This study reports a combined experimental–numerical investigation of hydraulic-assisted single point incremental forming (HA-SPIF) to enhance thickness uniformity in truncated quadrilateral parts formed from AA1050 aluminum alloy sheets. A finite element model is established and validated using experiments on circular blanks of 220 mm diameter and 1.0 mm thickness, forming components with a 60° wall angle and a target height of 55 mm. Material anisotropy is characterized through uniaxial tensile tests along 0°, 45°, and 90° to the rolling direction, and direction-dependent Voce hardening laws are calibrated from true stress–strain data, with flow stresses of 69.0–82.6 MPa and anisotropy coefficients of 0.45–0.86. The validated model is first employed to assess the effect of material orientation on thickness distribution, revealing pronounced differences in thinning behavior. A subsequent parametric study examines the coupled influences of hydraulic pressure, tool feed rate, and vertical step size on thickness variation and minimum residual thickness. Hydraulic assistance is shown to effectively suppress localized thinning and significantly improve thickness homogeneity relative to conventional SPIF. The main contribution of this work is the integrated consideration of anisotropic Voce hardening and hydraulic support for a non-axisymmetric geometry, providing a reliable predictive framework and practical guidelines for thinning control in high-precision forming applications.
Additive manufacturing with powder bed fusion of metals using a laser beam (PBF-LB/M) enables the manufacturing of complex medical implant geometries, such as patient-specific dental implants that replicate the shape of natural tooth roots. Such geometries may support improved osseointegration in the jawbone. This study investigates commercially pure titanium grade 2 (TiGd2) produced by PBF-LB/M as the basis of an additive-subtractive process for medical implants. Specimens were manufactured in different build orientations (BO) to assess material and surface properties. Micro computed tomography indicated a low internal defect density. Tensile testing showed BO dependent anisotropy, whereas compression testing showed no pronounced anisotropy. Across all BO, mean values were ultimate tensile strength RM = 772 MPa, elongation at break A = 24.4 %, and compressive yield strength R_dp = 746 MPa. Surface post processing enabled adjustment of roughness, with Ra decreasing from Ra_as = 6.3 µm in the as built state to Ra_pb = 5.1 µm after particle blasting, Ra_et = 3.6 µm after acid etching, and Ra_df = 0.36 µm after disc finishing. Overall, the results confirm the potential of PBF-LB/M processed TiGd2 for additive subtractive manufacturing of medical implants.
Industry 4.0 and the rise of collaborative value chains necessitate seamless data integration across organizational and technological boundaries. Yet the fragmented standardization environment prevents interoperability across systems and within frameworks like the Digital Twin (DT). While the Asset Administration Shell (AAS) and Open Platform Communications (OPC) Unified Architecture (UA) are established as core standards for digital asset representation and communication, their independent origins and differing metamodels create semantic gaps that hinder unified machine-level understanding. This systematic mapping study examines how Knowledge Graphs (KG) can bridge semantic gaps to create unified semantic models. Current research demonstrates progress in pairwise integrations, such as transforming OPC UA models into RDF graphs or synchronizing AAS repositories with graph databases. However, these efforts often remain isolated solutions that address specific integration challenges. Consequently, large-scale industrial deployment remains limited. This paper consolidates current knowledge on integrating AAS, OPC UA and KGs, contextualizing these efforts and highlights research gaps toward achieving interoperable manufacturing ecosystems.
Contact coordinate measuring systems determine the dimensions and forms of complex geometries. In the context of Industry 4.0, measurement integration into the manufacturing process, such as a machine tool, is sought. On-machine measurement (OMM) may enhance productivity and reduce costs, but it also presents challenges, including the presence of coolant and contaminants, as well as dynamic errors that impact measurement. The influence of scanning speed and sense, as well as the size and type of the measured ring gauges, on measurement is experimentally investigated based on bidirectional scans of five reference ring gauges at three speeds between 2116 mm/min and 15240 mm/min. The actual measurement deviations from the nominal value were isolated by subtracting the first harmonic from the raw machine and probe readings. The peak-to-peak values and root mean square (RMS) values of probe readings were used to assess the results. The peak-to-peak and RMS values increase as the speed and diameter of the ring increase. In the case of high scanning speed, for example, the RMS value is 0.126 mm, and for low scanning speed, the RMS value is 0.062 mm. The non-parametric Kruskal-Wallis test shows a significant impact of scanning speed (p = 7.75e-9) and ring size (p = 9.31e-4) on the magnitude of error.
