prediction and automatic optimization of complicated performance relationships.The proposed design and renovation strategies for elevated station facade PV systems encompass orientation selection,integrated shading component design,aperture parameter optimization,and PV component integration.It offers directly applicable parameter suggestions and decision-making references for collaborative design during the proposal phase of new stations and low-intervention upgrades of existing stations.This technological framework exhibits strong portability and can be extended to other urban elevated stations and similar transportation infrastructure.It plays a significant role in promoting the deep integration of PV technology in the transportation sector and facilitating the achievement of the"carbon emission peak and carbon neutrality"goals.