Alberto Jose Alvares

Semantic segmentation of metallographic micrographs is a key task for quantitative microstructural analysis in additive manufacturing, yet it remains challenging due to phase heterogeneity, complex morphologies, and the scarcity of annotated data. The MetalDAM dataset, composed of 42 labeled scanning electron microscopy images of steel microstructures, has been widely adopted as a benchmark, with U-Net commonly reported as the strongest supervised baseline. Nevertheless, the encoder–decoder structure of U-Net imposes architectural constraints that hinder the precise delineation of heterogeneous and irregular phase boundaries under severe data limitations. To address this limitation, this paper investigates a Fully Convolutional Network (FCN)-based architecture as an alternative approach for semantic segmentation on the MetalDAM dataset. The FCN is trained and evaluated under the same experimental protocol as the U-Net baseline, enabling a direct and fair comparison. Performance is assessed using multiple evaluation metrics, including Intersection over Union (IoU), precision, recall, and mean Average Precision at an IoU threshold of 0.5. The results show that the FCN achieves comparable overall IoU values (0.75) while delivering substantial improvements at the class level, particularly for minority and morphologically complex phases, with gains of up to 25–30% in class-specific IoU. Additional metrics confirm enhanced robustness, with consistently higher precision, recall, and mAP@0.5 values. These findings demonstrate that FCN-based architectures constitute a competitive and robust alternative to U-Net for metallographic segmentation in additive manufacturing scenarios characterized by limited annotated data.

In the context of Industry 4.0, new methods of manufacturing, monitoring, and data generation related to industrial processes have emerged. Over the last decade, a new method of part manufacturing that has been revolutionizing the industry is Additive Manufacturing, which comes in various forms, including the more traditional Fusion Deposition Modeling (FDM) and the more innovative ones, such as Laser Metal Deposition (LMD) and Wire Arc Additive Manufacturing (WAAM). New technologies related to monitoring these processes are also emerging, such as Cyber-Physical Systems (CPSs) or Digital Twins (DTs), which can be used to enable Artificial Intelligence (AI)-powered analysis of generated big data. However, few works have dealt with a comprehensive data analysis, based on Digital Twin systems, to study quality levels of manufactured parts using 3D models. With this background in mind, this current project uses a Digital Twin-enabled dataflow to constitute a basis for a proposed data analysis pipeline. The pipeline consists of analyzing metal AM-manufactured parts’ surface roughness quality levels by the application of a Deep Neural Network (DNN) analytical model and enabling the assessment and tuning of deposition parameters by comparing AM-built models’ 3D representation, obtained by photogrammetry scanning, with the positional data acquired during the deposition process and stored in a cloud database. Stored and analyzed data may be further used to refine the manufacturing of parts, calibration of sensors and refining of the DT model. Also, this work presents a comprehensive study on experiments carried out using the CW-GTAW (Cold Wire Gas Tungsten Arc Welding) process as the means of depositing metal, resulting in hollow parts whose geometries were evaluated by means of both 3D scanned data, obtained via photogrammetry, and positional/deposition process parameters obtained from the Digital Twin architecture pipeline. Finally, an adapted PointNet DNN model was used to evaluate surface roughness quality levels of point clouds into 3 classes (good, fair, and poor), obtaining an overall accuracy of 75.64% on the evaluation of real deposited metal parts.

