S Mohan

The persistent problem in reservoir operation is that the derived optimal releases fail to incorporate the decision maker or reservoir operators’ knowledge into reservoir operation models. The reservoir operators’ knowledge is specific to that particular reservoir and incorporating such an experienced knowledge will help to derive field reality based operation rules. The available historical reservoir operation databases are the representative samples of reservoir operators’ knowledge or experience. Thus, an attempt has been made that deals with the development of a methodological framework to recover or explore the historical reservoir operation database to derive the reservoir operators’ knowledge as operational rules. The developed methodological framework utilizes the strength and capability of recently developed predictive datamining algorithms to recover the knowledge from large historical database. Predictive data-mining algorithms such as a) classifier: Artificial Neural Network (ANN), and b) regression: Support Vector Regression (SVR) have been used for single reservoir operation data-mining (SROD) modelling framework to explore the temporal dependence between different variables of reservoir operation. The rules of operation or knowledge learned from the training database have been used as guiding rules for predicting the future reservoir operators’ decision on operating the reservoir for the given condition on the inflow, initial storage, and demand requirements. The developed SROD model was found to be efficient in exploring the hidden relationships that exist in a single reservoir system.

A combined optimization-simulation approach was used to develop and evaluate the alternate priority-based policies for operation of surface and groundwater systems and is demonstrated with a case study. An optimization model was used to find optimal cropping pattern with and without the conjunctive use of surface and groundwater, as well as with and without socio-economic constraints. The optimization model, based on linear programming, maximizes the net benefit from irrigation activities subject to various physical, economical, and social constraints. A simulation model was used to evaluate the conjunctive operation of the system using the optimal cropping pattern derived from the optimization model. The developed policies have been verified with long-generated stream flow sequences. Three alternate priority-based policies differing in level of groundwater pumping and area of cultivation of rice crop have been evaluated: (1) irrigation with surface water only (Policy-1); (2) irrigation with conjunctive use of surface and groundwater, without socio-economic constraints (Policy-2); and (3) irrigation with conjunctive use operation and with socio-economic constraints (Policy-3). It was found that the use of available groundwater within three meters below ground level (Policy-2) to be optimal, and these results were used in simulation for further evaluation of policies. It was also found that the policy-3 of conjunctive use operation with a priority of 75 percent of maximum possible rice area (using groundwater available within four meters below ground level) resulted in a better scenario. Thus the conjunctive use Policy-2 and Policy-3 with 75 percent of maximum possible rice crop area can be used as better policies for the system studied. © 2003 Taylor & Francis Group, LLC.

Due to the complexity of irrigation management problems, reliance on experience and experts is necessary for effective decision-making in this domain. Expert systems (ES) are efficient means for providing decision support to tasks that primarily require experience-based knowledge. This paper reviews the adoptability and suitability of ES applications in the domain of irrigation management. Core concepts of ES are briefly discussed. A detailed review of the existing applications of ES is presented under three classes of ES applications: (a) expert systems proper, (b) intelligent front-ends, and (c) hybrid systems. This review of literature shows that the ES approach is applied more recently to broader domain areas in contrast to the earlier systems that were focused on narrower domain problems. Additional research on ES application to domains such as real-time irrigation scheduling, reservoir operation involving stochastic nature of inflows and evapotranspiration demand, and integrated operation of irrigation system components is needed to evolve guidelines for optimal water use. The problem of handling multiple experts to evolve decisions that are less biased than an individual expert needs to be addressed. A methodology that takes into account the uncertainty of the ES decisions is also warranted. Further, there is a need for practical evaluation of the quality of recommendations made by the ES which would result in the successful implementation of the ES. © 1997 Elsevier Science B.V.

A heuristic simulation based optimization model has been developed to optimize the operation of a multiple reservoir using dynamic programming and soft computing techniques. Soft computing based tools including Adaptive Neuro Fuzzy Inference systems (ANFIS) and Genetic Algorithms (GA) are used here to tackle the complexity of deriving the operational policies. This heuristic approach involves three stages in the model development. In the first stage GA is used to develop initial trajectories for DP model and then DP model results are used to adopt training data set for ANFIS model. Finally, a general operating policy is developed for multi-reservoir system operations. The demonstration is carried out through application of Parambikulam Aliyar Project systems in India. The performance of the proposed GA-DP-ANFIS is compared with simulation based multi regression model. Rule curves are developed for different scenarios like current operating policy, rezoning pattern, improved irrigation management and increase of available water potentials. Copyright © 2007 IICAI.