附录A 外文文献译文
有效地发展太阳能热发电技术在中国是一个巨大的挑战。在该文件中的一个抛物面槽式太阳能集热器实验平台系统（抛物槽式太阳能集热器系统）的开发热发电，以及抛物槽式太阳能集热器系统性能进行了实验研究与合成油为散发有机热载体（导热油）。该太阳能集热器与太阳通量的变化与流速的导热油效率鉴定。该集热效率抛物槽式太阳能集热器系统可以在40％-60％之间。有人还发现，存在着一个为导热油温度响应太阳能通量，而在设计中发挥了重要作用抛物槽式太阳能集热器系统指定的延迟。对集热效率热损失的影响进行了研究，这对于用180℃之间的收藏家温度和环境温度的温差接收约220瓦/米，大约相当于在总数的10％，太阳能事件收藏家。这些令人鼓舞的结果可为开发抛物面槽式太阳能在中国火电厂的基本数据。 简介
抛物槽式太阳能热发电技术已经取得了三个太阳能热发电技术的广泛普及。大型抛物面槽式收藏家可以提供热能，可用于生产蒸汽涡轮发电机一朗肯蒸汽。经常采用的是抛物槽式太阳能集热器系统蒸汽发电由于其高集电极的效率在中间温度（约300℃）。作者在抛物槽式太阳能集热器系统性能起着重要作用的抛物面槽式太阳能电厂，它直接决定了投资成本。优化和改进的抛物槽式太阳能集热器系统性能，许多调查已经进行了自太阳能资源。在中国丰富，尤其是在如青海省，西藏，新疆地区，西部，抛物线槽式太阳能热发电技术已受到广泛关注最近。
实验平台和程序
为了研究太阳能热发电，实验平台的开发抛物槽式太阳能集热器系统。实验平台主要由四部分组成：太阳能集热器，换热流体（导热油）循环，冷却水循环，测量系统.找个实验平台流程图使用合成油作为流通导热油。该导热油不仅可以流通之间的油箱和水箱电热水器，还可以通过太阳能接收流动单位。在一个典型的实验中，首先被送入导热油电加热器坦克一泵，加热到一定温度。然后，导热油，抽到太阳能接收装置吸收太阳能。加热后的导热油离开接收机是由冷却的冷却水热交换器，最后回流到油箱，循环完成的。考虑到在高温下，一膨胀水箱的体积膨胀系数的导热油对实验平台。他的顶端固定值得注意的是，能源的电热水器提供的是不是在评估过程中考虑太阳能。与30平方米的通州区，北京，内置抛
物槽式太阳能集热器系统。实验共进行了近9:00至16:00从与不同的导热油流速。
结果与讨论
许多重要的因素，如太阳能通量，对导热油流量和热损失，影响热发电相转移性能。所以，他们考虑在9月21日获得这个研究成果.A典型结果，其中合成导热油体积流量率显示0.45立方米/秒钽，钛，并ηc为环境温度的立场，对导热油进口温度，出口的导热油和集电极效率，分别.图 4显示了两者之间的相转移效率和太阳能通量关系的温度下不同体积流率。仿真结果也已得到考虑太阳能通量从200个不同的实验条件，按照900 W/m2时。对导热油的流速为0.45立方米/小时，0.28立方米/小时和0.10立方米/小时，分别从图4固定的，它可以发现，收集效率在40％-60％之间。实验数据分散的模拟线路。可知，集电极效率非线性对太阳通量而定。当导热油流量为0.45立方米/小时，效率有一个由200％提高到7550W/m2时，只有3％，由550至900W/m2时，集电极效率和最大的可能是在最高所得太阳能通量的研究范围之内。一个原因是，该接收器的平均气温为这些值大幅上升，由导热油的热量慢慢得到了应有的热损失。它也可以看到从图4的集电极效率在流量非常低0.10立方米/小时由于雷诺数低于1000，之间的接收器和导热油传热成为严重恶化.图5显示之间的太阳能集热效率和通量从早晨到下午关系。实验进行了从10:30至14:30，2009年6月3,。结果发现，在下午集电极效率比上午高。一个可能的原因是，如金属支架，波纹管的温度，相转移，主体结构热容量都非常低的早晨，和一个较大的太阳热能百分之是由人体吸收的相转移。随着太阳通量的下降，在导热油温度不会立即下降，因为个别的主体结构相转移高温。因此，温度的反应落后于太阳通量下降。当太阳通量是可变的，确定的温度滞后的时间间隔是至关重要的。最简单的情况，核实现有的热容量的影响是没有的吸收剂太阳的情况下。实验已完成从14:50至16：00，下午和导热油0.40流量立方米/小时可以看出，之间的出口与进口的导热油温度差异仍然保持积极，并成为与时间小。这是因为，能量吸收的导热油，如接收器的相转移，波纹管和支持的。15分钟后，几乎在15:05，进口和出口温度趋于相同。这一特点在设计中扮演一个抛物槽和经营太阳能发电厂的重要作用。类似的性质也报告了Senthilkumar等人。作者：热输出相转移依赖于吸收的能量采集器上的太阳辐射减少事故损失的收藏家，包括热损失，光损失。所以这是一个关键问题是如何确
定热损失。如果接收器外壁温度，得到从资历名册可以很容易地计算出一个已知的风速和环境温度大约在接收器领域的抛物线槽热损失。该方法可以不依赖于接收条件，如真空，空气失去了真空。本文的PT100温度传感器被包围的接收器和铝箔被用来阻止直接吸收辐射的正常温度传感器。类似的措施，方法等进行了Liipfert。，谁使用了红外线摄像头，以评估在原址抛物槽接受.热性能显示了热损失和温度随时间的上述环境温度导热油温度差异。根据实验已完成了约2米/秒，并在29.5-31.5摄氏度的环境温度范围内流速条件可以发现，上述的时间减少了环境温度导热油温度。这主要是因为采集器上的太阳辐射事故停机的时间。值得注意的是，上述环境温度导热油温度从13:30至13:50略有减少。一个可能的原因是该集电极阴影的影响和支持的太阳辐射增加很少。正如从图7还看到，通过对流和辐射热损失均随时间，热损失约为220瓦/米180℃之间的收藏家温度和环境温度，温差米，抛物面槽式太阳能集热器可达约总共10％收集太阳能效率。 结论
在该文件中的一个抛物面槽式太阳能集热器实验平台系统（相转移）的开发热发电，以及抛物槽集热器的性能进行了实验研究与合成油为散发有机热载体（导热油）。该集热效率相转移可在40％-60％之间。有人还发现，存在之间的导热油的温度响应和太阳通量，而在设计中发挥了重要作用相转移指定的延迟。对集热效率热损失的影响进行了研究，这对接收约220瓦/米在1180℃之间的收藏家温度和环境温度的温差，就约达总数的10％，太阳能事件收藏家。令人鼓舞的成果可为开发抛物面槽式太阳能在中国火电厂的基本数据。
