中文摘要：

根据再生水消毒工艺的进水水质和出水总余氯浓度，通过联合使用主成分分析和Logistic回归方法，建立了工艺出水的总余氯风险预测模型，与实测的清水池出水余氯风险相比，模型预测准确度〉95％；通过分析不同余氯风险下各水质样本的pH值、水温、氨氮浓度和COD浓度，确定水温和NH3-N浓度是影响出水余氯风险的主要因素；结合风险预测和影响分析结果，划分了出水余氯风险等级，并为不同风险等级分别拟定了应对措施。

英文摘要：

Based on the influent and effluent quality of reclaimed water disinfection process, a risk prediction model for effluent residual chlorine concentration was developed using principal component analysis and Logistic regression methods. The model was tested with the measured residual chlorine concentration, and the accuracy on risk prediction was higher than 95%. The effects of influent parameters including pH value, water temperature, ammonia nitrogen concentration and COD concentration on the risk prediction model were investigated, and the results indicated that temperature and ammonia nitrogen were the main factors affecting the effluent residual chlorine concentration. The effluent residual chlorine risk was divided into three levels and a set of resolutions was proposed for the risk of each level.