基于贝叶斯理论校核管网压力驱动模型的研究  

Research on checking pressure-driven model of pipeline network based on Bayesian theory

作　　者：王彤[1] 朱多林 雷春元 崔红军 刘永佳 宋思怡 WANG Tong;ZHU Duolin;LEI Chunyuan;CUI Hongjun;LIU Yongjia;SONG Siyi(School of Civil Engineering,Chang'an University,Xi'an 710018,China;Xi'an Water Supply Co.,Ltd.,Xi'an 710082,China;Xi'an Water Group Planning and Design Research Institute Co.,Ltd.,Xi'an 710061,China)

机构地区：[1]长安大学建筑工程学院,西安710018 [2]西安市自来水有限公司,西安710082 [3]西安水务(集团)规划设计研究院有限公司,西安710061

出　　处：《给水排水》2023年第5期170-177,共8页Water & Wastewater Engineering

摘　　要：基于贝叶斯理论,利用粒子MCMC算法建立了供水管网压力驱动模型的校核方法。该方法采用一种双层嵌套循环结构,利用压力监测值在内层通过集合卡尔曼滤波原理更新节点流量,外层采用改进的马尔科夫链蒙特卡洛抽样算法校核模型的海曾-威廉系数和漏失系数。实际应用结果表明:该方法能够实现高维、非线性系统的估计,利用压力监测值通过内外层循环运行实现节点流量和两个系数的联合校核,避免补偿误差,提高了校核效率。Based on Bayes theory,the verification method of pressure-driven model of water supply network is established by using particle MCMC algorithm.The method uses a double-layer nested loop structure to update node traffic on the inner layer using pressure monitoring and the principle of collective Kalman filter,the model's Hazen-William coefficients and loss coefficients are checked using an improved Markov chain Monte Carlo sampling algorithm.The practical application results show that this method can realize the estimation of Galway and nonlinear system,and the joint check of node flow and two coefficients can be realized by using the pressure monitoring data through inner and outer loop operation,so as to avoid the compensation error,improve the efficiency of checking.

关 键 词：模型校核 压力驱动 贝叶斯理论 粒子MCMC算法 状态空间模型 

分 类 号：TU990[建筑科学—市政工程]