在中央空調系統中,冰水主機通常僅有on/off卸載控制邏輯,而此邏輯僅能使單一主機運轉於最佳點,如為多部主機並聯,此邏輯便無法有效的使多部主機達到最佳運轉點,因此在主機運轉控制上,應有相當多的調度與排程空間。 目前冰水主機負載分配的最佳化方法有平均負載法、拉格蘭傑法、基因演算法等,這些方法皆有其缺點。例如平均負載法 (Average Loading,AVL) 往往不是最佳運轉點、拉格蘭傑法 (Lagrangian Multiplier Method,LGM) 其λ值初值設定不當則會有發散的問題。雖然基因演算法(Genetic Algorithm , GA)可用來克服此缺點,但其求解過程需應用到複製、交配、突變等機制及編碼、解碼等運算,使程式的撰寫較為麻煩且費時。 本文以三次多項式來表現冰水主機之功率(Kilowatt)及部分負載率(Partial Load Ratio,簡稱PLR)性能曲線,並提出進化規劃法(Evolutionary Programming , EP)進行冰水主機最佳運轉排序,並以系統總耗電為目標函數,在相關限制條件及滿足系統負荷需求下,應用此演算法以機率方式擇優選取下一代,且運用高斯突變的技巧可使搜尋範圍擴大,以避免陷入局部最小值,使主機運轉於最大效率以達到節約能源的目的。
In the air-conditioning system, Chillers usually have on/off Load-shedding control logic. This logic only makes one of the chillers work in the optimal condition. If the chillers are parallel, this logic can not make the chillers work in the optimal condition. Accordingly, there are many deployments and sequence spaces in the control of the chillers operation. Recently, the optimal chiller loading methods include Average Loading (AVL) method, Largrangian Multiplier (LGM) method and Genetic Algorithm (GA). These methods have some shortcomings. For example, AVL method is not optimal. LGM method will diverge if the initial condition isn’t suitable. Although the GA method overcomes these shortcomings of Lagrangian method, reproduction, crossover, mutation, encoding and decoding are applied to the process of evolution. It makes the process of evolution is very complicated and makes the coding of program more difficult. This thesis uses a cubic equation to simulate the chiller’s curve. It applies the Evolutionary Programming to Optimal Chiller Sequencing and finds a set of chiller output which doesn’t violate the operating limits and satisfy the demand of system load. It applies Evolutionary Programming (EP) choose Filial Generations by the probability method. Besides, it applies Gauss Mutation to amplify searching range and avoid caving in Local Optimal. The purpose of this thesis makes the operation of machine achieve the most efficiency in order to attain economize energy.