基于字符级CNN的调水工程巡检文本智能分类方法

日常安全巡检是维护长距离调水工程安全运行的重要手段。目前巡检采集的非结构化文本数据主要依靠人工进行安全等级评判,在工作效率和准确率方面存在明显不足。本研究基于自然语言处理技术,提出了一种面向字符层面的卷积神经网络的巡检安全文本智能分类方法。该方法通过引入预训练的单个字符向量改进卷积神经网络的输入层,使得分类模型直接从原始文本中提取特征信息,不仅避免了传统分类方法对专业词库的依赖,而且不易受文本中出现的口语化表达和错别字的影响。以国内南水北调工程的巡检文本为案例,通过与多种深度学习算法进行全面比较,对比验证了所提方法的有效性和优越性。结果表明,字符级的分类方法明显优于传统基于词的分类方法,且卷积神经网络在巡检文本分类方面明显优于其他深度学习网络。该方法具有较高的分类准确率,以此为调水工程安全维护提供新的智能化手段。     

高流速含沙浑水体流变特性试验研究

加速行进中的洪水具有高含沙、高流速和粗粒径的特点,而针对洪水的流变特性研究关注点多为高含沙量,忽略了高流速即高剪切速率的特点。以黄河下游泥沙为研究对象,系统地研究了高流速剪切条件下含沙量、剪切模式、温度和盐度对流变特性的影响。结果表明,在高流速剪切条件下（剪切速率≥ 1200 s-1,对应线性流速为3.6 m/s）,含沙量在106.64 ~ 957.83 kg/m3范围内的浑水体符合牛顿流体的应力本构关系,不存在屈服应力,表观黏度与含沙量呈指数关系。温度降低并未改变浑水体的流变特性,但会使其表观黏度增加,且温度越低黏度增幅越大,浑水体黏度增大梯度约为等温度清水的数倍至十几倍。盐度变化对该测试浑水体的表观黏度和流变特性影响甚微,可忽略不计。本研究结果可为高流速水流挟沙和远距离输沙研究提供理论参考。     

Fouling characteristics of 90° elbow in high salinity wastewater from coal chemical industry

Due to the high salt content of coal chemical wastewater, pipeline fouling often occurs during wastewater treatment. Fouling will cause the diameter of the pipe to shrink or even block, which is not conducive to the safe and stable operation of the wastewater treatment process. In this paper, the experimental device was designed by using FLUENT software and the fouling deposition mechanisms at different flow velocities and different positions in a 90 deg bend were studied. The experimental results show that when the flow velocity is between 0.2 m·s~(-1) and 0.3 m·s~(-1), the thickness of fouling layer was positively correlated with the flow velocity; when the flow velocity is equal to 0.4 m·s~(-1), the formation of fouling is the most serious; when the flow velocity is between 0.4 m·s~(-1) and 0.7 m·s~(-1), the thickness of fouling layer was negative correlation with the flow velocity; with the increase of inlet velocity, the time for sediment point to develop into sediment surface is shortened. The fouling layer is easy to fall off because of the large shear force on the wall surface of the inner bend of the 90° elbow, so the density of sediment at this position is high.     

基于国外现货市场与四川实际的现货市场建设分析

现货市场是电力市场建设的关键环节,同时对推进国家电力体制改革、打破电力行业垄断、实现电力资源市场化的合理配置具有重大意义。本文基于分析国外电力现货市场建设经验,结合四川水电清洁能源大省、电网结构复杂的特点,从市场模式、市场主体、交易品种、电价机制、结算机制以及弃水量大、阻塞断面较多、弃水期与非弃水期市场差异化较大等问题出发,对四川电力现货市场建设进行了思考与分析,为建设具有四川特色的电力现货市场方案提供建议。     

40万t/a煤制乙二醇项目气流床气化技术比较研究

煤气化技术对煤化工项目的经济效益和稳定生产具有重要意义。以某40万t/a煤制乙二醇项目为例,选取有代表性的某干煤粉气流床煤气化技术和某水煤浆气流床煤气化技术,在原料煤可同时适用的前提下,通过对原料制备、原料输送、气化和灰水处理、变换等主要工艺单元的对比,分析了两种技术在能量消耗和利用、操作安全稳定、环保等方面的差异。结果表明,水煤浆气化技术的原料制备和输送环节能耗较低、操作安全环保,虽然气化工段煤耗、氧耗较高,但是可通过回收高温粗煤气的显热副产大量高压饱和蒸汽,能量利用率高,气化过程不需要加入蒸汽。综合对比,水煤浆气化技术生产成本低于干煤粉气流床气化技术。     

多孔普鲁士蓝类似物的合成及电催化析氧性能

采用一种简单易行的共沉淀法合成了前驱体镍铁普鲁士蓝类似物NiFe-PBA（NF）,然后通过溶剂热处理获得了镍铁普鲁士蓝纳米多孔材料（NFP）。通过XRD、SEM、TEM、XPS、BET及电化学方法对所得材料进行了结构表征和析氧性能测试。结果表明,NFP相对于前驱体NF,电化学比表面积增大、催化活性位点增多,电催化析氧反应（OER）性能显著提高。在浓度1 mol/L KOH水溶液中,达到10 mA/cm2电流密度时,NFP所需过电位仅为260 mV,比NF（320 mV）前驱体低18.75%,也优于大多数已报道的非贵金属催化剂和商用贵金属催化剂,显示出良好的应用前景。     

