An official website of the United States government
Abstract Inkjet printing (IJP) is one of the promising additive manufacturing techniques that yield many innovations in electronic and biomedical products. In IJP, the products are fabricated by depositing droplets on substrates, and the quality of the products is highly affected by the droplet pinch-off behaviors. Therefore, identifying pinch-off behaviors of droplets is critical. However, annotating the pinch-off behaviors is burdensome since a large amount of images of pinch-off behaviors can be collected. Active learning (AL) is a machine learning technique which extracts human knowledge by iteratively acquiring human annotation and updating the classification model for the pinch-off behaviors identification. Consequently, a good classification performance can be achieved with limited labels. However, during the query process, the most informative instances (i.e., images) are varying and most query strategies in AL cannot handle these dynamics since they are handcrafted. Thus, this paper proposes a multiclass reinforced active learning (MCRAL) framework in which a query strategy is trained by reinforcement learning (RL). We designed a unique intrinsic reward signal to improve the classification model performance. Moreover, how to extract the features from images for pinch-off behavior identification is not trivial. Thus, we used a graph convolutional network for droplet image feature extraction. The results show that MCRAL excels AL and can reduce human efforts in pinch-off behavior identification. We further demonstrated that, by linking the process parameters to the predicted droplet pinch-off behaviors, the droplet pinch-off behavior can be adjusted based on MCRAL.
more »
« less
Model Calibration in Inkjet Printing Process
Abstract Inkjet printing (IJP) is an additive manufacturing process capable to produce intricate functional structures. The IJP process performance and the quality of the printed parts are considerably affected by the deposited droplets’ volume. Obtaining consistent droplets volume during the process is difficult to achieve because the droplets are prone to variations due to various material properties, process parameters, and environmental conditions. Experimental (i.e., IJP setup observations) and computational (i.e., computational fluid dynamics (CFD)) analysis are used to study the droplets variability; however, they are expensive and computationally inefficient, respectively. The objective of this paper is to propose a framework that can perform fast and accurate droplet volume predictions for unseen IJP driving voltage regimes. A two-step approach is adopted: (1) an emulator is constructed from the physics-based droplet volume simulations to overcome the computational complexity and (2) the emulator is calibrated by incorporating the experimental IJP observations. In particular, a scaled Gaussian stochastic process (s-GaSP) is deployed for the emulation and calibration. The resulting surrogate model is able to rapidly and accurately predict the IJP droplets volume. The proposed methodology is demonstrated by calibrating the simulated data (i.e., CFD droplet simulations) emulator with experimental data from two distinct materials, namely glycerol and isopropyl alcohol.
more »
« less
- Award ID(s):
- 2134409
- PAR ID:
- 10471396
- Publisher / Repository:
- American Society of Mechanical Engineers
- Date Published:
- ISBN:
- 978-0-7918-8723-3
- Format(s):
- Medium: X
- Location:
- New Brunswick, New Jersey, USA
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Orb weavers produce webs that trap prey using a capture spiral formed of regularly spaced glue droplets supported by protein fibers. Each droplet consists of an outer aqueous layer and an adhesive, viscoelastic glycoprotein core. Organic and inorganic compounds in the aqueous layer make droplets hygroscopic and cause droplet features to change with environmental humidity. When droplets contact a surface, they adhere and extend as an insect struggles. Thus, a droplet’s extensibility is as important for prey capture as its adhesion. Cursory observations show that droplets can adhere, extend, and pull off from a surface several times, a process called cycling. Our study cycled individual droplets of four species—Argiope aurantia, Neoscona crucifera, Verrucosa arenata, and Larinioides cornutus. Droplets were subjected to 40 cycles at two humidities to determine how humidity affected droplet performance. We hypothesized that droplets would continue to perform, but that performance would decrease. Droplet performance was characterized by filament length and force on droplets at pull-off, aqueous volume, and glycoprotein volume. As hypothesized, cycling decreased performance, notably extensibility and aqueous volume. However, humidity did not impact the response to cycling. In a natural context, droplets are not subjected to extensive cycling, but reusability is advantageous for orb-weaving spiders. Moreover, the ability to cycle, combined with their environmental responsiveness, allows us to characterize orb weaver droplets as smart materials for the first time.more » « less
-
