中央空調系統之運轉參數主機冰水出水溫、二次側冰水流量、空調箱送風量,分別影響主機耗電量、二次側泵浦耗電量、空調箱送風機耗電量。在相同負載條件下,此三個運轉參數值有眾多種組合,如何求出最佳組合是本論文主要目的。 本研究先紀錄主要影響空調系統耗電量之各種參數值,利用迴歸分析建立主機耗電量模型、二次側泵浦耗電量模型、空調箱送風機耗電量模型。採用基因規劃法搜尋空調系統運轉於最低耗電量時的運轉參數組合,且空調系統提供的冷凍噸符合空調空間所需的負載量。基因規劃法演算過程先隨機產生冰水出水溫、二次側冰水流量、空調箱送風量,代入各個設備耗電量模型,分別產生主機耗電量、二次側水泵耗電量、空調箱耗電量,此三個參數設為基因規劃之終點端,利用數學符號(+)作為基因規劃法之函數端,再以基因規劃法之運算子作迭代運算,經由實驗結果指出,最終得到最低耗電量之運轉參數組合,其空調系統於此運轉參數組合下可提供符合空調空間所需的負載量,且較固定運轉參數模式之耗電量低,而達到節能目的。
Operation parameter of central air-conditioning system: the temperature of chilled water from the chiller, secondary chilled water flow, and supply air flow of the air handing unit affect the chiller power consumption, secondary pump power consumption, and air handing unit fan power consumption respectively. There are a range of combinations relating to these three operation parameters for suppling the same cooling load. The objective of this research is to derive the best combination. This research first records the various parameters that affect the air-conditioning system before establishing the chiller power consumption model, secondary pump power consumption mode, and air handing unit fan power consumption model through regression analysis. Genetic Programming is used to search for the operation parameter combinations generated while the air-conditioning system is operating at the lowest point of power consumption, and the refrigeration ton supplied by the air-conditioning system meets the cooling load requirement for the configured cooling space. The Genetic Programming process first produces data of chilled water temperature, secondary chilled water flow, and supply air flow of the air handing unit randomly. The derived data is then input the power consumption model of each facility to generate three parameters: chiller power consumption, secondary zone pump power consumption, and air handing unit power consumption. The three parameters are then configured as the terminals for the genetic programming, the mathematical symbol “+” is used as the functions for the genetic programming, and the option of GP which contains ‘crossover’, ‘reproduce’, and ‘mutation’ for iterant computing the population is composed of functions and the terminal of GP. The results show that an air-conditioning system operating under the derived operation parameter combination supplies the cooling load needed for the configured cooling space, as well as achieved the objective of power-saving as it indicates lower power consumption than an air-conditioning system operating under a constant operation parameter.