Title: Computational Mechanics of Particle-Functionalized Fluid and Solid Materials for Additive Manufacturing and 3D Printing Processes

Location: UC Berkeley, Sibley Auditorium, Bechtel Engineering Center
Date: May 30-31, 2017
Organizers: Tarek I. Zohdi (UC Berkeley) and Eric Shaqfeh (Stanford)
Logistical Coordinators: Maxwell Micali and Zeyad Zaky
Sponsored by the National Academy of Sciences in conjunction with

TRAVEL: OAK and SFO are recommended airports, and they are both accessible from UC Berkeley by short cab/rideshare services and the local public transit system, B.A.R.T.

LOCAL ACCOMMODATIONS: The Hotel Durant and Hotel Shattuck Plaza are both short walks to the UC Berkeley campus and Sibley Auditorium.

PARKING DURING SYMPOSIUM: There is no free parking on campus, and parking on the local streets is subject to parking meters and hourly limits. As such, walking and/or the use of public transportation and campus shuttles are great alternatives if possible. UC Berkeley has several pay-to-park garages, and the closest two garages (in order of preference) are: Please note parking is subject to first come first serve. More information on these garages and other visitor parking in general can be found on the UC Berkeley Parking and Transportation website.

SCIENTIFIC COMMITTEE: Pierre Alart-France, Paul Cleary-Australia, Hans Herrmann-Switzerland, Sergio Idelsohn-Argentina, Eugenio Onate-Spain, Jerzy Rojek-Poland, Peter Wriggers-Germany
BACKGROUND: Within the last decade, several industrialized countries have stressed the importance of advanced manufacturing to their economies. Many of these plans have highlighted the development of robust additive manufacturing techniques, such as 3D-printing, which are still in their infancy. The combination of rigorous material modeling and particle-laden rheological theories, coupled with the dramatic increase of computational power, opens up the possibility that scientific computing can play a significant role in the analysis, control, and design of many emerging additive manufacturing processes. However, for these goals to be realized, a deep understanding of the essential ingredients comprising the materials involved in additive manufacturing is needed. Particle-functionalized materials play a central role in this field, frequently in the form of:
  • heated filament comprised of particles in a binding matrix,
  • inks comprised of particles in a solvent/lubricant, which cure when deposited and
  • dry particles (powders) which are deposited onto a surface and then heated in very targeted areas
  • potentially by a laser or other external source, in order to fuse them into place.
The role of these processes is primarily to build surface structures, coatings, etc., which are extremely difficult to construct using classical manufacturing methods.
GENERAL OBJECTIVES: To bring together researchers in materials, rheology, manufacturing and computation with the hope of developing cross-fertilization between these fields, which should have a natural synergy. In particular, interest is in exploring methodologies which are or can be applied to emerging advanced additive manufacturing techniques related to 3D-printing, and new materials processing. One key for next generation manufacturers and materials processors to succeed is to draw upon rigorous mathematical theory, cutting edge experimentation and high-performance computation to guide and simultaneously develop design rules for scaling up to industrial-level, mass-production and precision manufacturing. A primary mission of the workshop is to seek to enhance greater cooperation and collaboration ("harmonization") between the sub-disciplines known traditionally as materials, rheology, manufacturing and computational mechanics, by assembling a diverse set of researchers from academia and industry. A particular emphasis of the workshop is modeling and simulation of fluid and solid systems involving powders, particles and particle-fluid systems – with the latter focused on complex fluid suspensions. Particle-based materials and numerical methods have become wide-spread in the natural and applied sciences, engineering, and biology. The term "particle methods/mechanics" has now come to have several different meanings to researchers in the 21st century, including:
  • (a)Particles as a physical unit in granular media, particulate flows, plasmas, etc.,
  • (b)Particles representing material phases in continua at the meso-, micro-and nano-scale,
  • (c)Particles as a discretization unit in continua and discontinua in numerical methods.
The workshop will focus on all of the above topics. We welcome contributions in a variety of applications including, but not limited to:
  • Particulate-laden suspension and granular flow rheological and transport problems motivated by high-tech industrial processes, such as those stemming from extrusion, spray, deposition and printing processes
  • Fluid-structure interaction problems including, but not limited to, the same particle-laden and granular flows including free surface flow effects
  • Coupled multiphysical and rheological phenomena involving solid, fluid, thermal, electromagnetic and optical systems
  • Material design/functionalization using particles to modify base materials including their rheology and processability
  • Manufacturing processes involving forming, cutting, compaction, material processing as well as the effect of complex particle-laden viscoelastic rheology on these processes
  • Impact resulting in fracture, fragmentation and material robustness
FORMAT: This workshop is by invitation only.We estimate that approximately 35-45 persons from academia and industry will attend. There are no parallel sessions, no keynote and no plenary talks. Everyone will be given equal time to speak, and we will also have ample time for deep roundtable discussions.

39 Talks-20 minutes each (see numbering legend at the end of the schedule)

May 30

8:40-9:00: Zohdi and Shaqfeh opening remarks

9:00-9:20 #39
9:20-9:40 #1
9:40-10:00 #38

10:00-10:20 #2
10:20-10:40 #37
10:40-11:00 #3

11:00-11:20 #36
11:20-11:40 #4
11:40-12:00 #35

Working lunch and roundtable discussion 12:00-1:00

1:00-1:20 #5
1:20-1:40 #34
1:40-2:00 #16

2:00-2:20 #33
2:20-2:40 #7
2:40-3:00 #11

3:00-3:20 #8
3:20-3:40 #31
3:40-4:00 #9

Roundtable discussion 4:00-5:00

5:00-5:20 #30
5:20-5:40 #10
5:40-6:00 #26

Onsite Banquet: 6:00-7:30

May 31

9:20-9:40 #28
9:40-10:00 #12

10:00-10:20 #27
10:20-10:40 #13
10:40-11:00 #29

11:00-11:20 #14
11:20-11:40 #25
11:40-12:00 #15

Working lunch and roundtable discussion 12:00-1:00

1:00-1:20 #24
1:20-1:40 #6
1:40-2:00 #23

2:00-2:20 #17
2:20-2:40 #22
2:40-3:00 #18

3:00-3:20 #21
3:20-3:40 #19
3:40-4:00 #20

Roundtable discussion and wrap up 4:00-5:00

ACKNOWLEDGEMENT: Generous financial support from the Robert M. and Mary Haythornthwaite Foundation through the US National Committee for Theoretical and Applied Mechanics is greatly appreciated.



