Structurally adaptive manufacturing systems require machine tools capable of modifying not only process parameters but also their intrinsic structural properties. This study presents a lathe spindle equipped with a hydraulically actuated adjustable bearing-preload system that enables controlled tuning of static and dynamic stiffness. The objective was to experimentally quantify the influence of preload variation on spindle mechanics and to evaluate its implications for turning performance. The spindle was first characterized on a dedicated test stand. Increasing preload produced a monotonic rise in static stiffness, with translational stiffness gains up to approximately 16 % over the range investigated. Experimental modal analysis revealed mode-dependent frequency shifts, with absolute increases of 2-12 Hz (up to 3.2 %), while mode shapes remained consistent, indicating a stiffness-driven eigenvalue shift rather than structural reconfiguration. Receptance measurements showed reduced resonance amplitudes and a preload-dependent damping increase, with the loss factor rising by up to 38 % for the most sensitive mode. After integration into a production lathe, cutting tests demonstrated that preload variation can alter machining stability under constant cutting parameters, including transitions between stable cutting and chatter in representative turning scenarios. No systematic dimensional offset in workpiece geometric features attributable to preload was observed within the applied accuracy assessment strategy. The results establish hydraulic preload control as a practical structural control variable for tuning spindle dynamics and influencing turning stability, supporting the development of structurally adaptive machine tools for advanced manufacturing applications.

This work presents an automated identification procedure for robot quasistatic compliance under external loads, together with a validation framework that accounts for uncertainty. Controlled tensile forces are applied through a cable routed to a single ground-fixed pivot while TCP deflection is measured with a laser tracker, enabling scalable and largely unattended acquisition of pose–load data for parameter identification. A kinematic compliance model is then identified with practical parameter reduction, and parameter as well as input uncertainties are propagated to TCP deflection prediction intervals via Monte Carlo simulation. Validation on a MABI MAX100 robot along circular trajectories shows overall agreement with the predicted uncertainty bands, while also revealing direction-dependent effects not captured by the current model.

The evaluation of machine tool characteristics and their impact on surface quality is challenging, often requiring disruptive traditional methods. This study introduces a novel, non-invasive approach using optical camera images for rapid and accurate assessment. Data robustness was ensured by acquiring initial images outside the machining chamber with consistent external illumination, focusing on detailed intensity profile analysis. Machined surfaces were processed using intensity profile extraction and Fast Fourier Transform (FFT). The dominant spatial wavelength (0.1833 mm) consistently showed excellent agreement (within 1.85%) with the theoretical feed per revolution (0.1800 mm). This robustly validates the method's ability to precisely capture primary kinematic tool marks. Temporal information, inferred from spatial frequencies, underwent subsequent FFT to identify periodic phenomena and harmonics. The comprehensive spatial and temporal FFT analyses offer detailed, quantitative surface characterizations. The clear distinctions in temporal harmonic patterns provide robust, frequency-domain signatures informing machining system performance and process integrity.

The introduction of new products necessitates a shift in manufacturing systems from high volume/low diversity to high volume/high diversity production. A fundamental element of the manufacturing system is the machine tool which has traditionally been acquired for specific well-defined operation. However, evolving requirements demand that machine tools be capable of producing a wider range of products particularly with multi-axis machining capabilities. The main focus of existing machine tool test methods is to assess and verify the geometric and kinematic accuracy of the machine, ensuring that positioning, motion, and spindle performance meet specified tolerances. Error separation techniques are often employed within these methods to distinguish and quantify different sources of error, such as those arising from the machine, the measurement system, or the test piece. This research presents a novel systematic methodology for evaluating machining systems through machining tests using a modular scalable test piece designed from industry requirements. The methodology encompasses a process based on industry-specific requirements where critical features from actual components are translated into a test piece design that replicates real-world machining challenges while maintaining measurability and comparability. The approach was examined through experimental machining of test pieces in Spheroidal Graphite Iron under two process conditions balanced quality/cycle time versus minimum cycle time using a five-axis machining centre. Results demonstrated significant differences in geometric accuracy particularly in datum establishment and positional features with aggressive cutting parameters showing substantially larger deviations. This study provides a practical methodology for assessing machine tool adaptability and capability in high-diversity production environments contributing to improved asset management and manufacturing system design.

