INTEGRATED PLANNING AND OPERATION OF A RESERVOIR

Abstract

Isolated studies on various aspects of planning/operation of a reservoir have been carried out and presented in various studies, using systems analysis techniques. A few studies on combined planning and operation of a reservoir are available in literature. Systematic studies towards the various water resources planning and management aspects of integrated planning and operation of a reservoir, i.e., determination of (i) conservation/flood control storages (ii) annual firm/targeted/secondary requirements for irrigation and hydropower (iii) seasonal/year to year or over-year carry-over reservoir capacity/storages and (iv) annual reservoir operation rules, all using systems analysis and (v) selection of annual flow of a given dependability for project planning depending on water use have not been presented. The specific decision problem under consideration is to solve the integrated planning and operation of a reservoir (single purpose and/or multipurpose) using systems analysis techniques. For this four different single reservoir projects have been selected on different rivers and are located at different locations in India, these are (i) Badanala irrigation project under normal hydrological conditions (ii) Kalluvodduhalla irrigation project under drought conditions (iii) Bodhghat hydroelectric project and (iv) Bargi multipurpose project. The approach now is to formulate suitable models for various aspects of planning and operation of a reservoir. Often these problems are very complex with different objective, scope, and scale. In such^case, theve is usually no unique model for the problem. Then a set of linked models may be used. These models may be nested in such a fashion that outputs of one models are inputs to another or two models are run in tandem. Although, the answer to how the model links should be arranged is iv ABSTRACT problem specific. Such use of nested models may be often quite useful. The use of optimization-simulation techniques as nested models is envisaged here to solve these integrated planning and operation problems of a reservoir. It is proposed to use linear programming and dynamic programming as optimization techniques. The preparation of input data for computation of linear programming by computer ^fe data-intensive and a Herculean task. To overcome this difficulty a suitable computer software for generating the input data matrix for linear programming optimization problems applicable to reservoir planning and operation, called MATGEN PACKAGE has been developed. Two categories of optimization models were used for reservoir planning as follows: Category-I Models: These models were used to account for short term variations in the river inflows. They used river flow data of one year length only. A concept of using annual flows of a probability of given occurrence was introduced in these models to account for the long term variations in the inflows. Category-II Models: These models were used to account for long term variations in the river inflows. They used river flow data of the entire length of the recorded historical flows. Reservoir operation was carried out using multi-rule curves based of the actual monsoon flows and the state of the reservoir at the end of monsoon period. In one of the rule curves a new concept of available over-year carry-over reservoir storage at the end of the year (non-monsoon period) is introduced. These rule curves were (i) Variable Upper Rule Curve to provide over-year carry-over reservoir storage ABSTRACT at the end of a year and the targeted demand in the non-monsoon period (ii) Middle Rule Curve to provide targeted demand in the non-monsoon period and (iii) Lower Rule Curves to provide firm demands in non-monsoon period. Category-I models require less computer memory and reasonably less computation time. Also, these models use river flow data of same length, i.e, of one year length only depending upon type of project, which makes this approach more uniform in terms of the length of data used. Category-II models require large computer memory and large computation time. These models use river flow data of different lengths, which makes this approach non uniform in terms of the length of data used. The multi-rule curves operation of a reservoir project with irrigation with the help of Variable Upper Rule Curve reduces water use deficits in early monsoon and the later non-monsoon periods. However, for a single purpose hydropower reservoir operation, use of only Lower Rule Curve is recommended. On the basis of the experience of applying the above models, a suitable and uniform methodology for integrated planning and operation of a reservoir using Category-I models with the help of nested models for planning and multi-rule curves for operation has been finally suggested. Also use of annual flows of probabilities of given occurrences is recommended.