P. D. Anderson
Eindhoven University of Technology

3D printing could be used to make components that are electrically (dielectric, conducting) and/or magnetically active. To that end, electric or magnetic particles are dispersed in the matrix material. The idea is that in the process of ‘writing’ the entire 3D structure, the dispersed particles have enough mobility so that they can be oriented and/or aligned, by the application of external electric / magnetic fields. As soon as the ‘focus of writing’ moves on, the left-behind material becomes solid, either by cooling or cross-linking, and so the written particle arrangement is fixated. By alteration of the imposed field during the process of writing, electrically or magnetically active structures can be generated that are imbedded in the 3D printed part. To model the rearrangement of particles under various externally applied fields ((in)homogeneous electric, (in)homogeneous magnetic), to get most efficient alignment one requires suspension-type calculations with explicit particles in a viscoelastic fluid. In this presentation two approaches as discussed: first we present simulations of the start-up of shear flow of suspensions of rigid particles in viscoelastic fluids. A novel numerical method is applied that makes use of bi/tri-periodic domains, which act as representative volume elements of the suspension. Local mesh refinement ensures that the method is both accurate and efficient, without the need for a repulsive potential between the particles. By averaging many simulations for random initial positions of the particles, we obtain the true rheological bulk response of particle-filled viscoelastic suspensions. Second, we demonstrate a thermodynamically consistent model for particle dynamics with constitutive relations and share some first results on the modelling the 3D printing of thermosets filled with electromagnetically active particles using Brownian dynamics. We demonstrate particle alignment under external fields.


José E. Andrade and Reid Kawamoto
California Institute of Technology

New computational and imaging developments are redefining---in an irreversible way---our capabilities to compute and observe behavior in granular materials. The second most manipulated materials on earth (only after water), granular materials play a pivotal role in our lives, appearing in our food, drugs, 3D-printed materials, and in geologic, civil, space and defense applications. In this talk, we will look at recent advances in computing and imaging that are moving the field from a heuristic practice towards a more mechanics-based science. These advances enable us to look at the granular structure as the fundamental building block in this materials and extract a wealth of knowledge that was not accessible before. We will also show that the old paradigm of computing and experimenting in isolation is now rendered insufficient to gain further understanding. A new tool called the level set discrete element method (LS-DEM) will be presented and its predictive capabilities will be evaluated using advanced X-ray imaging. Potential applications in additive manufacturing will be explored.


Davey Beard
FLIR Systems

FLIR makes thermal cameras that measure temperature and display thermal gradients through infrared imagery. Various size and resolution cameras are available off-the-shelf, designed for OEM integration. Since temperature control and thermal gradient development is a critical part of 3D printing, integration into a print system offers a significant advantage. Possibilities include monitoring, troubleshooting, or actively controlling a 3D print process. With a properly designed system capable of resolving temperature differences of less than 1C, FLIR camera integration can unlock the next level of 3D printing performance.


Matthew R Begley, Materials, UCSB.

This talk will describe the development novel 3D printing nozzles for direct deposition of two-phase materials, which exploit acoustic focusing to assemble colloidal solids and control composite microstructures on-the-fly during deposition. Since acoustic assembly is relatively material agnostic (i.e. not dependent on specific compositions or surface chemistry), it can be combined with chemical self-assembly to create hierarchical assembly platforms to fabricate macroscale specimens from nanoparticles. In the first half of the talk, printing regime maps will be presented that identify system parameters that lead to effective “on-the-fly” assembly during direct deposition, including nozzle geometry, acoustic pressures, viscosities, etc. The second half of the talk will describe the use of acoustics to achieve rapid assembly of millimeter-scale specimens of colloidal solid, consisting of surface functionalized gold nanoparticles and polymer micro- beads. The talk will conclude with a brief discussion of the implications of these results with respect to achieving specific microstructures to enhance the performance of functional materials and architected structural materials.


John F. Brady & Mu Wang
Divisions of Chemistry & Chemical Engineering and Engineering & Applied Science, California Institute of Technology, Pasadena, CA 91125

We study structural and stress evolution of hard-sphere suspensions during uniaxial compression achieved by moving one confining wall towards the other via Brownian dynamics simulations. Compression at both constant velocity and at constant normal stress are examined. The systems microstructure in monitored via correlational order parameters, the pair-distribution function, Voronoi tessellations, etc. When Pe >>1, where Pe characterizes the compression rate relative to the time scale for Brownian motion, the moving boundary not only causes geometric frustration by disrupting Brownian relaxation, but also introduces local ordering near the moving boundary. At constant compression velocity the normal stress diverges as the gap width decreases. When Pe > 1, the normal stress on the confining walls are proportional to Pe, and the stress distribution changes rapidly in a boundary layer near the moving wall. For compression at constant normal stress, the moving wall eventually fluctuates around an equilibrium position, and the structural and stress evolution depend on the wall mobility. Continuum models for both constant velocity and constant stress simulations are developed to describe the interplay between the moving boundary and Brownian motion. Without any fitting parameters, the model agrees quantitatively with the simulation results for the wall stresses and wall displacements, and qualitatively captures the stress and local volume fraction profiles across the gap throughout the compression process.