Vibration-based condition monitoring (VCM) has shown good performance in laboratory testbeds operating under stable conditions. In industrial production machinery, however, variable operating speeds and mechanically coupled subsystems lead to strong non-stationary vibration signals. Under such conditions, purely data-driven models often fail to generalize, struggling to distinguish mechanical fault signatures from speed-induced signal variations, resulting in degraded diagnostic robustness. To address this challenge, the present study proposes a hybrid diagnostic framework that integrates domain knowledge with data-driven deep learning. Adopting a knowledge-integrated methodology, features grounded in physical principles are embedded into the learning pipeline to provide invariant references that stabilize diagnostic performance under variable operating speeds. The proposed framework is evaluated against a purely data-driven baseline network (1D-CNN-BiLSTM) using vibration data collected from a pharmaceutical carton labelling machine. Results demonstrate that the hybrid model enhances diagnostic robustness by significantly reducing performance variance under stationary conditions and improving mean accuracy under variable-speed scenarios. Under high-speed intra-speed operation, the standard deviation of classification accuracy is reduced by a factor of four (from 4.7% to 1.1%), while approximately a 5% improvement in mean accuracy is achieved under the multi-speed validation protocol. These findings demonstrate that integrating physically grounded domain knowledge into data-driven pipelines can signiLicantly enhance diagnostic robustness in complex production machinery.

This study explores three methods for controlling a multi-arm robotic digital twin in an mixed reality (MR) environment on the Meta Quest 3: direct leader-follower operation, basic voice input interpreted by a large language model (LLM), and a conversational voice interface capable of understanding complex natural language commands. The system allows real-time robot control by transcribing speech and interpreting instructions through the LLM, with support for multilingual input. It includes a failure recovery mechanism that allows users to retry unsuccessful actions without repeating the original commands. A user study involving a precision pouring task in laboratory evaluated each method in terms of task completion time, success rate, reliability, and perceived workload as measured by the NASA-TLX questionnaire. While leader-follower control offered the fastest completion times, it also led to higher error rates and less consistent performance, especially among inexperienced users. By comparison, both voice-based interfaces offered superior precision, more consistent performance, and a significantly lower cognitive burden. Notably, the conversational interface with confirmation received the highest workload ratings, pointing to reduced cognitive demand and enhanced ease of use. These results indicate that LLM-driven natural language interfaces can significantly enhance the precision, performance, and reliability of AR-based robotic control, particularly in scenarios requiring high usability and consistency across users.

Accurate compensation of force-induced machining errors in subtractive processes such as milling, grinding, and polishing, whether performed on machine tools or articulated robots, depends on reliable prediction of tool deflection from machine stiffness and cutting forces. However, both inputs are constrained by measurement uncertainties. This paper introduces a measurement-oriented framework that applies uncertainty analysis to evaluate the reliability of tool deflection prediction. Specifically, an acceptance criterion for measurement process capability is proposed, enabling evaluation of whether current measurement systems can meet tolerance requirements in machining. A parametric analysis is carried out across representative ranges of cutting forces (50-800 N), machine stiffness values (2-30 N/μm), and standard tolerance grades. The results identify conditions under which existing approaches to stiffness and cutting force measurement are adequate, as well as scenarios – particularly for low-stiffness machines and tight-tolerance applications – where they fail due to excessive uncertainty. Cutting force measurement, supported by established calibration standards, is found sufficiently mature for most medium- and high-stiffness applications. In contrast, stiffness measurement lacks a universally accepted measurand, and interlaboratory validation, limiting its reliability for compensation tasks. The findings highlight that advancing stiffness metrology is critical to enabling practical and trustworthy compensation of force-induced machining errors in both machine tools and robots.

