臺灣博碩士論文加值系統

台灣地區年平均降雨量約2,500公厘，惟因降雨時空分配不均，豐枯懸殊，每人每年平均可用水量僅為世界平均值的七分之一，屬於水資源相對不足的地區。除受限於水文不均與不足現象外，近年來受全球氣候變遷、地區降異常雨及工業用水成長需求增加影響，如何加強掌握乾旱現象及水資源調度，以穩定供應各標的用水需求並兼顧水資源之永續利用，為水資源管理單位刻不容緩之艱鉅任務。
水庫在水資源利用上具有調洪濟枯之功能，在臺灣地區更具有其重要性，尤其在枯水旱季期間，各標的用水量往往需藉助於水庫的調蓄來滿足其需水量。在水庫系統中，一般以操作規線進行運轉，根據現況蓄水量，依其運轉規則調蓄利用，因此水庫現況蓄水量，決定了能否滿足未來用水需求，另外，水庫的調蓄利用與水文條件息息相關，因此水文洪枯變化對水庫所反映之訊息，不僅是水庫操作管理之重要資訊，更是水資源應用不可或缺之資訊。
本研究採用經濟部水利署依據民國94年2月經濟部頒布之「區域水資源預警機制」所建立的枯旱預警綜合指標中的水庫蓄水率及水庫的入流量指標，作為水庫供水決策參考指標因子，利用模糊理論“亦此亦彼”的特性與調適性網路模糊系統模式，探討曾文水庫的供水決策參考指標其與水庫操作決策關係。
依本研究結果顯示，曾文水庫於豐水期若能考量稍為減供10%以內各標的用水，即不一定要按操作規線完全供水，若逢枯水乾旱缺水時，則水庫將可多提供約10%到20%之供水，除了增加供水外，亦可疏緩供水壓力，尚不失為水庫操作的另類決策方案。

In Taiwan area, the annual average rainfall is about 2,500mm. Because the rainfall space-time assigns unevenly, abundant and withered and with a wide gap, the water volume that can be utilized is only about 1/7 of the average of the world on average for each person every year, it is the relatively insufficient area of water resources in the world.
For recent years, except that it is uneven with fewer than the phenomenon that limited to hydrological condition, global climate changes, area unusual rainfall and increasing water use for industrial demand are the main factors of water resources management. How to strengthen and grasp the phenomenon on land of the universe drought and schedule of water resources are becoming more serious. In order to supply the water demand of each goal and give consideration to the continuous utilization of the water resource steadily, it is a very urgent and difficult task for management authorities of water resources.
For helping and dealing with the withered function in water resources utilization, the reservoirs play an important role and have its importance even more in Taiwan. Especially during low water dry season, the management authorities often needed to make use of every amount of water stored in the reservoir to regulate and meet the water use required.
In general, for reservoir’s operation system, the operation rules is applied to go on, namely according to present situation of water demand to regulate and utilize in accordance with their operation rule. Therefore, the present situation of the reservoir has determined whether could meet the water demand in the future. In addition, it is closely linked with hydrological condition in the future that the reservoir regulation and utilization, so the big withered change of hydrology, to the information that the reservoir reflects, not only the important information of operating management in the reservoir, it is that the water resources uses indispensable information even more.
This research adopts two indicators, storage rate index and inflow index of reservoir of Tsengwen reservoir, as decision-making referenced factors. Above two indexes of reservoir are set up by WRA, which is issued by MOEA in February, 2005 for building a warning mechanism according the condition of drought and local areas of water resources. By using the fuzzy theory characteristic of “ that also this also “ and the models set up by adaptive network fuzzy inference system, we are trying to evaluate and get the better operation rules of Tsengwen reservoir. Reviewing and comparing with the traditional operation curves, we hope to find the best relation of two reservoir’s indicators and water supply policy of Tseng reservoir in order to obtain another view of water supply of reservoir for management authorities.
Accordance to the results of study, taking Tsengwen reservoir for example, the reservoir could be considered and reduced the water volume required for each goal’s demand less than 10% in wet season, namely should not be according to offering the complete water supply of operation rules. In This case, when meeting with the universe drought and lacking of water, then the reservoir would offer roughly about 10% of the water demand to 20% or more. It could not only dredge the pressure of management authorities, but also offer more water for all purposes, still could yet be regarded as the scheme of decision of different class operation rules for reservoirs.