太阳能系统设计师往往估计太阳能表现在网站上纬度为基础，平均天气简介及太阳能系统的特点。本文介绍了集成，这意味着预测日照与直接造成的损失太阳系底纹。一个自动化人工地平线采集和处理系统介绍。这种人为的地平线是太阳相比则在高时间分辨率的坐标，使底纹亏损可在任意时间总结。该方法适用于光电，太阳热能和建筑应用。一个子程序已制作了建模套件流行TRNSYS。
太阳能工程师将与作为一个有用的设计图熟悉的援助森帕特预测太阳发生收益制度。阻碍太阳的路径可能会不厌其烦地用测量仪器和测量在为系统区域设置和时代森帕特图叠加底纹推导从此。本文提供了一种替代人工捕捉的地平线部分自动化方法信息及其后续加工提供能源方面日晒造成的损失有意义的数据遮
阳系统
L亮度或亮度系数（0.0 <= L<= 1.0） R雷德颜色强度系数（0.0 <= R<= 1.0） G绿颜色强度系数（0.0 <=G<=1.0） B蓝颜色的亮度系数（0.0 <= B<= 1.0） j太阳方位角测量从真北（度） q测量太阳高度的垂直度（度） b人工地平线高程测量的水平度（度） D日期
K太阳常数K（W/m2时）
H型钢对地球表面的太阳曝晒在系统区域设置事件（W/m2时） M空气质量
人工捕获的地平线
低清晰度数码相机是用来捕获图像序列的地平线。数码相机安装在三脚架上的水平，由一方位驱动器驱动。此驱动器包括一个步进电机电子控制朝着一个方向方位后，一推到底60度按钮。每个动作后，数码照片，并采取在相机的记忆中。使用的驱动器允许六图像被捕获，从而覆盖整个地平线。此过程大约需要10分钟。该相机采用了一个大约64度的方位，海拔高度在48度视场。该三脚架是对齐，使图片的底部，是对0度仰角相对水平。对于遥远的地平线，这代表了海拔高程的地平线。后来出现的图像的宽度会跟60度，使他们配合邻近的图像。大多数相机镜头遭受非靠近图像的边缘线性度，通常被称为引脚缓冲作用。考虑到这一点，角的一格画面上绘制一个纸板（图片）采取了与叠加后的实际角网格。假设镜头是一个点的重点设备，人们可以很容易观察任何图像失真。图像失真被发现是近图像的界限具有重要意义。虽然其中有些是在加工过程中出现，没有进一步采取行动进行纠正。该一个数字图像线性化不是一件小事考虑到人工地平线是将检测到的像素的像素的基础。因此，必须了解定量可能引入错误这种近似。网格的图像分析产生错误的表1列出了这一特定数量摄像头。
该算法使用了一个方位，海拔度角分辨率的计算。因此相机必须有至少60帧像素宽。大多数数码相机使用的最小分辨率640 * 480像素，并利用这些额外
的决议是由歧视算法。这不是在CCD像素罕见设备失败，所以对一个邻近像素数均在鉴别使用。每个图像从相机下载到电脑使用的相机提供的软件。每个图片解压缩从JPEG形成为与设备无关位图（DIB）称为一个位图的形式。这种格式可以检查单个像素的元素。处理图像，需要检测的天空和人工地平线，一般差定义的树木和建筑物。彩色单是在许多情况下，贫穷歧视，并已发现更好地用于此目的的亮度。一旦转换为灰度，亮度或色彩亮度给予的关系
L ? 0.299 R ? 0.589 G ? 0.114 B
为遴选地平线的目的，必须是光明的天空和地平线的对象必须是代表 黑暗。它可能是有益的修改亮度信号色彩平衡，实现更大的反差。给出合理的对比与亮度的组成如下
L ? 0.3 R ? 0.4 G ? 0.3 B
现在可以扫描一个算法从顶部向底部的每个图像，一列的像素在一时间，这个搜索从黑暗的光亮度变化的人为改变，表示地平线。通过试验和错误是发现一对三组相邻的垂直像素加权平均了最好的分化和最佳宽容的地平线虚假的迹象。这种方法是不够好，让落叶树木被检测是在冬天。
太阳可以从一个坐标的算法所得的数字在当地的纬度，经度和时间（早发，1985）。太阳能坐标是由两个数量，方位和高度的代表在本文件。该太阳的轨迹在这个坐标系统可能是一一年和一天时间天之间绘制，引起到森帕特图极可能在形式或笛卡尔形式代表。人工地平线上还可以绘制一个森帕特图。重叠的地区之间的区域之外这种跟踪和太阳坐标代表在代表一个位置阴影的时期。这是方便写一个计算机算法来定位每个阴影事件。
上述结果的使用太阳能坐标作为预测对太阳系遮光时间的基础。虽然本身有用，遮光时间可能被用来预测太阳能人为造成的损失地平线，但日照时间的崩溃是众所周知的。该到达地球表面的太阳辐射常用近似依赖两种计算： 日照事件后，地球上的气氛与之间的距离变化地球与太阳（即季节性）。这是近似太阳常数k
K ? 1373 ( 1 ? 0.033cos ( 2p (d ? 10 ) / 365.25 ) ) W/m2
计算从太阳通过地球大气层的直接辐射的路径长度现场的兴趣。这被称为空气质量m M = 1 / cos (q)这种近似忽略了在低日照由于大气折射太阳的小变化过
程。这通常是因为太阳的强度较低，这些时间被忽视，然而这是当阴影是最有可能发生。因此，实证折射修正被列入。
M ? (1 / cos (q) ? 0.15 (3.885 ? q) ^ -1.253) * exp (-0.0001184 * Altitude)
大气吸收和网站的高度，但忽略的变化这个等式帐户在气压和温度因此，直接在两束跟踪已知纬度和日照时间收藏家轴事件可发现表达
H ? K? 0.7? (M ? 0.678) W/m2
这日晒可以随时改装成日照事件后的已知或其他平板方向跟踪模式
结合天气影响
读者将认识到，当地的天气条件将有一个大的太阳能量的影响能源失去了遮荫作用的结果。要结合天气的影响，人工地平线信息导入到一个模拟程序，利用气象资料。受欢迎的TRNSYS（克莱恩，1996年）的建模包是在这种情况下使用。 TRNSYS使用作为典型气象年的气象数据格式称为（TMY），这是一个有代表性的一小时平均气象数据集的时间分辨率。阿TMY文件位于为测试区域设置（莫里森等人，1988年）
阿新TRNSYS类型的构造，基本上是作为对日射光束信号数字滤波器的行为从现有TRNSYS辐射处理器。漫辐射不受影响。这种类型之间的连接辐射处理器类型和太阳能集热系统中的模拟类型。该算法很简单： IF [ (p/2 - q) > b(j) ] THEN H=K*M ELSE H=0 ENDIF
一个简单的TRNSYS模型来给予平均每年在天气条件下的定量结果在堪培拉接受调查的网站。这些结果可能是表列的任意时间间隔。每月摘要结果列于表1。为在堪培拉试验场模拟结果表明了约9％，原因是辐射年度亏损底纹从附近建筑物和树木。该森帕特图可用于追溯标识通过查找对象造成使用为指标，方位的原始图像对象的最大的损失。
结论
