Centro para la Optimización y Probabilidad Aplicada (COPA), Departamento de Ingeniería Industrial, Universidad de los Andes, Bogotá, Colombia.
Centro de Investigaciones en Ingeniería Ambiental (CIIA), Departamento de Ingeniería Civil y Ambiental, Universidad de los Andes, Bogotá, Colombia E-mail:
Water Sci Technol. 2019 May;79(9):1727-1738. doi: 10.2166/wst.2019.172.
The planning and scheduling of maintenance operations of large conventional sewer systems generate a complex decision-making environment due to the difficulty in the collection and analysis of the spatiotemporal information about the operational and structural condition of their components (e.g. pipes, gully pots and manholes). As such, water utilities generally carry out these operations following a corrective approach. This paper studies the impact of the spatiotemporal correlation between these failure events using Log-Gaussian Cox Process (LGCP) models. In addition, the association of failure events to physical and environmental covariates was assessed. The proposed methods were applied to analyze sediment-related blockages in the sewer system of an operative zone in Bogotá (Colombia). The results of this research allowed the identification of significant covariates that were further used to model spatiotemporal clusters with high sediment-related failure risk in sewer systems. The LGCP model proved to be more accurate in comparison to those models that build upon a fundamental assumption that a failure is equally likely to occur at any time regardless of the state of the system and the system's history of failures (i.e. a homogeneous Poisson process model).
由于大型传统污水系统的运行和结构状况的时空信息收集和分析具有难度，其维护操作的规划和调度会产生复杂的决策环境（例如，管道、雨水口和检查井）。因此，水公用事业公司通常会采取纠正措施来进行这些操作。本文使用对数高斯 Cox 过程（LGCP）模型研究了这些故障事件之间的时空相关性的影响。此外，还评估了故障事件与物理和环境协变量的关联。所提出的方法应用于分析哥伦比亚波哥大（Bogotá）一个作业区污水系统中的泥沙相关堵塞。这项研究的结果能够识别出重要的协变量，进一步用于对具有高泥沙相关故障风险的污水系统中的时空集群进行建模。与那些基于故障发生的可能性在任何时间都是相同的基本假设的模型（即均匀泊松过程模型）相比，LGCP 模型被证明更加准确。
