Cirp Annals-Manufacturing Technology最新文献

This paper aims to enhance the machining characteristics of sinking electrical discharge machining by superimposing an impulse current on a conventional rectangular pulse. The material removal was found to be more significant when the time point of superposition was set earlier after the dielectric breakdown, as the plasma diameter is smaller. However, superimposition immediately after the discharge ignition leads to a higher tool wear ratio. Through experimentation and simulation, it was determined that the optimal time point to increase material removal while keeping tool wear low is approximately 12.5 μs when the rectangular pulse duration was 50 μs.

Nowadays, the growing complexities of manufacturing processes and systems make it difficult to identify the root causes of critical deviations in performance. Conventional methods often fall short in capturing the multifaceted nature of these challenges, despite a wealth of diverse untapped manufacturing data. To harness the full potential of diverse data sets and transform them into a valuable asset to guide root cause exploration, this paper presents an innovative approach that combines multimodal predictive analysis and explainable artificial intelligence (XAI) to uncover insights into system dynamics. This work contributes to a paradigm shift in industrial decision-making regarding manufacturing diagnostics.

Feedstock from locally-sourced, recycled Ti-6Al-4V (Ti64) swarf is a low-cost alternative to atomized powder and encourages circular additive manufacturing. This study investigates the feasibility of Ti64 swarf as feedstock for laser directed energy deposition (L-DED) additive manufacturing. Ti64 swarf was recycled and ball-milled into irregular-shaped powder and compared to spherical plasma-atomized powder in powder flow, melt flow, and resulting microstructure. In situ monitoring showed that plasma-atomized powder had laminar flow during deposition and that ball-milled swarf powder had turbulent flow. Plasma-atomized powder caused steady melt pool dynamics and acicular microstructure. Ball-milled swarf powder caused melt pool fluctuation and equiaxed microstructure.

Flexible and rapidly adaptable automated production processes, e.g. with collaborative lightweight robots, are key aspects for the competitiveness of companies in globalised markets. However, software adaptation of the robot still requires specific programming skills. We have developed a human-centred programming by demonstration approach based on augmented reality to enable operators to intuitively adapt the robot programme. The developed hand-interaction model overcomes the challenge of object tracking during the assembly demonstration phase. This allows quick programme modifications by the operator using the head-mounted augmented reality device. The human-in-the-loop concept ensures a highly reliable and robust programming process.

Multiplexed Fused Filament Fabrication (MF3) combines a dynamic extrusion-based toolpath with a multi-extruder-single-gantry machine tool architecture to increase printing speed without sacrificing geometric capabilities or increasing hardware complexity. This work establishes a novel capability for geometrically generalizable synthesis of MF3 toolpaths and creates a new thermal model that incorporates the unique nature of material deposition. It is shown that MF3 can increase throughput by an order of magnitude or more via the atypical and nonlinear effects of part size, infill density and extruder spacing. Further, the unusual temperature history in MF3 is found to enhance the part's mechanical properties.

High-precision handling processes are essential for various high-tech industries and are typically realized using specialized high precision robots. Articulated robots are rarely used for such tasks due to their low stiffness and accuracy stemming from their kinematic structure. This work presents a methodology for designing handling processes that maximise repeatability with articulated robots. By considering their kinematic structures’ highly pose dependent properties as well as sensitivity to external disturbances in every step of the processes design, significantly improved repeatability can be achieved. The applicability of the new methodology is verified experimentally using the example of handling of large silicon dies.

While additive manufacturing (AM) provides design flexibility, challenges persist in handling intricate freeform shapes, especially those laden with fine details. Conventional AM processes, such as slicing stereolithography (STL) format models, generating line segment toolpaths, and polyline-based printing, prove costly and compromise accuracy. This paper proposes a solution: the spline scanning generative design method. Utilizing spline patterns to construct smooth toolpaths directly, it enables seamless curved printing, significantly reducing computational expenses while maintaining high accuracy through spline control points. Experimental implementation, supported by dedicated algorithms, attests to its efficacy, emphasizing its potential for intricate freeform structure design and printing.

To increase the accuracy of indirect position-controlled dual motor rack-and-pinion drives, a new approach is being proposed that uses the encoders of both motors on a test bench. By electrically preloading these motors, backlash does not occur simultaneously in both drive trains. Continuous contact between the pinion and rack and thus force-transmission to the table is ensured, by switching away from the signal with backlash. Backlash is therefore eliminated from the control loop and the system accuracy without direct table position sensing can be increased. Experiments show that the tracking error is reduced by 59 % compared to indirect control.

Accurate prediction of tool tip dynamics is vital for understanding machine tool behavior and chatter. Traditional methods involve several impact tests, finite element simulations, and the receptance coupling (RC) approach. However, substructure coupling necessitates multiple experiments and encounters difficulties due to complexities of capturing rotational dynamics. The intricate nature of RC inhibits its widespread industrial applicability in predicting tool tip dynamics. We introduce machine learning (ML)-based approach relying on a few experiments and computer vision to predict dynamics. Comparative analysis with direct experiments shows the ML-based method's potential to expedite dynamic identification with accuracy, chatter prediction, and machining process optimization.

Grey cast iron is known for its poor machinability directly after casting, but attains excellent machining performance after ageing. The present work explores the impact of cutting speed on the performance of pcBN machining for non-aged material. Findings suggest that tool wear can be minimized by identifying an optimal cutting speed that supports the formation of a stable Al₂O₃ and MnS build-up layer (BUL). Insufficient BUL protection accelerates pcBN wear by diffusion, while at very high speeds protective Al₂O₃ is replaced by weaker (Fe,Mn)₂SiO₄ and (Fe,Mn)O, and oxidation accelerates tool wear. The higher mechanical properties of aged GCI facilitate generation of high enough temperatures for stable deposition of Al₂O₃ BUL.