Water resources management for a multipurpose project: an operational perspective

Abstract

The study aims at looking into the operations of a multipurpose project with a view to efficient water management. Lkai-Kakrapal project, the largest in operation in Gujarat is taken for the purpose. Given the Ukai system, the research addresses itself to the problem of determining an operating policy that will maximize the benefits due to irrigation and power over a water year, and meet the requirements for both in a desired way. The problem is complex because: a)The objectives of maximizing irrigation supply and maximizing energy generated are not always complementary; b)Energy generated is not only dependent upon the depletion of storage, but also the reservoir Level; c)Evaporation loss as a function of storage level complicates the tradeoff between maintaining storage and releasing water for power and irrigation; and d) The supply of water is uncertain. A deterministic mathematical model, discrete in time, is developed and the resulting nonlinear programming problem is solved by an iterative procedure and the method of separable programming. Uncertainty is then introduced in the deterministic framework by random generation of different levels of supply of water from theoretical distributions fitted to the inflow data. The deterministic model is solved several times to gain insight into the process of optimal allocation. An attempt to contain uncertainty in the form of simple release rules is made from the above results. State-independent rules are derived for each period by considering the periodic release distributions, applying the criteria of minimization of the expected value of absolute deviations from optimal releases. Next release distributions together with the corresponding initial storage distributions and the inflow distributions are taken and linear; log- linear, log-log relationships are tried out for each period. The relationships explaining maximum variance for each period are chosen to represent the state-dependant rules. Finally, some performance measures and performance oases are identified and the policies are evaluated by simulating the reservoir operations. The study provides an explicit analytical framework for operations management for the Ukai project, and the methodology is general enough to be applied for other multipurpose projects. The solutions provided are simple to operate and found to yield better results than actual practice.