Search

Automatic vision-based inspection systems have played a key role in product quality assessment for decades through the segmentation, detection, and classification of defects. Historically, machine learning frameworks, based on hand-crafted feature extraction, selection, and validation, counted on a combined approach of parameterized image processing algorithms and explicated human knowledge. The outstanding performance of deep learning (DL) for vision systems, in automatically discovering a feature representation suitable for the corresponding task, has exponentially increased the number of scientific articles and commercial products aiming at industrial quality assessment. In such a context, this article reviews more than 220 relevant articles from the related literature published until February 2023, covering the recent consolidation and advances in the field of fully-automatic DL-based surface defects inspection systems, deployed in various industrial applications. The analyzed papers have been classified according to a bi-dimensional taxonomy, that considers both the specific defect recognition task and the employed learning paradigm. The dependency on large and high-quality labeled datasets and the different neural architectures employed to achieve an overall perception of both well-visible and subtle defects, through the supervision of fine or/and coarse data annotations have been assessed. The results of our analysis highlight a growing research interest in defect representation power enrichment, especially by transferring pre-trained layers to an optimized network and by explaining the network decisions to suggest trustworthy retention or rejection of the products being evaluated.

Nowadays, due to the decreasing use of traditional generators in favor of renewable energy sources, power grids are facing a reduction of system inertia and primary frequency regulation capability. Such an issue is exacerbated by the continuously increasing number of electric vehicles (EVs), which results in enforcing novel approaches in the grid operations management. However, from being an issue, the increase of EVs may turn to be a solution to several power system challenges. In this context, a crucial role is played by the so-called vehicle-to-grid (V2G) mode of operation, which has the potential to provide ancillary services to the power grid, such as peak clipping, load shifting, and frequency regulation. More in detail, EVs have recently started to be effectively used for one of the most traditional frequency regulation approaches: the so-called frequency droop control (FDC). This is a primary frequency regulation, currently obtained by adjusting the active power of generators in the main grid. Because to the decommissioning of traditional power plants, EVs are thus recognized as particularly valuable solutions since they can respond to frequency deviation signals by charging or discharging their batteries. Against this background, we address frequency regulation of a power grid model including loads, traditional generators, and several EVs. The latter independently participate in the grid optimization process providing the grid with ancillary services, namely the FDC. We propose two novel control strategies for the optimal control of the batteries of EVs during the frequency regulation service. On the one hand, the control strategies ensure re-balancing the power and stabilizing the frequency of the main grid. On the other hand, the approaches are able to satisfy different types of needs of EVs during the charging process. Differently from the related literature, where the EVs perspective is generally oriented to achieve the optimal charge level, the proposed approaches aim at minimizing the degradation of battery devices. Finally, the proposed strategies are compared with other state-of-the-art V2G control approaches. The results of numerical experiments using a realistic power grid model show the effectiveness of the proposed strategies under the actual operating conditions.

The efficient management of Heating Ventilation and Air Conditioning (HVAC) systems in smart buildings is one of the main applications of the Internet of Things (IoT) paradigm. In this paper we propose an IoT based architecture for the implementation of Model Predictive Control (MPC) of HVAC systems in real environments. The considered MPC algorithm optimizes on line, in a closed-loop control fashion, both the indoor thermal comfort and the related energy consumption for a single zone environment. Thanks to the proposed IoT based architecture, the sensing, control, and actuating subsystems are all connected to the Internet, and a remote interface with the HVAC control system is guaranteed to end-users. In particular, sensors and actuators communicate with a remote database server and a control unit, which provides the control actions to be actuated in the HVAC system; users can set remotely the control mode and related set-points of the system; while comfort and environmental indices are transferred via the Internet and displayed on the end-users’ interface. The proposed IoT based control architecture is implemented and tested in a campus building at the Polytechnic of Bari (Italy) in a proof of concept perspective. The effectiveness of the proposed control algorithm is assessed in the real environment evaluating both the thermal comfort results and the energy savings with respect to a classical thermostat regulation approach.

Investing in the optimal measures for improving the energy efficiency of urban street lighting systems has become strategic for the economic, technological and social development of cities. The decision-making process for the selection of the optimal set of interventions is not so straightforward. Several criticalities-such as difficulties getting access to credit for companies involved in street lighting systems refurbishment, budget constraints of municipalities, and unawareness of the actual energy and economic performance after a retrofitting intervention-require a decision-making approach that supports the city energy manager in selecting the optimal street lighting energy efficiency retrofitting solution while looking not only based on the available budget, but also based on the future savings in energy expenditures. In this context, the purpose of our research is to develop an effective decision-making model supporting the optimal multi-period planning of the street lighting energy efficiency retrofitting, which proves to be more effective and beneficial than the classical single-period approach and has never before been applied to the considered public lighting system context. The proposed methodology is applied to a real street lighting system in the city of Bari, Italy, showing the energy savings and financial benefit obtained through the proposed method. Numerical experiments are used to investigate and quantify the effects of using a multi-period planning approach instead of a single-period approach.