Remodeling of RC Circuit Building Thermodynamics Model with Solar Radiation Based on A Regularization-like Technique

Remodeling of RC Circuit Building Thermodynamics Model with Solar Radiation Based on A Regularization-like Technique
复制标题

DOI:
--
复制
发表时间:
2019-06
期刊:
2019 12th Asian Control Conference (ASCC)
影响因子:
--
通讯作者:
T. Hatanaka;Tomohiro Ikawa;Daichi Okamoto
T. Hatanaka;Tomohiro Ikawa;Daichi Okamoto
中科院分区:
其他
文献类型:
--
作者:
T. Hatanaka;Tomohiro Ikawa;Daichi Okamoto

文献摘要

相似文献

本文提出了一种利用太阳辐射对RC(阻容)电路建立热力学模型的改造方案。首先建立了东京工业大学一幢真实的建筑的三维结构模型,并在此基础上建立了详细的EnergyPlus仿真器,利用MATLAB的BRCM(Building Resistance-Capacitance Modeling)工具箱建立了其RC电路模型。然后证实,RC电路模型在存在太阳辐射的情况下并不总是适合EnergyPlus的温度轨迹,而在没有辐射的情况下,它工作得很好。因此,本文旨在通过EnergyPlus生成的数据重新确定太阳吸收系数,该系数量化了太阳辐射对区域温度的影响。我们首先应用标准的最小二乘法,并指出两个困难:(i)的秩的数据矩阵变得病态,和(ii)增加的数据量是禁止的,由于高维的RC电路模型。因此,我们采用正则化技术,而不增加计算工作量,并通过与EnergyPlus数据的比较来验证所得到的模型。
In this paper, we present a remodeling scheme of so-called RC (Resistance–Capacitance) circuit building thermodynamics models with solar radiation. We first build a 3-D building structure model of a real building in Tokyo Institute of Technology, a detailed EnergyPlus simulator associated with the structure model, and its RC circuit model identified by BRCM (Building Resistance-Capacitance Modeling) MATLAB toolbox. It is then confirmed that the RC circuit model does not always fit the temperature trajectories by EnergyPlus in the presence of solar radiation, while it works well in the absence of radiation. This paper is thus intended to reidentify the solar absorption coefficients, which quantifies the effects of solar radiation on zone temperatures, through data generated by EnergyPlus. We first apply the standard least square method and point out two difficulties: (i) the rank of the data matrix gets ill-conditioned, and (ii) increasing the volume of the data is prohibited due to high dimensionality of the RC circuit model. We thus employ a regularization technique without increasing the computational efforts and validate the resulting model through comparisons with the EnergyPlus data.