透過您的圖書館登入
IP:18.224.214.215
  • 期刊
  • OpenAccess

運用模糊序率動態規劃於水庫操作之研究

Study on Reservoir Operation by Using Fuzzy Stochastic Dynamic Programming

摘要


臺灣地區興建水庫日愈困難,而水資源的需求卻日益增加,致水資源的利用更加珍貴。傳統之水庫操作規線均是利用歷史流量以模擬法所制定,然當特殊水文事件發生時(如異常豐水年、乾旱年),利用該操作規則難以獲致滿意的操作結果。近年由於電腦的進步,以優選法規劃水庫多目標最佳運用成爲可行,本研究探討模糊序率動態規劃(Fuzzy Stochastic Dynamic Programming, 簡稱FSDP)求取滿足水庫供水率的最佳長期蓄水規線,發現藉由模糊集理論的引進,使得傳統序率動態規劃運用時處理目標函數的方式更加擴展。以FSDP模式與現行規線(M-5)之操作方式模擬放水的結果可知,FSDP模式雖在缺水月份較現行規線爲多,然對總缺水量及最大缺水量皆較現行規線爲小,即FSDP模式求得之長期水資源利用較現行規線為佳;另將異常豐水年、乾旱年之水位及缺水量作比較可知,現行規線較缺乏彈性,而FSDP模式利用時段的入流量機率資訊所建立之操作規線,可適當的以期望水位作水資源合理利用的調整,故整體而言,以FSDP所獲致的結果有較佳的表現。

並列摘要


Recently building a reservoir is much more difficult in Taiwan, while the demand of water resources is consistently increasing. Consequently, how to utilize the water resources becomes an important task. The traditional reservoir operation rules were obtained through the analysis of historical inflow records. It couldn't obtain an appropriate operation for the statistically special hydrologic events. The recent progress in computer technology making the optimal algorithms employed in multiple reservoir programming becomes possible and easier. In this study, the Fuzzy Stochastic Dynamic Programming (FSDP) is utilized for gaining the optimal operation rules. The results of the operating simulation reveal that the FSDP model is better than the traditional M-5 rule curves in terms of utilization of water resources. Although the months of water shortage simulated by the FSDP model are little more than those of M-5 rule curves, the total quantity and the maximum monthly water shortage simulated by the FSDP model are less than those of the M-5 rule curves. Meanwhile, comparing the results of the water level and lack quantity in special flood events and drought years obtained from the M-5 rule curves and the FSDP model, the former is inflexible, while the latter model can alter the water resources appropriately by the expected water level. The results, as a whole, demonstrate that the FSDP medel is better than M-5 rule.

延伸閱讀