A novel scheme of load forecasting pertaining to long term planning of a distribution system

In this paper, a novel approach is proposed to address the load-forecasting and expansion-planning process focused on long-range aspects. Depending upon land-use, five different user classes are considered. The existence of these classes and their associated boundaries are modeled with trapezoidal distribution. To find the relative share of one particular user class with remaining classes, a class composition vector is proposed. Further, dividing all the area segments into small square shaped cells of 40-acre size, the small area growth trend is estimated with 'S-shaped' curve for 20-year period. Here a piece-wise linear model is used to model the growth ramp segment of "S-shaped" growth characteristic. Finally two indices are proposed, one to estimate the new load demand, and other index to estimate the growth. The results obtained match the typical growth trend of different user classes for a practical distribution system.