人工神经网络在预测深基坑周边地表沉降变形中的应用研究

深基坑开挖引起的周边地表变形预测是一个复杂非线性问题,引起地表沉降的影响因素很多,各因素之间呈高度的非线性关系。传统的基坑用边地表沉降变形预测方法存在着一定的局限性,其预测精度有待提高,而人工神经网络是一种多元非线性动力学系统,可以灵活方便地对多成因的复杂未知系统进行高度建模,实现全面考虑各种主要影响因素的深基坑周边地表沉降变形预测。本文介绍了误差反向传播(BP)网络模型的结构、学习过程及其算法的改进,径向基函数(RBF)网络模型的结构及其学习过程;分析了影响深基坑开挖周边土体沉降变形的主要影响因素;以25个基坑工程的地表沉降实测资料为训练样本,建立了11个输入影响因素的BP神经网络模型和RBF神经网络模型,通过对样本的学习训练过程及对5个检验样本的预测精度,说明了人工神经网络用于预测基坑周边地表沉降的可行性和准确性。     

Migration Pathways and Remediation of Urban Runoff for PAH Control

The migration pathways for polyaromatic hydrocarbons (PAH) and the associated active attenuating mechanisms within the urban environment are examined. The interest in developing an improved understanding of PAH migration/attenuation is focused toward quantification of the cost-effectiveness of urban runoff control strategies for mitigation of PAH inputs to receiving water bodies.A case study application to the City of Sault Ste Marie demonstrates that snow removal from sub-basins with the largest deposition rates, and relatively small wet and/or dry stormwater management ponds, are the most cost-effective remediation strategies.     

基于改进遗传算法的河流水污染源反演方法

针对河流水污染应急响应过程中污染源排放历史迟知、未知的问题,结合多种群遗传算法和自适应遗传算法,利用一维河流水质模型和水质监测数据,研究建立了基于改进遗传算法的河流水污染定量源反演方法,实现了对河流污染源排放历史的识别与重构.将该方法应用于美国特拉基河流的3个不同流量下的示踪剂实验中,对示踪剂排放历史进行定量源反演分析.结果表明：IGA算法对高、中、低不同流量下的3次示踪剂实验均可以很好的重构和识别示踪剂排放历史,对于实际河流水污染源反演分析的误差均在可接受范围内.IGA算法在河流水污染源反演分析中具有一定的可靠性和稳定性,可为河流水污染精准溯源与治理提供科学的技术支撑.     

沱江沉积物中有机氯农药分布特征及其与藻类有机质的关系和生态风险评估

使用气相色谱-质谱联用仪分析了长江支流沱江流域48个表层沉积物中有机氯农药（organochlorine pesticides,OCPs）的残留水平,探讨了其分布和组成特征及其与总有机碳（total organic carbon,TOC）、藻类有机质之间的关系,以及评估其生态风险.结果表明沱江流域表层沉积物中OCPs的总含量为3.17~127 ng·g^-1,其中六六六（hexachlorocyclohexane,HCHs）类农药的含量为2.83~86.0 ng·g^-1,滴滴涕（dichlorodiphenyltrichloroethane,DDTs）类农药的含量为0.340~40.9 ng·g^-1.OCPs空间分布特点为：上游 < 中游 < 下游 < 支流.HCHs和DDTs的组成成分分析表明,沱江流域存在林丹输入的现象,主要来自于历史残留,这与绝大多数流域的DDTs的输入状况相似.OCPs含量与TOC、藻类有机质含量之间存在着极显著的正相关性关系,表明藻类有机质在TOC对沉积物中的OCPs分配中起更为重要的作用.生态风险评估表明,沱江流域表层沉积物的有机氯农药存在较大的生态风险,可能对河流的底栖生物及生态环境造成明显的影响.     

