Ph D Thesis Synopsis

Optimal Reservoir Operation for Irrigation of Multiple Crops

An integrated model for optimal reservoir operation for irrigation of multiple crops is developed. This work comprises of the following.

Optimization of sum of relative yield from different crops over a particular crop season as a function of Actual Evapo-transpiration considering the crop response in each growth stage. The crop season is divided into ten daily sub periods to decide the irrigation allocations for each crop. The optimization is subject to various constraint sets such as reservoir water balance, soil moisture balance and other release constraints with due consideration for the crop root growth over the time. Linear Programming is used as an optimization tool. A suitable code was developed for this purpose in Language for INteractive General Optimization (LINGO). The optimization is carried out for a given state of the system: Inflow into the reservoir over a crop season, Storage in the reservoir at the beginning and end of the crop season and the rainfall in the irrigated area over the crop season.

The reservoir steady state operating policy is determined using Stochastic Dynamic Programming to maximize the expected sum of relative yields over a year with seasonal inflow into the reservoir, storage in the reservoir at the beginning of the season and the seasonal rainfall in the irrigated area as the three state variables. Stochasticity of inflows and rainfall are considered through their seasonal transitional probabilities.

To adopt the derived operating policy in real time, forecasts of seasonal inflows and rainfall are required. For this purpose Auto Regressive Moving Average (ARMA) models were used. A Disaggregation Model based on conditional expectations was used to obtain the intra seasonal values for the given seasonal state of inflow and rainfall.

The developed model is demonstrated by applying it to an existing single purpose reservoir, Malaprabha reservoir in Karnataka state, India.