Development of Surrogate Models forLatent Thermal Energy Storage Systems with Macro-encapsulated Phase Change Material

宏观封装相变材料潜热储能系统替代模型的开发

• 批准号: 444616738
• 负责人: Dr.-Ing. Andreas König-Haagen
• 金额: --
• 依托单位: Escuela de Ingeniería de Bilbao
• 依托单位国家: 德国
• 项目类别: WBP Fellowship
• 财政年份: 2020
• 资助国家: 德国
• 起止时间: 2019-12-31 至 2021-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/444616738?language=en
• 关键词: Development; Surrogate; Models; forLatent; Thermal

So far, only highly simplified models are available for the simulation of a complete latent thermal energy storage with macro-encapsulated phase change material (PCM). However, extensive preliminary work has shown that single capsules of these storage units can already be described satisfactorily with the aid of CFD simulations. But the CFD simulation of a complete storage unit cannot be realized due to the required computing and storage capacity. The aim of this project is therefore to derive surrogate models based on CFD models which precisely predict the phase change processes in latent thermal energy storage systems with macro-encapsulated PCM, but require only a small fraction of the computational effort of a complete CFD simulation. The surrogate models to be developed are divided into two groups. The first group consists of data-fit surrogate models which can be subdivided into CFD surrogate models and real-time surrogate models. In the CFD surrogate models, the heat transfer fluid is explicitly simulated, whereas the capsules are considered with the help of boundary conditions and the data-fits coupled to them. The real-time surrogate models, on the other hand, approximate the behavior of the entire storage unit using data-fits. The second group of surrogate models is based on a projection of a high-resolution model onto a subspace, which greatly reduces the number of equations to be solved. This method is first applied to Stefan problems in MATLAB and then transferred to CFD simulations of latent thermal energy storage units with macro-encapsulated PCM. In the context of this project the following questions are to be answered:1. To what extent are adjustments to existing CFD models necessary in order to be able to derive surrogate models for latent thermal energy storage units with macro-encapsulated PCM?2. How large are the deviations between surrogate models and CFD models for the simulation of a single capsule?3. Are projection-based surrogate models suitable for application to latent thermal energy storage units with macro-encapsulated PCM and latent thermal energy storage devices in general?4. How close is the agreement between surrogate models and test results for a complete storage system?5. Are the deviations from simplified to detailed surrogate models significant in relation to the deviations to the experiments?6. How do the boundary and initial conditions affect the three points mentioned above?7. What influence does the choice of the transferred parameters have on the effort involved in creating the surrogate models and the computational effort as well as accuracy of the simulations performed with the surrogate models?

到目前为止，对于大封装相变材料（PCM）的完全潜热蓄能的模拟只有高度简化的模型。然而，大量的初步工作表明，这些存储单元的单个胶囊已经可以通过CFD模拟得到令人满意的描述。但由于计算和存储容量的限制，无法实现完整存储单元的CFD模拟。因此，该项目的目的是基于CFD模型推导替代模型，该模型可以精确预测具有宏观封装PCM的潜热储能系统的相变过程，但只需要完整CFD模拟的一小部分计算量。待开发的代理模型分为两组。第一组由数据拟合代理模型组成，可细分为CFD代理模型和实时代理模型。在CFD替代模型中，传热流体是显式模拟的，而胶囊是借助边界条件和与之耦合的数据拟合来考虑的。另一方面，实时代理模型使用数据拟合来近似整个存储单元的行为。第二组代理模型是基于高分辨率模型在子空间上的投影，这大大减少了需要求解的方程的数量。首先将该方法应用于MATLAB中的Stefan问题，然后将其应用于宏观封装PCM潜热储能装置的CFD仿真。在这个项目的背景下，需要回答以下问题：1。为了能够推导出具有宏观封装PCM的潜热储能单元的替代模型，需要对现有CFD模型进行多大程度的调整？对于单个胶囊的模拟，代理模型与CFD模型的偏差有多大？基于投影的替代模型是否适用于具有宏观封装PCM的潜热储能装置和一般的潜热储能装置？对于一个完整的存储系统，代理模型和测试结果之间的一致性有多接近？从简化到详细的替代模型的偏差与实验偏差是否显著？边界和初始条件如何影响上述三点？转移参数的选择对创建代理模型所涉及的工作量、计算工作量以及使用代理模型执行的模拟的准确性有什么影响？