孤岛微电网逆变器不平衡负载下的控制策略

提出一种孤岛微电网逆变器在不平衡负载下的控制策略,采用分序网络解耦控制,在虚拟同步发电机(VSG)控制的基础上引入自适应负序补偿环.对三相负载不平衡下的电路进行分析,得到逆变器分序网络等效电路和逆变器间负序环流产生机理.通过分序网络解耦控制得到电压环参考电压各序分量,引入分序网络虚拟阻抗改进逆变器各序输出阻抗.正序网络采用VSG控制实现正序调频调压和电流的自主分配,负序网络加入负序补偿环抑制微电网三相电压负序分量.详细分析了微电网电压负序分量与负序补偿系数和逆变器输出电流负序分量的关系.负序补偿环采用自适应控制,在实现微电网三相电压平衡的同时根据逆变器额定容量自主分配输出电流负序分量.建模仿真结果验证了所提控制策略的有效性.     

Knockout of Auxin Response Factor SlARF4 Improves Tomato Resistance to Water Deficit

Auxin response factors (ARFs) play important roles in various plant physiological processes; however, knowledge of the exact role of ARFs in plant responses to water deficit is limited. In this study, SlARF4, a member of the ARF family, was functionally characterized under water deficit. Real-time fluorescence quantitative polymerase chain reaction (PCR) and β-glucuronidase (GUS) staining showed that water deficit and abscisic acid (ABA) treatment reduced the expression of SlARF4. SlARF4 was expressed in the vascular bundles and guard cells of tomato stomata. Loss of function of SlARF4 (arf4) by using Clustered Regularly Interspaced Short Palindromic Repeats/Cas 9 (CRISPR/Cas 9) technology enhanced plant resistance to water stress and rehydration ability. The arf4 mutant plants exhibited curly leaves and a thick stem. Malondialdehyde content was significantly lower in arf4 mutants than in wildtype plants under water stress; furthermore, arf4 mutants showed higher content of antioxidant substances, superoxide dismutase, actual photochemical efficiency of photosystem II (PSII), and catalase activities. Stomatal and vascular bundle morphology was changed in arf4 mutants. We identified 628 differentially expressed genes specifically expressed under water deficit in arf4 mutants; six of these genes, including ABA signaling pathway-related genes, were differentially expressed between the wildtype and arf4 mutants under water deficit and unlimited water supply. Auxin responsive element (AuxRE) elements were found in these genes’ promoters indicating that SlARF4 participates in ABA signaling pathways by regulating the expression of SlABI5/ABF and SCL3, thereby influencing stomatal morphology and vascular bundle development and ultimately improving plant resistance to water deficit.     

采用VSG策略的MMC阻抗建模及并网稳定性分析

随着分布式能源并网的增多,局部电网逐渐呈现出弱电网特性,虚拟同步发电机（VSG）控制由于可以模拟同步发电机的特性,能够为弱电网提供惯性和阻尼支撑。本文采用谐波线性化方法对采用VSG控制策略的模块化多电平换流器(MMC)进行阻抗建模。所得到的阻抗模型在低频段主要呈容性,在高频段主要呈感性,在复杂电力系统中,可能会与电网阻抗存在谐振点。分析了VSG主要控制参数对MMC阻抗的影响,发现有功-频率下垂系数、阻尼系数和虚拟转动惯量主要影响50Hz附近的阻抗。根据阻抗稳定性判据对MMC并网系统进行稳定性分析,发现并网系统在高频段有振荡的风险。     

Virtual Screening of FDA-Approved Drugs against Triose Phosphate Isomerase from Entamoeba histolytica and Giardia lamblia Identifies Inhibitors of Their Trophozoite Growth Phase

Infectious diseases caused by intestinal protozoan, such as Entamoeba histolytica (E. histolytica) and Giardia lamblia (G. lamblia) are a worldwide public health issue. They affect more than 70 million people every year. They colonize intestines causing primarily diarrhea; nevertheless, these infections can lead to more serious complications. The treatment of choice, metronidazole, is in doubt due to adverse effects and resistance. Therefore, there is a need for new compounds against these parasites. In this work, a structure-based virtual screening of FDA-approved drugs was performed to identify compounds with antiprotozoal activity. The glycolytic enzyme triosephosphate isomerase, present in both E. histolytica and G. lamblia, was used as the drug target. The compounds with the best average docking score on both structures were selected for the in vitro evaluation. Three compounds, chlorhexidine, tolcapone, and imatinib, were capable of inhibit growth on G. lamblia trophozoites (0.05–4.935 μg/mL), while folic acid showed activity against E. histolytica (0.186 μg/mL) and G. lamblia (5.342 μg/mL).