The plasma electron temperature has been estimated starting from the spectroscopic analysis of the optical emission of the laser-generated plasma plume during quite diverse stainless steel welding procedures (c.w. CO2 and pulsed Nd:YAG). Although the optical emissions present different spectral features, a discrete contribution of several iron lines can be highlighted in both types of welding. We have found that the electron temperature decreases as the laser power is enhanced, in static as well as dynamic conditions. Such a result could be useful to develop a closed loop control system of the weld penetration depth

Load flexibility management is a promising approach to face the problem of balancing generation and demand in electrical grids. This problem is becoming increasingly difficult due to the variability of renewable energies. Thermostatically-controlled loads can be aggregated and managed by a virtual battery, and they provide a cost-effective and efficient alternative to physical storage systems to mitigate the inherent variability of renewable energy sources. However virtual batteries require that an accurate control system is capable of tracking frequency regulation signals with minimal error. A real-time control system allowing virtual batteries to accurately track frequency or power signals is developed. The performance of this controller is validated for a virtual battery composed of 1000 thermostatically-controlled loads. Using virtual batteries equipped with the developed controller, a study focused on residential thermostatically-controlled loads in Spain is performed. The results of the study quantify the potential of this technology in a country with different climate areas and provides insight about the feasibility of virtual batteries as enablers of electrical systems with high levels of penetration of renewable energy sources.

An unsupervised approach to classify surface defects in wire rod manufacturing is developed in this paper. The defects are extracted from an eddy current signal and classified using a clustering technique that uses the dynamic time warping distance as the dissimilarity measure. The new approach has been successfully tested using industrial data. It is shown that it outperforms other classification alternatives, such as the modified Fourier descriptors.