城市雨水管网模型参数优化及应用研究

发布时间：2018-09-08 13:08

【摘要】：在城市内涝灾害频发和海绵城市建设的大背景下,城市雨洪模型逐渐成为一种有效的雨水管理手段,意在为解决好城市雨水排水的本质——雨水的空间分配问题出谋划策。然而,人们在应用时缺乏相应的模型校核,致使模拟结果与实际情况出现较大偏差,参数优化率定作为其应用准确性和可靠性的关键性问题近年来受到越来越多的重视。本文基于众多城市雨水管网模型软件一维模拟内核SWMM模型和参数识别最优化思想,对其水文水力模块参数自动优化问题进行了系统研究。基于SWMM模型对城市雨水管网数学模型组成及其模拟原理进行了详细介绍。在系统研究了SWMM模型水文水力参数取值范围的基础上,以C市某区雨水管网为例,采用修正的Morris筛选法对14个参数进行了局部灵敏度分析,得到了各参数对不同目标变量(径流总量、峰值流量、排放口水力负荷)的灵敏度分布及其稳定性分析结果。并引入子排水区面积和不渗透率修正系数最终确定了10个优化参数集合。研究了自适应混沌粒子群优化算法(ACPSO)。引入按指数函数非线性递减的自适应惯性权重及逻辑自映射混沌搜索技术对标准粒子群算法(SPSO)进行了组合改进。采用三个经典的Benchmark问题对其进行的性能测试表明:该算法能有效摆脱早熟收敛,收敛性能及稳定性能均得到加强,可用于参数优化研究。研究了SWMM模型参数优化问题。采用C语言二次开发了共10个外部API函数用于本文研究。以水位和流量为匹配对象构建了SWMM参数优化数学模型,并用ACPSO算法对其进行了求解。结果表明:用于优化和验证的场次降雨中,两个目标变量的纳什效率系数(NSE)均在0.70以上,峰值流量相对误差绝对值均在10%以内,内涝溢流验证结果也与实际基本吻合,经过参数优化的SWMM模型准确性和可靠性令人满意。作为城市雨水管网模型的应用,基于以上研究结果开展了C式某区管网运行性能评估。其雨水管网整体达标率为57.69%(满足设计重现期P(28)2a及以上),但与之相对的系统整体耐受度达到98.73%,说明了在该区域进行海绵城市开发建设的必要性和可行性。
[Abstract]:Under the background of frequent waterlogging disasters and the construction of sponge cities, urban rainwater flood model has gradually become an effective means of rainwater management, aiming at solving the problem of spatial distribution of rainwater, the essence of urban rainwater drainage. In recent years, more and more attention has been paid to parameter optimization and calibration as the key problem of its accuracy and reliability. Based on SWMM model of one-dimensional simulation kernel and parameter identification optimization thought of many urban rainwater network model software, the automatic optimization of hydrological and hydraulic module parameters has been studied in this paper. Based on SWMM model, the mathematical model of urban rainwater pipe network and its simulation principle are introduced in detail. On the basis of systematically studying the range of hydrological and hydraulic parameters of SWMM model, taking a rainwater pipe network in C city as an example, the modified Morris screening method is used to analyze the local sensitivity of 14 parameters, and each parameter is obtained. The sensitivity distribution and stability analysis results of parameters for different target variables (total runoff, peak flow, hydraulic load at discharge outlet) are given. The area of sub-drainage area and the impermeability correction coefficient are introduced to determine 10 optimal parameter sets. The adaptive chaotic particle swarm optimization algorithm (ACPSO) is studied. Nonlinearity by exponential function is introduced. Three classical Benchmark problems are used to test the performance of the standard particle swarm optimization (SPSO). The results show that the algorithm can effectively get rid of premature convergence, and its convergence and stability are enhanced. It can be used in parameter optimization research. The optimization of SWMM parameters is studied. A total of 10 external API functions are developed in C language for this study. The mathematical model of SWMM parameters optimization is constructed by matching water level and flow rate and solved by ACPSO algorithm. The results show that Nash effect of two objective variables is used to optimize and validate the field rainfall. The ratio coefficient (NSE) is above 0.70, and the absolute relative error of peak discharge is less than 10%. The results of waterlogging and overflow validation are basically consistent with the actual situation. The accuracy and reliability of SWMM model optimized by parameters are satisfactory. As an application of urban rainwater pipe network model, the performance evaluation of C-type pipe network in a certain area is carried out based on the above results. It is estimated that the overall standard-reaching rate of the rainwater pipe network is 57.69% (meeting the design recurrence period P(28)2a and above), but the relative overall system tolerance is 98.73%. This shows the necessity and feasibility of developing and constructing sponge cities in this area.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU992

【参考文献】