一种新的遮阳太阳能系统的影响现实世界的简单分析方法已被提出。该系统
提供了一个简单的工具，工程师进行定量预测障碍物附近的阴影效果。一新的子程序可用于TRNSYS用户谁愿意加入这除了他们的模型。
摘要
太阳能热水系统的必然要求存储由于太阳之间的热水供应和需求的时间差。通常情况下，储存容器使用一个辅助加热器，以保持在低日照时间的热水量。这种热水器的运作是由被动控制装置称为athermostat。本文提出了一种先进的控制解决方案，使温控器操作谨慎，从而降低太阳能是由辅助加热器操作流离失所。该控制算法是基于预测的能量平衡，尽管其行为可能会被人输入修改。 引言
对国内存储性能优化热水系统已收到近倍，虽然有一个有利的技术，主机可用性有限的注意。尽管有一些这个问题（Furbo 2000年，2000年的来临普鲁德姆和吉莱）数实际工作解决方案已在发现文学。本文着重就加强了热水储罐恒温性能。
恒温器
自动调温器是一个自动温度控制开关，在一个强大的热水储罐加热器。它一般有两个固定的温度设定，一转水加热器，另一个切换OFF，读者可能确定这种方法的几个缺点： 1.设置固定的温度
2.自动调温器在存储船只的位置是固定的 3.恒温开关一个非常强大的加热器 4.自动调温器是indiscrete关于开关时间 5.恒温槽是反应温度下降到 6.自动调温器的作用是立即
对反应的恒温控制和侵略行为简单化的内涵是热水系统表现不佳从能源效率和成本角度来看。当与一个太阳能集热器配合使用，性能更差。经过一个典型的早晨或傍晚高峰负荷，恒温槽发生反应，在温度下降，立即切换加热器，从而取代从系统不那么强大的太阳能投入。发生类似的效果后，晚上高峰负荷。太阳能集热然后有要么没有工作做，将运行在一个适当的收集效率低的太阳能热水系统MKDennis水温的升高预测能量平衡2太阳能电力系统2002年 - 澳大利亚和
新西兰太阳能学会卷一现在目前在储存容器。这可能被视为一个机会的损失。此外，恒温器将继续定期切换没有需求，例如在一夜之间（次加热器）和保持一个不必要的大量的水在高温下，从而提高存储寄生损失。自动调温器，因为它仍然是一种简单，廉价和可靠的设备。本文的挑战是控制恒温行动提供一个有远见和耐心，其运作的程度。
描述算法
一个热水系统的工作是由它的主要功能即在提供一个给定的时间最低温度的热水量。聪明的算法节省了拖延的存储容器的水加热到最后一刻的能量。因此，控制器必须能够预言的一些坦克在未来的能量平衡。这一段时间记的贮存容器，由地平线之间的坦克数量和加热sources.Example权力关系决定的：对于一个300升罐和一个2.4kW加热器，加热约7.3小时地平线在热水负荷超出地平线发生后，可以决定开关NOW.The贮存容器能量平衡的热水器不具有效力的能源供应和需求之一：
电力/燃气加热+太阳能援助=负荷冲压+电网线损
一旦最后三个组成部分是从当前时间量化到船只的视野，任何差别最多可使用电/燃气热水器的补充能源。该控制器的理解，这不会发生立即加热和有效负载中投下阴影的影响力定义状态变量的控制算法旨在控制的能源数量的热水箱中的水的形式储存。作者：罐的状态指示的定义是在热水中的能量超过临界温度的临界温度通常是高温回火阀门数额体现。它的目的是将状态变量连续使用坦克配置传感器的基础上，目前正在开发测量。国家变量的预测，在propogating平衡负载反馈学习算法误差的重要作用算法（后面介绍）。让我们定义一个性能指标，K中的控制算法。 K是本质上是一个错误的数量和被定义为
K? ( Eavail Eload ) / Eload
完美的表现是用K = 0和K表示，只有正面的价值观是可以接受的，这表明有足够的能量储存在坦克，以应付即将到来的负载。正数值表示的K过剩能量储存。人们可以很容易看到，K是一个有用的反馈变量，该控制器可以用于学习和修改未来的行为。 能源供应
虽然热输入从气体或电加热器是容易量化的，太阳能的投入将更加困难。太
阳能系统可能是模仿或学习，使该控制器是日照之间的关系所知，入口温度收藏家，收藏家流量和能源回收。如果日照配置文件是可用，那么太阳能输入可能是一个预测的基础上量化。合适的地方日照时间分辨率和可预测的工作继续提供这项服务，在气象学局。如果天气信息没有及时传达到控制器，控制器不承担任何的能量平衡的目的，因此在太阳能输入一个保守的方式运作。有一些从坦克状态反馈变过去的业绩指标。这并没有为未来的能源平衡的迹象。
读者将注意到了这一信息通信控制器的需要。这些算法被设计为运行在分布式智能环境，使节点通过网络连接到中央智能服务器。这是发生在世界各地的许多城市，甚至对这些算法验证所使用的控制器是网络还启用。 能源需求
重要的是，该控制器能够预测多少能量将耗尽系统和在什么时间，以便它能够呼吁不加热，以满足长时间存放多余的热量，从而招致机会这一要求的正确金额和以前discussed.It寄生损失似乎是不明智的尝试提供一个预先定义的加载配置控制器，因为每个热水安装后会在某些方面有所不同。这将是最好的控制器能够学会预测此一按个别案件的负担。幸运的是，人工神经网络（ANN）也提供了solution.A的人工神经网络的详细解释是这个文件的范围之外。总之，输入数据集一天的时间组成，每周日和目前的坦克状态变量是提供给神经网络，这反过来又利用其内部的映射提供了坦克，从中预期的能源负荷预测时。在整个操作的坦克，坦克状态变量实际上是表明在能源平衡预测误差，这是作为反馈变量来协助神经网络学习。值得注意的是，该神经网络结构自适应共振理论（ART）（格罗斯伯格，1988年）为基础，使新的学习信息并不需要为ANN.The热水负荷新的培训制度是在特定的能量，而不是数量上和反映了之间的供水温度和热参考温差通常采取的是自动调温器关闭temperature.Furthermore，人工神经网络识别在系统负荷预测从坦克的损失和水等同抽奖。 升压阴影算法
现在，对于预测能量平衡的组件可用，一个从热水器所需的净能源，可确定的时间范围。时间范围内的每一装载有能力影响控制器是否就加热器开关，因为现在每个有效负载的决定蒙上阴影回来的时候对当前时间。阴影的长度是加热器的权力代表。图2显示了一个低功率加热器必须比大功率加热器，以应付即将到
来的早期负载切换。在地平线多个负载只是积累的阴影区域，如图3所示。在这个例子中，请注意如何Load1不会要求加热将于但如果是为了累积的阴影不会Load2（2也提供负载在地平线上出现的坦克切换）。因此，工程向后算法从油箱地平线，累积负荷供热的阴影，腐烂在加热器功率打分。如果余额大于零的当前时间，加热器必须开启。 建模预测控制
