基于聚类算法和数值模拟的建筑群平面优化

Optimization of Building Group Arrangementbased on Clustering Algorithm and Numerical Simulation

摘要:

为将建筑空间热舒适度评价应用于建筑群平面的设计优化上，提出基于Kmeans聚类的分组偏移变量设计方法，并结合该变量设计、数值模拟、遗传算法和通用热舒适评价指数，在Matlab软件上实现建筑平面优化流程，最后对集中式、分散式、边流式的水体配置下的建筑平面进行了优化实验。结果表明，优化后的建筑空间，平均热舒适度评价指数（UTCI）值降低了0.1~0.6°C，热舒适度均有所增强。在较强热应力的范围内，建筑空间内的平均热舒适度改善的主要目标是提高平面内的风速。

Abstract:

To apply the thermal comfort into the optimization of building arrangement, a variables design method based on Kmeans clustering was proposed. Then, it was combined with numerical simulation, genetic algorithm and universal thermal climate index(UTCI) to implement the building arrangement optimization on Matlab. Finally, the building arrangements with centralized type, decentralized type and edge flow type water displacement were optimized respectively. The results showed that after optimizing, a reduction of 0.1~0.6°C of UTCI was observed. It is alsoconcluded that in the moderate heat stress, the increasing of wind speed is the key of thermal comfort improvement.