• Unified Framework for PdM Algorithm Development: The pdm-tools Architecture

    pg(s) 130-133

    This paper proposes a novel architecture for pdm-tools, a dedicated library designed to streamline the development process of predictive maintenance (PdM) algorithms within the Industry 4.0 paradigm. pdm-tools tackles the time-consuming nature of PdM algorithm development by encompassing five key components within a unified framework. pdm-tools tackles the time challenge by offering a unified workflow: generate synthetic vibration data (gearboxes, bearings, shafts) for rapid prototyping, eliminating initial real-world data collection. Data preparation and feature extraction ensure readiness through scaling, filtering, and extracting diverse condition indicators, empowering robust algorithms. Finally, user-defined model development and evaluation allow training, optimization, and selection of optimal PdM algorithms (machine learning/deep learning). By integrating these functionalities within a single architecture, pdm-tools empowers rapid development and evaluation of PdM algorithms, ultimately leading to a more efficient and effective implementation of predictive maintenance strategies within the Industry 4.0 landscape.

  • Deep Eye – Intelligent Device for Creating Digital Models of Underwater Topography

    pg(s) 126-129

    This project endeavors to advance underwater mapping technologies through the development of algorithms tailored for mapping, analyzing, and creating digital models of bathymetry. A key aspect of this research involves the integration of advanced technologies, specifically High-Definition cameras and a purpose-designed sonar system. By leveraging these tools, high-resolution point cloud data and precise representations of submerged geographical structures can be obtained. The project also aims to conduct a waterproof device, operated by a diver, incorporating both HD cameras and sonar sensors. This device will enhance data collection and mapping capabilities in challenging underwater environments, addressing complexities associated with underwater terrains. Several critical questions are central to this research. Primarily, the investigation seeks to understand how artificial intelligence algorithms can be optimized to harness the collective potential of HD cameras and the specially crafted sonar system. Additionally, the research endeavors to evaluate the comparative efficacy and accuracy of the proposed sonar system and HD cameras in underwater mapping, particularly under adverse environmental conditions where visibility is compromised by high particle concentrations. Througth these efforts, this project aims to contribute to the advancement of underwater mapping technologies, offering insights into optimizing sensor integration and algorithm development for enhanced mapping accuracy in challenging underwater environments.

  • Application of convolutional networks to detect the operating phases of energy systems using a biomass boiler as an example

    pg(s) 122-125

    The development of neural algorithms opens new perspectives for the analysis of technological processes. Particularly relevant are strongly nonlinear and complex objects, such as power plants. One of the modern solutions enabling data analysis are convolutional neural networks (CNNs). The research presents the application of CNNs to monitor and optimize combustion processes in biomass boilers. The fuel analyzed was gray straw, which is difficult to control due to the nature of combustion. The proposed technique is based on the processing of temporal data, which represent different stages of the combustion process. The work examined the effectiveness of the model in identifying key operating parameters and detecting the stages of firing from ignition initialization to nominal operation. Analysis of images of parameter curves from the time waveforms makes it possible to capture repeatable relationships that enable faster response to future changes in the conditions of the combustion process. Determining the phase of the process, based on data and trends of selected parameters, allows the control system to react faster, without operator intervention. As a result of the study, the efficiency of process stage change detection by the convolutional network, expressed by means of an error matrix, through the F1-score parameter (harmonic mean between precision and sensitivity) was achieved at a level close to 96%. The proposed solution can be effectively applied to a number of technological processes including those that are part of Industry 4.0 effectively influencing technological transformation..

  • Fabrication and application of internet of things device using sigfox network and at commands

    pg(s) 118-121

    In this study, the Sigfox network and the Internet were employed to develop an Internet of Things device equipped with temperature, humidity, barometric pressure, and other sensors to measure different physical data. In addition, we attempted to control Sigfox communication by employing AT commands using the Arduino UNOR4 Sigfox module, and social implementation of the proposed system.

  • Review of feature selection methods for Predictive Maintenance Systems

    pg(s) 97-100

    The development of Industry 4.0 and Predictive Maintenance Systems allows for effective utilization of equipment by incorporating ML methods for identifying tool condition. However, including large number of Condition Indicators for machinery monitoring increases computational complexity, hence the response of the system elongates. Therefore, it is important to check the utility of indicators and reduce them. In this paper, we investigate different feature selection methods: AIC (Akaike Information Criterion), BIC (Bayesian Information Criterion), Random Forest, Lasso Regression (L1 Regularization) for NASA Gearbox Fault Detection Dataset, PHM 2009. We processed the raw data and calculated CI from time domain, frequency domain and envelope. An SVM Classifier model was trained on full collection of indicator and reduced, then performances were compared. The obtained results highlight the advantage of feature selection, proving that effective PdM systems can be based on diminished number of Health Indicators.