该TRNSYS建模套件（克莱因1996年）是用来分析一个典型的热水系统的性能和没有预测控制器。该系统模拟了太阳能热水系统设在堪培拉服务在国内居住的四口之家。仿真使用的典型气象年的气象资料（莫里森等人1988年）和假设的预测日照完美的可用性。它还承担了即将到来的负载事件不完美的知识和人工神经网络模型。此负载是来自一个随机概率分布，并有相同的澳大利亚标准AS4234的总能量，虽然不同光谱的内容。仿真从而为真正的控制器的性能上限。然而，这是一个有用的工作，以取得此潜在想法。生动地展示了在行动算法。紫色线表示该单位罐温度，圆形的蓝线表示日晒而浅蓝色尖峰负荷表明正在从体系绘制红线指示的一号加热器立即通知开关温控器提供一个过多的热量，和不必要的隔夜加热是显而易见的。也可见，是加热后的水负荷事件和小太阳的贡献明显速度。该预测算法使水箱温度下降之间的了解，这是任何能才能满足需求，只是在时间负载。该模型证实，该做的工作提出意见，并提供比传统恒温器的潜在能源，节省成本的太阳能储存罐迹象。
仿真结果
结果表明提出的商业300升的存储连接到堪培拉，澳大利亚太阳能集热3.8平方米船只的年度业绩。一个热水3900kW.h总额每年来自系统。读者应该明白，这些结果是控制算法的性能指标，其结果是敏感的系统配置和系统如何被使用。负载配置文件是偏重在清晨和傍晚太阳落山后的高负荷。这大大不同负荷曲线从澳大利亚标准剖面，是更多的作者认为现实。一些身体的修改将需要储存罐，以实现这些好处，包括多种有效加热器的位置。这些修改将在未来的讨论文件。 讨论
这“只是在时间”的做法意味着，热水的最低产量以满足即将到来的负载在任何时候都应该保持。由于储罐并不总是充满热水，热分层将导致一些，这意味着
太阳能集热器将运行温度较低给水靠近槽底更有效率。事实上，任何效果，提高储存容器内的热分层将有助于改善效率。刚刚在时间的方法也意味着用完热水并铭记这一更大的风险，算法必须有能力评估当攻击或保守行为是适当的。这种现象是由调制精度，能量平衡，可预测。有人可能会争辩说，最终只是实时系统是一个瞬时热水器。这才是真正的所有，但不一定需要存储供应和需求之间的缓冲时间抵消太阳能系统，并在复杂的关税结构中存在。每一种能源在能源平衡的一个组成部分，是伴随着数量的不确定性，这决定了该算法变得咄咄逼人尽量减少额外的热水。这种“保守的能量”可以体现在两方面 ?截至恒温量增加额外的能量储存温度 ?在最小化的体积增加额外的能量储存温度。
模型显示，这些计划提供非常类似能源的额外能量储存效率处罚。然而，温度升高的情况似乎导致减少去层损失.这种保守主义意味着该系统的用户有一些控制。事实上用户可能希望重写控制器（开关，加热器或关闭行为的影响）学习算法和负载会无意中拿起作为一个大错误指示其预测，在坦克状态变量的变化。许多热水服务装有回火阀门，以确保烫水不能被传递到热水服务。卫生标准要求的贮存容器保持最低温度设定，以防止细菌生长，一般55℃至60℃的温度的调节阀门设置通常低于温控器设定，因此任何在此温度以上蓄水被认为是有用的，因为它的体积可能有助于热水负荷。因此，撤销，分层水箱之间的负载程度可能并不总是有害的。 结论
一个能源平衡预测算法，即赋予热水的耐心和远见美德系统恒温。更准确的能量平衡，可确定的，更为积极的控制器可在最大限度地节约成为同时保持合理的预测算法。这种有效性是敏感的初始热水系统适用性设计 并从中得出其负载的适用性。随着模拟随机载荷，建模表明能源和30％，为了节约成本都可用，虽然未必可以实现的。构建了一个控制器和基础设施的各个组成部分正在组装验证算法和怎么积蓄，其实可以意识到。这种办法将工作气体或电加热水箱，但提供了最大的太阳能辅助加热生活用水效益系统。
附录Ｂ 外文文献原文
Developing solar thermal power technology in an effective manner is a great challenge in China. In this paper an experiment platform of a parabolic trough solar collector system (PTCS) was developed for thermal power generation, and the performance of the PTCS was experimentally investigated with synthetic oil as the circulate heat transfer fluid (HTF). The solar collector’s efficiency with the variation of the solar flux and the flow rate of the HTF was identified. The collector efficiency of the PTCS can be in the range of 40%–60%. It was also found that there existed a specified delay for the temperature of the HTF to response to the solar flux, which played a significant role in designing the PTCS. The heat loss effect on collector efficiency was also studied, which was about 220 W/m for the receiver with a 180°C temperature difference between the collector temperature and the ambient temperature, amounting to about 10% of the total solar energy incident on the collector. The encouraging results can provide fundamental data for developing the parabolic trough solar thermal power plant in China.
