ESTIMATION OF HEAT TRANSMISSION THROUGH WINDOW FOR CFD SIMULATION OF INDOOR ENVIRONMENT USING VARIATIONAL CONTINUOUS ASSIMILATION METHOD
ESTIMATION OF HEAT TRANSMISSION THROUGH WINDOW FOR CFD SIMULATION OF INDOOR ENVIRONMENT USING VARIATIONAL CONTINUOUS ASSIMILATION METHOD
复制标题
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
T. Matsuo;A. Kondo;H. Shimadera;A. Komatsu;S. Shiochi
中科院分区：
文献类型：
--
作者：
T. Matsuo;A. Kondo;H. Shimadera;A. Komatsu;S. Shiochi
在利用计算流体力学（CFD）对室内热环境进行评价时，边界条件的不确定性会影响评价的准确性。本研究应用变分连续同化法（VCA）来估计边界场，以提高CFD模拟的精度。通过数值实验验证了该方法的有效性，将VCA方法应用于计算窗内的温度场、流场和传热。实验按照以下步骤进行：首先，通过正确的边界条件下的CFD模拟来创建“真实”温度场和流场，然后，从“真实”温度场中提取“观测数据”，最后，通过CFD模拟来创建“初始条件”，而不需要关于通过窗口的热传递的边界条件，最后，通过计算流体力学模拟来创建“初始条件”，并通过计算流体力学模拟来创建“初始条件”。然后用VCA方法将“观测数据”同化为“初始条件”，最后将同化结果与“真实”温度场、流场和边界条件进行比较。实验结果表明，VCA方法能够以可接受的精度估计通过窗户的热传递、温度场和流场。引言为了适当地管理室内热环境，有必要了解温度场和流场。计算流场的方法有几种，大致分为观测和CFD模拟两种。观测可以获得关于温度和流量的准确数据，但从观测成本的角度来看，很难对整个房间进行观测。另一方面，CFD模拟可以很容易地估计温度和流场，但它需要精确的边界条件才能获得准确的数据。对于室内热环境的估计，很难设置精确的边界条件，因为例如通过窗户的热传递取决于纬度、季节、天气和时间。此外，还有许多不确定的边界条件，如灯光，设备和居住者的加热。因此，如果我们能从少量的观测数据中估计出精确的边界条件，这是值得的。本研究利用资料同化方法，从观测资料中估计透过窗户的热传递。在以往的研究中，发展了许多资料同化方法。其中一些是为估计室内环境而开发的。例如，一些方法通过求解输运方程来执行源估计，而另一些方法计算潜在源位置与观测浓度之间的关系。虽然这些方法对源估计是有用的，但它们不能用于温度和流场的估计，因为这些方法假设精确的流场是已知的。有几种方法可以对流场进行修正。Nakagawa等人使用了一种成本函数法，通过最小化剩余的控制方程以及观测值与CFD计算值之间的差异来校正温度场和流场。然而，该方法不能用于估计边界条件，因为该方法直接校正温度场和流场。Sasamoto等人发展了另一种方法，即计算室内气候贡献率（CRI）来评估室内热因子对温度分布的贡献，并利用CRI从实测气温中估计各热因子对温度分布的影响。由于该方法不仅可以估计温度场和流场，而且可以估计热因子的边界条件，因此如果有足够的计算资源来计算每个因子的CRI，则该方法看起来是有吸引力的。在这项研究中，使用了另一种数据同化方法。变分连续同化（VCA）方法是Derber等人发展的，作者对它进行了改进。该方法通过在CFD控制方程中加入修正项来修正CFD模拟。修正项可以假定为源项，因此该方法可以用于边界条件的估计。VCA方法的细节将在下一节中描述。
In estimation of indoor thermal environment using computational fluid dynamics (CFD), uncertainty of boundary conditions will affect the accuracy of the estimation. In this study, the variational continuous assimilation (VCA) method was applied to estimate the boundary filed and to improve the accuracy of the CFD simulations. The method was validated by the numerical experiment, which applied the VCA method to estimation of the temperature field, flow filed, and heat transmission through the window. The experiment was performed according to the following procedure: first, the “true” temperature and flow fields were created by a CFD simulation with correct boundary conditions; second, “observation data” was extracted from the “true” temperature field; third, “initial conditions” was created by a CFD simulation without the boundary condition about heat transmission through the window; fourth, the “observation data” was assimilated into the “initial condition” by the VCA method; finally, the result of the assimilation was compared to the “true” temperature field, flow field, and boundary conditions. As a result of the experiment, it was confirmed that the VCA method could estimate the heat transmission through the window, the temperature field, and the flow field, with acceptable accuracy. INTRODUCTION In order to manage indoor thermal environments appropriately, it is necessary to understand temperature and flow fields. There are several methods to estimate the fields, which roughly divided into the two methods: observations and CFD simulations. Observations can obtain accurate data about temperature and flow, but it is difficult to observe whole room from the view point of observation costs. CFD simulations, on the other hand, can easily estimate temperature and flow fields, but it requires accurate boundary conditions to obtain accurate data. For the estimation of indoor thermal environments, it is difficult to set up accurate boundary conditions, because heat transmission through a window, for example, depends on the latitude, season, weather and time. In addition, there are many uncertain boundary conditions such as heating of lights, equipment, and occupants. Thus, it is worth if we can estimate accurate boundary conditions from a few observation data. In this study, a data assimilation method was used to estimate heat transmission through window from observation data. In previous studies, many data assimilation methods were developed. Some of them were developed for estimation of indoor environment. For example, some methods performing source estimation by solving transport equation reversely, and other methods calculate the relationship between potential source location and observed concentration. Although these methods are useful for source estimation, they can not be used for the estimation of temperature and flow field because these methods assume that the accurate flow field is known. A few methods can correct flow field. Nakagawa et al. used a cost function method which corrects the temperature and flow fields by minimizing the reminder of governing equations and the differences between observed values and CFD calculated values. This method, however, cannot be used for estimation of boundary conditions because the method correct temperature and flow fields directly. Sasamoto et al. developed another method which calculates the contribution ratio of indoor climate (CRI) to evaluate the contribution of indoor heat factors to temperature distribution; and the CRI was used to estimate effect of each heat factor from observed air temperature. Since this method can estimate not only temperature and flow field but also boundary conditions of heat factors, it looks attractive method if there is enough calculation resource to calculate the CRI of each factor. In this study, another data assimilation method was used. The method called variational continuous assimilation (VCA) method was developed by Derber, and modified by authors. The method correct CFD simulations by adding a correction term into the governing equations of CFD. The correction term can be assumed as the source term, thus the method can be used for estimation of boundary conditions. The detail of the VCA method is described in next section.