高湿烟气PCDD/Fs净化体系中同类物去除率差异的机理分析与模型研究

The combination of activated carbon injection and fabric filter is widely used in flue gas cleaning and dioxin removal, however, the removal efficiencies of PCDD/Fs congeners are heterogeneous. Highly-chlorinated PCDDs have apparently lower removal efficiencies than lowly-chlorinated congeners, even with superfluous injection of activated carbon. The reason was not elaborated in a unified and reasonable way in the existing studies, with no attention paid to the impact of high moisture in flue gas. This study aims to find the mechanism of heterogeneous removal efficiencies of congeners in depth, with the aid of field sampling and analysis and subsequent modeling. After separating the sorption features of congeners into multiple solid phases including particles and actived carbon, the sorption variation of congeners resulted from changing the polarity of particles is investigated, and the competitive adsorption of water and PCDD/Fs into actived carbon is focused on. The competitive sorption of PCDD/Fs in flue gas with high moisture is analyzed by constructing a gas-particle partition model for congeners in all phases and a quantitative structure-activity relationship (QSAR) model for congener adsorption in gas phase. Hence, comprehensive analyses will be given to identify the most important factors with consideration of influence factors such as molecular structure, gas moisture content, and operating conditions. As a result, the key mechanism for explaining the heterogeneous removal efficiencies of congeners are determined. Findings, analyses, and models are beneficial to optimize the operating conditions and obtain higher removal efficiencies of PCDD/Fs congeners.

垃圾焚烧等行业高湿烟气的二噁英类（PCDD/Fs）净化会出现同类物去除率差异显著、高氯代PCDDs去除率偏低等问题，制约了烟气净化效果。现有研究没有考虑烟气含水率对PCDD/Fs相际迁移和去除的影响，对相关机理研究不够深入。本项目结合现场采样和机理建模，分离PCDD/Fs同类物在多种固相（颗粒物、活性炭）上的附着特征，研究高湿烟气中颗粒相极性变化对于同类物附着特性的影响，考察水分子和PCDD/Fs分子对于活性炭含氧官能团的竞争作用。通过构建同类物的气相-颗粒相分配模型和活性炭吸附气相PCDD/Fs的定量构效关系模型，分析高湿烟气PCDD/Fs同类物在多种固相上的附着竞争机制，从而全面考察同类物分子结构、烟气含水率、工况条件等因素对PCDD/Fs同类物去除率的影响，识别主要影响因素，最终探明PCDD/Fs同类物去除率出现差异的关键机制，为优化工艺参数、提高烟气PCDD/Fs的去除率提供指导。

通过采样分析、模型构建、机理分析等方式，考察影响垃圾焚烧烟气PCDD/Fs同类物相际迁移、净化去除的主要因素，研究同类物去除率差异的内在机理，分析同类物多元因子特征，构建同类物的回归预测模型。主要包括：（1）烟气PCDD/Fs同类物相际分配特点和去除率差异的机理分析。同类物的正辛醇/空气分配系数、饱和蒸汽压等物化属性的对气相/颗粒相分配行为的影响占比仅为30%，运行工况、同类物分子结构的影响更大。完善同类物分子描述符体系，以改善同分异构体分子描述符的差异性，分别以同类物初始浓度、相际分配为因变量，以分子描述符为自变量，构建了多元线性回归QSAR模型。发现PCDD/Fs产生时以高氯代同类物为主，同类物初始浓度比例与氯取代个数、结构修正因子、分子轨道能隙相关性较好，但QSAR模型参数受运行工况影响。同类物初始占比、饱和蒸汽压共同影响它们在活性炭喷射+袋式除尘体系中的去除率：同类物占比越大，则去除率越高，同类物饱和蒸汽压越大，则越不容易去除，但不同烟气中同类物初始占比差异性较强，每种同类物的去除率难以用固定数值率定。（2）烟气PCDD/Fs同类物的多元因子特征解析。排放烟气中PCDD/Fs仍以高氯代同类物为主。使用正定因子矩阵解析模型作降维分解，分析各同类物在主因子上的因子载荷，发现高氯代PCDDs或PCDFs生成机制的关联性较强，而低氯代PCDDs与低氯代PCDFs的生成机制不同。（3）烟气PCDD/Fs同类物与运行工况的回归模拟。分别应用多元线性回归、支持向量回归，建立PCDD/Fs物质浓度及毒性当量浓度与工况参数、常规污染物排放浓度之间的回归预测模型。支持向量回归的预测性能更好，对烟气PCDD/Fs物质浓度、毒性当量浓度的预测误差分别可控制在10%和20%左右。物质浓度预测的特征变量组合为炉膛温度、NOX、HCl、颗粒物，毒性当量浓度预测的特征变量组合不包括NOx。