Introduction
Parabolic trough solar thermal power technology has obtained wide popularity in the three solar thermal power technologies. Large scale parabolic trough collectors can supply the thermal energy that can be used to produce steam for a Rankine steam turbine generator. The PTCS is frequently employed for steam generation due to its high collector’s efficiency at the middle temperatures (around300°C). The performance of the PTCS plays an important role in the parabolic trough solar power plant, which directly determines the investment cost. Many investigations on optimizing and improving the performance of the PTCS have been carried out .Since the resource of solar energy is abundant in China,especially in the west, such as Qinghai Province, Tibet, Xinjiang Autonomous Regions, the parabolic trough solar thermal power technology has received extensive attention recently.The main motive of this paper is to develop an experimental platform of the PTCS, and evaluate the
performance of the PTCS. Furthermore, some key parameters that influence the performance of the PTCS, such as the solar flux,the flow rate of the HTF, heat loss, are also investigated.The encouraging results have been obtained, which can provide fundamental data for developing the parabolic trough solar thermal power plant in China.
Experiment platform and procedure
In order to investigate the PTCS for solar thermal powergeneration, an experiment platform was developed. The experiment platform mainly consists of four parts: the solar collectors, the heat transfer fluid (HTF) circulation, the cooling water circulation, and the measurement system.The flow diagram of the experimental platform using synthetic oil as the circulate HTF is shown in Figure 1. The HTF not only can circulate between the oil tank and the electric heater tank, but also can flow through the solar receiver unit. During a typical experiment, the HTF was firstly fed into the electric heater tank by a pump, and heated to a certain temperature. Then the HTF was pumped to the solar receiver unit to absorb the solar energy. The heated HTF leaving the receiver was cooled by the cooling water in a heat exchanger, and finally flowed back to the oil tank, the loop being finished. Considering the coefficient of volume expansion of the HTF at a high temperature, an expansion tank was fixed on the top of the experimental platform.It is worth noting that the energy provided by the electric heater is not considered in the process of evaluating the solar energy.The PTCS with 30 m2 was built in Tongzhou District,Beijing, as shown in Figure 2. Experiments were carried out from nearly 9:00 to 16:00 with different flow rates of the HTF.
