Ceramic water filters (CWFs) are manufactured worldwide using local materials and infrastructure. In this study, we assessed flow rate (FR) and nominal pore diameter (NPD) values as parameters to predict the microbial removal of CWFs. Two empirical models (flow rate model, FRM, and nominal pore diameter model, NPDM) were developed based on the log removal values (LRVs) for total coliform obtained from the operation of CWFs manufactured under controlled conditions in the United States. The proposed empirical models were validated using CWFs manufactured in Nepal and India. The models and principal component analysis (PCA) were used to identify materials and processes in CWF manufacturing with the highest impact on LRV. Our results showed that both the FRM and NPDM have a good predictive capability with high coefficients of determination, R², and statistical significance (p) values less than 0.05. PCA showed that the burnout material was the most important variable in the manufacturing process. Even though both models are appropriate to predict LRV from a CWF, the FRM and NPDM could have slightly different applications at CWF factories. The FRM could be used to predict LRVs of CWFs manufactured in an already operational factory as it fits with existing quality control procedures, such as measurement of flow rate. We propose that the NPDM could be applied in the research and performance enhancement of CWFs as additional filter characteristics can be tested for the potential to improve LRV. The use of either model could help factories improve the removal of bacteria by their CWFs.

中文翻译：

---

评估流量和标称孔径作为预测陶瓷滤水器去除微生物的参数

陶瓷滤水器（CWF）在全球范围内使用当地材料和基础设施制造。在这项研究中，我们评估了流速（FR）和标称孔径（NPD）值作为预测CWF微生物去除的参数。基于美国在受控条件下生产的CWF的操作获得的大肠菌群的对数去除值（LRV），开发了两个经验模型（流速模型FRM和标称孔径模型NPDM）。拟议的经验模型已使用尼泊尔和印度制造的CWF进行了验证。使用模型和主成分分析（PCA）来识别CWF制造中对LRV影响最大的材料和过程。我们的结果表明，FRM和NPDM都具有良好的预测能力，具有较高的测定系数R²，且统计显着性（p）值小于0.05。PCA表明，燃尽材料是制造过程中最重要的变量。即使这两种模型都适合从CWF预测LRV，但FRM和NPDM在CWF工厂的应用可能会略有不同。FRM可以用来预测在已经运行的工厂中生产的CWF的LRV，因为它符合现有的质量控制程序，例如流量的测量。我们建议将NPDM应用于研究和CWF的性能增强，因为可以测试其他过滤器特性来改善LRV。使用这两种模型都可以帮助工厂提高其CWF对细菌的去除率。