Jerry Cabalo US Army

A range of animal models are used in toxicological studies to determine the hazards to humans. However, these studies encompass a wide range of species that have different respiratory physiology, so that for inhalation studies it is difficult to determine the actual mass of material deposited in the lungs. To understand the true toxicological threat with the simplest animal model, it is critical to understand the differences between individual humans, as well as between humans in general and animal models. The Edgewood Chemical Biological Center is currently performing deposition studies of aerosols in 3D printed models of individual human respiratory tracts as well as animal models of various species. However, we encounter a number of limitations in the current 3D printed technology. We intend to discuss how advances in the technology could benefit the core research.


Eduardo Campello
University of Sao Paulo

This work investigates the suitability of some cohesion force models for the simulation of particle deposition processes via the discrete element method (DEM). The results shown are partial conclusions of a bigger effort in the direction of identifying or proposing an adhesion force model that is well suited to cohesive granular materials, with the possible feature of temperature-dependency. We assess the models in their capability of capturing inter-particle bonging at the micro level such that the cohesive behavior of the material at the macro level is well characterized. Bonding initiation and occasional breakage (leading eventually to material disaggregation) are also of interest, since these phenomena are key aspects on the mechanics of cohesive granular media. Considering we are working under a DEM framework, cohesion force models need to be computationally efficient and yet render reliable results at the macro scale. We hope we may furnish results from numerical investigations that may subsidize the pursue for good adhesion force models on this regard. In our opinion, consistent DEM models may be a useful tool for the simulation of particle deposition processes and, in a broader sense, many other particle systems wherein bonding and agglomeration are observed.


P.W. Cleary (CSIRO), M.D. Sinnott, S.C. Cummins, G.W. Delaney, V. Lemiale, G.G. Pereira, and S.M. Harrison

Particle and multiphase modelling developed for processes such as comminution and material forming is well suited to the challenges of simulation of additive manufacturing processes. In this talk we focus on the use of DEM, SPH and coupled DEM+SPH and their application to both traditional particle processing domains (such as mineral processing and extrusion) and emerging areas (such as additive manufacturing and food digestion). Examples of simulation will include:
  • Grinding mills and crushers to demonstrate DEM (including with particle breakage) for the prediction of size reduction processes
  • Extrusion (of metals and food materials) to demonstrate use of SPH in solid deformation manufacturing processes
  • Cold spray which is a form of additive manufacturing where very high speed metallic particles collide with and adhere to a substrate that then build up to form controlled surfaces with desirable properties
  • Powder additive manufacturing in which examples of DEM simulation of the raking and compaction of the powder bed, and then SPH simulation of the partial melting of powder by laser or electron beam, are used for prediction of microstructure properties.
A specific additive application that places larger demands on the manufacturing process is that of food materials. For these, it is also important to understand how these food materials break down in the body – first in the mouth during oral digestion and then how they are broken down and absorbed within the body. Examples of DEM and SPH application to food digestion will also be discussed. Particle modelling of powder additive manufacturing, comminution, extrusion and digestion processes


Fernando E. Garcia, Ph.D. Candidate, UC Berkeley
Faculty Advisor: Jonathan D. Bray, Ph.D., P.E., NAE
Contact: estefan31@berkeley.edu

Permanent ground deformation associated with earthquake surface fault rupture poses a significant hazard to the built environment in near-fault regions. Linear infrastructure such as pipelines, roads, and railways often cannot avoid traversing an active fault trace, and life-threatening damage can result if large fault offsets occur directly beneath a structure. The phenomenon of surface fault rupture propagation through sand deposits involves granular material response to boundary deformation at the fundamental level. It is also inherently three-dimensional. Thus, it is well-suited for analysis using the Distinct Element Method (DEM). DEM simulations of surface fault rupture through a dilatant assembly of non-spherical particles are evaluated against closed-form solutions for the shape and location of the failure surface propagating from a distinct bedrock fault. The capability of DEM to capture dilatant and contractive behaviors in particulate assemblies with different void ratios is shown using direct shear test simulations, and the influence of soil ductility on fault rupture propagation is analyzed using particle assemblies with different void space distributions. Macroscopic and particle-scale mechanisms of shear band formation are analyzed through homogenized stresses and strains, particle rotations, frictional dissipation, and contact forces. A mechanism of graben formation during shallow-dipping normal faulting is related to the stress-arching phenomenon that commonly occurs over trapdoor displacements.


Wolfgang Gerlinger

Fiber reinforced composites are promising materials for lightweight applications. Materials with a high solids content and long fibers can be achieved by additive manufacturing, e.g. by Resin Transfer Molding (RTM). The RTM technology is very challenging especially when small cycles times are necessary, and the process conditions can have a tremendous impact on the mechanics of the materials.This contribution shows a holistic multiscale simulation approach including rigorous simulation of the process at the different length scales as well as virtual material testing of the composites by micromechanical simulations.


Costas P. Grigoropoulos
Laser Thermal Laboratory
Department of Mechanical Engineering
University of California, Berkeley CA 94720-1740

This presentation reviews recent and on-going research at the Laser Thermal Laboratory. Interactions of pulsed laser radiation with nanostructures have been investigated. Laser annealing of nanoscale precursors is utilized to produce single crystalline domains on non-participating substrates. New methods have been introduced for the localized growth, assembly and functionalization of nanostructures using laser-assisted additive manufacturing methods. Maskless fabrication of functional devices on flexible substrates has been conducted by using nanoparticles in conjunction with laser processing and printing/nanoimprinting. High-performance devices have been realized on flexible substrates. Ultrafast laser radiation drives nonlinear interactions that can be utilized for nanoprocessing using even far field optics. A new concept combining laser fabrication of three-dimensional nano-scaffolds and directed self-assembly of nanostructures using block co-polymer technology is outlined


Oliver Harlen (University of Leeds), Claire McIlroy, Wouter Mathues, Christian Clasen

