In the simulation of building overheating risks, the use of typical meteorological years (TMY) can greatly reduce the simulation workload and accurately reflect the distribution of simulation results according to the weather conditions over a given period. However, all meteorological parameters in most current TMY methods use a uniform weighting factor which may make the simulation results against the actual simulation results of the period and negatively affect the accuracy of the evaluation results. In addition to differences in climate characteristics between climate zones, the sensitivity of different simulation results to external parameters will also be different. Therefore, a TMY method based on the Finkelstein-Schafer statistical method is proposed, which considers the climatic characteristics of different regions and the correlation with the output parameters of indoor simulation to select the typical month. The proposed method is demonstrated in the three future scenarios for the three cities in different climate zones in China. The results show that the traditional TMY method has an overestimated weight of solar radiation and wind speed and an undervalued weight of dry bulb temperature when indoor temperature-related indicators are the output target. Compared with the traditional TMY method, the TMY generated by the improved method is closer to the distribution characteristics of the long-term outdoor weather data. Furthermore, when using the improved TMY data to evaluate the overheating performance of the passive residential buildings, the difference of the results of the unmet degree hours, indoor overheating degree, and the overheating escalation factor between the long-term projected data and the TMY data can be reduced by 63%–67% compared with the traditional TMY data.

In the building with many transparent envelopes, solar radiation can irradiate on the local surface of floor and cause overheating. The local thermal comfort in the room will be dissatisfactory and the thermal performance of radiant floor will be strongly affected. However, in many current calculation models, solar radiation on the floor surface is assumed to be uniformly distributed, resulting in the inaccurate evaluation of the thermal performance of the radiant floor. In this paper, a calculation model based on the theory of discretization and the RC thermal network is proposed to calculate the dynamic thermal performance of radiant floor with the consideration of unevenly distributed solar radiation. Then, the discretization model is experimentally validated and is used to simulate a radiant floor heating system of an office room in Lhasa. It is found that with the unevenly distributed solar radiation, the maximum surface temperature near the south exterior window can reach up to 35.6 ℃, which exceeds the comfort temperature limit and is nearly 8.5 ℃ higher than that in the north zone. Meanwhile, the heating capacity of the radiant floor in the irradiated zone can reach up to 171 W/m², while that in the shaded zone is only 79 W/m². The model with the assumption of uniformly distributed solar radiation ignores the differences between the south and north zones and fails to describe local overheating in the irradiated zones. By contrast, the discretization model can more accurately evaluate the thermal performance of radiant floor with the influence of real solar radiation. Based on this discretization model, novel design and control schemes of radiant floor heating system can be proposed to alleviate local overheating and reduce heating capacity in the irradiated zone.