论文部分内容阅读
中庭作为一种大空间建筑,其热环境的设计和组织控制十分复杂,现有的空调设计方法难以满足实际工程设计的需要。本文从中庭建筑热环境的物理特性入手,首先分析了影响中庭热环境的关键性因素,明确了中庭内部空气流动传热模型和动态热过程模型之间的耦合关系,并在此基础上提出了一种全新的基于神经网络的中庭热环境全年动态模拟方法。该方法采用神经网络算法提取热环境参数分布CFD模拟的关键信息,使其既能够覆盖全年不同的中庭热环境状况,又能够保证合理的工况计算数量以提供给神经网络建模。而后在软件中实现了神经网络与热过程模拟耦合迭代的中庭热环境全年动态模拟,并运用此方法针对实际工程中的中庭空调系统设计问题进行了合理的模拟分析,给出了改进建议。
As a big space building, the design and organization control of the thermal environment in the atrium is very complicated. The existing air conditioning design method can not meet the needs of the actual engineering design. This paper begins with the physical characteristics of atrium thermal environment. Firstly, the key factors that affect the thermal environment of atrium are analyzed. The coupling between the air-flow heat transfer model and the dynamic thermal process model is defined. Based on this, A New Method of Annual Dynamic Simulation of Atrium Thermal Environment Based on Neural Network. This method uses neural network algorithm to extract the key information of thermal environment parameter distribution CFD simulation so that it can cover the thermal environment conditions of different atriums throughout the year and can also ensure the reasonable calculation of working conditions to provide neural network modeling. Then, a year-round dynamic simulation of atrium thermal environment with neural network and thermal process simulation coupled iteration is implemented in the software, and a reasonable simulation analysis of the design of atrium air conditioning system in practical engineering is made with this method, and suggestions for improvement are given.