謝誌辭……………………………………………………………i
中文摘要…………………………………………………………ii
SUMMARY …………………………………………………………iii
目 錄…………………………………………………………v
圖 目 次…………………………………………………………vii
表 目 次…………………………………………………………viii
符 號 表…………………………………………………………ix
第一章 緒論
1.1 前言…………………………………………………………1
1.2 研究動機……………………………………………………1
1.3 研究目的與範圍 …………………………………………2
1.4 研究方法 ……………………………………………………3
1.5研究流程 ……………………………………………………5
第二章 文獻回顧
2.1 模糊理論 ……………………………………………………6
2.2 類神經網路 …………………………………………………9
2.3 台灣乾早問題………………………………………………11
第三章 理論分析
3.1 模糊理論……………………………………………………17
3.2 模糊聚類……………………………………………………21
3.3 類神經網路…………………………………………………26
3.4 調適性網路模糊推論系統…………………………………30
3.5 模式評鑑指標………………………………………………40
3.6 MATLAB軟體簡介……………………………………………41
第四章 供水決策參考指標
4.1 供水決策參考指標之選定…………………………………43
4.2 乾旱預警綜合指標…………………………………………43
4.3 水庫蓄水率決策參考指標之建立…………………………49
4.4 水庫進水量決策參考指標之建立…………………………52
第五章 水庫操作決策庫之建立
5.1 水庫操作運用規線…………………………………………55
5.2 水庫操作決策規則庫之建立 ……………………………58
第六章 模式建構及驗證
6.1 模式建構原理………………………………………………63
6.2 輸出及輸入變數定義………………………………………65
6.3 模糊推論系統建構…………………………………………66
6.4 適應性網路模糊推論系統模式建置………………………70
6.5 模式成果與檢討……………………………………………74
第七章 模式應用與檢討………………………………………81
第八章 結論與建議
8.1 結論…………………………………………………………85
8.2 建議…………………………………………………………87
參考文獻 ……………………………………………………… 89
附錄一、經濟部「農業用水調度使用協調作業要點」………91
附錄二、ANFIS模式參考程式 …………………………………93

中文部份
[1] 張斐章(2003),「類神經網路—理論與實務」,松岡圖書,台北。
[2] 井上 洋．天笠美知夫(2000),「模糊理論」,五南書局,台北。
[3] 藎壚(1991)，「實用模糊數學」，亞東書局，台北。
[4] 張躍、鄒壽平、宿芬(1992)，「模糊數學方法及其應用」，煤炭工業出版社，北京。
[5] 經濟部水利署南區水資源局(2004-2006),「南部地區水資源決策支援系統建置計畫報告」。
[6] 經濟部水利署南區水資源局(2002)，「台灣省南區水資源統籌調配管理系統規劃總報告」。
[7] 經濟部水利署水利規劃試驗所(2004),「乾旱潛勢分析與預測模式之建置計畫」期中及期末報告。
[8] 經濟部水利署(2006),「台灣地區水資源決策支援系統建置報告」。
[9] 吳正逵、鄭義重(2004),經濟部水利署南區水資源局,「曾文水庫乾旱預警探討報告」。
[10] 陳昶憲、楊朝仲、王益文(1996)，類神經網路於烏溪流域洪流預報之應用，中華水土保持學報，第27卷，第4期。
[11] 黃文政、劉佳明(1997),「高屏溪、曾文溪流域可用水量及用水現況之探討」 ,經濟部水資源局研究報告。
[12] 黃文政、吳建民、謝宏智(1994)，模糊聚類模式之應用，第7屆水利工程研討會。
[13] 張斐章、黃浩倫、張麗秋，2001，回饋式類神經網路於河川流量推估之應用，農業工程學報，第47卷，第2期。
[14] 經濟部水資源局(1999), 「乾早指標之建立-中部及南部」，經濟部水資源局研究報告。
[15] 黃文政、楊富堤(1996)，「河川設計流量之模糊決策分析」，中國環境工程學刊，VOL. 6 NO. 4。
[16] 經濟部水利署南區水資源局(2006),「南區水資源營運管理系統」網站HTTP://210.69.129.211/WRASB/LOGIN.ASPX。
西文部份
[1] Reservoir Operation Modeling with Fuzzy Logic D.P.Panigrahi and P.P. Mujumdar, Water Resources Management 14:89-109,2000.
[2] ANFIS：Adaptive-Network-Based Fuzzy Inference System Jyh-Shing Roger Jang, 1993.
[3] Hopfield, J. J.,”Neural Networks and Physical Systems with Emergent Collective Computional Abilities,”Porc.Natl. Academy Sci,Vol.79,pp. 2554-2558,1982.
[4] Hopfield, J. J. and Tank,D.,”Neural Computation of Decision in Optimization Problems,”Biological Cybernetics,Vol.51,pp.141-152,1985.
[5] McCulloah,W.S. and Pitts., W.H.,”A logical Calculus of the Ideas Imminent in Nervous Activity,” Bull.Math. Biophysics,Vol.5,pp. 115-123,1943
[6] Hebb,D.O., The Organization of Behavior, A Neuropsy-chological Theory,Ney York:John Wiley,1949.
[7] Rosenblatt, F.,”The Perception: A Probabilistic Model for information Storage and organization,” Pysch., Rev., Vol. 65,pp.386-408,1958.
[8] Minsky, M., and Papert, S., Perceptrons, MIT press., Cambridge, Mass,1969.
[9] Chen, S. J. & Hwang,C.L. (1992) “Fuzzy Multiple Attribute Decision Making - Methods and Applications” Springer-Verlag, New York.