A Neuro-Fuzzy Algorithm for Middle-Term Load Forecasting

The forecasting method presented in this paper is intended for industrial customers who mainly have two types of consumers in the distribution grid: consumers for climate control and consumers from production activities. Forecasting model is divided into two subsystems: 1. Consumption forecasting using a neural network with back-propagation and using data from previous months and 2. Forecast adjustment using temperature and production load differences for air conditioning and production consumption between the previous months and months for which the forecasting is done.