.png)
The knowledge of a day ahead load is necessary for a utility in a competitive electricity market for fuel purchase scheduling, planning for energy transactions and to maintain their power reserve close to the minimum as required by Independent System Operator. Previous researches do not consider the effect of wind direction on load forecasting, however this paper investigates the effect of wind direction and weather event on load requirements and accordingly presents a novel Neuro-Fuzzy based approach to Short term load forecast (STLF) i.e. a day ahead average load forecast utilizing parameters identified as historical load, temperature, weather event (for e.g. fog and snow) and wind direction. Four different input structures, three using Neuro-Fuzzy approach and one using only Neural network (NN) are tested. Among the four input structures, structure utilizing Neuro-Fuzzy approach with wind direction as one of the input parameters gives impressive result, with an average error of 1.735 %. The model is trained and tested on load and weather data pertaining to Norwalk/Stamford in Connecticut Valley Electric Exchange.