金属表面涂装行业VOCs排放特征及排放系数

对比了浙江省2014和2018年金属表面涂装企业的有机废气排放及治理情况,分析了该行业涂料及稀释剂的使用、主要污染因子,测算了溶剂型、水性涂料的挥发性有机物（VOCs）产生系数和排放系数.结果表明：2018年VOCs治理水平明显高于2014年,水性涂料使用企业比例由18%上升至36%,纯溶剂型企业由82%下降至64%;金属表面涂装行业的主要排放污染物为二甲苯、丁醇、乙酸乙酯、乙酸丁酯、甲苯、丙二醇、乙苯、苯乙烯等8种有机物.溶剂型和水性涂料的VOCs产生系数分别为0.72和0.31kg/kg;溶剂型和水性涂料2014年VOCs排放系数为0.64和0.29kg/kg,2018年为0.48和0.21kg/kg.     

底泥原位洗脱过程中氮磷含量与形态变化特征

随着城市化进程的不断推进,城市河流底泥中内源氮磷污染问题日益突出.为研究原位洗脱技术对污染底泥中氮、磷物质和成分的作用效果,以北京市凉水河作为研究对象,分别比较了现场试验洗脱与对照组0~30 cm分层底泥以及实验室模拟洗脱处理前后0~10 cm底泥中pH、E_h（氧化还原电位）、容重、含水率、OM（有机质）、TN、TP和氮、磷形态分布的变化特征.结果表明:①现场洗脱对0~5 cm厚度底泥的处理和模拟洗脱处理均对底泥中pH、E_h、含水率的改变以及其中TN、TP的去除效果表现显著,其中,现场试验洗脱组0~5 cm范围底泥中w（TN）、w（TP）分别较对照组降低了66.4%±17.4%、40.8%±24.4%,模拟洗脱处理对底泥中OM、TN、TP的去除率分别达70.1%±4.8%、66.8%±2.0%、43.1%±3.1%.②原位洗脱技术对底泥中TN的去除主要通过对PON（颗粒态有机氮）的去除来实现,现场和模拟试验过程中PON对TN去除的贡献率分别达55.0%和73.6%.③原位洗脱对底泥中TP的去除主要通过对Al-P（铝结合态磷）、OP（有机磷）的去除来实现,现场和模拟试验中Al-P、OP对TP去除的贡献率分别达37.0%、66.2%和31.3%、43.7%.研究显示,原位洗脱技术可有效去除城市河流底泥处理层中的氮磷物质,并以有机氮、磷物质的去除为主.      

Reactive uptake of NO2 on volcanic particles: A possible source of HONO in the atmosphere

The heterogeneous degradation of nitrogen dioxide (NO₂) on five samples of natural Icelandic volcanic particles has been investigated. Laboratory experiments were carried out under simulated atmospheric conditions using a coated wall flow tube (CWFT). The CWFT reactor was coupled to a blue light nitrogen oxides analyzer (NO_x analyzer), and a long path absorption photometer (LOPAP) to monitor in real time the concentrations of NO₂, NO and HONO, respectively. Under dark and ambient relative humidity conditions, the steady state uptake coefficients of NO₂ varied significantly between the volcanic samples probably due to differences in magma composition and morphological variation related with the density of surface OH groups. The irradiation of the surface with simulated sunlight enhanced the uptake coefficients by a factor of three indicating that photo-induced processes on the surface of the dust occur. Furthermore, the product yields of NO and HONO were determined under both dark and simulated sunlight conditions. The relative humidity was found to influence the distribution of gaseous products, promoting the formation of gaseous HONO. A detailed reaction mechanism is proposed that supports our experimental observations. Regarding the atmospheric implications, our results suggest that the NO₂ degradation on volcanic particles and the corresponding formation of HONO is expected to be significant during volcanic dust storms or after a volcanic eruption.     

