• 关键词：厌氧处理  支持向量机(SVM)  粒子群算法(pso)  分类  挥发性脂肪酸  COD去除率

• 基金项目：国家自然科学基金资助（No.31570568，31670585）；中国制浆造纸工程国家重点实验室（No.201535）；广东省高层次人才基金项目（No.201339）；广州市科技计划项目（No.201607010079，201607020007）；广东省科技计划项目（No.2016A020221005）

• 刘林
• 华南理工大学环境与能源学院, 广州 510006

• 马邕文
• 1. 华南理工大学环境与能源学院, 广州 510006;2. 华南理工大学教育部工业聚集区域污染控制与修复重点实验室, 广州 510006;3. 华南理工大学制浆造纸国家重点实验室, 广州 510006

• 万金泉
• 1. 华南理工大学环境与能源学院, 广州 510006;2. 华南理工大学教育部工业聚集区域污染控制与修复重点实验室, 广州 510006;3. 华南理工大学制浆造纸国家重点实验室, 广州 510006

• 王艳
• 1. 华南理工大学环境与能源学院, 广州 510006;2. 华南理工大学教育部工业聚集区域污染控制与修复重点实验室, 广州 510006;3. 华南理工大学制浆造纸国家重点实验室, 广州 510006

• 谢彬
• 华南理工大学环境与能源学院, 广州 510006

• 武书彬
• 华南理工大学制浆造纸国家重点实验室, 广州 510006

• 摘要：由于厌氧消化过程的复杂性和厌氧菌的敏感性，保持厌氧消化体系的稳定和高效性是比较困难的.本文在实验室采用IC反应器构建了一套厌氧废水处理系统处理人工合成废水，基于支持向量机（SVM）提出了一种预测废水厌氧处理系统出水挥发性脂肪酸（VFA）浓度和COD去除率的软测量模型.为了提高模型的精确性和鲁棒性，加入pso算法（粒子群算法）优化SVM模型，并引入了分类策略对元数据集进行有效分类.仿真结果表明，基于pso-SVM模型的软测量模型对厌氧废水处理系统出水VFA浓度和COD去除率具有较好的预测能力，模型预测系统COD去除率及出水总VFA浓度测试样本数据相关系数分别为65.86%、85.25%；加入分类策略后，元数据集分成两类，模型预测系统COD去除率测试样本数据相关系数分别为92.34%、83.41%；模型预测系统出水总VFA浓度测试样本数据相关系数分别为99.14%、99.59%，系统预测精度明显提高.引入分类策略对元数据集进行有效分类，基于pso-SVM的软测量模型可为监控、优化和理解厌氧消化过程提供指导.

• Abstract：Anaerobic digestion is a complex,nonlinear, biochemical process. In particular, methanogenic bacteria are very sensitive to changing surroundings, so it is difficult to maintain the stability and efficiency of anaerobic digestion. In this study, a soft-sensing model was developed to simultaneously predict the volatile fatty acid (VFA) concentration and COD removal rate of a lab-scale AD reactor system based on support vector machine (SVM) model. In order to improve the performance of the SVM model, the particle swarm optimization (PSO) was applied to optimize the model, and a data-classification strategy was introduced to deal with the relatively large amount of data. The results demonstrate that the hybrid model gave a relatively poor performance before introducing the data-classification strategy. The correlation coefficient (R) is 65.86% and 85.25% for the COD removal rate and total VFA concentration prediction, respectively. After the application of data-classification strategy, the metadata set was divided into two sets, and the performance of the hybrid model was significantly improved. The R2 of the COD removal rate modeling fitting is 92.34% and 83.41%, respectively. And The R2 of the total VFA concentration model fitting is 99.14% and 99.59%, respectively. Thus, introducing the data-classification strategy, the hybrid PSO-SVM based model can give a successful window, which is a good reference for optimizing the AD process.

• Key words：anaerobic system  support vector machine mode (SVM)  particle swarm optimization  classification  volatile fatty acid  COD removal rate

• 
  摘要点击次数： 1126 全文下载次数： 11027