Million cubic meters from abandoned mines worldwide could be used as subsurface reservoirs for large scale energy storage systems, such as adiabatic compressed air energy storage (A-CAES). In this paper, analytical and three-dimensional CFD numerical models have been conducted to analyze the thermodynamic performance of the A-CAES reservoirs in abandoned mines during air charging and discharging processes. Unlike other research works, in which the heat transfer coefficient is considered constant during the operation time, in the present investigation a correlation based on both unsteady Reynolds and Rayleigh numbers is employed for the heat transfer coefficient in this type of application. A tunnel with a 35 cm thick concrete lining, 200 m³ of useful volume and typical operating pressures from 5 to 8 MPa were considered. Fiber-reinforced plastic (FRP) and steel were employed as sealing layers in the simulations around the fluid. Finally, the model also considers a 2.5 m thick sandstone rock mass around the concrete lining. The results obtained show significant heat flux between the pressurized air and the sealing layer and between the sealing layer and concrete lining. However, no temperature fluctuation was observed in the rock mass. The air temperature fluctuations are reduced when steel sealing layer is employed. The thermal energy balance through the sealing layer for 30 cycles, considering air mass flow rates of 0.22 kg s⁻¹ (charge) and −0.45 kg s⁻¹ (discharge), reached 1056 and 907 kWh for FRP and steel, respectively. In general, good agreements between analytical and numerical simulations were obtained.

The drill-and-blast method is widely used for the excavation of hard rock tunnels. Toxic gases such as carbon monoxide and nitrogen oxides are released immediately after blasting by the detonation of explosives. To provide a safe working environment, the concentration of noxious gases must be reduced below the threshold limit value according to health and safety regulations. In this paper, one-dimensional mathematical models and three-dimensional CFD numerical simulations were conducted to analyze the concentration, propagation and dilution of the blasting fumes under different operating conditions. Forced, exhaust and mixed ventilation modes were compared to determine the safe re-entry times after blasting in a 200 m-long tunnel excavated using the top-heading-and-benching method. Based on the numerical simulations, carbon monoxide was the most critical gas, as it required a longer ventilation time to reduce its concentration below the threshold limit value. The safe re-entry time reached 480 s under the typical forced ventilation mode, but was reduced to 155 s when a mixed ventilation system was used after blasting, reducing the operating costs. The reduction of the re-entry time represents a significant improvement in the excavation cycle. In addition, the results obtained show that 1D models can be used to preliminary analyze the migration of toxic gases. However, to reliably determine the safe re-entry times, 3D numerical models should be developed. Finally, to verify the accuracy of the CFD results, field measurements were carried out in a railway tunnel using gas sensors. In general, good agreements were obtained between the 3D numerical simulations and the measured values.

It is well known that working photovoltaic (PV) plants show several maintenance needs due to wiring and module degradation, mismatches, dust, and PV cell defects and faults. There are a wide range of theoretical studies as well as some laboratory tests that show how these circumstances may affect the PV production. Thus, it is mandatory to evaluate the whole PV plant performance and, then, its payback time, profitability, and environmental impact or carbon footprint. However, very few studies include a systematic procedure to quantify and supervise the real degradation effects and fault impacts on the field. In this paper, the authors first conducted a brief review of the most frequent PV faults and the degradation that can be found under real conditions of operation of PV plants. Then, they proposed and developed an innovative Geographic Information System (GIS) application to locate and supervise them. The designed tool was applied to both a large PV plant of 108 kWp and a small PV plant of 9 kWp installed on a home rooftop. For the large PV plant, 24 strings of PV modules were modelized and introduced into the GIS application and every module in the power plant was studied including voltage, current, power, series and parallel resistances, fill factor, normalized PV curve to standard test conditions (STC), thermography and visual analysis. For the small PV installation three strings of PV panels were studied identically. It must be noted that PV modules in this case included power optimizers. The precision of the study enabled the researchers to locate and supervise up to a third part of every PV cell in the system, which can be adequately georeferenced. The developed tool allows both the researchers and the investors to increase control of the PV plant performance, to lead to better planning of maintenance actuations, and to evaluate several PV module replacement strategies in a preventive maintenance program. The PV faults found include hot spots, snail tracks, ethylene vinyl acetate (EVA) discoloration, PV cell fractures, busbar discoloration, bubbles and Si discoloration.

Mining is an essential activity for obtaining materials necessary for the well-being and development of society. However, this activity produces important environmental impacts that must be controlled. More specifically, there are different soils near new or abandoned mining productions that have been contaminated with potentially toxic elements, and currently represent an important environmental problem. In this research, a contaminated soil from the mining district of Linares was studied for its use as a raw material for the conforming of ceramic materials, bricks, dedicated to construction. Firstly, the contaminated soil was chemically and physically characterized in order to evaluate its suitability. Subsequently, different families of samples were conformed with different percentages of clay and contaminated soil. Finally, the conformed ceramics were physically and mechanically characterized to examine the variation produced in the ceramic material by the incorporation of the contaminated soil. In addition, in this research, leachate tests were performed according to the TCLP method determining whether encapsulation of potentially toxic elements in the soil occurs. The results showed that all families of ceramic materials have acceptable physical properties, with a soil percentage of less than 80% being acceptable to obtain adequate mechanical properties and a maximum of 70% of contaminated soil to obtain acceptable leachate according to EPA regulations. Therefore, the maximum percentage of contaminated soil that can be incorporated into the ceramic material is 70% in order to comply with all standards. Consequently, this research not only avoids the contamination that contaminated soil can produce, but also valorizes this element as a raw material for new materials, avoiding the extraction of clay and reducing the environmental impact.

Benzene is a pollutant which is very harmful to our health, so models are necessary to predict its concentration and relationship with other air pollutants. The data collected by eight stations in Madrid (Spain) over nine years were analyzed using the following regression-based machine learning models: multivariate linear regression (MLR), multivariate adaptive regression splines (MARS), multilayer perceptron neural network (MLP), support vector machines (SVM), autoregressive integrated moving-average (ARIMA) and vector autoregressive moving-average (VARMA) models. Benzene concentration predictions were made from the concentration of four environmental pollutants: nitrogen dioxide (NO2), nitrogen oxides (NOx), particulate matter (PM10) and toluene (C7H8), and the performance measures of the model were studied from the proposed models. In general, regression-based machine learning models are more effective at predicting than time series models.