Existing algorithms for predicting milling chatter have not been widely adopted in industry since they require specialized instruments to measure the stability inputs. This study describes how the machining process for a meter-scale aluminum aerostructure was optimized using a physics-guided Bayesian stability model. The study was performed in collaboration with an industrial partner on production machines to evaluate the practicality of the proposed method under real-world conditions. For each cutting tool, the Bayesian approach automatically selected a small number of cutting tests, which were monitored using a microphone to observe the chatter frequency. The algorithm learned the system dynamics, cutting forces, and stability map from these test results. A novel algorithm for predicting tool bending stress was incorporated into the test selection algorithm to avoid tool breakage. On average, each set of optimized cutting parameters required less than six tests to identify and were 97% more productive than baseline parameters from the cutting tool manufacturer. The machining program was then further optimized using commercial feedrate scheduling software to remove cutting force spikes and reduce air cutting time. Five components were machined using the optimized process. These results demonstrate the potential for physics-guided Bayesian models to improve productivity in industrial settings.

A common aerospace and defense industry challenge is low volume production of components for legacy aircraft due to compromised casting and forging supply chains. A hybrid manufacturing approach is presented to address this challenge that uses additive friction stir deposition, structured light scanning, and CNC milling. The paper describes a novel slicing and toolpath development strategy for additive friction stir deposition of a relevant aerospace geometry, post deposition measurement of the two-sided preform and identification of the machining work coordinate system, and five-axis CNC machining to obtain the final part geometry while ensuring stable machining behavior.

Process damping can provide improved machining productivity by increasing the stability limit at low spindle speeds. While the phenomenon is well known, experimental identification of process damping model parameters can limit pre-process parameter selection that leverages the potential increases in material removal rates. This paper proposes a physics-informed Bayesian method that can identify the cutting force and process damping model coefficients from a limited set of test cuts without requiring direct measurements of cutting force or vibration. The method uses time-domain simulation to incorporate process damping and provide a basis for test selection. New strategies for efficient sampling and dimensionality reduction are applied to lower computation time and minimize the effect of model error. The proposed method is demonstrated, and the identified cutting and damping force coefficients are compared to values obtained using machining tests and least-squares fitting.

Milling is a key manufacturing process that requires the selection of operating parameters that provide efficient performance. However, the presence of chatter, a self-excited vibration causing poor surface finish and potential damage to the machine and cutting tool, makes it challenging to select the appropriate parameters. To predict chatter, stability maps are commonly used, but their generation requires expensive data, making it difficult to employ these maps in industry. Therefore, there is a pressing need for an approach that can accurately predict stability maps using limited experimental data. This study introduces the new Encoder GAN (EGAN) approach based on Generative Adversarial Networks (GANs) that predicts stability maps using limited experimental data. The approach consists of the encoder, generator, and discriminator subnetworks and uses the trained encoder and generator to predict the target stability map. This versatile method can be applied to various tool setups and can accurately predict stability maps with limited experimental data (five to 10 cutting tests) even when there is little information available for unknown parameters. The study evaluates the proposed approach using both numerical data and experiments and demonstrates its superior performance compared to state-of-the-art benchmarks.

Accurately simulating machining operations requires knowledge of the cutting force model and system frequency response. However, this data is collected using specialized instruments in an ex-situ manner. Bayesian statistical methods instead learn the system parameters using cutting test data, but to date, these approaches have only considered milling stability. This paper presents a physics-based Bayesian framework which incorporates both spindle power and milling stability. Initial probabilistic descriptions of the system parameters are propagated through a set of physics functions to form probabilistic predictions about the milling process. The system parameters are then updated using automatically selected cutting tests to reduce parameter uncertainty and identify more productive cutting conditions, where spindle power measurements are used to learn the cutting force model. The framework is demonstrated through both numerical and experimental case studies. Results show that the approach accurately identifies both the system natural frequency and cutting force model.

An integrated, multi-system research platform has been developed to explore the fabrication of large-scale metal components though hybrid manufacturing. The resulting research cell is highly flexible as it incorporates multiple robot stations, a multi-axis part transfer system and a five-axis CNC machine tool. Current process capabilities include multi-material wire-arc additive manufacturing, fringe projection scanning metrology, robotic part handling and finish machining. A geometric digital twin is used to establish and transfer part positions, datums and coordinate axes across these processes. Mechanical and electrical system integration is complete and a sequential process flow has been demonstrated by fabricating a monolithic, single wall part geometry.

