Machining Research Laboratory

Metal fused filament fabrication (MFFF) has gained attention as a cost-effective and safe additive manufacturing route for metallic components. However, sintered MFFF parts often require post-process machining to achieve the surface finish and dimensional accuracy needed for functional applications. The layer-wise deposition inherent to the MFFF process introduces a unique anisotropic microstructure and defect profile that fundamentally alters material removal mechanisms. The machining behavior of MFFF metals remains insufficiently understood, particularly the influence of raster orientation on cutting forces and chip formation. This study investigates the dynamic material response and machinability of MFFF-produced 17–4 PH and 316 L stainless steels across raster orientations of 0°, ±45°, and 90°. A combined framework including quasi-static tensile testing, high-strain-rate Split Hopkinson Pressure Bar experiments, and orthogonal cutting tests were used to derive Johnson-Cook (J-C) parameters. To reduce reliance on resource-intensive dynamic testing, an inverse parameter identification approach was used to estimate J-C constants from quasi-static tests and orthogonal cutting experiments. The calibrated models were coupled with Oxley’s machining theory to predict cutting forces, which were validated against experimental measurements. SEM analysis linked material resistance and chip formation to internal discontinuities and interlayer bonding, establishing a foundation for predictive modeling of MFFF components.

Fused filament fabrication (FFF) has evolved from a technology for prototyping into a functional manufacturing method; however, its industrial viability is often hindered by challenges in process predictability and part fidelity, particularly during high-speed cornering motions. Traditional trajectory planning relies on linear segments (G1 commands), where sudden directional changes necessitate abrupt decelerations to avoid exceeding mechanical limits, can result in local over-extrusion, dimensional inaccuracies, and increased print times. To address these limitations, this study proposes a path smoothening approach using quintic Pythagorean-Hodograph (PH) curves to replace abrupt direction changes with smooth, curvature-continuous blends. Unlike established polynomial interpolations, PH curves allow for exact arc-length calculation, enabling precise synchronization between nozzle motion and material extrusion without the need for numerical approximation. The proposed method blends curvature-continuous corner smoothing with a hybrid feed rate scheduling strategy that keeps the commanded acceleration continuous while keeping the jerk bounded and maintaining uniform deposition in corner regions. The methodology was validated through experiments conducted on an industrial machine tool benchmark and a retrofitted, screw-driven FFF system, utilizing test trajectories with varying corner angles. Experimental results demonstrate that the PH-based blending strategy significantly reduces acceleration peaks at trajectory junctions compared to sharp corner trajectories. This capability allows the machine kinematics to maintain higher non-zero velocities through corners, while reducing mechanical vibrations and preventing the instantaneous acceleration and deceleration demands associated with unoptimized transitions. Capabilities of the PH curves allow to increase the quality and fidelity of the parts printed with fused filament fabrication method for end use applications.

One of the most significant faults that can occur during the fused filament fabrication (FFF) additive manufacturing process is a loss of power. After such an event there is often no straightforward way to resume the 3D print, and the incomplete part must be discarded. Existing solutions depend on a battery back-up or non-volatile memory but may be insufficient when the interruption occurs during an active print motion. In the present paper, a new method is proposed and tested which utilizes computer vision to determine the precise layer the failure occurred on, and the precise location of the nozzle at the time of loss of power. Print failure detection is carried out using various functions in the OpenCV library of computer vision processing tools running on an independent Raspberry Pi 5 microcomputer. This makes the system retrofittable to a large variety of machines and allows for a high degree of automation in the process. After determining the precise point of failure at the top layer, a new G-Code file is automatically generated that trims the original file and adds corrective actions that allow the print to be resumed seamlessly. Testing and experimentation showcased how the proposed system can robustly function across a variety of geometries, including straight lines, circles, curves, independent islands, and a complex puzzle piece. The puzzle piece was further tested to validate the system's functionality under various lighting scenarios, such as low-contrast conditions and high background noise. Overall, this paper demonstrates that full print resumption is possible using inexpensive, off-the-shelf components and software. The uncertainty of the system in detecting the failure point is bounded by the camera's spatial resolution, approximately 0.055 mm per pixel in the presented configuration, which confines the detected failure location to well within a single extrusion width.

Material extrusion additive manufacturing, particularly metal fused filament fabrication (MFFF), is advancing rapidly in industry. However, MFFF parts require post-process machining to meet quality standards. The layer-by-layer nature of MFFF introduces anisotropy, which affects machining performance depending on print orientation. This paper investigates the machining characteristics of MFFF 17-4PH and 316 L stainless steels. Quasi-static tensile tests, split Hopkinson pressure bar experiments, and finite element analysis were performed to determine the Johnson–Cook plasticity and damage parameters for force modeling.

This work investigates the combined effects of multi-walled carbon nanotubes (MWCNTs) weight%(wt%), aspect ratio (AR), and surfactant content on tool wear during MQL machining using nanofluids. While advancements have been made in the application of nanofluids in machining processes, the interplay between these parameters and their influence on both nanofluids stability and tool wear has not been systematically examined. The turning operation of AISI 304 austenitic stainless steel revealed that nanofluid combinations with lower wt% and AR of MWCNTs improved tool wear performance.

Metal additive manufacturing (AM) is an emerging technology for producing metallic parts, with metal fused filament fabrication (FFF) technique gaining attention due to its cost-effectiveness. In FFF, a filament composed of metal powder and polymeric binder is deposited layer by layer, followed by debinding and sintering to produce the final part. However, FFF parts often contain microstructural defects, with porosity being one of the most critical ones, as it significantly impacts material properties. Accurate porosity measurement is therefore essential for ensuring part quality.

The shift of additive manufacturing from pure prototyping technology to producing end-use components has raised a new challenge for mechanical design. Efficient design requires a method for predicting the strength of a fused filament fabricated (FFF) material to replace slow and costly trial and error methods. Previous studies have developed models for failure prediction in FFF materials but often only for in-plane stress analysis, not for full 3D. In the current paper, a modified Tsai-Wu (TW) criterion for full 3D failure prediction of FFF poly-ethylene terephthalate glycol (PETG) polymer was implemented both analytically and numerically.

Fused filament fabrication (FFF) is a promising additive manufacturing method; nevertheless, the mechanical properties of its final products, particularly for end-use applications, still require enhancements. Combining FFF's low cost and well-established technology with enhanced mechanical properties would increase its competitiveness among other additive manufacturing methods. Similar to the well-established subtractive manufacturing methods, the majority of the trajectory planning algorithms developed for FFF, prioritize print time and dimensional accuracy. However, the effect of trajectory planning on the strength of parts produced through FFF has not received adequate attention.