The goal of this task is to accurately predict the energy demand in heating or cooling of the network in order to save energy, reduce costs and greenhouse gases. This prediction will be utilised by the controller to schedule the production. For this purpose, Machine Learning models are under development to predict energy demand in the same prediction horizons as the weather forecasting models. Those models are dependent on the weather predictions as they are correlated to the energy demand of the consumers. The training methodology relied on the historical observations of the correlation of weather with the past heating or cooling production of the plants. This constituted the input of the models and the output is the future energy load. Different machine learning algorithms are being tested currently, including Dense Neural Networks, Deep DNNs, Recurrent NNs, Random Forests and different Ensemble Learning Regressors. The deployed models will receive as inputs: current energy consumption, historical data and the outputs from the weather forecasting models (predicted weather parameters).