An official website of the United States government
We developed and applied a model-based feedforward control approach to reduce temperature-induced flaw formation in the laser powder bed fusion (LPBF) additive manufacturing process. The feedforward control is built upon three basic steps. First, the thermal history of the part is rapidly predicted using a mesh-free graph theory model. Second, thermal history metrics are extracted from the model to identify regions of heat buildup, symptomatic of flaw formation. Third, process parameters are changed layer-by-layer based on insights from the thermal model. This technique was validated with two identical build plates (Inconel 718). Parts on the first build plate were made under manufacturer recommended nominal process parameters. Parts on the second build plate were made with model optimized process parameters. Results were validated with in-situ infrared thermography, and materials characterization techniques. Parts produced under controlled processing exhibited superior geometric accuracy and resolution, finer grain size, and increased microhardness.
more »
« less
Monitoring and control of biological additive manufacturing using machine learning
The goal of this work is the flaw-free, industrial-scale production of biological additive manufacturing of tissue constructs (Bio-AM). In pursuit of this goal, the objectives of this work in the context of extrusion-based Bio-AM of bone tissue constructs are twofold: (1) detect flaw formation using data from in-situ infrared thermocouple sensors; and (2) prevent flaw formation through preemptive process control. In realizing the first objective, data signatures acquired from in-situ sensors were analyzed using several machine learning approaches to ascertain critical quality metrics, such as print regime, strand width, strand height, and strand fusion severity. These quality metrics are intended to capture the process state at the basic 1D strand-level to the 2D layer-level. For this purpose, machine learning models were trained to classify and predict flaw formation. These models predicted print quality features with accuracy nearing 90%. In connection with the second objective, the previously trained machine learning models were used to preempt flaw formation by changing the process parameters (print velocity) during deposition—a form of feedforward control. With the feedforward process control, strand width heterogeneity was statistically significantly reduced, reducing the strand width difference between strand halves to less than 50 µm. Using this integrated process monitoring, detection, and control approach, we demonstrate consistent, repeatable production of Bio-AM constructs.
more »
« less
- Award ID(s):
- 1739696
- PAR ID:
- 10401515
- Date Published:
- Journal Name:
- Journal of Intelligent Manufacturing
- ISSN:
- 0956-5515
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The goal of this work to mitigate flaws in metal parts produced from laser powder bed fusion (LPBF) additive manufacturing (AM) process. As a step towards this goal, the objective of this work is to predict the build quality of a part as it is being printed via deep learning of in-situ layer-wise images obtained from an optical camera instrumented in the LPBF machine. To realize this objective, we designed a set of thin-wall features (fins) from Titanium alloy (Ti-6Al-4V) material with varying length-to-thickness ratio. These thin-wall test parts were printed under three different build orientations and in-situ images of their top surface were acquired during the process. The parts were examined offline using X-ray computed tomography (XCT), and their build quality was quantified in terms of statistical features, such as the thickness and consistency of its edges. Subsequently, a deep learning convolutional neural network (CNN) was trained to predict the XCT-derived statistical quality features using the layer-wise optical images of the thin-wall part as inputs. The statistical correlation between CNN-based predictions and XCT-observed quality measurements exceeds 85%. This work has two outcomes consequential to the sustainability of additive manufacturing: (1) It provides practitioners with a guideline for building thin-wall features with minimal defects, and (2) the high correlation between the offline XCT measurements and in-situ sensor-based quality metrics substantiates the potential for applying deep learning approaches for the real-time prediction of build flaws in LPBF.more » « less
-