The surface tension driven thinning and break-up of liquid bridges is a fundamental mechanism controlling jet breakup and droplet formation, and its prediction is vital to processes such as inkjet drop and spray formation. For Newtonian fluids the final stages of break-up follow a progression of similarity thinning laws depending upon the value of the Ohnesorge number. However, experimental studies of non-Brownian particle suspensions find that the approach to break-up depends upon the particle size, as well as particle concentration and so cannot be predicted from the bulk properties of the suspension alone. In particular once the radius of the filament shrinks to a few particle diameters there is a transition from continuum thinning to an “accelerated” regime in which the rate of decrease of the filament radius becomes faster even than that of the suspending fluid alone. This is followed by a further transition to a regime corresponding to the viscosity of the continuous phase. To elucidate the mechanism responsible for this accelerated thinning, we have constructed a simple model in which the viscosity is determined from the local particle density, found by tracking individual particles within the suspension. The results of this model are compared with capillary thinning experiments for suspensions with three different particles sizes. Despite the simplicity of the model it is not only able to reproduce the sequence of thinning regimes seen in experiments, but also provides a quantitatively accurate reproduction of the accelerated regime and its transitions


Neil E. Hodge
Lawrence Livermore National Lab

Selective Laser Melting (SLM) is a manufacturing process which can realize significant benefits over traditional manufacturing processes, including significantly shortened time between design and manufacture of parts, and the ability to create parts with much more geometric complexity than has previously been possible. However, the extreme sensitivity of the results to input parameters results in a process that is difficult to predict/control. Indeed, it is not uncommon for the resulting parts to vary significantly from their as-designed geometry, due to the influence of extreme and inhomogeneous thermal gradients. The goal of this research is to develop a part-scale model of the process, in order to aid in part qualification. This presentation will describe the formulation and implementation of a continuum-based model, as well as comparisons between model and experimental results. It will also discuss ongoing implementation tasks, as well as the researchers’ thoughts on necessary future work.


Peter Hoseman
UC Berkeley

It has been shown that additive manufacturing opens the door to novel shapes and geometries produced in an efficient fashion. In addition to these advances one can also envision to manufacture components with multiple materials to tailor the materials property to the geometry. This work features the graded materials from stainless steel to a high strength Maraging steel within one single shaft used for space applications. Detailed microstructural and mechanical properties show that a gently graded composition leads to a multi phase materials (fcc and bcc mixed) but also experiences a sharp transition of hardness at a specific composition likely due to a threshold for precipitate formation.


Antonio Souto Iglesias.
Associate Professor
Technical University of Madrid (UPM)

Viscosity dominated free-surface viscous flows are relevant for various areas of Engineering, in particular additive manufacturing and 3D printing, the focus of this symposium. Particle methods can be a competitive alternative to model these kinds of flows, but issues about convergence and stability of some schemes have hindered a further development of the use of such methods for modeling these problems. Smoothed particle hydrodynamics (SPH) is one of such methods. Its formulation deserves attention due to some inconsistencies occurring when considering free-surface flows, arising from kernel truncantion. In SPH formulations one usually assumes that free-surface conditions are implicitly verified. The referrred inconsistency and this assumption are discussed in the talk for Newtonian flows. In particular, the principle of virtual work is used to demonstrate the verification of the free-surface boundary conditions in a weak sense, with the method rendering accurate energy dissipation. Numerical verification of this analysis is provided, suggesting the method can be considered an alternative for modeling certain processes relevant in the additive manufacturing and 3D printing field.


Ken Kamrin

To simulate deforming solids of general shape and constitutive behavior, one typically uses Lagrangian techniques, commonly the finite element method. However, fluid flows tend to be easier to solve using Eulerian-frame techniques like finite-volume or finite-difference methods. To solve particle-laden flow problems with soft, highly-deformable particles, the methodology chosen must mitigate this dichotomy. While a number of approaches exist, here we present a new method called the Reference Map Technique (RMT), which permits one to simulate, on a single Eulerian background grid, many soft generally-shaped particles immersed in a Navier-Stokes-obeying fluid phase. The particles are treated as arbitrary solid bodies obeying finite-deformation constitutive behavior (e.g. hyperelasticity, plasticty, etc.). As is needed to simulate particle-laden flows in the dense limit, the routine also permits blunt contact interactions between particles. We also show how RMT can be used to model soft "active" solids, such as swimmers. Throughout, several benefits of the method being fully-Eulerian are emphasized such as the computational efficiency and simplicity of operating on a single fixed grid, as well as certain numerical advantages regarding contact detection and the enforcement of constraints like incompressibility.


Yannis Kevrekidis (Princeton University) and Len Pismen

We explore the use of data mining techniques to produce informative embeddings for trajectories of active particles of different kinds (such as Janus particles, defects in active nematics, or microswimmers) in combined self-generated and externally imposed flows. The emphasis is on gauge invariant embeddings that will not depend on the measuring instrument and allow one to infer the mechanism of active flow from experimental data or particle-based simulations.


Petros Koumoutsakos (ETH Zurich, Switzerland) and Costas Papadimitriou (University of Thessaly, Greece)

Particles are use for simulations of systems ranging from Nano-devices to Galaxies. What is the value of the predictions of such simulations ? We address this question by distinguishing errors and remedies for particle discretisations of continuum equations and by quantifying uncertainties of discrete particle models. We discuss in particular a Bayesian inference framework for quantifying data driven predictions of nano/micro-fluidics and granular materials using particle models.


Adrian Lew (Stanford). This is a collaboration with Wei Cai (Stanford), Brandon McWilliams (ARL), and Daniel Galles (ARL).

In this presentation I will describe a series of experiments designed with the aim of calibrating and evaluating the accuracy of a material model for Cu-Zn strips that are thermally treated with a laser beam, as well as the material and numerical model we are in the process of validating. The ultimate goal of this exercise is to serve as a precursor to modeling selective laser melting manufacturing processes on the same material.