The objective of the wire arc additive manufacturing (WAAM) hybrid cell is to significantly reduce lead time associated with large scale parts. The WAAM process utilizes a 6 degree-of-freedom (DOF) robot manipulator in addition to a 2 DOF part positioner. After printing, the part is translated into position for scanning to inform the subtractive manufacturing process. Scanning in a manual setting can be a long and arduous process to generate scans of sufficient quality for the basis of informed machining. Effective path planning for scanning with the intent to reduce scanning time, increase quality of scans, and move toward a full automation of the hybrid manufacturing cell is investigated in this paper. Simulation software is used to create and verify path plans prior to importing and implementing on a 6 DOF robotic manipulator to which a GOM ATOS Q 3D scanner is mounted. The quality, amount of time necessary to produce, and number of scans required to produce a sufficient representation for machining are compared using three methods. The first method is a manual scanning configuration, the second is using a generalized path plan, and the third is using a geometry-based path plan.

This paper describes a hybrid manufacturing approach for silicon carbide (SiC) freeform surfaces using binder jet additive manufacturing (BJAM) to print the preform and machining to obtain the design geometry. Although additive manufacturing (AM) techniques such as BJAM allow for the fabrication of complex geometries, additional machining or grinding is often required to achieve the desired surface finish and shape. Hybrid manufacturing has been shown to provide an effective solution. However, hybrid manufacturing also has its own challenges, depending on the combination of processes. For example, when the subtractive and additive manufacturing steps are performed sequentially on separate systems, it is necessary to define a common coordinate system for part transfer. This can be difficult because AM preforms do not inherently contain features that can serve as datums. Additionally, it is important to confirm that the intended final geometry is contained within the AM preform. The approach described here addresses these challenges by using structured light scanning to create a stock model for machining. Results show that a freeform surface was machined with approximately 70 µm of maximum deviation from that which was planned.

Structured light scanning is used to create a digital twin of a manufactured part, where features are extracted to determine if the part meets the designer's intent and required tolerances. This paper describes repeatability and reproducibility analyses for a commercially-available structured light scanning system and measurement artifact. The repeatability study used five repeated scans at 15 measurement positions. Repeatability was assessed by randomly selecting one of the five scans at each of the 15 positions and creating a part mesh. This process was performed 50 times and the statistics for the dimension variations were calculated to isolate the scanning effects only. The same sequence was then performed for 10 of the 15 positions and five of the 15 positions to evaluate the repeatability sensitivity to the number of measurement positions. Reproducibility was assessed by selecting 15 positions to create a mesh and repeating the 15-position measurement sequence 10 times using different positions for each mesh construction. The statistics for the dimension variations were then calculated. This incorporated the effects of both scanning and the position and orientation of the part relative to the scanner. This sequence was repeated for 10-position and five-position scans to evaluate the corresponding sensitivity. Finally, the artifact dimensions from structured light scanning were compared to coordinate measuring machine measurements of the same features.

Hybrid manufacturing consisting of metal additively manufactured preforms and computer numerical control (CNC) machining has been established to be an effective method for high material use rates. However, hybrid manufacturing introduces unique challenges. Near-net shape designs are typically selected, which result in a smaller margin for part placement within the stock and stringent requirements for work coordinate system identification. Additionally, less stock material reduces the preform stiffness, which limits the material removal rates during machining. This paper demonstrates a digital twin for CNC machining of a wire arc additively manufactured preform that implements: 1) structured light scanning for stock model identification and tool path generation; 2) a fused filament fabrication apparatus to attach temporary fiducials and scan targets to the preform that enable coordinate system definition for both the CAM and CNC machine; 3) preform and tool tip frequency response function measurements to enable stable milling parameter selection; and 4) post-manufacturing measurements of geometry, surface finish, and structural dynamics to confirm designer intent. These efforts define key components of the machining digital twin for hybrid manufacturing.