The objective of this work is to detect process instabilities in laser wire directed energy deposition additive manufacturing process using real-time data from a high-speed imaging meltpool sensor. The laser wire directed energy deposition process combines the advantages of powder directed energy deposition and other wire-based additive manufacturing processes, such as wire arc additive manufacturing, as it provides both appreciable resolution and high deposition rates. However, the process tends to create sub-optimal quality parts with poor surface finish, geometric distortion, and delamination in extreme cases. This sub-optimal quality stems from poorly understood thermophysical phenomena and stochastic effects. Hence, flaw formation often occurs despite considerable effort to optimize the processing parameters. In order to overcome this limitation of laser wire directed energy deposition, real-time and accurate monitoring of the process quality state is the essential first step for future closed-loop quality control of the process. In this work we extracted low-level, physically intuitive, features from acquired meltpool images. Physically intuitive features such as meltpool shape, size, and brightness provide a fundamental understanding of the processing regimes that are understandable by human operators. These physically intuitive features were used as inputs to simple machine learning models, such as k-nearest neighbors, support vector machine, etc., trained to classify the process state into one of four possible regimes. Using simple machine learning models forgoes the need to use complex black box modeling such as convolutional neural networks to monitor the high speed meltpool images to determine process stability. The classified regimes identified in this work were stable, dripping, stubbing, and incomplete melting. Regimes such as dripping, stubbing, and incomplete melting regimes fall under the realm of unstable processing conditions that are liable to lead to flaw formation in the laser wire directed energy deposition process. The foregoing three process regimes are the primary source of sub-optimal quality parts due to the degradation of the single-track quality that are the fundamental building block of all manufactured samples. Through a series of single-track experiments conducted over 128 processing conditions, we show that the developed approach is capable of accurately classifying the process state with a statistical fidelity approaching 90% F-score. This level of statistical fidelity was achieved using eight physically intuitive meltpool morphology and intensity features extracted from 159,872 meltpool images across all 128 process conditions. These eight physically intuitive features were then used for the training and testing of a support vector machine learning model. This prediction fidelity achieved using physically intuitive features is at par with computationally intense deep learning methods such as convolutional neural networks.more » « less
-
The goal of this work is to detect flaw formation in wire arc additive manufacturing (WAAM). This process uses an electric arc as the energy source in order to melt metallic wire and deposit the new material, similar to metal inert gas (MIG) welding. Industry has been slow to adopt WAAM due to the lack of process consistency and reliability. The WAAM process is susceptible to a multitude of stochastic disturbances that cause instability in the electric arc. These arc instabilities eventually lead to flaw formation such as porosity, spatter, and excessive deviations in the desired geometry. Therefore, the objective of this work is to detect flaw formation using in-situ acoustic (sound) data from a microphone installed near the electric arc. This data was processed using a novel wavelet integrated graph theory approach. This approach detected the onset of multiple types of flaw formations with a false alarm rate of less than 2%. Using this method, this work demonstrates the potential for in-situ monitoring and flaw detection of the WAAM process in a computationally tractable manner.more » « less
-
Abstract Aerosol jet printing (AJP) is a direct-write additive manufacturing (AM) method, emerging as the process of choice for the fabrication of a broad spectrum of electronics, such as sensors, transistors, and optoelectronic devices. However, AJP is a highly complex process, prone to intrinsic gradual drifts. Consequently, real-time process monitoring and control in AJP is a bourgeoning need. The goal of this work is to establish an integrated, smart platform for in situ and real-time monitoring of the functional properties of AJ-printed electronics. In pursuit of this goal, the objective is to forward a multiple-input, single-output (MISO) intelligent learning model—based on sparse representation classification (SRC)—to estimate the functional properties (e.g., resistance) in situ as well as in real-time. The aim is to classify the resistance of printed electronic traces (lines) as a function of AJP process parameters and the trace morphology characteristics (e.g., line width, thickness, and cross-sectional area (CSA)). To realize this objective, line morphology is captured using a series of images, acquired: (i) in situ via an integrated high-resolution imaging system and (ii) in real-time via the AJP standard process monitor camera. Utilizing image processing algorithms developed in-house, a wide range of 2D and 3D morphology features are extracted, constituting the primary source of data for the training, validation, and testing of the SRC model. The four-point probe method (also known as Kelvin sensing) is used to measure the resistance of the deposited traces and as a result, to define a priori class labels. The results of this study exhibited that using the presented approach, the resistance (and potentially, other functional properties) of printed electronics can be estimated both in situ and in real-time with an accuracy of ≥ 90%.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1007/s10845-023-02092-6