Gang-yu Liu, Jiali Zhang
Department of Chemistry, University of California, Davis, CA 95616
Victoria Piunova, Jane Frommer
IBM Almaden Research Center, 650 Harry Road, San Jose, CA 95120

3D printing that reaches the nanometer scale reproducibly is particularly challenging. Difficulties arise in the printing process from the high spatial precision required for materials delivery and registry between components, physically and temporally. Using polymer nanoparticles and AFM-based instrumentation, we have investigated the accuracy of material delivery and spatial control. Combining the scanning probe’s spatial precision with advanced local delivery methodologies, including microfluidics (Figure 1), this work reveals how our molecules assemble from the nano- to the meso-scale. The behavior of our polymer nanoparticles in a dynamic and spatially confined local environment differs from the anticipated electrostatics- and polarity-driven conventional behaviors of the micro and macro scales. 2D and 3D nanolithography with layer-by-layer nanoprinting of various structures with designed functionalities will be presented and potential applications will be discussed.


Fuduo Ma, Vince Battaglia, and Ravi Prasher
Energy Storage and Distributed Resources Division
Lawrence Berkeley Laboratory

New materials hold the key to developing technologies that can eliminate the use of fossil fuels and reduce carbon emission. Better materials can help, for example, batteries store more energy, solar cells become more efficient, and fuel cells become less expensive. While there is an active and robust research activity on discovering new materials, history shows that the translation of each new material from lab to market can take more than a decade. For example, in batteries, one challenge lies in going from the small quantity, low-yield experiments that are ample for publications to the large volumes of materials needed at high purity and low cost. Moreover, lab-scale experiments do not focus on issues of scalability, process-ability, impact of defects and impurities, and the cost associated with translation; all important issues in the real world. The first step of the materials development process, the design of novel compounds with highly-optimized properties, has undergone major improvements in the last ten years. Today, advances in computational materials modeling are accelerating us towards a future where most properties of real and virtual compounds can be available on demand, enabling rapid screening in material design efforts. These successes in accelerated materials design have moved the bottleneck in materials development towards the synthesis of novel compounds and their incorporation into components and devices. This delay in going from promising materials concept to validation, optimization, and scale-up is a significant burden to the commercialization of novel energy materials. For example in Lithium Ion Battery, electrode fabrication involves mixing three essential components, i.e. electrochemically active materials, conductive additives, and polymer binders in a solvent. The mixing is followed by casting and drying on a current collector. The sizes of active material particles and conductive additive particles typically range from several nanometers to about tens of micrometers, indicating the electrode slurry is a colloidal system. In this paper we will discuss how colloidal science can be applied in developing battery processing physics.


Simo A. Mäkiharju, UC Berkeley

While computed tomography (CT) is widely utilized to inspect additively manufactured (AM) parts after the AM process is completed, high-speed [O(kHz)] X-ray densitometry and computed tomography techniques developed for multiphase flows could enable observing defects forming in real-time, monitoring of internal part dimensions during the AM process and may provide useful data for process control. While applicable to free form, filament and most of the AM techniques, particularly in case of electron beam AM the utilization of X-ray measurements might prove cost-effective to implement, as the electron beam melting the material can also be utilized to generate X-rays.


John Michopoulos, John Steuben, Athanasios Illiopoulos
Computational Multiphysics Systems Laboratory
US Naval Research Laboratory
4555 Overlook Ave. SW
Washington, DC 20375, USA

The recent developments and applicability of Additive Manufacturing (AM) technologies for a broad range of scientific and industrial purposes have introduced the opportunity for Integrated Computational Material Engineering as the means for connecting process parameters with functional performance of as-produced parts. The drastic microstructural and multiscale differences between materials produced via AM and conventional methods has motivated the development of computational tools that model and simulate AM processes realistically in order to enable their control for the purpose of optimizing the desired outcomes. We are presenting recent advances in the development of the Multiphysics Discrete Element Method (MDEM) for the simulation of AM processes based on polydisperse powder feedstock within the ICME context. This multiphysics enhanced particle-based method elegantly encapsulates the relevant physics of powder-based AM processes. In addition to a brief description of the overall computational framework developed at NRL we will focus on its recent enrichments from various perspectives. Specifically, we will focus on the underlying constitutive behavior implemented for including thermoplasticity as it relates to the solidification process and the resulting residual strains and on the particle-air interactions and on meta-ball based methodologies for modeling the melting and re-solidification of the feedstock materials. Algorithmic improvements that increase computational performance will also be discussed. We finally demonstrate the ability of MDEM to enable the simulation of the AM process of scaled-up macro-scale components in a manner that includes the interaction of the AM process with the applicable chamber environment including the interaction of the recoater with the under process part. Concluding remarks will be presented on the tasks required for the future evolution of the MDEM as well as the required experimental validation.


Garrett Nelson, PhD.
Defense Threat Reduction Agency
Edgewood Chemical Biological Center

Two-Photon Polymerization (2PP) 3D printing was recently used to print submillimeter-scale complex liquid droplet generators with submicron resolution, 150 to 250 µm stitching intervals, and a printing time of about 2.3 hours per device. These injectors matched the performance of their hand-crafted predecessors in serial crystallography at the LCLS XFEL at SLAC, and the 2PP approach has opened doors for previously unattainable sample delivery capabilities. The Edgewood Chemical Biological Center is looking to enhance piezo-driven on-demand aerosol generation systems using a 2PP approach, but we desire 1) stitchless printing to reduce gas permeability and 2) variable resolution printing to enable larger, cm-scale devices.


Mehdi Nourbakhsh, Ph.D. (Autodesk Research)
Massimiliano Moruzzi (Autodesk Research)

Generative design mimics nature’s evolutionary approach to design. Designers or engineers input design goals into generative design software, along with parameters such as materials, manufacturing methods, and cost constraints. Then, leveraging distributed computing and advanced functionally graded materials, the software explores all the possible permutations of a solution, quickly generating design alternatives. In this talk, we'd like to present some of the research findings on applying manufacturability constraints and innovative numerical models for functionally graded materials to our generative design platform "Dreamcatcher".