  • Trends in non-linear MIMO Objects Control in the Era of Industry 4.0: The Use of Artificial Neural Networks

    pg(s) 94-96

    The Industry 4.0 revolution has significantly influenced the control of non-linear Multiple Input Multiple Output (MIMO) systems, particularly through the application of artificial neural networks (ANNs). This paper explores current trends in the control of non-linear MIMO objects, emphasizing the role of ANNs in enhancing performance and efficiency. Key developments, methodologies, and case studies are reviewed to illustrate the impact of ANNs on non-linear MIMO control

  • Modeling solar data using artificial neural networks for solar applications in transport infrastructure

    pg(s) 90-93

    This paper examines an approach using artificial neural networks for innovative modeling of solar data that is needed to realize solar applications for transport infrastructure purposes. Through this modeling, detailed solar data are generated by geographical positions, and monthly and annual maps are created for the territory of Bulgaria for horizontal solar irradiation with its diffuse and direct components and for inclined and reflected solar irradiation, according to Norio Igawa’s model. Diffuse fraction and horizontal and inclined solar irradiation can be helpful in designing solar applications in road infrastructure, such as power signaling systems and street lighting. By demonstrating the capabilities of accurate modeling and analysis of solar data, this paper highlights the importance of applying artificial neural networks in planning and improving the resilience of transport infrastructure against climate change. Using solar energy in transport infrastructure reduces carbon emissions and strengthens environmental sustainability.

  • Simulation experiment for the follow-up controller of the MIMO system

    pg(s) 86-89

    Controlling Multi-Input Multi-Output (MIMO) systems, such as portal conveyors, poses significant challenges due to their inherent complexity and variability. Traditional control methods often fall short in handling the dynamic and nonlinear nature of these systems. This paper presents a novel reinforcement learning (RL) approach, leveraging the twin-delayed deep deterministic policy gradient (TD3) algorithm, to develop a follow-up controller that is robust to changes in system parameters. Our simulation experiments demonstrate the effectiveness of this method.

  • A single-board computer based distributed system for internal and external sound monitoring

    pg(s) 60-62

    This work showcases a distributed noise monitoring system used for measuring noise levels both inside and outside of buildings in urban environments. The system utilises single board computers as main operating units which handle the measurement of parameters and communication. The noise monitoring stations can measure instantaneous and long-term acoustic noise levels, as well as atmospheric parameters such as atmospheric pressure, air temperature and humidity. Measurement stations are connected to an external server via the MQTT protocol. The external server allows for the recording of data into a secure database, as well as for providing end users with historical and instantaneous data on request. A test run of the measurement system was performed and has shown that the measurement stations can function in their design capacity. The system shows great promise in use in internal monitoring of industrial plants both as a precautionary measure for worker health and safety and as an easily modifiable sound monitoring platform in the industry 4.0 standard.

  • Enhancing OCR Accuracy for ID-1 Documents with Security Features through Machine Learning-driven Image Optimization

    pg(s) 56-59

    OCR technology is widely used in various applications, including document digitization, data extraction, and document management systems. The OCR technology has seen significant advancements in recent years, especially with the integration of machine learning and artificial intelligence techniques. These advancements have led to substantial improvements in accuracy, particularly for standard fonts and clear document images. However, challenges still exist, especially when dealing with low-quality images, noisy images, handwritten text, or documents with unusual fonts. Some documents like ID cards, driving licenses, etc. use some security features like deliberate errors, OVI (Optical Variable Ink), Rainbow print, Guilloche pattern, fine line, and microprint to protect the documents from being counterfeited. These security elements also generate noise on the image to perform the OCR. In this paper, we present a way to enhance the OCR accuracy for ID-1 documents with security features through machine learning-driven image optimization. Albanian driving license images on the personalization process are used as a dataset to train the model. During the training process, the model of the ID-1 card is presented with a dataset containing input features (such as images, texts, or numerical data) along with corresponding labels or outcomes. After training, the model and the implemented algorithm to optimize the image for the OCR process are implemented in real-life application.

  • Application of Neural Networks in Underwater Drone Control

    pg(s) 52-55

    The integration of neural networks has revolutionised various technological domains, including the control mechanisms of underwater drones, especially in the context of Industry 4.0. This review explores the application of neural network architectures to improve the navigation, stability and overall performance of underwater drones. A systematic analysis of current methods is provided, focusing on their effectiveness in addressing challenges such as dynamic underwater environments, sensor noise, and real-time decision making. Key advances in neural network-based control strategies are discussed. Furthermore, the integration of these networks with traditional control systems to achieve robust and adaptive control frameworks is highlighted. Through the examination of case studies and experimental results, this review identifies potential areas for future research and development to further the advancement of autonomous underwater vehicles through intelligent control systems.

  • Expanding the Capabilities of Data Transmission Technology to Improve the Efficiency of Multi-Channel Communication Systems

    pg(s) 48-51

    The article, been a continuation of previous studies, proposes and analyzes one of the ways to increase the efficiency of multichannel communication systems through the simultaneous use of several derivative systems of Walsh functions with different generating functions, which are necessary for constructing pseudo-random sequences (PSR). These sequences, in turn, are used as spreading codes when generating noise-like signals in multi-channel data transmission systems with channel separation (SDMA) according to the signal shape or code. The correlation properties of the PSR are studied and it is shown that, in the contrast of the known ones, the proposed approach makes it possible to increase the number of channels in the system without increasing the channel bandwidth that improves the efficiency of a multi-channel communication system.