Small and Medium-sized Enterprises operating legacy stamping presses often lack affordable diagnostics, risking undetected tool damage. This study proposes a frugal vibroacoustic retrofitting methodology using a low-cost piezoelectric sensor and edge processing. Experiments compared normal blanking (0.5 mm brass) with simulated double-hits (0.1 mm spacer). Analyses revealed that abnormal events generate distinct, highly impulsive transients with shorter post-impact signal persistence and elevated broadband energy including mid- and high-frequency components. Unsupervised clustering confirmed reliable detection using simple features despite partial signal overlap. The approach enables cost-effective tool protection and sustainable asset life-extension, supporting Zero-Defect Manufacturing in resource-constrained environments.
Near-net-shape (NNS) cold forging offers significant potential for material savings compared to processes like machining from solid billets. However, this process introduces residual stress (RS) fields that can cause distortions during subsequent machining. Predicting and controlling these distortions is essential for ensuring dimensional accuracy and minimising scrap rates in industrial applications. This study investigates the distortions caused by targeted milling operations on extruded aluminium cross-shaped specimens made from an EN AW-6082 aluminium alloy. The aim is the determination of the predictive capabilities of machining-induced distortions from RS liberation by commercial FEM software. Elastoplastic FEM simulations of the complete operation chain are developed using DEFORM 3D. Boolean operations combined with RS relaxation are used to simulate the milling operations in a simplified way. The numerical results are validated experimentally using systematic milling tests and software-aligned 3D scans of the specimens at all tested milling depths. The results are in good agreement, with simulations showing an average discrepancy of 22% with respect to experiments, good distortion mode prediction and the capture of observed trends in the influence of milling depth. However, a more accurate and repeatable alignment strategy is required for a higher confidence in the predictions.
This research paper examines the impact of post-review cooling on the fatigue resistance of medium carbon steel (AISI 1045 S45C) through the application of various heat treatments and quenching in water or oil at temperatures of 275 °C, 475 °C and 675 °C. It also compares the effects of the two quenching media on the mechanical properties and fatigue resistance of the steel. Fatigue tests were performed under steady stress conditions with a stress ratio of -1. The results of the experiments indicate that water quenching at a holding temperature of 275°C for one hour provides the greatest fatigue resistance for the steel. This is due to the formation of diluted martensite. Long fatigue cracks were accurately measured using a scanning electron microscope (SEM). Two models were selected to evaluate the fatigue life of AISI 1045 S45C medium carbon steel tempered and hardened at different cooling temperatures. The first model was selected from the crack growth rate equation (dE/dN), and the second from the stress intensity factor equation (ΔS).
Ti6Al4V is characterized by its superior strength and exceptional corrosion resistance. Due to its high strength-to-weight ratio, this alloy is extensively utilized in fabricating high-precision components within the aerospace industry, such as turbine blades and engine casings as well as in the medical sector for artificial joints and dental implants. This study explores the multi-objective optimization of the grinding process for this specific alloy using SiC abrasive wheels under Ester-based MQL conditions. The Taguchi method was implemented to develop the experimental matrix, focusing on four primary input parameters: fluid pressure, workpiece velocity, feed rate, and depth of cut. To evaluate the process performance, surface roughness and three cutting force components were analyzed. Five distinct ranking methodologies were integrated with four weighting techniques to determine the optimal processing conditions. Experimental analysis reveals that Ester oil in the MQL system effectively penetrates the cutting zone to mitigate friction, maintaining a minimum surface roughness of 0.828 \mu m and suppressing the material plowing effect associated with SiC wheels. The findings further confirm the process stability, as surface quality remained consistently controlled even at low normal forces (Fz = 7.39 N), allowing for greater flexibility in selecting cutting regimes. The final ranking results demonstrated high convergence across different MCDM methods and weighting techniques, validating the reliability and objectivity of the identified optimal solution.