This paper describes a milling stability identification approach that simultaneously considers: physics-based models for the tool tip frequency response functions and stability predictions; the binary result from a milling test (automatically labeled as stable or unstable based on frequency content); chatter frequency when an unstable result is obtained; and user risk tolerance. The algorithm applies probabilistic Bayesian machine learning with adaptive, parallelized Markov Chain Monte Carlo sampling to update the probability of stability with each milling test. The result is a robust solution for rapid convergence to optimized milling parameters for maximum metal removal rate using all available information.

This paper describes the workforce development activities supported by America’s Cutting Edge (ACE), a national initiative for machine tool technology development and advancement. ACE is supported by the Department of Defense Industrial Base Analysis and Sustainment (IBAS) program from the Office of Industrial Policy. Both the online and in-person components of the computer numerically controlled (CNC) machining and metrology training programs are summarized and participation information is provided.

This paper describes coordinate system definition and transfer for five-axis machining of additively-manufactured preforms. In this method, a set of fiducials are attached to the temporarily attached to the part, and their location relative to the preform geometry is calibrated using a structured light scanner. Those fiducials can then be measured in the machine tool to determine the location and orientation of the part. The method is demonstrated by finish-machining a carbon fiber layup mold from an additively manufactured Invar preform. In addition to showing the coordinate transfer methods necessary to machine the part, several key challenges with machining additively-manufactured preforms are discussed and potential solutions are proposed. Unfortunately, the final part was ultimately unusable due to porosity inside the part left from the additive process. Future work will remanufacture this part while taking steps to avoid porosity and other challenges encountered.

In this paper, it is shown that hybrid manufacturing can produce a freeform surface on silicon carbide (SiC) preforms printed using binder jet additive manufacturing (BJAM). While additive manufacturing methods, such as BJAM, can fabricate complex geometries, machining or grinding is still required to achieve the desired surface finish and geometry. Hybrid manufacturing has been proven to be an efficient method to address these issues. However, hybrid manufacturing faces its own issues dependent on the combination of processes. When the subtractive and additive manufacturing steps are completed in two separate systems, for example, a common coordinate system must be defined for part transfer. This is challenging, because AM preforms do not inherenly contain features that can be used for accurate part location. Additionally, it must be confirmed that the intended final geometry is contained within the AM preform. This paper addresses issues for AM + machining.

The design and construction of a hybrid manufacturing work cell is described. The intent is to enable large-scale additive metals manufacturing by combining robotic wire arc additive manufacturing (WAAM), five-axis machining, supporting metrology, and part transfer. Currently, large metal parts produced using WAAM do not generally offer the required surface finish necessary or geometric accuracy for industrial use. To provide the required dimensional accuracy and finish, the additive preforms are typically machined. Machining WAAM preforms presents several challenges, including final part containment within the preform, datum identification (if available), and coordinate system transfer from WAAM to the machining center. The work cell addresses hybrid manufacturing challenges by linking the additive and machining processes through material handling, metrology, and supervisory system control and monitoring.

This paper describes a method for establishing and transferring coordinate systems through multiple hybrid manufacturing operations. To demonstrate the approach, an additively manufactured preform is finish machined to produce the desired part geometry. A set of external fiducials is temporarily attached to the preform using a polymer frame. The assembly is inspected using a structured light scanner and the resulting scan is used to define an alignment and coordinate system which respects the physical requirements of the manufacturing processes. The coordinate system is then used to program subsequent machining operations. Once the part is set up on the milling machine, the fiducials are used to establish the shared coordinate system for the machining operation using standard on-machine probing. After the part is machined, the same fiducial/scanning process is repeated for a second machining operation to complete the part (i.e., some features could not be accessed in the first setup). Finally, the method performance is assessed.

This paper describes a physics-guided Bayesian framework for identifying the milling stability boundary and system parameters through iterative testing. Prior uncertainties for the parameters are identified through physical simulation and literature reviews, without physical testing of the actual milling system. Those uncertainties are then propagated to the stability map using a physics-based stability model, which is used to suggest a test point. The uncertainties are updated based on the new information acquired from the cutting test to form a new probability distribution, called the posterior. Finally, the posterior are compared to measured values for the stability boundary and system parameters to evaluate the approach. Based on experimental observations, the advantages and disadvantages of using a physics-guided model are discussed.