Performance enhancement of the artificial neural network–based reinforcement learning for wind turbine yaw control

The yaw angle control of a wind turbine allows maximization of the power absorbed from the wind and, thus, the increment of the system efficiency. Conventionally, classical control algorithms have been used for the yaw angle control of wind turbines. Nevertheless, in recent years, advanced control strategies have been designed and implemented for this purpose. These advanced control strategies are considered to offer improved features in comparison to classical algorithms. In this paper, an advanced yaw control strategy based on reinforcement learning (RL) is designed and verified in simulation environment. The proposed RL algorithm considers multivariable states and actions, as well as the mechanical loads due to the yaw rotation of the wind turbine nacelle and rotor. Furthermore, a particle swarm optimization (PSO) and Pareto optimal front (PoF)‐based algorithm have been developed in order to find the optimal actions that satisfy the compromise between the power gain and the mechanical loads due to the yaw rotation. Maximizing the power generation and minimizing the mechanical loads in the yaw bearings in an automatic way are the objectives of the proposed RL algorithm. The data of the matrices Q (s,a) of the RL algorithm are stored as continuous functions in an artificial neural network (ANN) avoiding any quantification problem. The NREL 5‐MW reference wind turbine has been considered for the analysis, and real wind data from Salt Lake, Utah, have been used for the validation of the designed yaw control strategy via simulations with the aeroelastic code FAST.     

Color inhibition of phenolic antioxidants in Ziegler‐Natta polyethylene. I. In‐situ polymer studies

Although the level of transition‐metal catalyst residues in polyethylene (PE) has been drastically reduced over the years, they can still give rise to discoloration, particularly when associated with other additives such as antioxidants. This first of this series of papers screens a variety of candidate color suppressants featuring a range of functional groups, including alcohols, amine/sulfur compounds, and acid‐containing species. These candidate color suppressants were melt‐blended into a Ziegler‐Natta linear low‐density PE in combination with 2,2′‐isobutylidenebis(4,6‐dimethylphenol) (a highly discoloring hindered bisphenol antioxidant) and zinc stearate antacid. Yellowness index measurements made after multiple extruder passes indicated that dipentaerythritol (DPE) and triisopropylamine (TIPA) gave good color inhibition and, in some cases, outperformed established phosphites. The DPE and TIPA were found (via melt flow rate measurement) not to affect melt stability, and hydroperoxide determination revealed that DPE had no peroxide decomposition activity. The latter results indicate that the color‐suppression mechanism of DPE and TIPA is different from that associated with phosphites. J. VINYL ADDIT. TECHNOL., 2009. © 2009 Society of Plastics Engineers     

Updating Polygonizations*

In this paper we consider polygonizations that are robust when faced with changes in the vertices that are present or in their position. We analyze the dynamic maintenance of different types of polygonizations (monotone, star-shaped…) and we introduce monotone half-convex polygonizations that are specially interesting because they provide minimum cost per insertion or deletion. If we had to delete not only one point but several external layers of the set, then the onion polygonizations would be suited, because they can be updated in constant time. We also consider the case of points that can be moved to contiguous positions and we show how to polygonize the set for updating in linear time. We deal too with security problems for a polygon: What is the maximum distance the vertices of a polygon could be moved away of their position in such a way that the topology on the boundary of the polygon (or its convexity) remains the same?.     

The impact of malnutrition on brain development, intelligence and school work performance

The findings from several authors confirm that undernutrition at an early age affects brain growth and intellectual quotient. Most part of students with the lowest scholastic achievement scores present suboptimal head circumference (anthropometric indicator of past nutrition and brain development) and brain size. On the other hand, intellectual quotient measured through intelligence tests (Weschler-R, or the Raven Progressives Matrices Test) has been described positively and significantly correlated with brain size measured by magnetic resonance imaging (MRI); in this respect, intellectual ability has been recognized as one of the best predictors of scholastic achievement. Considering that education is the change lever for the improvement of the quality of life and that the absolute numbers of undernourished children have been increasing in the world, is of major relevance to analyse the long-term effects of undernutrition at an early age. The investigations related to the interrelationships between nutritional status, brain development, intelligence and scholastic achievement are of greatest importance, since nutritional problems affect the lowest socioeconomic stratum with negative consequences manifested in school-age, in higher levels of school dropout, learning problems and a low percentage of students enrolling into higher education. This limits the development of people by which a clear economic benefit to increase adult productivity for government policies might be successful preventing childhood malnutrition.     