This study investigates the microstructural and mechanical characteristics of a high-strength steel thin wall fabricated via wire arc additive manufacturing (WAAM). A 20-layer wall was produced using a bidirectional deposition strategy with fixed processing parameters to ensure repeatability. Microstructural analysis revealed significant spatial heterogeneity along the build direction. The bottom region exhibited the finest grain size (0.96±0.23 μm) due to the substrate heat sink effect, while the top region reveals the coarsest grain size (1.84±0.36 μm) due to slower cooling rates. XRD analysis confirmed the deposited material consists entirely of the ferrite phase. Mechanical testing showed a microhardness gradient ranging from 256±17.04 HV0.1 at the top to 288±16.78 HV0.1 at the bottom that directly correlates with grain refinement. Tensile tests revealed exceptional performance with ultimate tensile strength (UTS) exceeding 980 MPa, yield strength (YS0.2%) in range of 560–620 MPa, and elongation above 20%, meeting industrial requirements for structural applications. Fracture surface morphology confirmed a ductile micro-void coalescence mechanism, indicating high plastic deformation capability. These results demonstrate the WAAM capability to produce high-strength, structurally reliable steel components, supporting its application in large-scale manufacturing for shipbuilding, heavy engineering, and load-bearing structural systems.
This study investigates the use of Acoustic Emission (AE) for monitoring cutting tool wear during turning. Experiments were carried out using a Vallen AMSY-6 system and a VS370-A2 sensor. The work focuses on optimizing AE system settings, including filter bandwidth, detection threshold, and impulsive signal acquisition under varying cutting conditions. A systematic procedure for configuring the AE measurement process was established. Results show that employing a 20 kHz–800 kHz band-pass filter and a detection threshold of ~90 dB effectively isolates AE events associated with cutting and reveals a pronounced increase in AE activity when the tool is worn. These findings confirm AE as a sensitive technique for detecting cutting edge wear and underline its potential for real-time tool condition monitoring.
Retrieval-Augmented Generation (RAG) is widely used for manufacturing assistance, but its effectiveness depends on selecting retrievable text units. We test whether humans or Large Language Models (LLMs) can judge which case descriptions are better suited as RAG inputs. We constructed 100 synthetic manufacturing service cases, each paired with a realistic query and two comparable problem–solution variants differing in contextual completeness, granularity, and quality. Five engineers and five LLMs chose the variant expected to be more retrievable and useful. As a reference, both variants were indexed in a minimal retrieval setup with one chunk per case and evaluated with MRR@3, treating the case-matching chunk as the only relevant item among distractors. LLMs showed much higher within-group agreement than humans, yet neither cohort consistently matched retrieval-derived winners. Ties were frequent; on non-tied cases, majority decisions fell below chance and were significantly worse than random guessing in one embedding setting, while no individual rater achieved above-chance performance. Overall, the findings indicate that perceived RAG-fitness is not a reliable proxy for retrieval performance and should be grounded in retrieval-based evaluation under the target deployment setup.
Three-dimensional (3D) printing via fused deposition modeling (FDM) enables the fabrication of complex bio-inspired structures with high design flexibility; however, achieving an optimal balance among mechanical performance indicators remains challenging. This study proposes a comparative multi-criteria decision-making (MCDM) framework for optimizing the mechanical performance of FDM-fabricated bio-inspired structures, including Gyroid (G), I-graph-wrapped package (IWP), Fischer–Koch–S (FKS), and Primitive–Gyroid–Modified–Y (PMY). Eight specimens manufactured using polylactic acid (PLA) and PLA reinforced with short carbon fibers (PLA–CF) were experimentally evaluated under compression to determine mass, deformation, load-bearing capacity, and elastic modulus. Objective criterion weights were determined using the MEREC method, and three MCDM techniques, SAW, TOPSIS, and EAMR were employed for performance ranking and comparison. The results consistently identify PMY printed with PLA–CF as the optimal design across all methods, demonstrating the robustness and reliability of the proposed framework. Overall, the optimal structure exhibits a favorable combination of low mass, high load capacity, minimal deformation, and superior elastic modulus. This study provides a systematic and experimentally validated approach for multi-objective mechanical optimization of bio-inspired structures in additive manufacturing.