Machined surfaces contain rich information about machining conditions and system behavior and are typically assessed using off-line, small-area metrology. This study developed and validated an image-based methodology for process-oriented surface texture analysis of end-milled Spheroidal Graphite Iron (SGI), enabling scalable, non-contact monitoring suitable for in-line deployment. End milling trials were conducted under optimized and aggressive cutting conditions and in two orthogonal feed directions (X,Y). Surface topography from White Light Interferometry (WLI) was complemented by Charge-Coupled Device (CCD) microscope imaging. Image processing comprised automatic orientation correction, intensity profile extraction, and frequency-domain analysis via Fast Fourier Transform and power spectral density estimation. Texture metrics (RMS amplitude, skewness, kurtosis, dominant wavelength) were derived from intensity profiles, and two spectral indices were introduced: a Change Index (CI), capturing high-frequency content linked to process disturbances, and a Surface Anisotropy Metric (SAM), quantifying texture directionality. Aggressive cutting increased RMS by 28.5% and shifted skewness by 274% with strong statistical significance. Directional analysis showed 22% higher texture amplitude in Y than X, indicating axis-dependent machine behavior. CI correlated with the machining parameters and stability, while SAM reflected the machine and setup characteristics. Trends were consistent with WLI, supporting the method as a rapid, complementary tool for surface quality and machine condition monitoring.

Data-driven systems have demonstrated significant value in real-time surface roughness evaluation and pre- diction for machining processes. This study presents a comprehensive methodology for evaluating three core elements critical to decision-making in machining monitoring: sensor-based dynamic signal selection and pro- cessing, sensory fusion scenarios, and Machine Learning (ML) models. Unlike previous research that often an- alyzes these elements in isolation, this study emphasizes their collective impact on unified datasets. The proposed methodology is validated using experimental data from machining trials on two distinct machines during the finishing process of hard turning. Hard turning is a widely employed finishing operation in manufacturing that directly affects dimensional accuracy, surface integrity, and surface finish, which are key characteristics of machined parts. Surface roughness, as an essential indicator of surface quality, plays a pivotal role in the functional performance of end products, necessitating accurate monitoring and assessment throughout the process. To optimize performance, Bayesian Optimization was employed for automatic hyperparameter tuning, facilitating efficient exploration of optimal parameters. The predictive capabilities of trained ML models were subsequently evaluated using performance metrics and quantitatively assessed for uncertainty through Predic- tion Intervals (PIs), calculated via non-parametric Kernel Density Estimation. Results indicate that the Wavelet Packet Transform method significantly enhances the predictive performance across all ML models. Among the evaluated models, Support Vector Regression and K-Nearest Neighbors demonstrated superior predictive accu- racy and minimal uncertainty across all signal processing methods. This work provides actionable guidelines for researchers and manufacturers in selecting optimal combinations of methodologies for developing accurate, reliable, and generalizable data-driven surface quality monitoring and prediction systems, particularly for hard part turning applications.

Advanced manufacturing systems demand the utilization of technologies, methods and capabilities to improve production efficiency or productivity, while ensuring environmental and societal sustainability. Digitalization emerges as an alternative solution for improving the monitoring capabilities of manufacturing systems and consequently enhance the decision-making process. However, the widespread adoption of digital solutions introduces complexities in measurement reliability, data management, and environmental concerns in terms of e-waste and data storing. Therefore, enhancing monitoring capabilities while minimizing resource consumption is crucial for ensuring system reliability in a sustainable way. This research introduces a methodology for assessing the monitoring condition of manufacturing systems. By integrating functional and dysfunctional analysis, approaches that focus on identifying critical functions and potential failure modes of a system, the proposed methodology provides a comprehensive system perspective and targeted directives for improvement. The effectiveness and versatility of the methodology are demonstrated and discussed through its application to various manufacturing systems at a component, machine, and line level.