Applications of Neural Networks and Decision Trees to Energy Management System Functions

Artificial neural network (ANN) and decision tree (DT) applications are proposed to contribute to control room Energy Management System (EMS) functions respectively aimed to assess critical conditions leading to inter-area oscillations, to evaluate the loading margin, and to support state estimation. To facilitate the design, the setting of the parameters, the training of ANNs and the growing of DTs, a tool has been developed within Matlab environment, that allows the user to design and train in an effective way ANNs and DTs with no need to write any procedure or line of code. The tool has been applied in the analysis of the above problems providing a valid support in training and testing both ANNs and DTs. oscillations in transferring large amounts of power; to determine the loading margin; to support the state estimation process. To facilitate the design and the training of the ANNs and the growing of the DTs and to optimize the choice of their parameters, a tool has been developed in the Matlab environment, that allows the user to operate efficiently the required analyses and processes with no need to write computer procedures or lines of code, thus saving time and resources.