Hybrid photovoltaic/thermal (PV/T) solar systems simulation with Simulink/Matlab

The purpose of this work consists in thermodynamic modeling of hybrid photovoltaic–thermal (PV/T) solar systems, pursuing a modular strategy approach provided by Simulink/Matlab.PV/T solar systems are a recently emerging solar technology that allows for the simultaneous conversion of solar energy into both electricity and heat. This type of technology present some interesting advantages over the conventional “side-by-side” thermal and PV solar systems, such as higher combined electrical/thermal energy outputs per unit area, and a more uniform and aesthetical pleasant roof area. Despite the fact that early research on PV/T systems can be traced back to the seventies, only recently it has gained a renewed impetus. In this work, parametric studies and annual transient simulations of PV/T systems are undertaken in Simulink/Matlab. The obtained results show an average annual solar fraction of 67%, and a global overall efficiency of 24% (i.e. 15% thermal and 9% electrical), for a typical four-person single-familiar residency in Lisbon, with p-Si cells, and a collector area of 6 m². A sensitivity analysis performed on the PV/T collector suggests that the most important variable that should be addressed to improve thermal performance is the photovoltaic (PV) module emittance. Based on those results, some additional improvements are proposed, such as the use of vacuum, or a noble gas at low-pressure, to allow for the removal of PV cells encapsulation without air oxidation and degradation, and thus reducing the PV module emittance. Preliminary results show that this option allows for an 8% increase on optical thermal efficiency, and a substantial reduction of thermal losses, suggesting the possibility of working at higher fluid temperatures. The higher working temperatures negative effect in electrical efficiency was negligible, due to compensation by improved optical properties. The simulation results are compared with experimental data obtained from other authors and perform reasonably well.The Simulink modeling platform has been mainly used worldwide on simulation of control systems, digital signal processing and electric circuits, but there are very few examples of application to solar energy systems modeling. This work uses the modular environment of Simulink/Matlab to model individual PV/T system components, and to assemble the entire installation layout. The results show that the modular approach strategy provided by Matlab/Simulink environment is applicable to solar systems modeling, providing good code scalability, faster developing time, and simpler integration with external computational tools, when compared with traditional imperative-oriented programming languages.     

Role of peroxy chemistry in the high-pressure ignition of n-butanol – Experiments and detailed kinetic modelling

Despite considerable interest in butanol as a potential biofuel candidate, its ignition behaviour at elevated pressures still remains largely unexplored. The present study investigates the oxidation of n-butanol in air at pressures near 80 bar. Ignition delays were determined experimentally in the temperature range of 795–1200 K between 61 and 92 bar. The time of ignition was determined by recording pressure and CH-emission time histories throughout the course of the experiments. The results display the first evidence of the influence of negative temperature coefficient (NTC) behaviour which was not observed in earlier ignition studies. The high-pressure measurements show that NTC behaviour is enhanced as pressures are increased. The experimental results were modelled using an improved chemical kinetic mechanism which includes a simplified sub-mechanism for butyl-peroxy formation and isomerisation reactions currently incompletely accounted for in n-butanol kinetic models. The detailed mechanism validated with the high-pressure ignition results for realistic engine in-cylinder conditions can have significant impact on future advanced low-temperature combustion engines.     

Microencapsulation of Propolis in Protein Matrix Using Spray Drying for Application in Food Systems

Propolis presents several health benefits due to the presence of bioactive compounds, mainly phenolic compounds; however, its application in food is limited due to undesirable odor and low water solubility. The bioactive compounds are usually susceptible to degradation by exposure to light, heat, or oxygen or by interaction with other compounds, which may limit its biological activity. The study aimed the propolis extract microencapsulation using rice, pea, soybean, and ovoalbumin proteins as wall material by spray drying and to analyze their in vitro digestion. The propolis extract presented a high concentration of apigenin. Encapsulation efficiency was greater than 70%, and it was maintained the antioxidant activity of propolis (88% inhibition of DPPH for propolis extract and >?73% for the microparticles). The DSC, ATR-FTIR, and X-ray diffraction techniques confirmed the encapsulation. The microparticles showed different shapes, sizes, and physical characteristics. The microparticles encapsulated with pea protein could be used in formulations of Minas Frescal cheese due to the controlled released, whereas the other microparticles could be used in pudding formulations.     

A deep learning approach for the forensic evaluation of sexual assault

Despite the existence of patterns able to discriminate between consensual and non-consensual intercourse, the relevance of genital lesions in the corroboration of a legal rape complaint is currently under debate in many countries. The testimony of the physicians when assessing these lesions has been questioned in court due to several factors (e.g., a lack of comprehensive knowledge of lesions, wide spectrum of background area, among others). Therefore, it is relevant to provide automated tools to support the decision process in an objective manner. In this work, we evaluate the performance of state-of-the-art deep learning architectures for the forensic assessment of sexual assault. We propose a deep architecture and learning strategy to tackle the class imbalance on deep learning using ranking. The proposed methodologies achieved the best results when compared with handcrafted feature engineering and with other deep architectures.