Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Turquie → Data mining approach for deriving Eleviyan irrigation reservoir operating rules

Ankara University (2009)

Data mining approach for deriving Eleviyan irrigation reservoir operating rules

SATTARİ, Mohammad Taghi

Titre : Data mining approach for deriving Eleviyan irrigation reservoir operating rules

Eleviyan sulama rezervuarında veri madenciliği yaklaşımları ile işletme kurallarının belirlenmesi

Auteur : SATTARİ, Mohammad Taghi

Université de soutenance : Ankara University

Grade : Doctor of Philosophy (PhD) 2009

The rise in the population in Iran, which has arid and semiarid climate and an erratic precipitation distribution in space and time, has increased demand to water much more than ever. Because of rapid growth in the population, water is getting a scarce resource. Especially, supplying of water need is becoming more serious problem in drought periods. Therefore, effective use of water resources called renewable is very important. At recent years, the optimization techniques and models based data in water resources and reservoirs with irrigation purpose have been commonly used. On the other hand, advance in the field of computational intelligence techniques allowed to solve water-related problem. These techniques became an alternative approach to the conventional control techniques. The Eleviyan dam with an active capacity of 60 hm3 was constructed for the irrigation and municipal water needs of the Maraghan region in Northwestern Iran. This thesis was aimed to determine operation policy belonging to the Eleviyan dam by alternative approaches. For this purpose, after allocating the municipal water need, the optimization models that maximize the released water for irrigation were developed for three condition, namely, the inflows measured in the 21 years prior to the construction of the reservoir, the inflows generated by the Monte Carlo simulation method and in the third phase, and the inflows after the construction of the reservoir. The results based on the optimization models implied that the capacity calculated during planning was right. And then, a data mining approach was applied to detect the operating rules in before and after dam construction periods. The data from the optimization for three conditions were used as input of the data mining process, and determined water to release from the reservoir as a decision trees or rules set. The data mining approach used in this work is the decision tree technique, named as C5.0 and developed by Quinlan (1993). The decision trees were validated for time periods before and after dam construction. The results showed that the operating rules based on data mining approach were effectively used water to release from the reservoir. Moreover, when optimization techniques with data mining approaches were taken into consideration in reservoir operation together, the reservoir operation was more successful.

Mots clés : Optimizasyon ve Simülasyon Modelleri, Rezervuar İşletmeciliği, Doğrusal Programlama, Monte Carlo Yöntemi, Veri Madenciliği, Rezervuar İşletme Kuralları, Karar Ağaçları, Optimization and Simulation models, Reservoir operation, Linear programming, Monte Carlo method, Data mining, Reservoir operation rule, Decision tree.


Version intégrale

Page publiée le 13 novembre 2016, mise à jour le 17 juin 2017