基于灰色支持向量机的城市用水量预测研究

研究城市用水量预测问题,城市需水量具有非线性和随机波动性,城市供水系统是复杂系统,传统灰色模型或支持向量机均无法进行准确预测。为更准确预测城市用水量,建立灰色支持向量机的城市用水量预测模型,以提高预测精度。首先采用灰色模型和支持向量机对城市用水量预测,然后采用多元线性回归确定模型权重系数,最后得到了灰色支持向量机的城市用水量预测结果。仿真结果表明,灰色支持向量机提高了城市用水量的预测精度,预测误差小于单一灰色模型和支持向量机。灰色支持向量机更好地反映出城市用水量变化的总体趋势,可为城市供水系统预测提供参考。     

席夫碱荧光探针对水系Zn2+的实时荧光标记与定量分析

为了实现水系Zn2+的实时检测与定量分析,设计以水杨醛苯甲酰腙作为Zn2+识别基团,引入具有特征性吸收的C=N键作为荧光发色团,合成得到一类新型席夫碱Zn2+荧光探针.通过IR、1H NMR、MS表征探针结构,采用FS分析探针光学性质、探针-Zn2+特异性识别和水系探针-Zn2+荧光标记性能.结果显示,该荧光探针能够特异性识别Zn2+,具有快速响应、光学信号稳定等优势,可用于水系Zn2+的定性检测和定量分析.在体外ECV304细胞荧光共聚焦显微成像实验中,细胞形态完整且荧光成像效果良好,说明探针对Zn2+的荧光标记产物具有较好的透膜性与低毒性,具有进一步应用于生命体中Zn2+标记示踪的前景.     

Magic Peptide: Unique Properties of the LRR11 Peptide in the Activation of Leukotriene Synthesis in Human Neutrophils

Neutrophil-mediated innate host defense mechanisms include pathogen elimination through bacterial phagocytosis, which activates the 5-lipoxygenase (5-LOX) product synthesis. Here, we studied the effect of synthetic oligodeoxyribonucleotides (ODNs), which mimic the receptor-recognized sites of bacterial (CpG-ODNs) and genomic (G-rich ODNs) DNAs released from the inflammatory area, on the neutrophil functions after cell stimulation with Salmonella typhimurium. A possible mechanism for ODN recognition by Toll-like receptor 9 (TLR9) and RAGE receptor has been proposed. We found for the first time that the combination of the magic peptide LRR11 from the leucine-rich repeat (LRR) of TLR9 with the CpG-ODNs modulates the uptake and signaling from ODNs, in particular, dramatically stimulates 5-LOX pathway. Using thickness shear mode acoustic method, we confirmed the specific binding of CpG-ODNs, but not G-rich ODN, to LRR11. The RAGE receptor has been shown to play an important role in promoting ODN uptake. Thus, FPS-ZM1, a high-affinity RAGE inhibitor, suppresses the synthesis of 5-LOX products and reduces the uptake of ODNs by neutrophils; the inhibitor effect being abolished by the addition of LRR11. The results obtained revealed that the studied peptide-ODN complexes possess high biological activity and can be promising for the development of effective vaccine adjuvants and antimicrobial therapeutics.     

Adenosine Receptor Agonist HE-NECA Enhances Antithrombotic Activities of Cangrelor and Prasugrel in vivo by Decreasing of Fibrinogen Density in Thrombus

Blood platelets’ adenosine receptors (AR) are considered to be a new target for the anti-platelet therapy. This idea is based on in vitro studies which show that signaling mediated by these receptors leads to a decreased platelet response to activating stimuli. In vivo evidence for the antithrombotic activity of AR agonists published to date were limited, however, to the usage of relatively high doses given in bolus. The present study was aimed at verifying if these substances used in lower doses in combination with inhibitors of P2Y₁₂ could serve as components of dual anti-platelet therapy. We have found that a selective A_2A agonist 2-hexynyl-5’-N-ethylcarboxamidoadenosine (HE-NECA) improved the anti-thrombotic properties of either cangrelor or prasugrel in the model of ferric chloride-induced experimental thrombosis in mice. Importantly, HE-NECA was effective not only when applied in bolus as other AR agonists in the up-to-date published studies, but also when given chronically. In vitro thrombus formation under flow conditions revealed that HE-NECA enhanced the ability of P2Y₁₂ inhibitors to decrease fibrinogen content in thrombi, possibly resulting in their lower stability. Adenosine receptor agonists possess a certain hypotensive effect and an ability to increase the blood–brain barrier permeability. Therefore, the effects of anti-thrombotic doses of HE-NECA on blood pressure and the blood–brain barrier permeability in mice were tested. HE-NECA applied in bolus caused a significant hypotension in mice, but the effect was much lower when the substance was given in doses corresponding to that obtained by chronic administration. At the same time, no significant effect of HE-NECA was observed on the blood–brain barrier. We conclude that chronic administration of the A_2A agonist can be considered a potential component of a dual antithrombotic therapy. However, due to the hypotensive effect of the substances, dosage and administration must be elaborated to minimize the side-effects. The total number of animals used in the experiments was 146.     