Results and discussion
Many important factors, such as the solar flux, the flow rate of the HTF and the heat loss, influence the performance of the PTCS for thermal power generation. Therefore, they are taken into account in this study as follows.A typical result obtained in September 21, 2009 is shown in Table 2, in which the volume flow rate of the HTF is 0.45 m3/s. Ta, Ti, To and ηc stand for the ambient temperature,the inlet temperature of HTF, outlet temperature of the HTF and collector efficiency, respectively.Figure 4 shows the relationship between the efficiency of the PTCS and
the solar flux under different volume flow rates. The simulation result has been also obtained considering the solar flux varying from 200 to 900 W/m2 in accordance with the experimental condition. The flow rates of the HTF were fixed at 0.45 m3/h, 0.28 m3/h and 0.10 m3/h, respectively.From Figure 4, it can be found that the collector efficiency was in a range of 40%–60%. The experimental data scattered the simulation line. Analyzing Figure 4, itcan be found that the collector efficiency nonlinearly depends on the solar flux. When HTF flow rate was 0.45 m3/h,the efficiency had an increase of 7% from 200 to 550 W/m2,while only 3% from 550 to 900 W/m2, and the maximum of collector efficiency could be obtained at the highest solar flux within the study range. A reason is that the mean temperature of the receiver increased drastically for these values,the heat gained by the HTF went up slowly due to the heat loss. It can be also seen from Figure 4 that the collector efficiency was very low at the flow rate of 0.10 m3/h. Since the Reynolds number is below 1000, and the heat transfer between the receiver and the HTF becomes seriously worsened.Figure 5 shows the relationship between the collector efficiency and the solar flux from morning to afternoon. The experiment was carried out from 10:30 to 14:30, June 3,2009. It was found that the collector efficiency in the afternoon was higher than in the morning. A possible reason is the heat capacity of the main structure of the PTCS, such as the metal brackets, bellows temperatures, were very low in the morning, and a larger percent of the solar thermal energy was absorbed by the body of the PTCS. With the decrease of the solar flux, the temperature of the HTF would not fall immediately due to high temperature of the main body structure of the PTCS. Therefore, the temperature response lagged behind the decrease of the solar flux. When the solar flux is variable, identifying the time interval of the temperature lag is essential. The simplest case available for verifying the influence of the heat capacity is the case with no-sun on the absorber, as shown in Figure 6. The experiment was done from 14:50 to16:00 in the afternoon, and the flow rate of the HTF was0.40 m3/h. It can be seen that the temperature difference between the outlet and inlet of the HTF still kept positive and became smaller with the time. This was because that the energy absorbed by the HTF was from the PTCS, such as receivers,
bellows and supports. 15 min later, nearly at 15:05,the inlet and outlet temperatures tended to the same. This characteristic plays an important role in the design and operation of the parabolic trough solar power plants. The similar property was also reported by Senthilkumar et al.The thermal output of the PTCS depends on the energy that the absorbed solar radiation incident on the collector reduced the losses of the collector, including the heat loss,optical loss. So it is a key issue how to determine the heat loss. If the outside wall temperatures of the receivers are obtained, the heat loss in the parabolic trough field from the receivers Qr can be easily calculated approximately for a known wind speed and ambient temperature. The method cannot depend on the receiver condition, such as vacuum,air, lost vacuum. In this paper, temperature sensors of PT100 were surrounded the receivers, and aluminum foils were used to prevent the temperature sensors from directly absorbing the normal irradiation. The similar measure method was performed by Liipfert et al., who used an infrared camera to evaluate the in-situ thermal performance of parabolic trough receivers.Figure 7 shows the heat loss and temperature difference of the HTF temperature above the ambient temperature with time. The experiment was done under the condition of the velocity of about 2 m/s and the ambient temperature in the range of 29.5–31.5°C. It can be found that the HTF temperature above the ambient temperature decreased with the time. It is mainly because the solar irradiation incident on the collector goes down with the time. It is noteworthy that the HTF temperature above the ambient temperature has a small decrease from 13:30 to 13:50. A possible reason is the effect of the shading of the collector supports and a little increase of solar irradiation. As also seen from Figure 7, the heat loss by convection and radiation both decreased with the time, and the heat loss were about 220 W/m for the PTCS for a 180°C temperature difference between the collector temperature and the ambient temperature, which could amount to about 10% of the total solar energy incident on the collector.
Conclusion
In this paper an experiment platform of a parabolic trough solar collector system (PTCS) was developed for thermal power generation, and the performance of the
parabolic trough solar collector was experimentally investigated with synthetic oil as the circulate heat transfer fluid (HTF). The collector efficiency of the PTCS can be obtained in the range of 40%–60%. It was also found that there exists a specified delay between the temperature response of the HTF and the solar flux, which played a significant role in designing the PTCS. The heat loss effect on collector efficiency was also studied, which was about 220 W/m for the receiver at a 180°C temperature difference between the collector temperature and the ambient temperature, amounting to about 10% of the total solar energy incident on the collector. The encouraging results can provide fundamental data for developing the parabolic trough solar thermal power plant in China.
