壓水式反應器的燃料在運轉約18個月後需進行再裝填工作,新燃料採購案為購買三個週期,所需燃料組合由電力公司向國外廠家提出週期能量需求,廠家再依其需求提出燃料價格,但國內電力公司對於廠家燃料選擇方式詳情無從得知,無法評估其設計是否合適。另外依各週期所需能量,新燃料所使用的濃縮度及可燃毒物棒根數皆為之決定,且每個週期皆受前面兩週期的爐心狀態所影響,更增加其設計難度,故產生發展自動化搜尋新燃料組合的動機。 本研究為執行多週期新燃料組合自動化設計,使用和聲搜尋演算法進行新燃料組合搜尋。和聲搜尋演算法為創作音樂時所啟發的半啟發式最佳化演算法。較傳統演算法的優勢在於使用較少的數學條件、不需要設定初始解,在考量現有所有解的情況下產生新的解,以上特色展現該演算法的彈性,並可搜尋到更佳的解。 當程式挑選出符合條件之燃料組合後,再將燃料束填入固定佈局進行爐心計算,會對熱通道因子、緩和劑溫度係數與週期末硼酸水溶液濃度進行分析;經濟方面,程式則會計算低濃縮度新燃料根數,數量多為佳。當符合程式停止條件後,將該燃料組合再進行自動化爐心佈局調整以驗證符合安全限值。 本研究之程式以C語言編寫,並以三維節點爐心模擬程式SIMULATE-3進行爐心計算。本研究的驗證對象為馬鞍山核電廠一號機第18至21週期,研究結果顯示一次搜尋多週期與個別搜尋單一週期兩種方法,皆有能力搜尋出符合安全限值及週期能量需求之新燃料組合,程式總執行時間相近,無法定論何種方式取得的新燃料束組合較佳。
For about every 18 months, the pressurized water reactor needs to be reloaded. Usually, fuel procurement consists of three fuel cycles. Based on the provided energy demand of each cycle, the fuel vendor will propose the composition and price of the fresh fuel assemblies. But how to decide the composition of fresh fuel assemblies is not known. Depending on each cycle’s required energy, enrichment and integral fuel burnable absorber (IFBA) number of each cycle are determined. Moreover, the current cycle status is affected by the preceding two cycles’ core condition; it will increase the difficulty of design. Thus, the motivation of automatic design is conceived. In this study, the harmony search (HS) algorithm is applied to automatically search the composition of fresh fuel assemblies. HS is inspired by the composition of harmony, and it is a metaheuristic optimization algorithm. Advantages of HS are using less mathematical conditions, no requirement of an initial solution, and considering the existing solutions to produce new solutions. These features make HS more flexible and be able to search better solution. After the composition of fresh fuel assemblies are decided, the selected fuel assemblies was permuted according to the fixed LPs, which were calculated by SIMULATE-3 to obtain moderator temperature coefficient, hot channel factor, and boron concentration at the end of cycle. After the proper FA composition is determined, the accurate LP design is searched to guarantee the satisfactory requirements. The developed program was coded by using language C. SIMULATE-3 code was used to calculate core status. The Maanshan unit I cycle 18 to 21 was designed to verify the effectiveness of HS. The results show that searching for tandem cycles and searching for successive cycles are able to find the composition of fresh fuel assemblies which satisfy the constraints. Since the computing time of these two methods are almost the same, it is difficult to conclude which method is better.