接地铜排对变电站二次电缆防护效果研究

当变电站开关操作遭遇雷击时,接地网会产生地电位差而对二次电缆造成影响,敷设接地铜排能有效阻止电流对二次电缆的损害。基于ATP-EMTP软件建立了一个简化的变电站主接地网仿真模型,对接地铜排与主接地网的连接方式等情况进行研究,分析了接地铜排对二次电缆的防护效果。研究结果表明：增设接地铜排在一定条件下可以降低二次电缆的干扰电压,但接地铜排与主接地网有多个连接点时,在不同位置注入电流,会对电缆芯线与屏蔽层间的电压以及电缆接地点的地电位升产生不同的影响;从电缆2个接地点水平方向的内侧或外侧分别注入电流,接地铜排的影响效果相反,随着接地铜排与主接地网连接点个数的增加,电缆芯线与屏蔽层间的电压和电缆接地点地电位升分别呈增大或减小的趋势。     

环氧树脂绝缘站用变压器局部异常发热检测分析与消除

由于站用变压器长期运行在高电压、高场强和负载电流条件下,所处环境工况复杂,其绝缘性能失效事件时有发生.通过红外热像仪测温、长焦相机拍照的带电检测方法,结合离线电气试验,分析了一起35 kV干式绝缘站用变压器因局部机械受损、运行环境湿度大,从而导致其绕组环氧树脂绝缘性能下降并发生沿面放电,进而引起局部异常发热的案例,并介绍了修复方案和预防措施,为保障站用变压器的安全稳定运行和维修试验提供了参考.     

机动车辆装载问题的一种混合遗传算法实现*

针对货物装载地和运送目的地均已确定情况下的车辆装载问题,给出了基于遗传算法求解的数学模型,并对基本遗传算法的各个算子针对问题的特点提出了改进方法,同时引入启发式策略,形成了一种混合遗传算法.该算法的特点是各个算子都以提高算法收敛速度和解的最优性为目标,同时种群大小是动态变化的,以求在保证群体多样性的同时加快收敛速度.实验结果表明,该算法对求解此类问题是有效的.     

基于WinCC平台的台区线损异常监测系统开发

随着能源需求量的不断增大,能源短缺问题变得越来越严峻,电力行业中的线损异常数据分析存在巨大的需求.运用更加新兴、科学化的大数据分析技术和物联网技术对数据进行分析,有助于提高对台区线损异常排查治理的效率.基于WinCC平台开发了一套台区线损异常监控系统.该系统实现了对海量用电数据的智能分析,建立了线损异常数据库及典型案例库,并将异常台区用电数据与线损异常数据库进行拟合,以缩小检测范围,可及时核查出存在线损异常风险的用户,实现了线损异常的科学诊断,提高了监测工作效率,降低了线损异常情况分析的时间及成本.     

大型复杂结构件抛丸改进方案

为解决某大型焊接结构件抛丸时工件上面大量堆积钢丸并可能导致安全事故的问题,从改变工件吊挂状态和在工件上增加漏丸工艺孔两方面入手,令工件可安全地在已有抛丸线上进行自动除锈.     

基于S变换相关度和深度学习的配电网单相接地故障选线新方法

针对配电网发生单相接地故障时特征信息不明显,且现有选线方法易受故障条件和环境噪声影响的问题,基于S变换相关度和深度学习,提出一种具有强抗噪声能力和高泛化水平的配电网单相接地故障选线新方法.首先,利用S变换获取零序电流时频信息,基于各线路零序电流的全频段信息计算线路故障信息相关度;其次,为提高故障特征的可辨识度和抗干扰性,提出一种S变换相关度图形(SCF)构建方法;在此基础上,建立含SCF层的卷积神经网络深度学习模型(S-CNN),并利用Simulink仿真模型生成的故障数据对其结构参数和超参数进行分步训练;最后,通过S-CNN提取配电网故障零序电流深层特征,实现故障选线,并测试了S-CNN在配电网不同运行状况和故障条件下的选线效果.仿真结果和实际配电网故障数据测试表明:在强噪声干扰场景中,基于S-CNN的故障选线模型在不同故障位置、故障相角、过渡电阻条件下可实现高正确率选线,且在各线路零序电流采样不同步条件下仍具有较强的鲁棒性.