For many remote and under-privileged communities, water filtration is the first step of treatments before boiling and/or other higher level treatments. Clogging is a common unavoidable hindrance to such systems, which causes reduced efficiency with regard to the flow rate of filtered water. Backwashing is commonly used to recover the filtration capacity of clogged filter media. This paper presents an improvement of a previously developed simple clogging and backwashing efficiency model for the third-generation IHE family filter, which was developed by Unesco-IHE Institute for Water Education. The earlier simple model considered a linear decay of flow rate due to clogging in a weekly timestep. With the aim of enhancing the model's accuracy, this work considered three other models (a linear model with a daily timestep, a power function model and an exponential model). The performance of the models was assessed using the root mean squared error and chi-squared (χ²) values derived from predicted and measured flow rates through the device during successive clogging and backwashing periods for up to 4 months. The improvements using the linear 1 day model over the linear 7 day model were insignificant. Among the tested models, the power function model provided the best predictions, with an average χ² improvement of 33% over the previously developed linear model.

中文翻译：

---

改进的堵塞和反冲洗模型，用于过滤砷污染的水

对于许多偏远和贫困的社区，在煮沸和/或其他更高级别的处理之前，水过滤是处理的第一步。堵塞是此类系统的常见不可避免的障碍，导致过滤水的流量效率降低。反洗通常用于恢复堵塞的过滤介质的过滤能力。本文介绍了由联合国教科文组织-IHE水教育学院开发的第三代IHE系列过滤器先前开发的简单堵塞和反冲洗效率模型的改进。较早的简单模型考虑了由于每周时间步长阻塞导致的流量线性衰减。为了提高模型的准确性，这项工作考虑了其他三个模型（具有每日时间步长的线性模型，幂函数模型和指数模型）。使用均方根误差和卡方（χ ²）值从预测和测量流率通过该装置连续堵塞和长达4个月反洗段期间导出。使用线性1天模型相对于线性7天模型的改进并不明显。在这些测试的模型，幂函数模型提供了最好的预测，平均χ ²在先前开发的线性模型改进的33％。