An Approach to Energy Management in Chittagong Hill Tracts: Prediction Based Analysis

The most prevalent justification for Demand Side Management in the power sector is that managing energy demand via efficiency and other demand side measures is typically more cost effective and socially good than increasing power supply or transmission capacity. In this paper, an Artificial Neural Network (ANN) is used to predict the energy demand of CHT for demand side management with sustainability. To forecast monthly solar irradiation, a multilayered feed forward ANN model with four layers and five independent input variables (Power source, minimum and maximum temperature, solar irradiance, connected load) was used. For training, data from selective areas was observed, with the remaining two areas being evaluated for testing and validation. This paper supports a model with Mean Square Error value of 0.0088 and Regression value of 0.905. The projected data appear to be in excellent agreement with the observation. This means that this model may be used to forecast CHT’s energy demand and offer enough information on the viability of solar-powered projects. Artificial Neural Network (ANN) is focused in this paper as a method of load forecasting using MATLAB toolbox and all these analyses are done with MATLAB software.