With the growing interest in urban scale simulation, urban building energy modelling tools (UBEM) are being developed. The purpose of these UBEM tools is to tackle multiple issues that can range from buildings retrofit potential to the sizing of thermal and/or electrical network or to the assessment of renewable energy sources potential. Unlike the modelling of a single building, collecting accurate building information data at district scale is nearly impossible and therefore the uncertainties increase. In this context, the selection of the most relevant and adequate models with regard to the available data and simulation objectives is challenging, and often a trade-off between accuracy and parametrization feasibility has to be found. A novel approach is proposed to choose the right level of model complexity for thermal heating demand simulation in buildings at district or city scale, by assessing in particular the impact of different thermal zoning methods with the UBEM tool DIMOSIM. This method for assessing the parsimony in modelling is applied to a selection of models from the literature in order to find a compromise between the available data, the different modelling levels of detail, the expected output and the computation time, according to district and buildings characteristics. By applying this approach, it has been found that a division per floor for building modelling is in most cases the most parsimonious choice for thermal heating demand simulation at district scale. Eventually, the coupling between thermal zoning models and solar shading models is analysed through the same methodology for assessing the modelling parsimony. The results show that the combination of simpler models (in particular a thermal zone per floor or mono-zone buildings for thermal zoning models, and a solar shading coefficient calculated per floor or facade for solar models) can be parsimonious enough to be used to determine the annual heating demand, speeding up strongly the simulation at the same time.

随着人们对城市尺度模拟的兴趣日益浓厚，正在开发城市建筑能源建模工具 (UBEM)。这些 UBEM 工具的目的是解决多个问题，范围从建筑物改造潜力到热力和/或电力网络的规模或可再生能源潜力的评估。与单个建筑物的建模不同，在区域范围内收集准确的建筑物信息数据几乎是不可能的，因此不确定性增加。在这种情况下，选择与可用数据和模拟目标最相关和最合适的模型是具有挑战性的，并且通常必须找到准确性和参数化可行性之间的权衡。提出了一种新方法，通过特别是使用 UBEM 工具 DIMOSIM 评估不同热分区方法的影响，为区域或城市规模的建筑物中的热供暖需求模拟选择合适的模型复杂度级别。这种评估建模简约性的方法适用于从文献中选择的模型，以便根据地区和建筑物特征在可用数据、不同建模细节级别、预期输出和计算时间之间找到折衷方案. 通过应用这种方法，我们发现建筑建模的每层分区在大多数情况下是区域尺度热供暖需求模拟的最简洁的选择。最终，通过与评估建模简约性相同的方法来分析热分区模型和遮阳模型之间的耦合。结果表明，简单模型的组合（特别是热分区模型的每层热区或单区建筑物，以及太阳能模型的每层或每立面计算的遮阳系数）可以足够简洁，可用于确定年供暖需求，同时强烈加速模拟。