A research study (12 credits) submitted in partial fulfillment of the requirement for the degree of Master of Engineering, School of Environment, Resources and Development

Research Studies Project Report (M. Eng..) - Asian Institute of Technology, 1987

Electricity consumption of a country is a typical indicator of its economic growth . Hydro-energy has been one of the promising solutions in electricity generation both in the developed and developing world. As a result , cascaded reservoirs have become a common feature in most of the power systems . The present study analyses a method for deter mining the monthly operating policy of a power system of n reservoirs in cascade, with a special application to the Mahaveli River scheme in Sri Lanka. The method consists of rewriting the stochastic non -linear optimization problem of n state variables as n -1 problems of two state variables, which are solved by Dynamic Programming, as the De composition method. The optimal policy obtained b y this method for the reservoirs is a function of the water content of that reservoir and of the tot al amount of potential energy stored in the down stream reservoirs. In this study, a strict correlation among natural inflows at different dam-sites, was assumed . The times of water travel between upstream and down stream are not taken in to account, assuming that is to be shorter than a month . Water losses in reservoirs due to seepage and evaporation are neglected . It is further assumed that there are no competing uses for the water in storage . In the initial stage of the study, a computer programme was developed, based o n the simplified version of the Decomposition algorithm. The sample data provided in the hypothetical system of TURGEON' s work( 1 9801, were used to check the programme. The same met h o d was applied for the four selected reservoirs in the Mahaveli River Scheme in Sri Lan ka, in order to obtain the optimal releases of individual reservoirs.