【摘要】 为更好地识别山地城市雨水管道淤积情况，以重庆市长寿区分流制雨水排水系统为例，提出通过数值模型与机器学习算法相结合的手段对雨水管道进行分类，建立了沉积物厚度、粒径分布等特性与各分类管段之间的关系。结果表明，基于SWMM模型及聚类分析的方法可以将研究区域的雨水管道分为7类，只有管径<DN1 200的第1、2、5类管段存在较严重的淤积风险，平均厚度为37.75 mm以上，淤积占比>4.9%，这主要是受管径小、坡度比大、转弯角度大等管段自身属性的影响导致冲刷时间较短，进而发生淤积。另外发现，山地城市雨水管道沉积物的粒径分布有着明显的两极分化现象，以>1 mm和<75μm的粒径居多。更多还原

【Abstract】 A combination of numerical modeling and machine learning algorithms was proposed to classify rainwater pipes belonged to the drainage system of Changshou District in Chongqing, aiming to better identify siltation in mountainous cities. The study also established the correlation between each classification pipe section and the properties such as sediment thickness and particle size distribution.Utilizing the SWMM and cluster analysis method, the rainwater pipes within the research area could be categorized into 7 categories. Only categories 1, 2, and 5 with pipe diameter less than DN1 200 exhibited significant risk of silting, characterized by an average thickness exceeding 37.75 mm and a silting ratio surpassing 4.9%. This was primarily attributed to inherent characteristics of the pipe such as small pipe diameter, steep slope ratio, and sharp turning angle, leading to rapid erosion and sediment accumulation.Secondly, it was observed that the particle size distribution of sediment in rainwater pipes in mountainous cities exhibited a distinct polarization phenomenon, with majority of the particles greater than 1 mm and less than 75 μm.更多还原