24-Reservoir operating rule for Narmadasagar dynamic programming approach.

Abstract

This work relates to application of dynamic programming model to Narmadasagar reservoir across river Narmada in the state of Madhya Pradesh . The model is formulated to determine the operating rule for maximizing power after meeting irrigation. The irrigation demand pertains to the development scenario 1991-92 while the full development of irrigation utilisation takes place only during 2023-24 (final scenario).The inflow into Narmadasagar for the historic record of 32 years, for the 1991-92 scenario of development is obtained from a simulation model (I.I.M.,Bangalore, June,1985). Eventhough, the reservoir is to be built to a height of 262.13 m (12212 Mm ) with a live storage of 10855 Mm , the model brings out the performance for FRLs namely256.0 m , 257.6 m, 259.0 m , 260.67 m and 262.13 m. The power generation for installed capacities namely 875 MW, 1000 MW, 1125 MW and 1250 MW are also discussed. In Dynamic programming problems, the accuracy of the results depends on the fineness of the state variables. As the number of states increase, the computer time increases and vice versa.The effect of variation of number of states on the performance and on the CPU time required are analysed. Number of states considered are 9, 15, 20, 25 and 40. The CPU time needed for a run with 40 states is about 20 times that of a run with 9 states. Benefit from irrigation is taken as 8.8 paise per cubic metre of water and the power benefit at bus bar is taken as 30 paise/KW hr. Efficiency of power generation is taken as 85 percent.