Design optimization of air distribution systems in non-residential buildings

作     者：Jorens, Sandy Verhaert, Ivan Sorensen, Kenneth 

作者机构：Univ Antwerp EMIB HVAC Campus GroenenborgerBldg ZGroenenborgerlaan 171 B-2020 Antwerp Belgium Univ Antwerp Operat Res Grp ANT OR City CampusPrinsstr 13 B-2000 Antwerp Belgium 

出 版 物：《ENERGY AND BUILDINGS》 (能源与建筑物)

年 卷 期：2018年第175卷

页      面：48-56页

核心收录：

学科分类：0820[工学-石油与天然气工程] 08[工学] 0813[工学-建筑学] 0814[工学-土木工程] 

主　　题：Air distribution system design Ductwork layout Duct and fan sizing Heuristic optimization algorithm Multi-start local search Test case 

摘      要：Centralized air distribution systems in non-residential buildings are characterized by an extensive air distribution network, that has to be built in a building environment with finite degrees of freedom. The ductwork layout, i.e., the network structure of the ducts, as well as the number and location of the fans, has a large impact on the total cost and performance of the air distribution system. Nevertheless, existing air distribution system design methods are limited to the sizing of each duct (and fan) in the network. The layout itself is considered predetermined, and thus not explicitly taken into account for optimization. In this paper, we meet this shortcoming by presenting the air distribution network design optimization method, that is able to calculate the optimal air distribution system configuration, i.e., the optimal layout and duct and fan sizes, while minimizing the total cost of the air distribution system. A multi-start local search algorithm is developed, consisting of a constructive and a local search phase. In the first phase, multiple air distribution system configurations are generated, and evaluated for feasibility. In the local search phase, all feasible solutions are further optimized in terms of material costs by decreasing and increasing the duct diameters following the steepest descent/mildest ascent approach. An application of the algorithm on a realistic test case demonstrates its usefulness in practice. (C) 2018 Elsevier B.V. All rights reserved.