Effects of pulp and paper mill effluents on the microplankton and microbial self-purification capabilities of the Biobío River, Chile

Most studies focus on the ecotoxicity of pulp and paper mill effluents, rather than on how they affect the physicochemical and biological structure and the intrinsic ecological capabilities of the receiving watercourses. We investigated the impact of such effluents on the water quality, microplankton system and microbial self-purification capacity (degradation of polymeric organic compounds via extracellular enzymes) of the Biobío River in Chile. The physicochemical impact on the water quality was indicated by raised conductivity, by the pollution of the water body with nitrate, nitrite and soluble reactive phosphorus, by the appearance of tannin and lignin, and by the steady accumulation of inorganic and organic suspended matter (SPM) along the river. From the biological structure of the microplankton system, very low and declining concentrations of chlorophyll a and heterotrophic flagellate densities were determined. The pulp and paper mill effluents introduced high bacterial abundances and biomass concentrations into the river water. This reflects the effective use made of the abundantly available inorganic and organic nutrients within this industrial and municipal process water by bacteria adapted to these extreme environments, additionally supported by concomitant low grazing pressure derivable from low heterotrophic flagellate abundances. Indeed, in one section of the river affected by a pulp mill, the plant was found to significantly contribute to the self-cleaning capacity of the river. However, this elevated degradation capacity was not enough to compensate for the additionally discharged organic material which, together with the toxic effects of the paper plant effluents, significantly interferes with the ecological status of the Biobío River.     

Variable neighborhood search to solve the generalized orienteering problem

This paper presents a variable neighborhood search (VNS) algorithm to solve the extension of the orienteering problem known as the generalized orienteering problem (GOP). Our algorithm aims to use a reduced number of neighborhoods without compromising the quality of the results. This reduced number of neighborhoods, together with the precalculation of scores associated with points of interest, allows us, in most cases, to outperform all previous metaheuristics proposed for this problem. This is the first time a VNS is being applied to the GOP, and it provides promising computational results. In particular, in the case studies considered in the paper, we were able to find 35 new best solutions, all of which were found using a shorter computational time. Furthermore, the information regarding other best‐known solutions provided in the literature has also been improved, with corrections to some previously published errors regarding scores and distances. In addition, the benchmark has been extended with the incorporation of new case studies based on real data from three of the most popular tourist cities in Spain.     

A microvalve for lab-on-a-chip applications based on electrochemically actuated SU8 cantilevers

Increasing the complexity and functionality of Lab-on-a-Chip devices requires integrated valves for internal flow regulation. The device introduced in this work is an electrochemical SU8 microvalve, featuring a cantilevered structure that seals an adjacent channel under the action of a growing electrolysis bubble. The valve is based on previously reported movable microcantilevers embedded in microchannels. As opposed to membrane valves moving perpendicular to the substrate, the cantilever valve is patterned next to the channel and closes it by moving horizontally, dramatically reducing footprint. Its compact, monolithic and 2D construction reduces assembly problems and results in negligible increase in dead volume (14.5 nl). Besides, its location in the way of the channels avoids the need for auxiliary filling ports or separate working media for actuation. Effective sealing with negligible leakage at 20 kPa has been achieved. The valve also displays passive flow regulation induced by the pressure differential generated across the valve chamber under flow conditions (42-72% flow decrease).     

A comparison between optimization-based human motion prediction methods: data-based, knowledge-based and hybrid approaches

In this paper an optimization-based hybrid dynamic motion prediction method is presented. The method is hybrid as the prediction relies both on actually performed motions for reference (following a data-based approach) and on the definition of appropriate performance measures (following a knowledge-based approach). The prediction is carried out through the definition of a constrained non-linear optimization problem, in which the objective function is composed of a weighted combination of data-based and knowledge-based contributions. The weights of each contribution are varied in order to generate a battery of hybrid predictions, which range from purely data-based to purely knowledge-based. The results of the predictions are analyzed and compared against actually performed motions both qualitatively and quantitatively, using a measure of realism defined as the distance of the predicted motions from the mean of the actually performed motions. The method is applied to clutch pedal depression motions and the comparison between the different approaches favors the hybrid solution, which seems to combine the strengths of both data- and knowledge-based approaches, enhancing the realism of the predicted motion.