Aerospace industry requires the production of complex, multi-functional components with very low geometrical tolerances using difficult to machine materials such as Ti-6Al-4V. One of the main challenges when machining such materials is the resulting part distortion after milling due to the release of residual stresses generated in previous manufacturing steps such as forging or heat treatment. To prevent costly scrap parts and manual rework, finite element method (FEM) can be utilized to predict resulting part distortion during the machining process development phase. This paper presents a novel approach for part distortion simulation by directly integrating relevant machining data from computer aided manufacturing (CAM) system for the FEM using Boolean subtraction operations. A developed software interface enables a step-by-step mechanical material removal simulation providing manufacturers with an efficient and flexible tool for automated process evaluation. The approach is implemented in a laboratory setup using commercial CAM and FEM systems and evaluated by using aerospace relevant demonstration parts.
This study offers an updated overview of existing approaches used to model geometric deviations in additive manufacturing. It also proposes a new framework for representing part deviations by discretizing an ideal planar surface and accounting for both deterministic and stochastic sources of variation. Deterministic deviations are described through two main components: surface waviness and overall orientation. In contrast, stochastic deviations are introduced through automatically generated variations following a normal probability distribution. The second section of this work presents a numerical investigation of a prismatic component featuring a functional planar surface produced using the Fused Deposition Modeling (FDM) technique. As an initial step, a reference specimen is fabricated to verify essential parameters associated with the mathematical models used for the FDM process. The geometric deviation model relies on converting the nominal planar surface into a mesh of nodes, after which a deformed surface—representing the actual manufactured geometry is generated using the deviations computed by the proposed approach. Finally, a Monte Carlo analysis is performed to examine how these geometric variations influence the evaluation of surface parallelism tolerance. To support interpretation of the results, a correlation is established between the simulated deviations, the specified tolerance limits, and the resulting non conformity rate calculated for each tolerance range.
Arc-based additive manufacturing (DED-Arc) is a widespread, well established, near-net-shape manufacturing process. However, achieving the desired geometry is still a challenging task. Interlayer machining is a common method to overcome such challenges. Therefore, multi-axis kinematics in the form of a pentapod are predestined because of their high stiffness and accuracy. Extensive research has concluded that monitoring the manufacturing process is crucial to increase the level of automation in manufacturing towards full automation and also to combine the additive and subtractive process in a more intelligent way. Laser scanners are suitable tools for this purpose. In this article, the authors present how a laser scanner can be integrated into a pentapod architecture and how its kinematic peculiarity has to be considered. At first, the resulting measurement error is shown if the implicit forced rotation of the laser scanner caused by the pentapod’s mechanical concept is neglected and how to consider the rotation using homogeneous coordinates. Finally, the achieved accuracy of the measurement system is evaluated.
Reliable quality control and fault diagnosis are essential for ensuring machine reliability and preventing unexpected failures. One of the critical machine components for which such a diagnosis enables failure-free, long-term exploitation is gearboxes. Conventional vibration-based monitoring often depends on expert interpretation of signal patterns and gear-mesh behaviour, which limits scalability and consistency. In this work, physics-informed machine-learning framework for binary gearbox health classification using engineered vibration features. Time and frequency domain descriptors capturing impulsiveness, gear-mesh spectral characteristics, and modulation effects were extracted from tri-axial acceleration signals. To account for direction-dependent dynamics, separate models were developed for left (RPM0) and right (RPM1) rotational conditions. We employ a unsupervised Isolation Forest trained exclusively on healthy data for anomaly detection, and a supervised Logistic Regression classifier trained on both healthy and faulty samples. Predefined decision thresholds were applied to ensure methodological transparency and minimize overfitting. Evaluation on independent test cases demonstrates that direction-specific modelling combined with physically interpretable features enables robust gearbox fault detection. The proposed framework provides a reproducible and industrially applicable strategy for automated condition monitoring. Such an approach will provide precise solutions for early fault detection, predictive maintenance scheduling, and real-time performance optimization of gearboxes and machinery systems.