Designers of solar systems often estimate solar performance based on the site latitude, averaged weather profiles and the characteristics of the solar system. This paper presents a means of integrating this predicted insolation with losses due to direct shading of the solar system. An automated artificial horizon capture and processing system is presented. This artificial horizon is then compared to the solar coordinates at high time resolution so that shading losses may be summarised over an arbitrary period.The approach is applicable to photovoltaics, solar thermal and building applications. A subroutine has been produced for the popular TRNSYS modelling package.
Solar engineers will be familiar with the sunpath diagram as a useful design aid to predict the occurrence of solargain on a system. Obstructions to the solar path may be tediously measured using surveying equipment and superimposed upon the sunpath diagram for the locale of the system and the times of shading deduced henceforth. This paper provides an alternative partially automated means of capturing the artificial horizon information and its subsequent processing to provide meaningful data regarding insolation energy losses due to shading of the system
L Luminance or brightness coefficient (0.0 <= L <= 1.0) R Red colour intensity coefficient (0.0 <= R <= 1.0) G Green colour intensity coefficient (0.0 <= G <= 1.0)
B Blue colour intensity coefficient (0.0 <= B <= 1.0) J Solar azimuth measured in degrees from true North (deg) Q Solar altitude measured in degrees from the vertical (deg)
B Artificial horizon elevation measured in degrees from the horizontal (deg) D Ordinal day of year K Solar constant (W/m2)
H Beam insolation incident on Earth’s Surface at locale of solar system (W/m2) M Air mass
Capturing the Artificial Horizon
A low resolution digital camera is used to capture a sequence of images of the horizon.The digital camera is mounted on a level tripod and is actuated by an azimuthal drive. This drive consists of a stepper motor controlled electronically to move exactly 60 degrees in an azimuthal direction upon the push of a button. After each movement, a digital photograph is taken and stored in the camera’s memory. Use of the drive allows six images to be captured and thus cover the entire horizon. This procedure takes about 10 minutes.The camera used had a field of view of approximately 64 degrees in azimuth and 48 degrees in altitude. The tripod is aligned so that the bottom of the image represents an elevation of 0 degrees relative to the horizontal.For distant horizons, this elevation represents the elevation of the horizon. The images are later cropped in width to be exactly 60 degrees wide so that they align with neighbouring images.Most camera lenses suffer a degree of non-linearity near the edges of the image, often referred to as the pin cushion effect. To account for this, an image of an angular grid (drawn on a cardboard screen) was taken and superimposed upon an actual angular grid. Assuming that the lens is a point focus device, one may readily observe any image distortion. The image distortion was found to be significant near the boundaries of the image.
Although some of these areas are cropped during processing, no further corrective action was taken. The linearisation of a digital image is not a trivial matter considering that the artificial horizon is to be detected on a pixel by pixel basis. It is
therefore important to understand quantitatively the errors that might be introduced by this approximation. Analysis of the grid image produces the error quantities listed in Table 1 for this particular camera.
The algorithm uses an angular resolution of one degree in azimuth and altitude for its calculations. Thus the camera must have a frame at least 60 pixels wide. Most digital cameras use a minimum resolution of 640*480pixels and this extra resolution is utilised by the discrimination algorithm. It is not uncommon for pixels in CCD devices to fail and so an average of a number of neighbouring pixels is used in the discriminator.
Each image is downloaded from the camera to the PC using the software provided with the camera. Each picture is decompressed from JPEG form into a bitmap form known as Device Independent Bitmap(DIB). This format allows inspection of individual pixel elements.Processing the images requires detecting the difference between sky and the artificial horizon, commonly defined by trees and buildings. Colour alone is a poor discriminator in many circumstances and it has been found better to use luminance for this purpose. Once converted to gray scale, the luminance or brightness of a colour is given by the relation
L ? 0.299 R ? 0.589 G ? 0.114 B
For the purpose of selecting the horizon, the sky must be bright and objects representing the horizon must be dark. It may be beneficial to modify the colour balance of the luminance signal to achieve greater contrast.Reasonable contrast is given with the following luminance composition
L ? 0.3 R ? 0.4 G ? 0.3 B
An algorithm may now scan each image from top to bottom, one column of pixels at a time, searching for this change from light to dark luminance change that represents the artificial horizon. Through trial and error it was found that a weighted average of groups of three neighbouring vertical pixels gave best differentiation and best tolerance to false indication of the horizon. This method was good enough to allow deciduous trees to be detected in winter.
The solar coordinates may be obtained from a number of algorithms given the
local Latitude, Longitude and time(Rabl, 1985). Solar coordinates in this paper are represented by the two quantities, azimuth and altitude. The locus of the sun in this coordinate system may be plotted for a range of days of year and times of day, giving rise to a Sunpath Diagram that may be represented in polar form (Figure 2) or cartesian form (Figure 3).
The artificial horizon may also be plotted on a sunpath diagram. The region of overlap between the area outside of this trace and the solar coordinates represents a period of shading at the represented location. It is convenient to write a computer algorithm to locate each shading event.
The above results use the solar coordinates as the basis of predicting the timing of shading on the solar system.Although useful in itself, the timing of shading may then be used to predict solar energy loss due to the artificial horizon, provided that temporal breakdown of insolation is known.A commonly used approximation for the insolation reaching the surface of the Earth relies on two calculations:
Insolation incident upon the Earth’s upper atmosphere which varies with the distance between the Earth and Sun (i.e. seasonal). This is approximated by the Solar Constant, K.
( 1 ? 0.033cos ( 2p (d ? 10 ) /3 65. )2 )5 K ? 1373W/m2
Calculation of the path length of direct radiation from the sun through the Earth’s atmosphere to the locale of interest. This is referred to as the Air Mass, M
M ? 1 / cos (q)
This approximation ignores a small change in insolation due to atmospheric refraction at low solar elevation. This is usually ignored since the sun’s intensity is low at these times, however this is when shading is most likely to occur. Thus the empirical refraction correction (Hu and White,1983) is included.