Eugenio Oñate, Francisco Zarate and Miguel Angel Celigueta
International center for Numerical Methods in Engineerinc (CIMNE)
Technical University of Catalonia , 08043 Barcelona, Spain

We present a multiscale particle-to-continuum computational approach for studying the key steps of the additive manufacturing (AM) of particulate-functional materials and the structural behavior of AM components manufactured using the Powder Bed technology. Melting and sintering AM processes using either laser or electron beam heat sources are considered. The so-called P2CAM approach integrates an improved Discrete Element Method (DEM) [1] for modeling the particle deposition phase and the interaction with the heat source. The DEM is coupled to enhanced versions of the Particle Finite Element Method (PFEM) [2] and multiscale analysis techniques for predicting the melt/sinter pool effects at micro and meso-structure level, leading to the curing and construction of the material layers to form an AM component. An innovative FEM-DEM technique [3] fed with the material information from the manufacturing process, is used for analysis of the structural behavior and ultimate strength of the AM component. We summarize next the key steps of the P2CAM approach: The starting point for the formation of each material layer is the lay out of the powder bed containing a set of particles placed on top of the previously formed solid layers. Particles are modelled with the DEM accounting for cohesion between discrete particles. The traveling laser/electron beam acts on the powder bed and defines the heat source influence zone where melting or sintering of the discrete particles will occur. A procedure for modelling the heat source and for defining the melt/sinter influence zone is developed (Fig 1.1). The melt/sinter influence zone and the solid layers are discretized via the PFEM, while particles laying outside that zone are discretized with the DEM. The meso-mechanics of the influence zone is studied accounting for its interaction with the adjacent powder bed and solid zones (Fig 1.2). Once the laser/electron beam has completed its path, the process is restarted by deposition of the next powder bed layer until the manufacturing of the component is finished (Fig 1.3).
[1] Oñate E., Franci A., Carbonell J.M. (2014c), A particle finite element method for analysis of industrial forming processes, Computational Mechanics, Vol. 54 (1), pp. 85-107.
[2] Oñate E., Zárate F., Miquel J., Santasusana M., Celigueta M.A., Arrufat F., Gandikota R., Valiullin K. and Ring L. (2015), A local constitutive model for the discrete element method. Application to geomaterials and concrete, Comput. Particle Mech., 2(2), pp. 139-160. Zarate F. and Oñate E. (2015) A simple FEM-DEM technique for fracture prediction in materials and structures. Comput. Particle. Mech. 2, 301-31


Andrew Rutter, Type A Machines

The considerations and challenges of FDM printing. From the fundamentals of extruder design, to the basic parameters that effect the suitability of polymers and polymer composites for the process. With particular focus on the areas that lack clear general understanding outside of academia, yet have very significant impacts on the results. Exploring the challenges of an industry driven by enthusiastic amateurs. And discussing the concerns as more exotic polymers become both available and in demand.


P. Randall Schunk, Sandia National Laboratories/University of New Mexico
Kristianto Tjiptowidjojo, University of New Mexico
Robert Malakhov, University of New Mexico
Nelson Bell, Sandia National Laboratories
Adam Cook, Sandia National Laboratories
Vivek Subramanian, UC Berkeley
Will Scheideler, UC Berkeley

Ink-jet, aerosol ink-jet, and gravure printing are promising routes for meeting the increased demand for printed functional materials, such as electronic and photonic devices and structural metamaterials. While gravure and ink-jet printing are substantially different forms of additive manufacturing in terms of scalability, overlay registration and feature size, they share common challenges related to processing particulate inks. Advancing these processes to achieve higher performing, more reliable and cost-effective devices is substantially limited by abrasion/wear, nozzle clogging, and print defects all traceable to the presence of particles. These problems are exacerbated when the particles themselves become a significant fraction of the feature sizes being printed, thus providing a fundamental limit. In this presentation we will discuss some of the problems and solutions related to processing particulate inks experienced at our institutions, and pose some challenges to the computational mechanics community. Specifically we will focus on fundamental effects of particle size, volume fraction and the like on printability through the effects on solution rheology, stability and consolidation by drying and sintering. We will demonstrate some of our current computational technology towards addressing these challenges, and pose challenges and requirements for future modeling and simulation approaches.


Ryohei Seto
Okinawa Institute of Science and Technology

From a macroscopic perspective, concentrated particle suspensions can be viewed as complex fluids. Constitutive laws for such fluids generally differ from that for Newtonian fluids. In particular, they incorporate features of non-Newtonian response such as shear thickening and normal stress differences. We introduce a general constitutive framework to characterize viscous fluids, a framework that goes beyond the conventional reliance of viscometric functions. We then use Stokesian Discrete Element simulations with periodic boundary conditions to compare the material functions obtained in simple shear and planar extensional flows to show the effectiveness of our framework in capturing flow-type dependence in the behavior of concentrated particle suspensions.


Eric S. G. Shaqfeh, Department of Chemical Engineering, Stanford University
Sreenath Krishnan, Department of Mechanical Engineering, Stanford University
Gianluca Iaccarino, Department of Mechanical Engineering, Stanford University

There are no comprehensive simulation-based tools for engineering the flows of viscoelastic fluid-particle suspensions in fully three-dimensional geometries. On the other hand, the need for such a tool in engineering applications is immense. Suspensions of rigid particles in viscoelastic fluids play key roles advanced manufacturing applications. In the present work, we describe the development of an Immersed Boundary Method (IB) to simulate the viscoelastic flow in suspensions of particles of arbitrary shape and deformability in complex flows. Since the phenomomena of interest occur typically at O(1) values of the flow Weissenberg or Deborah number, we describe the methods necessary to obtain accurate resolution of the stress boundary layers near the particle surface even in the IB framework. Since the code is massively parallel, we demonstrate the simulation of a few hundred particles with the code, and examine in detail three problems where the multi-particle viscoelastic interactions provide unique physical results: 1) The sedimentation of rigid spheres in orthogonal shear and 2) The rheology of a rigid sphere suspension in a viscoelastic fluid in a parallel plate device and (3) the rheology of deformable particles in the same device. We examine these suspensions up to 5% volume fraction and demonstrate that, in each case, the dilute approximation is poor even at very low volume fraction because of the finite Wi wake interactions between particles. We believe this tool can be used to examine flows of suspensions in nozzles at finite volume fractions in AM applications for example.


