Abstract
In this paper we describe a new approach to identify the combination of tap transformer positions, capacitor bank steps together with the minimum amount of toad to be shed that assures one to obtain a specified security degree of a power system. The basic approach is designed to identify the most adequate actions to be taken for a given contingency. This identification procedure uses Genetic Algorithms given their adequacy to model discrete actions. However, Genetic Algorithms are known for their usually large computation time. In order to address this issue and having in mind the objective of developing a real time tool, we incorporated a classification procedure based on Neural Networks. The paper includes results obtained using the developed approach both to evaluate the quality of the solutions for a number of contingencies and the quality of the overall performance when using the Neural Network tool. Results obtained for a reduced version of the Brazilian Mate Grosso transmission system are presented and discussed.

Abstract
The application of artificial neural networks or other techniques in load forecasting usually outputs quality results in normal conditions. However, in real-world practice, a remarkable number of abnormalities may arise. Among them, the most common are the historical data bugs (due to SCADA or recording failure), anomalous behavior (like holidays or atypical days), sudden scale or shape changes following switching operations, and consumption habits modifications in the face of energy price amendments. Each of these items is a potential factor of forecasting performance degradation. This paper describes the procedures implemented to avoid the performance degradation under such conditions. The proposed techniques are illustrated with real data examples of current, active, and reactive power forecasting at the primary substation level.

Abstract
Quality prediction of load evolution at different levels of distribution network is a basic requirement for adequate operation planning of modern power systems. This paper describes the models, based on artificial neural networks, developed for active and reactive power forecasting at primary substations' transformers. The main goal consists on defining a regression process characterized by good quality estimates of those future values, based on historical data. Anticipation interval shall include from the next hour to one week in advance. The implemented forecasting tool is able to deal with noisy data, holidays and special occasions and adapts forecasts in case of power network reconfiguration whenever planned. Used techniques and implementation foundations of selected forecasting models are reported. Finally, the potential of the adopted approach is sustained by illustrative examples. © 2003 IEEE.

Abstract
Voltage stability is an important concern of power system managers not only in the net planning phase but also in operation. This issue has become especially critical in recent years due to the deregulation phenomenon because of new exploration policies complying a system operation closer to its security limits. In particular, voltage collapse distance may approach emergency values or, in the worst case, make the system collapse. As voltage profile is extremely dependent on reactive power compensation, most common approaches integrate both objectives in the operation setting phase, trying to optimize reactive power production taking voltage profile into consideration. In this paper, authors propose an evolutionary approach application to the same problem but in the planning phase. It is shown that the cooperative procedure of planning and preventive control provides better solutions that if one deals with these issues one at a time. © 2001 IEEE.

Abstract
This paper deals with a problem of identification of the best subset of variables that should be used for dynamic security assessment of a power system, when this task is pro-vided by artificial neural networks (ANN)- The approach de-scribed here exploits ANN output sensitivities relatively to the inputs and correlation degrees, to identify the most relevant system variables to be used for an effective security assessment task. The ANNs are initially trained with all low-correlated candidate features, which enables the sensitivity analyses for the initial set of system features. Derivatives of the ANN output relatively to each input are obtained by exploiting the chain rule, similar to the one used for weights adaptation on Back-propagation Algorithm. A description of the application of this approach in a real system is present in the paper. Results obtained in the dynamic security assessment problem of the network of the island of Crete were quite successful. © 2001 IEEE.

Abstract
The paper describes the on-line dynamic security assessment functions developed within the European Union, DGXII programme, CARE. These functions are based exclusively on the application of machine learning techniques. A description of the problem and the data set generation procedure for the Crete island power system are included. Comparative results regarding performances of Decision Trees, Kernel Regression Trees and Neural Networks are presented and discussed.The paper describes the on-line dynamic security assessment functions developed within the European Union, DGXII programme, CARE. These functions are based exclusively on the application of machine learning techniques. A description of the problem and the data set generation procedure for the Crete island power system are included. Comparative results regarding performances of Decision Trees, Kernel Regression Trees and Neural Networks are presented and discussed.