Abstract
This paper presents a new approach to incipient fault diagnosis in power transformers, based on the results of dissolved gas analysis. A set of autoassociative neural networks or autoencoders are trained, so that each becomes tuned with a particular fault mode. Then, a parallel model is built where the autoencoders compete with one another when a new input vector is entered and the closest recognition is taken as the diagnosis sought. A remarkable accuracy is achieved with this architecture, in a large data set used for result validation. © 2011 IEEE.

Abstract
This paper analyzes the application of clustering techniques for wind power scenario reduction. The results have shown the unimodal structure of the scenario generated under a Monte Carlo process. The unimodal structure has been confirmed by the modes found by the information theoretic learning mean shift algorithm. The paper also presents a new technique able to represent the wind power forecasting uncertainty by a set of representative scenarios capable of characterizing the probability density function of the wind power forecast. From an initial large set of sampled scenarios, a reduced discrete set of representative scenarios associated with a probability of occurrence can be created finding the areas of high probability density. This will allow the reduction of the computational burden in stochastic models that require scenario representation. © 2011 IEEE.

Abstract
Maintenance decisions in electricity markets are one of the most important strategic conflicts among power players. A generation company, for instance, is responsible for its assets but it is the Transmission System Operator's task to schedule the most suitable period for maintenance, preserving system reliability. This relationship is sometimes conflicting and can be seen as an obstacle to the maximization of profits of generation companies, as well as to the transmission system operator, since it is mandatory to keep the adequate risk-based level on overall system. In this paper, the attention is focused on this relationship. Moreover, it highlights the proposal of building a dedicated environment in order to improve and clarify the discussions about maintenance outage schedule for large equipments, using the agent-based technology. In order to illustrate the proposal, results with the IEEE-RTS are presented. © 2009 IEEE.

Abstract
A novel optimization methodology is proposed for the design of transmission line grounding systems, taking into account technical as well as economical considerations. The problem of designing the grounding systems of transmission fines is stated as a linear-integer programming problem in terms of the construction characteristics and the particular requirements of the tower grounding schemes at the supports of each of the different line sections, in order to minimize the variable investment costs, subject to the maximum allowed line outage rate due to the lightning activity. The mathematical statement of the problem allows solutions in which the transmission tower footing resistance changes along the line, depending on the cost and on the particular characteristics of each tower grounding, assuring however, that the average behavior enforces the desired outage rate due to lightning activity, selecting the complementary electrode scheme required at each tower. The methodology is tested on a real case consisting of a 230 kV transmission line, 85.4 Km long, with 180 towers. The linear programming branch and bound mathematical technique was applied for the solution of the test case. Two different simulation approaches for the calculation of the behavior of the fine subject to lightning phenomena were evaluated without loss of generality: the approach proposed in [1], selected as an initial test due to its simplicity, and the improved version presented in [2]. Results are presented and compared to the design obtained through conventional tower design approaches with important reductions in the investment costs, encouraging the use and further development of the methodology.

Abstract
Decision problems cannot be fully represented without underlying assumptions about the Decision Maker motivations and behavior. This paper describes one technique to build a rule model representing the interaction of preferences of a Decision Maker, by training a Fuzzy Inference System based on examples. © 2005 ISAP.

Abstract
This paper presents a comparison in performance of 3 variants of Genetic Algorithms (GA) vs. 2 variants of Evolutionary Particle Swarm Optimization (EPSO), made in the extremely complex context of a multi-energy market simulation where the behavior of energy retailers is observed. The simulations are on JADE, a FIPA compliant platform based on intelligent autonomous agents running in a cluster of PCs. Each agent formulates its strategy by an inner complex simulation process using a meta-heuristic that tries to define optimum decisions. The results suggest that an EPSO approach is more efficient than GA. © 2005 ISAP.