The double-ballbar (DBB) is a cost-effective, easy-to-automate, and fast measurement device for the kinematic calibration of machine tools and robotic systems. To extend its bandwidth of applications, a DBB with a larger measuring range has been developed. This device has now been comprehensively examined for the first time with regard to its systematic and stochastic measurement uncertainty. This paper focuses on the DBB’s static behaviour – i. e. deformation due to its own weight – and on thermally induced length measurement deviations, arising from environmental conditions and self-heating due to electrical power losses. For the evaluation of a DBB with 150 mm stroke, an Abbe-compliant reference measurement arrangement with a laser interferometer was implemented in an air-conditioned measuring room. Based on the results, an approach for correcting the systematic temperature-dependent length measurement deviation was implemented and validated. The DBB can thus achieve a measurement uncertainty significantly below 10 µm over a 150 mm stroke under known environmental conditions. This enables its application in the kinematic calibration of machine tools.
Manufacturing data typically flows one way (CAD→CAPP→CAM→CNC), expanding as it approaches the shop floor through technologists’ know-how and parameter choices. Using the CAM-Connect module integrated with Mastercam, we automatically extract these CAM settings and export a structured process description (STEP/STEP-NC XML) for scalable analysis. This paper introduces a practical CAM-driven method to identify and quantify non-productive motions in 3-axis milling. We focus on vertical auxiliary motions: Z-axis approaches, withdrawals, retracts and safe-height traverses. Auxiliary distances and times are reconstructed by interpreting CAM option logic, including clearance modes and retract/feed-plane definitions. The method is applied to 242 industrial milling operations comprising 2912 toolpaths. After excluding special drilling cycles and extremely short operations, auxiliary motion averages 7.96% of total operation time. Moreover, 14.5% of operations exceed a 15% auxiliary-time share. While rapid and feed auxiliary distances are similar, feed-executed auxiliary segments generate about 96% of auxiliary-motion time. The findings provide actionable diagnostics tied to CAM decisions and support systematic auditing and improvement of machining processes.
The design of two-stage helical gearboxes inherently involves trade-offs between structural compactness and transmission efficiency. Traditional design methods often fail to capture these competing objectives simultaneously. This study presents a comprehensive multi-objective optimization approach using the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to identify optimal trade-offs between minimizing the cross-sectional area and maximizing the efficiency of a two-stage helical gearbox. Drawing from the strengths of evolutionary algorithms and integrating insights from prior literature on hybrid and decision-making-based optimization methods, the proposed model formulates the gearbox design problem with realistic constraints and evaluates the Pareto front of optimal solutions. The results demonstrate that NSGA-II provides a well-distributed set of non-dominated solutions, offering engineers greater flexibility in balancing performance and structural requirements. Comparative analysis with existing approaches highlights the effectiveness of the proposed method in simultaneously achieving compactness and energy efficiency in gearbox systems.
The use of robots in pick-and-place operations has been a well-known and widely used approach in production processes practically since the inception of robotic systems. Currently, however, increasingly stringent requirements are being placed on the control of robotic pick-and-place processes. This requires real-time robot control and the collection of objects of various classes in a random order and location within the robot's workspace. Collaboration between robots and vision systems monitoring objects appearing in front of the robot on a conveyor belt is becoming commonplace in industrial environments. This also necessitates the use of appropriate pick-and-place algorithms in robot control. This article presents an overview of component picking algorithms for robotic production lines. Key features of these algorithms are described. The effect of local object density for a selected object class on sorting performance at varying conveyor belt speeds was analysed. The study focused on the interaction between input flow characteristics and conveyor dynamics, assuming constant robot kinematic parameters and a fixed object placement point location. Furthermore, an approximate relationship describing the maximum allowable conveyor belt speed as a function of object density was derived, defining the stability limit of the SPT algorithm.