Logan Pensinger (Universal Filaments) and Brendan Mcsheehy

There is strong interest within filament deposition modeling (FDM) with improving physical properties of the formed part, particularly high flexural modulus and high heat deflection temperatures. One way to achieve this goal is through addition of fibers in to a high glass transition temperature polymer. Specific systems of commercial interest include carbon fiber-filled polycarbonate and nylons. FDM filament is typically made through polymer extrusion. The filament must be of uniform diameter and roundness along the length of the filament to maintain consistent feed rates in FDM. Production extrusion speeds are limited due to wall slippage in the extrusion die resulting in process instabilities and subsequent variability in filament diameter. The use of slip and adhesion additives or die material of construction obviously may influence stick-slip phenomenon. Additives used must not interfere with interlayer adhesion during its subsequent use in FDM. Abrasiveness of the carbon fibers limits the choice of materials of construction of the extrusion die. The melt viscosity of the polymer compound also plays an important role in filament formation.


Andrew J. Szeri
Deptartment of Mechanical Engineering
University of California, Berkeley
6119 Etcheverry Hall, Berkeley, CA 94720-1740
E-mail: aszeri@berkeley.edu

Advanced manufacturing methods often employ heterogeneous materials in their design, and occasionally produce unwanted heterogeneities as a consequence of imperfections in the manufacturing process. A challenge in advanced manufacturing is to characterize the manufactured items either during or following a processing step, in a non-invasive way. The purpose may be to spot flaws or to determine material properties associated with, for example, a desired heterogeneous microstructure. This presentation concerns a methodology for accomplishing this goal, which relies on the careful processing of thermal images.


Roger I.Tanner
University of Sydney, Sydney 2006, Australia

For practical use in manufacturing it would be convenient to have a model, albeit approximate, of particle-laden materials (suspensions) that would not need large amounts of computing and/or experimentation in every case. There are now adequate models of the pure matrix fluid behaviour, but there are no such models for suspensions, especially for suspensions with large particles (non-colloidal suspensions). One of the obstacles has been the single-minded devotion to shearing motions; experience with the matrix modelling has shown that it is not possible to formulate widely usable models if only shear is considered. Here some results of axially-symmetric elongational tests on suspensions are compared with shearing data, and differences emerge. Also, the dramatic effects of friction and particle roughness on suspension rheology need to be explored thoroughly; we found that roughness in spherical particles induces large increases in viscosity in both shear and elongation. Some suggestions for modelling based on these observations are presented.


Jun Tao, Zhen Chen, Yonggang Zheng, and Hongwu Zhang
State Key Laboratory of Structure Analysis for Industrial Equipment, Department of Engineering Mechanics, Dalian University of Technology, Dalian, Liaoning 116024, China
Department of Civil & Environmental Engineering, University of Missouri, Columbia, MO 65211, USA
Corresponding author: chenzh@missouri.edu

Additive manufacturing (AM) aims at building 3D objects by adding layer-upon-layer of material, and becomes increasingly popular in recent years. Selective laser melting (SLM) is an AM process in which layers of metal powders are heated via laser in order to build an engineering part [1]. To better control this AM process, a fully coupled thermo-mechanical modeling and simulation procedure should be developed with a higher level of fidelity. To simulate multi-phase (solid-fluid-gas) interactions without invoking master/slave nodal treatment, the material point method (MPM) has evolved over the last two decades, and been applied to many areas in Simulation-Based Engineering Science, as shown in a recent comprehensive literature survey [2]. Recently, an effort has been made to develop the generalized interpolation material point (GIMP) method with a weakly coupled thermo-mechanical formulation [3]. However, both thermal state and deformation state are dependent on each other in the SLM process. Based on the previous study by Hodge et al. for uncoupled thermo-mechanical cases [1], hence, the GIMP method for strongly coupled thermo-mechanics is being developed for multi-physics model-based simulation. Except for heat transfer, the phase change phenomenon is also considered using the Stefan-Neumann equations such that the influence of the deformation state on the thermal state could be evaluated. Verification and validation of the proposed procedure with available experimental data and other numerical solutions will be presented in the workshop.
[1] Hodge, N.E., Ferencz, R. M., and Solberg, J.M., “Implementation of a thermomechanical model for the simulation of selective laser melting,” Computational Mechanics, Vol. 54, pp. 33–51, 2014.
[2] Zhang, X., Chen, Z., and Liu, Y., The Material Point Method – A Continuum-Based Particle Method for Extreme Loading Cases, Academic Press, Elsevier, 2016.
[3] Tao, J., Zheng, Y.G., Chen, Z., and Zhang, H.W., “Generalized interpolation material point method for coupled thermo-mechanical processes,” International Journal of Mechanics and Materials in Design, Vol. 12, pp. 577-595, 2016.


Hayden Taylor
Mechanical Engineering, UC Berkeley

There has been rapid recent growth of the use of fused deposition modeling filaments containing solid particles in a thermoplastic binder. These composite filaments can offer enhanced mechanical stiffness and strength compared with pure thermoplastics, as well as higher electrical and thermal conductivities. The particulate phase may be metallic or ceramic, or be composed of graphene or carbon nanotubes. These materials present particular challenges for the extrusion of filaments, with abrasion of the extrusion nozzle being greatly accelerated and filament diameter becoming more difficult to hold within desired tolerances. We develop analytical and numerical models to assess alternative filament formation techniques, including multi-stage extrusion and micro-scale rolling.


