【摘要】 目前，空调系统的优化设计存在巨大的问题，一方面传统优化算法不能适应多变量高维度的末端优化问题，另一方面，空调的负荷分布存在极大的随机性，一旦设计选型偏离设计工况，会使空调系统性能的下降，造成极大能源浪费。针对这些问题，以能源利用效率为目标函数，使用随机走步变步长的近似优化方法对末端选型进行优化，并通过次优理论寻求对不同负荷分布适用性最好的末端选型。结果表明：随机走步变步长的近似优化算法对不同的负荷分布的平均能源利用效率的优化增量可达0.014。对于不同的随机负荷波动，反向验证后次优解的平均能源利用效率增量可达0.087，且优化后95%置信区间的下限比传统设计95%置信区间的上限高0.04，证明近似优化算法可以提高中央空调冷冻水系统的能源利用效率以及对随机负荷的适应性，对末端的选型优化有重要指导作用。更多还原

【Abstract】 At present, there are huge problems in the optimization design of air conditioning system. On the one hand, the traditional optimization algorithm cannot adapt to the end-of-pipe optimization problem with high multi-variable dimension. On the other hand, the load distribution of air conditioning has great randomness. Once the design selection deviates from the design condition, the performance of air conditioning system will decline, resulting in great energy waste. In view of these problems, this paper takes energy utilization efficiency as the objective function, uses the approximate optimization method of random walking variable step size to optimize the air conditioning terminal, and seeks the air conditioning terminal selection with the best applicability to different load distributions through the suboptimal theory. The results show that the optimization increment of average energy utilization efficiency for different load distributions can reach 0.014. For different random load fluctuations, the average energy efficiency increment of the sub-optimal solution after reverse verification can reach 0.087, and the lower limit of the95 % confidence interval after optimization is 0.04 higher than the upper limit of the 95 % confidence interval of the traditional design. It is proved that the approximate optimization algorithm can improve the energy efficiency of the central air conditioning chilled water system and the adaptability to random load, which has an important guiding role for the selection and optimization of air conditioning terminals.更多还原