Abstract
Over time, the electricity price and energy consumption are increasingly growing their weight as prime foundations of the electrical sector, with their analysis and forecasts being targeted as key elements for the stable maintenance of electricity markets. The advent of smart grids is escalating the importance of forecasting because of the expected ubiquitous monitoring and growing complexity of a data-rich ever-changing milieu. So, the increasing data volatility will require forecasting tools able to rapidly readjust to a dynamic environment. The Generalized Regression Neural Network (GRNN) approach is a solution that has recently re-emerged, emphasizing good performance, fast run-times and ease of parameterization. The merging of this model with more conventional methods allows us to obtain more sturdy solutions with shortened training times, when compared to conventional Artificial Neural Networks (ANN). Overall, the performance of the GRNN, although slightly inferior to that of the ANN, is suitable, but linked to much lower training times. Ultimately, the GRNN would be a proper solution to blend with the latest smart grids features, which may require much reduced forecasting training times.

Abstract
The price evolution in electricity market with large share of renewables often exhibits a deep volatility, triggered by external factors such as wind and water availability, load level and also by business strategies of market agents. Consequently, in many real applications, the performance of electricity price is not appropriate. The goal of this article is to analyze the available market data and characterize circumstances that affect the evolution of prices, in order to allow the identification of states that promote price instability and to confirm that class segmentation allows increasing forecast performance. A regression technique (based on Artificial Neural Networks) was applied first to the whole set and then to each class individually. Performances results showed a clear advantage (above 20%) of the second approach when compared to the first one.

Abstract
The increasing integration of distributed renewable energy sources, such as photovoltaic (PV) systems, requires adequate regulatory schemes in order to reach economic sustainability. Incentives such as Feed-in Tariffs and Net Metering are seen as key policies to achieve this objective. While the Feed-in Tariff scheme has been widely applied in the past, it has now become less justified mainly due to the sharp decline of the PV system costs. Consequently, the Net Metering scheme is being adopted in several countries, such as Brazil, where it has is in force since 2012. In this context, this paper aims to estimate the minimum monthly residential demand for prosumers located in the different distribution concession areas in the interconnected Brazilian system that ensures the economic viability of the installation of PV systems. In addition, the potential penetration of PV based distributed generation (DG) in residential buildings is also estimated. This study was conducted for the entire Brazilian interconnected system and it demonstrates that the integration of distributed PV systems is technical-economic feasible in several regions of the country reinforcing the role of the distributed solar energy in the diversification of Brazilian electricity matrix.

Abstract
This paper presents a new concept for an approach to deal with measurements contaminated with gross errors, prior to power system state estimation. Instead of a simple filtering operation, the new procedure develops a screen-and-repair process, going through the phases of detection, identification and correction of multiple gross errors. The method is based on the definition of the coverage of the measurement set by a tiling scheme of 3-overlapping autoencoders, trained with denoising techniques and correntropy, that produce an ensemble-like set of three proposals for each measurement. These proposals are then subject to a process of fusion to produce a vector of proposed/corrected measurements, and two fusion methods are compared, with advantage to the Parzen Windows method. The original measurement vector can then be recognized as clean or diagnosed with possible gross errors, together with corrections that remove these errors. The repaired vectors can then serve as input to classical state estimation procedures, as only a small noise remains. A test case illustrates the effectiveness of the technique, which could deal with four simultaneous gross errors and achieve a result close to full recognition and correction of the errors.

Abstract
This paper proposes a methodology using Hybrid Control System (HS) to manage the integration of Variable Renewable Electricity Sources (VRES). The HS define a combination of discrete and continuous signals, in this case, discrete by Pump-Hydro-Storage (PHS) and continuous performance is the Wind Power (WP). The coupling of Wind Power and PHS to produce a dispatchable power output could be a significant benefit to those in an energy trading system. Improving VRES prediction reduces system dispatch errors, however does not give total economic opportunities to the generator. Increased dispatchable backup power generation can improve the system's ability to handle deviations of WP, as verified when the proposed approach is applied to Brazilian and Portuguese power system. © 2018 IEEE.

Abstract
This paper presents the concept of a tapered deep neural network, subject to unsupervised training layer by layer, under a criterion of maximum entropy, to perform the estimation of breaker status in the absence of a specific sensor signal. The almost perfect prediction power of the model confirms the conjecture that the knowledge of the topology of a network is hidden in the electric measurement values in the network. A test case is presented with computing speed accelerated by using a GPU (graphics processing unit). The comparison with a previous model illustrates the superiority of the novel approach. © 2018 IEEE.