M ? (1 / cos (q) ? 0.15 (3.885 ? q) -1.253) exp (-0.0001184 Altitude)
This equation accounts for atmospheric absorption and the altitude of the site but ignores changes in air pressure and temperature Thus the direct beam insolation incident on a two axis tracking collector with known Latitude and time can be found
from the expression
H ? K ?0.7(M 0.678)W/ m2
This insolation may be readily modified into insolation incident upon a flat plate of known orientation or other mode of tracking (Rabl, 1985)
Incorporating Weather Effects
The reader would recognise that local weather conditions will have a major influence on the amount of solar energy lost as a result of shading effects. To incorporate weather effects, the artificial horizon information is imported into a modeling program that utilises weather data. The popular TRNSYS (Klein, 1996) modelling package is used in this case. TRNSYS uses a format of weather data known as Typical Meterological Year (TMY) which is a representative average weather dataset of one hour time resolution. A TMY file was located for the test locale (Morrison et al ,1988)
A new TRNSYS Type was constructed that essentially acts as a digital filter on the beam insolation signal from the existing TRNSYS Radiation Processor. The diffuse radiation is unaffected. This Type connects between the Radiation Processor Type and the Solar Collector Type in the system simulation. The algorithm is simply: IF [ (p/2 - q) > b(j) ] THEN H=K*M ELSE H=0 ENDIF
A simple TRNSYS model was constructed to give quantitative results for average annual weather conditions at the surveyed site in Canberra. These results may be tabulated by arbitrary time interval. Monthly summary results are presented in Table 1.
The simulation results for the test site in Canberra indicate an annual loss of radiation of around 9% due to shading from nearby buildings and trees. The sunpath diagram may be used to retrospectively identify the objects causing the greatest losses
this technology.Figure 4. graphically demonstrates the algorithm in action. The flat purple line indicates tank temperature, the rounded blue line indicates insolation while the light blue spikes indicate load being drawn from the system.Finally, the red line indicates the switching of the heater.One notices immediately that a thermostat provides too much heat, and unnecessary overnight heating is apparent. Also visible is the rapidity of reheating of the water after a load event and the small solar contribution evident. The predictive algorithm allows the tank temperature to drop between loads knowing that it is able tomeet any demand just-in-time. The model confirms that the ideas presented do work and provides an indication of potential energy and cost savings over a conventional thermostat based solar storage tank.
Simulation Results
The results presented indicate annual performance for a commercial 300L storage vessel connected to 3.8m2 of solar collector in Canberra, Australia. A total of 3900kWh of hot water is drawn from the system annually. The reader should understand that these results are indicative of the control algorithm performance and that the results are sensitive to system configuration and how the system is used. The load profile is biased towards high loads in the early morning and evening after sunset. This load profile differs substantially from the Australian Standard profile and is more realistic in the author’s view. Some physical modifications would be required in the storage tank to realise these benefits, including a variable effective heater location. These modifications will be discussed in a future paper.
DISCUSSION
This “just-in-time” approach implies that the minimum volume of hot water to meet an upcoming load should be maintained at all times. Since the storage tank is not always full of hot water, some thermal stratification will result and this means that the solar collectors will run more efficiently from lower temperature feed water near the bottom of the tank. Indeed, any effects that enhance thermal stratification within the storage vessel will assist in improving efficiency.The just-in-time approach also implies a greater risk of running out of hot water and with this in mind, the algorithm must have the ability to assess when aggressive or conservative behaviour is
appropriate. This behaviour is modulated by the accuracy to which the energy balance may be predicted. One might argue that the ultimate just-in-time system is an instantaneous heater. This is true for all but solar systems that necessarily require storage to buffer the time offset between supply and demand, and where complex energy tariff structures exist.Each component in the energy balance is accompanied by an uncertainty quantity and this determines how aggressive the algorithm becomes in minimising additional hot water. This “conservative energy” could manifest in two ways
? Additional energy storage as additional volume at the thermostat temperature ? Additional energy storage as additional temperature in the minimised volume.
Models suggest that these schemes provide very similar energy efficiency penalties for the additional energy storage. However the increased temperature scenario seems to result in less destratification loss.This conservatism implies that the user of the system has some influence on the behaviour of the controller.Indeed the user may wish to override the controller (switching the heater either ON or OFF) and the load learning algorithm will inadvertently pick up the change in tank state variable as an indication of a large error in its prediction.
Many hot water services are fitted with tempering valves to ensure that scalding water cannot be delivered to a hot water service. Health standards demand that the storage vessel maintain a minimum temperature setting to prevent bacterial growth, typically 55°C to 60°C. The temperature setting for the tempering valve is typically lower than the thermostat settings and so any water in the store above this temperature is considered useful in that it’s volume may contribute to the hot water load. Thus, a degree of de-stratification in the tank between loads may not always be detrimental.
CONCLUSION
A predictive energy balance algorithm is presented that endows hot water system thermostats with virtues of patience and foresight. The more accurately the energy balance may be determined, the more aggressive the controller may become at maximising savings while maintaining reasonable reliability.The effectiveness of the predictive algorithm is sensitive to the suitability of the initial hot water system
designand its suitability for the load drawn from it. With simulated stochastic loads, modelling suggests energy and cost savings in the order of 30% are available, although not necessarily achievable. A controller has been constructed and the various components of the infrastructure are being assembled to validate the algorithm and to asses what savings can actually be realised.This approach will work on gas or electrically heated tanks, but provides greatest benefit in solar assisted domestic water heating systems.