DOI: 10.3390/buildings16214225 ISSN: 2075-5309

Synergistic Optimization of Daylighting and Thermal Performance and Heterogeneous Design Strategies for Teaching Buildings with Different Exterior-Corridor Orientations

Shuyuan Mai, Xiongjun Tao, Ming Yang

Reducing operational energy use in teaching buildings is an important way to lower carbon emissions from buildings. For teaching buildings with exterior corridors in Guangzhou, corridor orientation changes solar exposure, shading, and daylight conditions, making it difficult to improve daylighting and cooling performance using the same design strategy. This study examines typical linear teaching buildings with exterior corridors in Guangzhou through parametric modeling, building performance simulation, multi-objective optimization, and machine-learning classification. Compared with the baseline, the overall optimal north-facing-corridor solution increased spatial daylight autonomy and useful daylight illuminance by 0.104 and 0.097, respectively, and reduced cooling energy consumption intensity by 12.35%. The corresponding south-facing-corridor solution increased the two daylighting measures by 0.068 and 0.018 and reduced cooling energy consumption intensity by 19.40%. The high-performance solutions occupied different parameter ranges: classroom depth was mainly 8.5–9.0 m and corridor width 2.6–2.8 m for the south-facing corridor, where the wider corridor provided self-shading; for the north-facing corridor, classroom depth was 8.5–10.0 m and corridor width mainly 2.2–2.6 m. The multilayer perceptron classifier achieved a harmonic mean of precision and recall of approximately 0.98, supporting rapid screening during early-stage design. Feature-attribution analysis showed that classroom depth made the largest contribution to high-performance classification for the north-facing corridor, whereas corridor-side window-to-wall ratio contributed most for the south-facing corridor. Thus, corridor orientation was associated with differences in performance values, favorable parameter distributions, and design-parameter priorities. Within the sampled design space of the Guangzhou case, the same parameter-adjustment sequence should not be applied to both corridor configurations.

More from our Archive