Author(s)

Quan Wang

Corresponding Author

Quan Wang

Abstract

Under the background of China's "double carbon" strategy, the traditional "quick discharge" drainage system is in urgent need of low-carbon transformation due to high energy consumption, high emission and ecological fragmentation effect. In this study, a multi-objective collaborative optimization framework of "low-carbon-elasticity-economy" was constructed: ① A dynamic carbon footprint evaluation system covering the whole life cycle was established, and the carbon emission of "grey-green-blue" facilities was quantified in real time by SWMM-LCA (storm water management model-life cycle evaluation) coupling model, and the concept of "carbon sensitivity" was put forward to reveal its collaborative mechanism; ② In the planning and design stage, SWMM-LCA iteration is driven by NSGA-II algorithm to optimize the layout and scale of pipe network and green infrastructure, and realize Pareto optimization of carbon emission, waterlogging risk and cost; ③ In the operation stage, an intelligent control platform based on deep reinforcement learning (DRL) is developed, with the real-time hydrological state as the input, the start-stop of the pumping station and the opening of the gate are dynamically adjusted, so as to reduce the operation energy consumption. The empirical study shows that compared with the traditional scheme, the carbon emission of the optimized "gray-green combination" scheme is reduced by 40%, the waterlogging amount is reduced by 72%, and the life cycle cost is only increased by 3.3%. DRL real-time control can further save energy by 23.5% in a single rainstorm, and the annual emission reduction benefit is remarkable. The integrated framework of "optimal design-intelligent operation" proposed in this study provides feasible technical path and decision support for the low-carbon transformation of urban drainage system.

Keywords

Low carbon, green design, operation optimization, urban drainage system, SWMM-LCA, deep reinforcement learning