Fate of secondary effluent dissolved organic matter during soil-aquifer treatment

The reduction of mass and trihalomethane formation potential (THMFP) of dissolved organic matter (DOM) and its fractions from secondary effluent during laboratory-scale soil-aquifer treatment (SAT) soil columns were studied. Reduction in dissolved organic carbon (DOC), absorbance of ultraviolet light at 254 nm (UV-254), biodegradable dissolved organic carbon (BDOC) and nonbiodegradable dis- solved organic carbon (NBDOC) for the bulk DOM averaged 72.35%, 53.98%, 97.49% and 35.33% across the soil columns, respectively. Using XAD-8 and XAD-4 resins, DOM was fractionated into 3 fractions: hydrophobic acid (HPO-A), transphilic acid (TPI-A) and hydrophilic fraction (HPI). HPO-A was removed by 61.06%, TPI-A by 54.86% and HPI by 74.95% as DOC as a consequence of the laboratory-scale SAT, respectively. The reduction of THMFP from HPO-A, TPI-A and HPI was 27.24%, 26.24% and 36.08%, respectively. Proton nuclear magnetic resonance (1H-NMR) spectra revealed that the HPO-A isolated from the secondary effluent contained more aromatic functional groups than the corresponding TPI-A. Fourier-transform infrared (FT-IR) spectrum analysis illustrated that TPI-A had decreased hydrocarbon and increased aromatics content in the SAT columns. Specific ultraviolet light absorbance (SUVA) and specific THMFP for each DOM fraction increased across the soil columns and HPI exhibited greater increase in both than HPO-A and TPI-A. The most problematic THM precursor was found to be HPO-A with its high quantity present in recharged water and high chlorine reactivity.     

基于改进狼群算法-深度置信网络(IGWO-DBN)模型的旋风分离器压降预测

针对目前旋风分离器压降计算模型在准确性和实用性上的不足,为更好地指导旋风分离器的结构设计和性能优化,采用深度学习方法对其压降进行了预测。选取了影响压降的7个几何参数,采用深度学习中的深度置信网络（deep belief network,DBN）对旋风分离器压降数据进行预测,并利用改进的狼群算法（improved grey wolf optimizer,IGWO）对DBN模型的初始化权重和偏置参数进行寻优,构建IGWO-DBN组合模型,同时与几种传统计算模型和机器学习模型的预测结果进行对比。结果表明,IGWO-DBN模型在计算精度上优于Shepherd-Lapple模型、Casal模型等传统计算模型,并优于反向传播神经网络（back propagation neural network,BPNN）、支持向量机（support vector machine,SVM）、极限学习机（extreme learning machine,ELM）等机器学习模型,计算效率大幅提升,且具有较好的泛化性和鲁棒性,可用于旋风分离器压降参数的预测。     

一种基于证据融合的故障诊断方法

结合故障诊断工作的实际场景,通过计算故障信息观测值与参考值之间的差异,生成对应的故障信度分配函数。然后基于证据分类的思想,将故障证据分为可信证据和冲突证据,并定义每个证据的可信度。最后,利用证据可信度修正最初的信度分配,再利用证据融合公式得到综合的故障诊断结果。电机转子故障诊断的实例表明本文提出的方法能够准确快速地识别产品故障,基于证据分类的修正方法能够加速故障信度分配的收敛。     

吸附-电解协同法处理铜氨络合废水

针对铜氨络合废水处理难度大、工艺复杂等问题,提出了吸附-电解协同处理的新方法。以模拟铜氨络合废水为研究对象,考察了电解时间、电流密度、电解温度等因素对电解法的影响;并以此为基础,将改性油酸吸附剂涂覆在阴极表面,采用正交试验方法研究了吸附-电解联合处理铜氨络合废水的协同作用。结果表明,吸附-电解协同法在电流密度为0.01A/cm2时对铜氨络合废水处理15 min,即可使铜离子去除率达到93%以上,电流密度小、处理时间短且电流效率明显提高。     

Field emission characteristics of ZnO nanotetrapods and the effect of thermal annealing in hydrogen

Large-scale tetrapod-like ZnO nanostructures have been synthesized using a thermal chemical vapor deposition method on a silicon substrate. The high-purity nanotetrapods show sharp tips geometry with a wurtzite structure. The field emission properties of the uniform ZnO nanostructural material are investigated at different anode-cathode distances. The turn-on field for the ZnO nanotetrapods is found to be about 3.7 V/μm at a current density of 1 μA/cm^2. The field emission behavior obeys Fowler-Nordheim relationship. More importantly, the field emission properties are improved after annearing in hydrogen, and therefore high emission current and low turn-on field are obtained. These results indicate that tetrapod-like ZnO nanostructures are a promising candidate for cold cathode emitters.     

带有模糊流动性约束的均值-方差-偏度-正弦熵投资组合优化模型

通过引入可信性偏度及模糊流动性约束,分别建立了同时满足随机不确定和模糊不确定情形下的均值-方差-偏度-正弦熵（M-V-S-SE）的投资组合模型和带有模糊流动性约束的均值-方差-偏度-正弦熵（M-V-S-L-SE）投资组合优化模型。然后运用马尔科夫方法求解模糊收益率,利用上海证券交易所数据进行实证研究。结果表明：模糊流动性约束的引入使得模型更加稳定,在提高收益、控制风险等方面更具有优势。