APPLICATION OF ANN IN INFLOW PREDICTION AND RESERVOIR OPERATION

Abstract

Artificial Neural Networks (ANNs) are new computing architectures in the area of artificial intelligence. The recent resurgence of interest in Artificial Neural Networks has its roots in the recognition that the brain performs computations in a different manner than do conventional digital computers. A human information processing system is composed of neurons switching at speeds about a million times slower than the computer gates. Yet humans are more efficient than computers at computationally complex tasks. Artificial Neural Network is the result of long standing effort for mimicking the computational capability of brain system.This derives its inspiration from two distinct yet related fields - Associationist psychology and Neuroscience. Quite contrast to the conventional digital computers, these ANN based systems can acquire, store and utilize experiential knowledge. For this reason, the scope of its applicability is being explored in many disciplines including Hydrology. The present study aims at application of this Neural Network baged computational paradigm in reservoir operation and inflow prediction. Upper Indravati hydro-electric project, which is a large multi purpose Water Resources Project in the Nowrangpur and Kalahandi districts of Orissa has been selected as the problem area for this study. On completion, the project shall comprise of four dams and a combined reservoir of 110 km2 area. The project has primarily two objectives: to provide irrigation to 1,28000 ha. of agricultural land and to provide 600 MW of electric power through four numbers of Francis turbines of 150 MW each. The present study capitalises on 32 years of monthly inflow record of the project. Specifically the objecti-:e is to maximise total project benefits from hydro power generation, simultaneously aiming to minimise the irrigation deficit and water losses through spilling. This objective was aimed to be attained through the following steps.