The ongoing digitalization of manufacturing is transforming how products are designed, produced, and optimized, driven by the convergence of the Digital Thread (DTh) and Digital Twin (DTw) paradigms. However, achieving seamless integration across these technologies remains a major challenge due to persistent issues of data interoperability, consistency, collaboration, and intelligent data exchange throughout the product lifecycle. Moreover, the limited intelligence at the machine level and the fragmentation of data across heterogeneous systems hinder the realization of fully connected, autonomous manufacturing environments. To address these challenges, this work conducts a comprehensive review of the literature on DTh and DTw technologies, focusing on their application in additive manufacturing (AM) and their alignment with international standards for data exchange and system interoperability. Building on this foundation, a unified digital ecosystem for contextualized intelligence is proposed, aiming to integrate DTh and DTw through standardized, semantically rich, and interoperable data flows. Furthermore, a standards-based DTh–DTw framework is presented, leveraging key industrial standards, including STEP/STEP-NC, MTConnect, QIF, OPC UA, MQTT, and ISO 23247, to ensure traceability, real-time synchronization, and data-driven decision-making across the AM lifecycle. Two implementation scenarios validate the proposed approach: (i) an FDM-based AM process using STEP-NC and MTConnect for integrated process planning and monitoring, and (ii) a robotic wire-based LMD cell featuring three DTw implementations compliant with ISO 23247 for real-time simulation, predictive maintenance, and process visualization. These implementations demonstrate the feasibility of constructing interoperable, data-centric manufacturing workflows using open standards. The results underscore the potential of the proposed ecosystem to enhance interoperability, data consistency, and intelligence across manufacturing processes, while also revealing current limitations in AM-specific standardization and cross-platform integration. Although focused on AM, the framework is inherently extensible to other manufacturing domains, paving the way toward standards-driven, intelligent, and generative manufacturing ecosystems.

Digital Twins (DTs) are transforming manufacturing by bridging the physical and digital worlds, enabling real-time insights, predictive analytics, and enhanced decision making. In Industry 4.0, DTs facilitate automation and data integration, while Industry 5.0 emphasizes human-centric, resilient, and sustainable production. However, implementing DTs in robotic metal additive manufacturing (AM) remains challenging because of the complexity of the wire-based laser metal deposition (LMD) process, the need for real-time monitoring, and the demand for advanced defect detection to ensure high-quality prints. This work proposes a structured DT architecture for a robotic wire-based LMD cell, following a standard framework. Three DT implementations were developed. First, a real-time 3D simulation in RoboDK, integrated with a 2D Node-RED dashboard, enabled motion validation and live process monitoring via MQTT (message queuing telemetry transport) telemetry, minimizing toolpath errors and collisions. Second, an Industrial IoT-based system using KUKA iiQoT (Industrial Internet of Things Quality of Things) facilitated predictive maintenance by analyzing motor loads, joint temperatures, and energy consumption, allowing early anomaly detection and reducing unplanned downtime. Third, the Meltio dashboard provided real-time insights into the laser temperature, wire tension, and deposition accuracy, ensuring adaptive control based on live telemetry. Additionally, a prescriptive analytics layer leveraging historical data in FireStore was integrated to optimize the process performance, enabling data-driven decision making.

In modern manufacturing, Digital Thread and Digital Twin technologies integrate and orchestrate data throughout the production lifecycle. To fully realize their potential, it is crucial to address challenges in data integration, management, and interoperability while enhancing system intelligence and contextual awareness. STEP-NC (STandard for the Exchange of Product model data - Numerical Control) emerges as a potential solution, enriching these digital systems with contextual information about products, processes, and machines. While STEP-NC has been extensively studied in machining, its application in Additive Manufacturing (AM) is still emerging. The current reliance on legacy formats like STL and G-code fails to meet the industry’s advancing needs. This work offers a comprehensive review of the current state of STEP-NC development for AM, highlighting a growing interest in its application, despite limited publications and advancements. To bridge these gaps, we propose STEP-NC entity definitions for managing data related to process parameters and operations in Fused Deposition Modeling (FDM) and Laser Metal Deposition (LMD) technologies. STEP-NC facilitates refined data management at the individual layer level, encompassing geometry, process parameters, material properties, and other key aspects. Additionally, we introduce a cyber-physical architecture for STEP-NC in manufacturing, aligned with the ISO 23247 framework. This architecture envisions a robust digital ecosystem driven by standards-based Digital Thread and Digital Twin technologies, with STEP and STEP-NC at its core, supported by essential open standards such as QIF, MTConnect, OPC-UA, and MQTT. By integrating contextual information, it enhances Digital Twin development for virtual monitoring, optimization, knowledge generation, and informed decision-making in manufacturing environments.