燃煤污染源排放颗粒物采样方法实验室比对研究

为研究适用于我国燃煤污染源排放颗粒物的分级采样标准测试方法，采用民用小煤炉排放装置结合烟气采集系统，对比研究了直接采样法(双级虚拟撞击PM₁₀/PM_2.5采样器、旋风采样器、总烟尘采样器)和稀释采样法(低压荷电撞击器ELPI配备稀释系统)对烟气中不同粒径颗粒物测试结果的稳定性、相关性、仪器可操作性等特点，并分析了煤质对颗粒物分级测试结果的影响. 结果表明:①从决定系数和残差平方和角度分析，稀释采样法测试结果自身拟合性相对直接采样法差，实测数据点相对分散，95%置信带较宽. 直接采样法中，基于本文确定的清洗和收集方式，旋风采样器测试得到的PM₁₀和PM_2.5浓度拟合度高达0.999；双级虚拟撞击PM₁₀/PM_2.5采样器测试结果稳定性也较高，其测试得到的PM₁₀和PM_2.5浓度拟合度也为0.999. ②不同颗粒物采样器对烟气中PM_2.5、PM₁₀、TSP的浓度测试结果均表明，稀释采样法与直接采样法测试结果相关性较低；而各直接采样法之间呈高度相关，Pearson相关系数在0.993~0.999之间. ③稀释采样法分级测试结果显示，当颗粒物排放浓度较低时，各级滤膜称量误差叠加可导致PM₁₀、PM_2.5浓度测试结果误差较大. 因此，针对常温且颗粒物浓度较低的烟气，建议采用直接采样法；针对高温烟气，稀释采样法可捕集稀释降温过程中形成的可凝结颗粒物. ④各采样方法测试结果表明，相对于燃用无烟煤(如蜂窝煤)，燃用烟煤产生的烟气中颗粒物浓度相对较高，且PM_2.5占比较高. 研究显示:稀释采样法能模拟燃煤污染源高温烟气排入大气环境中可凝结颗粒物的形成过程，测试结果更接近真实排放情况；对于常温颗粒物浓度较低的烟气，更适宜采用直接采样法. 

Laboratory Comparison of Particle Sampling Methods for Coal-Fired Pollution Sources

In order to study the grading sampling standard method of particles emitted from coal-fired stationary sources in China, two sampling methods, including the dilution sampling method (Dekati ELPI in Finland combined with FPS-4000 dilution systems) and the direct sampling method (virtual impactors, cyclone samplers and total dust sampler), were compared in measuring PM_2.5, PM₁₀, TSP from coal-fired sources. The stability and correlation of the test results, the instrument operability of the four samplers were assessed using a coal stove with a self-designed flue gas collection system. The influence of coal quality on the test results of particle graded sampling was also studied. The results showed that: (1) The fitting degree of the test results of the dilution sampling method was lower than the direct sampling method based on the coefficient of determination R-square and residual sum of squares. The points from the dilution sampling method were relatively scattered, with a wide 95% confidence interval. Among the direct sampling methods, the goodness of fit of PM₁₀ and PM_2.5 mass concentrations measured by cyclone sampler was as high as 0.999, by adopting the cleaning and collection methods of particles in this study, so was the virtual impact sampler. Namely the test results of the direct sampling methods were stable. (2) The test results of PM_2.5, PM₁₀ and TSP in the flue gas by different particle samplers showed that the correlation between the test results of the dilution sampling method and the direct sampling method was slightly low, while the test results of the direct sampling methods were highly correlated. The Pearson correlation coefficient was between 0.993 and 0.999. (3) The results of the dilution sampling method showed that it was possible that the weighing error of all levels of the filter membrane caused a large error in the mass concentration of PM₁₀ and PM_2.5 when the concentration of particulate matter in flue gas was low. Hence, for the flue gas with ambient temperature and low particle concentration, the direct sampling method was recommended. For the flue gas with high temperature, the condensable particulate matters formed during dilution and cooling process could be collected by the dilution sampling method. (4) The test results by all sampling methods showed that the concentration of particulate matter in the flue gas of burning bituminous coal was higher than that from the anthracite (such as honeycomb briquette), and the proportion of PM_2.5 was also higher. The research indicated that the dilution sampling method could simulate the formation process of the condensable particulate matter when high temperature flue gas from coal-fired pollution source was discharged into the atmosphere, and the test results were closer to the actual emissions. The direct sampling method was more suitable for the determination of flue gas with low particle concentration at room temperature.