The phase transition from subcritical to supercritical conditions, referred to as transcritical behavior, significantly impacts the evaporation and fuel–air mixing in high-pressure liquid-fuel propulsion systems. Transcritical behavior is characterized as a transition from classical two-phase evaporation to a single-phase gas-like diffusion regime as surface tension and latent heat of vaporization reduce. However, the interfacial behavior represented by the surface tension coefficient and evaporation rate during this transition which are crucial inputs for Computational Fluid Dynamics (CFD) simulations of practical transcritical fuel spray is still missing. This study aims at developing new evaporation rate and surface tension models for transcritical n-dodecane droplets using molecular dynamics (MD) simulations irrespective of the droplet size. As MD simulations are primarily limited to the nanoscale, the new models can bridge the gap between MD and continuum simulations and enable the direct application of these findings to microscopic droplets. A new characteristic timescale, i.e., “undroplet time,” is defined which marks the transition from classical two-phase evaporation to single-phase gas-like diffusion behavior. The undroplet time indicates the onset of droplet core disintegration and penetration of nitrogen molecules into the droplet, which occurs after the vanishment of the surface tension. By normalizing the time with respect to the undroplet time, the rate of surface tension decay, evaporation rate, and the rate of droplet mass depletion become independent of the droplet size. Calculation of pairwise correlation coefficients for the entire MD results shows that both surface tension coefficient and evaporation rate are strongly correlated with the background temperature, while pressure and droplet size play a less significant role past the critical point. Therefore, new models for surface tension coefficient and evaporation rate spanning from sub- to supercritical conditions are developed as a function of background pressure and temperature, which can be used in continuum simulations. The identified phase change behavior based on the undroplet time shows a good agreement with the phase change regime maps obtained using microscale experiments and nanoscale MD predictions.more » « less
-
Material in HED regimes are at very high pressures and temperatures but can often still be modeled in the plasma-fluid approximation. Historically HED regimes were created using large laser/ion- beam drivers heating solid targets. Exciting data was obtained from these single shot experiments. In recent years there has been a shift to obtain HED related data from a large number of shots by using high-repetition-rate drivers. For high-repetition-rate experiments a series of droplet targets are often used to have a fresh target/droplet for each shot. However, one must make sure that target debris from the previous shot does not degrade the target for subsequent shots. This is a challenging CFD problem as one needs to model the initial dynamics of the heated droplet and the subsequent interaction with the following droplets. We use the CFD modeling code PISALE to study this complex problem. We discuss results for liquid hydrogen droplets heated by an x-ray free electron laser (XFEL). We first show 2D results for single heated droplet then 3D results for a heated droplet interacting with two unheated droplets.more » « less
-
Abstract In order to improve the quality of products during additive manufacturing, we developed a novel freezing sublimation-based method for inkjet-based three-dimensional (3D) printing technology, which can significantly improve the uniformity of material distribution in printed products. In our previous studies, we used a laboratory prototype with single droplets of inkjet solution containing colloidal particles to prove the concept of this study. However, understanding the interaction between droplets on the printing substrate surface is also crucial for determining the printing resolution and accuracy of this method, which cannot be fully investigated through single droplet-based experimental studies. To fill this knowledge gap, we conducted a series of experiments on colloidal droplet impingement, freezing, and sublimation on substrates using dual droplets. The experimental setup allowed the release of two droplets in quick succession from a modified nozzle with two needles. These droplets coalesced on the substrate surface due to spreading during their impingement processes. Observations revealed that the coalescence pattern of these two droplets varied depending on the time interval between their release. When the second droplet was released immediately after the first, their coalescence was governed by fluid dynamics. However, when the second droplet was released after the first droplet had frozen on the substrate, it spread above the ice surface of the first droplet in a relatively slower process. This observation provides new insights for the continued study and optimization of the proposed novel freezing sublimation-based 3D printing method.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1115/MSEC2023-105134