Veena Tikare, Theron Rodgers, Jon Madison and John Mitchell
Sandia National Laboratories

Additive manufacturing (AM) is of tremendous interest given its ability to realize complex, non-traditional geometries in engineered structural materials. However, microstructures generated from AM processes are very different and often more complex than their conventionally processed counterparts with spatial heterogeneity that is unique to AM parts and depends on the details of the process used. We have developed a numerical model that can simulate microstructural evolution in AM parts produced by the Laser Engineered Net Shaping, LENS, and Powder Bed Fusion, PBF using a kinetic Monte Carlo model to predict three-dimensional grain structure in additively manufactured metals. The model simulates solidification from the molten pool and grain growth in the complex transient thermal field as it rasters to build the AM component. The utility of this model to design microstructure for optimal engineering performance will be presented and discussed by showing examples of its predictive capabilities for LENS and PBF processes.


Henning Wessels, Christian Weissenfels, Peter Wriggers
Institute of Continuum Mechanics, Leibniz University Hannover, 30167 Hannover, Germany

Selective Laser Melting (SLM) is an additive manufacturing (AM) process, where a powder bed is partially melted. Layer by layer, complex three dimensional geometries including overhangs can be produced, because non-melted powder acts as support structure. Up to date the multiple interacting physical phenomena are not yet fully understood. This is why the material and process development mainly relies on experimental studies that are time and cost intensive. Novel simulation tools such as meshless methods offer the potential to gain a deeper understanding of the process -structure - property interaction. This can help to find optimal process parameters and to individualize AM manufactured parts by locally altering material properties. Using conventional FEM methods, extremely large deformations of the mesh lead to ill-shaped elements and per consequence to degenerate computations. Meshfree methods eliminate the mesh dependency by employing a more exible formulation to relate a point of integration to its neighboring nodal points. This requires exible shape functions that depend on the nodal positions only [1]. The Optimal Transportation Meshfree (OTM) Method is a meshless method based on the weak formulation of the differential equations and can be downscaled to the Finite Element Method. It accounts for a broad variety of materials ranging from solids to fluids [2] including thermo-mechanical behavior. This exibility makes the OTM an optimal tool to simulate the melting of powder particles and the motion of the melt flow. An approach to account for the phase transition and the fusion of particles using OTM will be presented. Furthermore, the behavior of the metal during solidification is assessed. Releasing the induced residual stresses can yield undesired deformations and destroy AM parts.
[1] Idelsohn, SR, Onate, E, Calvo, N, Del Pin, F. The meshless finite element method. International Journal for Numerical Methods in Engineering 2003. 58(6):893-912
[2] Li, B, Habbal, F, Ortiz, M. Optimal transportation meshfree approximation schemes for uid and plastic flow. International Journal for Numerical Methods in Engineering 2010. 83(12):1541-1579


Roseanna N. Zia
Cornell University

Understanding kinetically arrested phase transition in complex media, and its influence on structure-property relationships, has been identified as one of the Grand Challenges for the future of soft matter science. Colloidal gels and glasses are an important class of such materials that are central to additive manufacturing processes, and are the subject of an emergent field of study in which much focus is placed on predicting yield behavior. More fundamentally, the physico-chemical nature of inter-particle colloidal attractions has permitted the construction of colloidal phase diagrams via molecular theories, where metastable and unstable phase separation closely parallels that in molecular systems. However, colloidal gels represent “arrested” states of phase separation, where the same interparticle attractions that promote phase separation also inhibit it, freezing in a non-equilibrium microstructure to form a viscoelastic network. In contrast to attempts to place them on equilibrium phase diagrams, we argue that such gels must exit the equilibrium phase diagram. We show that when interparticle bonds are O(kT), thermal fluctuations enable ongoing particle migration and a (logarithmically) slow march toward full phase separation. Our work reveals the surprising result that gel yield can occur with no network rupture; rather, localized re-entrant liquid regions permit yield and flow, and subsequent resolidification. Analysis of the evolving osmotic pressure and potential energy reveals the interplay between bond dynamics and external stress that underlies mechanical yield, and provides a compelling connection to stress-activated phase separation. I will show that external fields and forces open a pathway of escape from arrested phases toward equilibrium, and I will propose a ‘non-equilibrium phase diagram as the foundation for “phase mechanics”, a new view of states of arrested colloidal matter.


Tarek Zohdi
Dept. of Mech. Eng., UC Berkeley
6117 Etcheverry Hall, UC, Berkeley, CA 94720-1740
Email: zohdi@berkeley.edu Phone: (510) 642-9172

Within the last decade, several industrialized countries have stressed the importance of advanced manufacturing to their economies. Specialized materials and the precise design of their properties are key factors in the processes. Specifically, particle-functionalized materials play a central role in this field, in three main ways: (1) to endow filament-based materials by adding particles to a heated binder (2) to ``functionalize'' inks by adding particles to freely flowing solvents and (3) to directly deposit particles, as dry powders, onto surfaces and then to heat them with a laser, e-beam or other external source, in order to fuse them into place. The goal of these processes is primarily to build surface structures, coatings, etc., which are extremely difficult to construct using classical manufacturing methods. The objective of this presentation is to introduce the audience to basic techniques which can allow them to rapidly develop and analyze particulate-based materials needed in new additive manufacturing processes. This presentation is broken into two main parts: continuum and discrete element approaches. The materials associated with methods (1) and (2) are closely related types of continua (particles embedded in a continuous binder) and are treated using continuum approaches. The materials in method (3), which are of a discrete particulate character, are analyzed using discrete element methods.