Coupling mechanism of activated carbon mixed with dust for flue gas desulfurization and denitrification

To clarify the effect of coking dust, sintering dust and fly ash on the activity of activated carbon for various industrial flue gas desulfurization and denitrification, the coupling mechanism of the mixed activated carbon and dust was investigated to provide theoretical reference for the stable operation. The results show that coking dust had 34% desulfurization efficiency and 10% denitrification efficiency; correspondingly, sintering dust and fly ash had no obvious desulfurization and denitrification activities. For the mixture of activated carbon and dust, the coking dust reduced the desulfurization and denitrification efficiencies by blocking the pores of activated carbon, and its inhibiting effect on activated carbon was larger than its own desulfurization and denitrification activity. The sintering dust also reduced the desulfurization efficiency on the activated carbon while enhancing the denitrification efficiency. Fly ash blocked the pores of activated carbon and reduced its reaction activity. The reaction activity of coking dust mainly came from the surface functional groups, similar to that of activated carbon. The reaction activity of sintering dust mainly came from the oxidative property of Fe₂O₃, which oxidized NO to NO₂ and promoted the fast selectively catalytic reduction (SCR) of NO to form N₂. Sintering dust was activated by the joint action of activated carbon, and both had a coupling function. Sintering dust enhanced the adsorption and oxidation of NO, and activated carbon further promoted the reduction of NO_x by NH₃; thus, the denitrification efficiency increased by 5%-7% on the activated carbon.     

特殊区域地表水环境背景值表征技术方法研究与应用

选取黑龙江省源头水保护区作为研究对象,根据DEM数据对保护区内随机布设的水质监测点控制单元进行划分并提取单元内的相关指标建立监测点背景特性量化模型,基于量化结果提出地表水环境背景值监测方案,建立背景值数据库.选取研究区自然属性指标与空间属性指标,以空间叠置技术与聚类分析方法对研究区地表水进行背景值分区,最终将黑龙江省分为六大地表水水质背景值地理分区,并计算了各分区地表水环境背景值表征范围.表征范围显示：依据现行标准进行水质评价,保护区内水质背景值已超出Ⅱ类水质标准限值.因此,基于水质背景值研究成果提出了考虑背景值影响下的水环境质量评价方法,并将方法应用于研究区,结果表明方法切实可行且优于单因子评价法.研究成果可为区域制定背景值影响下的水环境评价方法提供科学的数据支撑与理论依据.     

海绵城市地块汇水区颗粒污染物的传输

目前我国海绵工程建设多集中在地块汇水区单元内开展,通过多个低影响开发(LID)设施协同完成地表径流水质水量的调控,但基于地块汇水区尺度下城市面源污染的产生和控制效果鲜有报道.本研究比较分析了不同硬化率地块汇水单元内的面源颗粒污染物晴天累积、降雨冲刷、地表径流及径流输出负荷状况.结果表明,地块汇水单元内硬质路面是面源颗粒污染物贡献的最主要的下垫面类型,中硬化率(61.1%)地块和高硬化率(73.6%)地块路面街尘累积量分别约占汇水区单元的88.4%(2.22～12.51 g·m~(-2))和90.1%(4.99～33.43 g·m~(-2)),对径流SS的输出贡献比率分别约为91.7%(0.97～7.34 g·m~(-2))和90.5%(0.92～18.77 g·m~(-2)),降雨径流SS污染负荷占比分别约为95.2%和83.1%,经LID设施处理后输出径流污染负荷约为地表径流的24.0%和40.2%.硬质路面的街尘晴天累积及降雨冲刷以150μm为主,地表径流及输出径流则以50μm粒径段为主,同时地块不透水比例的增加,细粒径(105μm)颗粒物的累积及冲刷分布增大(24.4%和106.4%),而粒径50μm的颗粒物在路面径流中的分布减小(12.4%).屋面的街尘累积、冲刷及降雨径流的粒径分布状况与硬质路面大致相似,但中硬化率地块(1 000μm)和高硬化率地块(250～450μm、45μm)在3个粒径段范围的颗粒物累积和冲刷相较于路面街尘粒径分布明显增加(1 000μm:58.1%和108.5%; 250～450μm:72.9%和41.8%;45μm:59.2%和64.8%).以上结果揭示了颗粒污染物在地块汇水区尺度下的污染全过程(累积-冲刷-输出)分布及LID设施对地块整体SS污染负荷的控制效果,可为地块汇水单元内LID设施工程绩效的科学评估提供重要参考.