近年来大量区域供冷系统在我国得到实际工程应用。通过案例实测看到，这些区域供冷系统的实测供冷量、管网散热量和系统效率都与设计预期有较大差异。现有的区域能耗模拟方法中大多首先进行单体建筑需冷量计算，然后将区域中各类建筑简单叠加以获取区域总体需冷量，且在系统模拟过程中，忽略了区域大型管网内冷冻水温度的时空分布，因而无法全面反映区域供冷系统随服务规模增大所带来的建筑需冷量、系统供冷量和运行能耗特征变化。因此，亟需提出一套科学准确的区域供冷系统动态能耗模拟方法，使得模拟结果能够更加全面地反映区域供冷系统在需冷量、管网散热量和系统效率等方面的实际运行特征，从而更好地为实际工程提供技术支撑。本研究以此为题，开展了以下研究工作：一、针对区域供冷系统开展大量实测案例调研与分析工作，总结归纳了区域供冷系统中用户的空调使用方式、建筑需冷量特征和系统运行能耗等。通过调研测试可以看到，由于区域中用户空调使用行为的巨大差异，引起了各户空调需求的不同步，从而造成建筑整体需冷量特征随着系统服务规模增大而发生巨大变化。二、基于详细的定量计算案例显著性分析，发现人行为的多样性和随机性是区域供冷系统需冷量的主要影响因素。本研究所提出的基于人群位移和空调使用行为分布的建筑需冷量动态模拟方法，首先基于大量问卷调研结果提炼为若干种典型人行为，继而基于人行为的强度分布将区域中的人行为扩展为“各个相似、各有不同”的行为模式，进而全面反映出不同系统规模下区域供冷系统需冷量的特征及变化。与常规方法采用单一、确定的人员和空调使用作息相对比，所提出的新方法全面体现了人行为多样性和随机性对区域供冷系统需冷量特征的影响。三、基于用户侧整体水力热力特性、管道动态热平衡方程和管网水力热力平衡方程，构建了区域供冷系统动态能耗计算模型。该模型通过详细模拟管网内冷冻水温度和流量的时空分布，可以更为准确地获得在不同系统运行策略和建筑需冷量条件下的系统的供冷量和能耗。利用以上研究所提出的区域供冷系统动态能耗模拟方法，更好地反映区域规模建筑需冷量的整体规律和个体差异，使得计算结果更加贴近实际需冷量特征，同时，更为准确地反映冷冻水温度和流量的时空分布对系统供冷量和能耗的影响。可为实际区域供冷系统工程设计、运行调适等工作提供更为科学准确的技术支撑。
In recent years, district cooling systems （DCS） have been used in many practical projects in China. Existing results on measurements in some practical DCSs have shown that there is a significant gap between design and operation of DCS on cooling consumption, heat loss of network and system efficiency. Conventional simulation methods of DCS energy consumption usually first predict the cooling load of one single building, and then duplicate the result for serval times to obtain cooling load of the whole district. In addition, most of these methods fail to calculate the temporal and spatial distribution of chilled water temperature in the large pipe network. As a result, existing simulation methods can hardly fully capture the change on features with increase of service area of DCS, i.e., the cooling load, supply cooling consumption, and energy consumption. Therefore, it is necessary to propose a new method for modelling dynamic cooling load and energy consumption of DCS in order to make the simulation results closer to the reality, which facilitates valuable simulation analysis work and provides technical support in practice as well. This research contains the following research work:First, this study summarized typical AC-use patterns, cooling load features and system operation features in DCS based on large-scale survey and practical measurements. Through these case studies, we found that huge diversity exists among users’ AC-use patterns, and it can cause the unsynchronization of cooling load in different rooms. This also leads to the change on features of DCS when system scale increases.Second, this study discovered that the diversity and randomness of occupant behavior （OB） are the main influencing factors of cooling load in DCS. We proposed a dynamic modeling method of cooling load based on occupant distribution with the following steps: 1） summarizing several typical OB patterns based on questionnaire survey; 2） using the intensity distribution of OB to change the occupants in serveral typical OB patterns into a group of people with completely different OB models. This method can fully reflect the features change of cooling load with the increase of system scale. Compared to existing methods, which used single and deterministic OB model, it can embody the OB diversity in district.Third, we established the dynamic energy consumption model of DCS by incorporating the overall thermo-hydraulic feature of the user-side, the temperature increase and time delay of the chilled water during supply process, and the hydraulic and thermodynamic equilibrium of distribution network. This model can reflect the temporal and spatial distribution of chilled water temperature and flow in the large pipe network. It can also be used to calculate the supply cooling and energy consumption more precisely.The proposed dynamic modeling method of cooling load and energy consumption for DCS provides a better description of global feature and individual difference on cooling load within a district. This method can also be applied to obtain temporal and spatial distribution of chilled water temperature and flow in the large pipe network. Preliminary results of this study have shown that the simulated results from the newly-proposed method are much closer to the real situation.
