线性扰动河流预报模型应用——南京站、镇江站潮水位预报

潮水河下游受潮水上灌,上游受河水顶托,其水流情况异常复杂,但它具有明显的季节特性.本文试图用线性扰动河流预报模型对长江南京站、镇江站的潮水位进行预报.效果较好,尝试是成功的.     

用场,路结合法分析感应钢包炉的等效电路

本文首先提出感应钢包炉的等效电路，而后从感应钢包炉的空载、负载和漏磁场的边值问题出发，利用有限元法，计算了这三种场，由此分析了等效电路中的各参数及性能的计算方法；其中在计算包壳中的涡流损耗时，利用了对偶有限元法．     

大学生创新实验项目教学改革和实践

大学生创新实验项目是改变传统教学方式、做好本科教学工作、培养学生创新意识和实践能力的重要手段和必由之路。合肥工业大学结合给排水科学与工程专业的特点以及实际教学情况,围绕大学生创新项目的申请、指导和监督、结题等教学环节,探讨了申请动员、实验内容及方案设计、科研带动作用、数据成果化等教学内容及方法;利用教师科研平台支持大学生创新实验项目。其结果有助于提高学生的文献查阅、实验方案设计和实施的能力,充分锻炼自身的创新和实践能力。     

东北地区中泥盆世岩相古地理研究

利用野外实测剖面资料和区域地质资料,在地层划分、对比和构造格局研究基础上,从点-线-面的沉积相分析入手,将中泥盆世分为艾费尔期、吉维特期两个编图单元,进行岩相古地理编图,恢复了东北地区中泥盆世的岩相古地理面貌。研究表明,中泥盆世东北地区艾费尔期沉积环境多为浅海陆棚相碎屑岩沉积及碳酸盐岩台地相沉积;佳木斯-松嫩微板块部分地区及额尔古纳-兴安微板块为古陆;内蒙古东乌珠穆沁旗哈诺敖包一带为滨海沼泽相沉积。吉维特期出现大规模海退,主要表现为:研究区南部沿西拉木伦缝合带一线未见沉积;延寿县马鞍山福兴屯一带发育福兴屯组湖相沉积;兴安海区西南段东乌珠穆沁旗一带的塔尔巴格特组表现为进积-加积型层序,由艾费尔期泥鳅河组浅海沉积过渡为滨海沉积为主。     

Discovery of gas-discharged pits (gas pit) in the plains of the Yellow River delta

A kind of special sedimentary structures are developed in the overwater plains of the Yellow River delta. They look like funnels: round or nearly round, concave of pit-like, with a diameter ranging from several centimeters to 20 or 30 cm, and depth from several millimeters to over 20 cm. There is a vertical pipe (called gas discharging conduit) in the center, with a diameter of several millimeters to 1 cm, depth of several centimeters to more than 10 cm. There may be a lip-like relief (called “lip-like relief”) on the periphery or some parts of the periphery, with ring structures showing horizontal stratification on the inner margin. Plant debris carbonized, plant fragments or dark-colored minerals may sometimes be found in the center of the pits. Such structures are usually developed in silt (with minor clay laminations), often underlain by one or more thin layers of mud matter. Our studies find that they are genetically related with gas discharging of organic matter during biological degradation. Therefore, they are called “gas-discharged pits”.     

主要栽培因子对小麦根腐病的影响

试验研究了主要栽培因子与小麦根腐病发生的关系。结果表明 :冀东冬麦区 9月 30日左右播种 ,播量为 1 5 0kg/hm2 ,稻麦轮作 ,氮磷肥配合施用 ,根腐病发生较轻。不同小麦品种对根腐病的抗感反应不同。供试品种中 ,未发现免疫品种     

Numerical simulation on mesoscale convective system along Mei-Yu front in Southern China

A mesoscale convective system (MCS) was generated over the South China Sea at around 115°E, 21°N on 0000 UTC 7 June 1998. Riding along the Mei-Yu front, the system moved through the Taiwan Strait and finally hit the southern part of the Taiwan Island, produced over 300 mm of rainfall over parts of Taiwan in the next 24 hours. It was found that the Penn State-NCAR Mesoscale Model Version 5 (MM5) did quite well in simulating the evolution of the MCS. Diagnostic studies on the mesoscale structure of MCS, moisture concentration process and the momentum budget were based on the model output. The concept model of the MCS in the mature phase can be concluded as the following: At the 850 hPa level, the mesoscale low level jet (mLLJ) was found to the southwest of the MCS, which was also associated with a mesoscale low generated by convection, the mesoscale upper level jet (mULJ) was found due east of the MCS (and an upper-level mesoscale high) at the 300 hPa level. Horizontal advection of momentum acted to accelerate the exit of the mLLJ but decelerate the entrance of the mLLJ where the pressure gradient force acted to compensate this lost and maintain its strength. For maintenance of the mULJ, vertical advection of momentum and the pressure gradient force were both found to be important.     

用基于等磁能法的有限元法计算二维非线性涡流场

本文讨论了用矢量磁位求解的二维非线性涡流场的有限元计算方法，包括平面场及轴对称场，采用了涡流场中铁磁材料磁化曲线的一种处理方法－等磁能法，并且给出了非线性涡流场的有限元计算格式，最后用该方法分析了一台感应钢包炉中的涡流场及炉料（钢水）的受力情况，并绘制了相应的曲线。     

五鹿山自然保护区白皮松群落物种组成与群落结构

以五鹿山国家自然保护区的白皮松群落为研究对象,通过典型样方调查,对其物种组成和群落结构进行了研究。结果表明:样地内植物共有32科45属61种,裸子植物2科2属3种,被子植物30科43属58种;从物种多度、平均胸径和重要值等指标来看,群落具有比较明显的优势种;群落垂直结构复杂,分为乔木层、灌木层和草本层,其中乔木层又可分为3个亚层;白皮松种群在每一个乔木亚层中都有出现,是该群落的优势种群;样地内所有树种的径级结构呈倒"J"形,几种优势种树种的径级结构可分为3种类型:倒"J"形、双峰形和倒"W"形。     

基于脑电信号深度学习的情感分类

情感脑电研究作为人工智能高级阶段的重要任务,近年来受到越来越多的关注。情感脑电分类广泛应用于人机交互、医学研究等领域。该文以轻量级的卷积神经网络为核心,设计了情感脑电分类模型,以DEAP（dataset for emotion analysis using physiologicalsignals）提供的情感脑电图数据为基础,将其中的观看视频划分为唤醒度和愉悦度2个维度。为了获得频域信息,提取了theta、alpha、beta和gamma波段的功率谱密度特征进行评估,并将功率谱密度矩阵表示为二维灰度图像。然后将该图像输入到卷积神经网络训练分类模型并完成2个维度的分类任务。实验结果表明,与传统机器学习相比,卷积神经网络具有更好的分类效果,唤醒度分类准确率达到了82.33%,愉悦度分类准确率达到了75.46%。