Reducing the emissions in the building sector requires reasonable renovation and integration of renewable energy sources to find the optimal trade-off between ecological benefit and economic effort. However, many buildings, different types of usage, needs of residents, and transient weather conditions contribute to a vast number of use cases, making the search for the optimal trade-off challenging. Therefore, a systematic reduction of use cases is necessary to target emission reductions in the building sector. Furthermore, emissions in the building sector mainly belong to heating and cooling purposes. Thus, many frameworks exist to estimate the heating and cooling demand at building and district scales. In addition, some contributions exist at a national scale, determining spatially resolved demands. However, the optimal integration of fluctuating renewable energy sources requires a detailed resolution in the time domain. Currently, no framework allows the calculation of spatially and temporally resolved demands at the national scale from which a meaningful use case reduction can be obtained.

This work proposes the ACoolHeaD framework for Automated Cooling and Heating Demand calculations for entire urban districts and countries based on spatially and temporally resolved building performance simulations. ACoolHeaD connects predefined formatted input data with the TEASER tool for automated parameterization of building simulation models. After parameterizing the building models, the framework selects representative weather data sets as boundary conditions for the simulation to ensure spatial and temporal resolution during computation. The calculation of the demands is fully automated. Finally, the k-medoids clustering algorithm is applied to systematically reduce the number of use cases, identifying representative demands. As a use case for ACoolHeaD serves Germany since sound data available. Thus, ACoolHeaD estimates the representative heating demand for about 19 million buildings in Germany in this work.

An extended input data preparation reduces the computational effort from about 19 million building simulations to 3,520. The average, maximum, and integral heating demand is determined in each simulation, which are classification indicators for the clustering algorithm. Germany’s integral overall (clustered) heating demand is estimated at 2,796 PJ (2,793 PJ), respectively, which is about 10 % higher than the current values of the Federal Ministry (2,557 PJ). The result shows a good agreement and a successful application of ACoolHeaD considering all necessary assumptions. Based on the maximum and average heating demand, the entire building stock is clustered by k-medoids clustering to five buildings, which can be used for further representative investigations considering Germany. As a next step, we recommend investigating the influence of occupancy, the selected weather data sets, and analyzing cooling for future scenarios to increase the detail level and thus improve the expressiveness of ACoolHeaD results.

减少建筑行业的排放需要对可再生能源进行合理改造和整合，以在生态效益和经济努力之间找到最佳平衡点。然而，许多建筑物、不同类型的使用、居民的需求和瞬息万变的天气条件会导致大量用例，这使得寻找最佳权衡具有挑战性。因此，有必要系统地减少用例，以实现建筑部门的减排目标。此外，建筑部门的排放主要属于供暖和制冷用途。因此，存在许多框架来估计建筑物和区域尺度的供暖和制冷需求。此外，一些贡献存在于全国范围内，决定了空间解决的需求。然而，波动的可再生能源的最佳整合需要时域中的详细分辨率。目前，没有任何框架允许在国家范围内计算空间和时间上解决的需求，从中可以获得有意义的用例减少。

这项工作提出了ACoolHeaD为框架一个utomated首席运营官玲HEA婷d emand计算基于空间和时间分辨的建筑性能模拟整个城区和国家。酷头将预定义格式的输入数据与 TEASER 工具连接起来，以实现建筑模拟模型的自动参数化。在对建筑模型进行参数化后，该框架选择具有代表性的天气数据集作为模拟的边界条件，以确保计算过程中的空间和时间分辨率。需求的计算是完全自动化的。最后，应用 k-medoids 聚类算法系统地减少用例的数量，识别有代表性的需求。作为一个用例，ACoolHeaD服务于德国，因为声音数据可用。因此，ACoolHeaD在这项工作中估计了德国大约 1900 万座建筑物的代表性供暖需求。

扩展的输入数据准备将计算工作量从大约 1,900 万次建筑模拟减少到 3,520 次。在每次模拟中确定平均、最大和积分加热需求，它们是聚类算法的分类指标。德国的整体（集群）供暖需求估计分别为 2,796 PJ (2,793 PJ)，比联邦部的当前值 (2,557 PJ) 高约 10%。结果显示了ACoolHeaD的良好一致性和成功应用考虑所有必要的假设。基于最大和平均供暖需求，整个建筑存量通过 k-medoids 聚类为五座建筑，可用于考虑德国的进一步代表性调查。作为下一步，我们建议调查占用率、选定的天气数据集的影响，并分析未来场景的冷却，以提高细节水平，从而提高ACoolHeaD结果的表现力。