DEVELOPMENT OF OPTIMAL OPERATING POLICY FOR PAGLADIA MULTIPURPOSE RESERVOIR
| dc.contributor.author | Ahmed, Juran Ali | |
| dc.date.accessioned | 2015-09-15T12:46:13Z | |
| dc.date.accessioned | 2023-10-19T12:32:16Z | |
| dc.date.available | 2015-09-15T12:46:13Z | |
| dc.date.available | 2023-10-19T12:32:16Z | |
| dc.date.issued | 2004 | |
| dc.description | Supervisor: Arup Kr. Sarma | en_US |
| dc.description.abstract | Pagladia multipurpose reservoir, located on the river Pagladia, a major north bank tributary of Brahmaputra, proposes to serve three purposes, namely, flood control, irrigation and power generation. In order to achieve these, a proper operating policy of the reservoir is imperative. Recent researches have revealed the potential of heuristic methods in deriving reservoir-operating policy. In this study, the potential of Genetic Algorithm (GA) and Artificial Neural Network (ANN) in deriving an optimal operating policy has been explored through their application in the Pagladia multipurpose reservoir. Efficiency of the policies derived by these recent techniques has been assessed through their critical comparison with policies derived by some long-established techniques. To have the advantage of using a long streamflow series in the development of a reservoir operating policy, an ANN based synthetic streamflow generation model has been developed and compared with the Thomas-Fiering and Autoregressive Moving Average (ARMA) models. Synthetic streamflow series generated by the ANN based model has been used in the development of operating policies, as its statistics have been found to be in better agreement with those of the observed historic series. For solving the reservoir optimization problem for Pagladia multipurpose reservoir, deterministic Dynamic Programming (DP) has first been solved. Both multiple linear regression and ANN have been used to infer general monthly operating policy from the DP results, and these models are being termed as DPR and DPN models respectively in this study. Stochastic Dynamic Programming (SDP) model, which uses an explicit stochastic optimization technique, has been developed next for deriving monthly optimal operating policy for the Pagladia multipurpose reservoir. Finally, GA, which is of recent origin and has the capability of solving complex optimization problem, has been used to derive optimal monthly operating policy for the reservoir. Performance of all the operating policies developed by different models has been analyzed on the basis of the reservoir simulation results for 228 months of historic streamflow series (1977-1996). For making a fair comparison among all the models, a total of eight performance criteria covering different aspects of reservoir operation, have been used. The study has shown that GA, which is a robust optimization technique, is quite capable of developing multipurpose reservoir operating policy and has been found to give the most efficient operating policy for the Pagladia multipurpose reservoir. Policies derived by SDP and GA have been found to be competitive in respect of some of the performance criteria. GA out performs the DPR and the DPN models developed in this study. Although performances of different models vary with different performance criteria, considering overall performance and giving priority to irrigation, the application of operating policy derived by GA model has been suggested to be the most appropriate for the Pagladia multipurpose reservoir. | en_US |
| dc.identifier.other | ROLL NO.01610402 | |
| dc.identifier.uri | https://gyan.iitg.ac.in/handle/123456789/52 | |
| dc.language.iso | en | en_US |
| dc.relation.ispartofseries | TH-0151; | |
| dc.subject | CIVIL ENGINEERING | en_US |
| dc.title | DEVELOPMENT OF OPTIMAL OPERATING POLICY FOR PAGLADIA MULTIPURPOSE RESERVOIR | en_US |
| dc.type | Thesis | en_US |