Pipelines are integral components for storing and transporting liquid and gaseous petroleum products. Despite being durable structures, ruptures can still occur, resulting not only in financial losses and energy waste but, most importantly, in immeasurable environmental disasters and possibly in human casualties. The objective of the ESTHISIS project is the development of a low-cost and efficient wireless sensor system for the instantaneous detection of leaks in metallic pipeline networks transporting liquid and gaseous petroleum products in a noisy industrial environment. The implemented methodology is based on processing the spectrum of vibration signals appearing in the pipeline walls due to a leakage effect and aims to minimize interference in the piping system. It is intended to use low frequencies to detect and characterize leakage to increase the range of sensors and thus reduce cost. In the current work, the smart sensor system developed for signal acquisition and data analysis is briefly described. For this matter, two leakage detection methodologies are implemented. A 2D-Convolutional Neural Network (CNN) model undertakes supervised classification in spectrograms extracted by the signals acquired by the accelerometers mounted on the pipeline wall. This approach allows us to supplant large-signal datasets with a more memory-efficient alternative to storing static images. Second, Long Short-Term Memory Autoencoders (LSTM AE) are employed, receiving signals from the accelerometers, and providing an unsupervised leakage detection solution.

The ability to exploit data for obtaining useful and actionable information and for providing insights is an essential element for continuous process improvements. Recognizing the value of data as an asset, marine engineering puts data considerations at the core of system design. Used wisely, data can help the shipping sector to achieve operating cost savings and efficiency increase, higher safety, wellness of crew rates, and enhanced environmental protection and security of assets. The main goal of this study is to develop a methodology able to harmonize data collected from various sensors onboard and to implement a scalable and responsible artificial intelligence framework, to recognize patterns that indicate early signs of defective behavior in the operational state of the vessel. Specifically, the methodology examined in the present study is based on a 1D Convolutional Neural Network (CNN) being fed time series directly from the available dataset. For this endeavor, the dataset undergoes a preprocessing procedure. Aspiring to determine the effect of the parameters composing the networks and the values that ensure the best performance, a parametric inquiry is presented, determining the impact of the input period and the degree of degradation that our models identify adequately. The results provide an insightful picture of the applicability of 1D-CNN models in performing condition monitoring in ships, which is not thoroughly examined in the maritime sector for condition monitoring. The data modeling along with the development of the neural networks was undertaken with the Python programming language.

Three-phase induction motors (IMs) are considered an essential part of electromechanical systems. Despite the fact that IMs operate efficiently under harsh environments, there are many cases where they indicate deterioration. A crucial type of fault that must be diagnosed early is stator winding faults as a consequence of short circuits. Motor current signature analysis is a promising method for the failure diagnosis of power systems. Wavelets are ideal for both time- and frequency-domain analyses of the electrical current of nonstationary signals. In this paper, the signal data are obtained from simulations of an induction motor for various stator winding fault conditions and one normal operating condition. Our main contribution is the presentation of a fault diagnostic system based on a hybrid discrete wavelet–CNN method. First, the time series of the currents are processed with discrete wavelet analysis. In this way, the harmonic frequencies of the faults are successfully captured, and features can be extracted that comprise valuable information. Next, the features are fed into a convolutional neural network (CNN) model that achieves competitive accuracy and needs significantly reduced training time. The motivations for integrating CNNs into wavelet analysis results for fault diagnosis are as follows: (1) the monitoring is automated, as no human operators are needed to examine the results; (2) deep learning algorithms have the potential to identify even more indistinguishable and complex faults than those that human eyes could.

Road safety is a major global concern, as millions of lives are lost every year because of road accidents. Towards an effort to increase road safety, several Internet-of-Vehicle systems have been developed over the last years in order to better monitor vehicle and driver behavior and issue warnings that effectively prevent life-threatening accidents. These systems face a number of challenges including connectivity issues and high installation and/or maintenance costs. The current work introduces the ODOS2020 system, an integrated Internet-of-Vehicles system aiming to increase road safety. The system comprises several On-the-Road Units for vehicle-related data collection from affordable, energy-efficient magnetometers and calculation of critical parameters, such as each passing vehicle’s speed and direction. A Road-Side Unit accumulates data from the On-the-Road Units, sends data to a cloud infrastructure for further analysis and sends dedicated warnings to the drivers based on their road behavior and/or specific traffic conditions via a dedicated Human–Machine Interface. The overall system architecture and the key features of its modules are being presented, as well as the evaluation results of specially designed tests performed in an actual motorway under real use case scenarios. The evaluation results showed both a very good technical performance of the system and a high level of user acceptance. This in turn means that the system can be employed for effective traffic control and road accident avoidance via monitoring of critical vehicle parameters and early warning of the drivers based on their and other drivers’ behavior, road conditions and real-time, unpredictable events.

A wireless monitoring system based on piezoelectric energy harvesting (PEH) is presented to provide fatigue data of wind turbine blades in operation. The system comprises three subsystems, each respectively providing the following functions: (i) the conversion of mechanical to electric energy by exploiting the bistable vibration of a composite beam with piezoelectric patches in post-buckling, (ii) harvesting the converted energy by means of a modified, commercial, off-the-shelf (COTS) circuit to feed a LiPo battery and (iii) the battery-powered acquisition and wireless transmission of sensory signals to the cloud to be elaborated upon by the end-user. The system was verified with ground tests under representative operation conditions, which demonstrated the fulfillment of the design requirements. The measurements indicated that the system provided 23% of the required power for fully autonomous operation when subjected to white noise base excitation of 1 g acceleration in the range of 1–20 Hz.

To preserve handwritten historical documents, libraries are choosing to digitize them, ensuring their longevity and accessibility. However, the true value of these digitized images lies in their transcription into a textual format. In recent years, various tools have been developed utilizing both traditional and AI-based models to address the challenges of deciphering handwritten texts. Despite their importance, there are still several obstacles to overcome, such as the need for scalable and modular solutions, as well as the ability to cater to a continuously growing user community autonomously. This study focuses on introducing a new information fusion architecture, specifically highlighting the Gateway API. Developed as part of the μDoc.tS research program, this architecture aims to convert digital images of manuscripts into electronic text, ensuring secure and efficient routing of requests from front-end applications to the back end of the information system. The validation of this architecture demonstrates its efficiency in handling a large volume of requests and effectively distributing the workload. One significant advantage of this proposed method is its compatibility with everyday devices, eliminating the need for extensive computational infrastructures. It is believed that the scalability and modularity of this architecture can pave the way for a unified multi-platform solution, connecting diverse user environments and databases.