Abstract

Abstract

Abstract
This paper addresses a single synchronous controller (SSC) for interfaced converters with high penetration of renewable energy resources (RERs) into a low inertia power grid. The SSC is modelled based on a comprehensive dependence between each operative feature of a synchronous generator (SG) and a power electronics converter. This can properly improve the performance of the power grid in such scenarios in which large-scale penetration of RERs are detected. The main contribution of this paper is representing an exhaustive relation between active components of the proposed SSC and SG features which enables the proposed SSC-based interfaced converter to more accurately mimic the behaviour of SGs during active power generating along with providing controllable inertia. Due to containing sufficient decoupling, both components of the proposed SSC have no impact on each other, also the proposed SSC has a superior operational flexibility within a wide range of inertia from very low to high values. Thus, two closed-loop control systems are considered to separately analyse the characteristic effects of SGs in active and reactive power sharing apart from the power grid stability challenges. In addition, the impacts of active power variations on reactive power are subsequently evaluated. To further analyse the operation of the system, the effects of the virtual mechanical power (VMP) error embedded in the SSC are considered as an alternative option for assessing the power grid stability. Also, the variations of the virtual angular frequency (VAF) error are carefully deliberated for more considerations associated with the active and reactive power performance of the SSC. Simulation results are presented to demonstrate the high performance of the SSC in the control of the power electronics-based SG when high-penetration renewable energy sources are integrated into the low inertia power grid.

Abstract
Lung volume segmentation is a key step in the design of Computer-Aided Diagnosis systems for automated lung pathology analysis. However, isolating the lung from CT volumes can he a challenging process due to considerable deformations and the potential presence of pathologies. Convolutional Neural Networks (CNN) are effective tools for modeling the spatial relationship between lung voxels. Unfortunately, they typically require large quantities of annotated data, and manually delineating the lung from volumetric CT scans can he a cumbersome process. We propose to train a 3D CNN to solve this task based on semi-automatically generated annotations. For this, we introduce an extension of the well-known V-Net architecture that can handle higher dimensional input data. Even if the training set labels are noisy and contain errors, our experiments show that it is possible to learn to accurately segment the lung relying on them. Numerical comparisons on an external test set containing lung segmentations provided by a medical expert demonstrate that the proposed model generalizes well to new data, reaching an average 98.7% Dice coefficient. The proposed approach results in a superior performance with respect to the standard V-Net model, particularly on the lung boundary.

Abstract
When dealing with electrocardiography (ECG) the main focus relies on the classification of the heart's electric activity and deep learning has been proving its value over the years classifying the heartbeats, exhibiting great performance when doing so. Following these assumptions, we propose a deep learning model based on a ResNet architecture with convolutional ID layers to classes the beats into one of the 4 classes: normal, atrial premature contraction, premature ventricular contraction and others. Experimental results with MIT-BIH Arrhythmia Database confirmed that the model is able to perform well, obtaining an accuracy of 96% when using stochastic gradient descent (SGD) and 83% when using adaptive moment estimation (Adam), SGD also obtained F1-scores over 90% for the four classes proposed. A larger dataset was created and tested as unforeseen data for the trained model, proving that new tests should be done to improve the accuracy of it.

Abstract
The building sector is responsible for a large share of Europe's energy consumption. Modelling buildings thermal behavior is a key factor for achieving the EU energy efficiency goals. Moreover, it can be used in load forecasting applications, for the prediction of buildings total energy consumption. The first phase of this work is the application of Artificial Immune Systems (AIS) for clustering buildings with similar physical characteristics and similar thermal efficiency. In the second phase, Artificial Neural Networks (ANN) are used to estimate the buildings heating and cooling loads. A final sensitivity test is performed to identify which building features have the most impact on the heating and cooling loads. The results obtained in the first phase revealed very distinct cluster prototypes, which demonstrates the AIS discriminating ability. The good estimation performance obtained in the second phase showed that this approach can be integrated in energy efficiency audits. Finally, the sensitivity analysis provided indications for actions (or legislation directives) in order to promote the design of more efficient buildings. © 2019 IEEE.