In this project, design sensitivity and numerical optimization methods are combined with finite element analysis of solidification processes to improve the design of castings. Criterion functions that describe casting soundness are evaluated. Sensitivities are used to modify cast shape and process parameters to improve the product.

In this project, analysis is performed to examine the solidification process in investment castings. After the analysis, the sensitivity of casting quality criterion functions to processing variables is used to improve the product design. Both 2-D and 3-D parts are being considered.

In this project, models for cast iron solidification are used to predict the behavior of casting processes. The information in these models is then used to systematically improve product design.

Various industrial sponsored research is being carried out for manufacturing applications of laser processing. These applications include laser cutting and the role of nozzle design on pressure distribution for laser cutting processes, laser heat treatment, earth-moving vehicle components, laser surface treatment of bearing steels, laser cladding with composite powder, low-energy input repair of nickel superalloys by laser cladding with convective nozzle, and welding and cladding of light alloys.

Next generation diode-pumped YAG lasers up to 6 kW maximum average power are being designed and built under this technology reinvestment program led by TRW. The major thrust of UIUC researchers is to develop process modeling, process diagnostics, and process development for future smart manufacturing systems. The paradigm for this effort is to apply atomic level understanding of the process to real-life application so that the lead time from concept to production is reduced.

Three-dimensional metallic objects are produced by laser cladding from the CAD/CAM electronic database. The materials are close to 100% density with mechanical properties similar to parent materials. This research focuses on developing a fundamental understanding of the relationship between the laser processing parameters, microstructures, mechanical properties, and deposit thickness. Preliminary results show that deposition dimension can be controlled up to 25 µm in thickness and 500 µm in width. To the best of our knowledge this is the only rapid manufacturing process which exhibited mechanical properties to the desired level for materials such as H13 steel, aluminum, and copper alloys.

Copper alloys are used extensively as electrical connectors in computers and cars. In a lifetime they may be subjected to cyclic loading over thousands of cycles. Therefore, besides electrical properties, they need to have better mechanical properties to withstand the service condition. Laser surface alloying with a Nd-YAG laser is being studied to develop surface alloys of Ni and Al to a depth of up to 80 µm for improved mechanical properties but retaining its electrical conductivity. The strategy is to charge the lathe parameters long enough by alloying to produce solid solution hardening of the surface, while the core remains pure enough to conduct electricity. Surface cladding and tin on copper alloys are also explored as a way to improve the connectors.

The objectives of this research are to identify the dominant motive forces for convection in weld pools for different ranges of process parameters and quantitative understanding of effects of fluid flow and heat flow on weld pool shape and microstructure. A laser-based visualization technique called ``reflective topography'' is developed to provide a clear image of the pool. Presently, laser absorption and fluorescence spectroscopy are employed to measure the metal vapor concentration in order to determine the discontinuity of the liquid-vapor interface (Knudsen layer). A 3-D mathematical model is developed for deep penetration welding whereas reflective topographic techniques are used to measure ``keyhole'' diameter and shape.

The chemical vapor deposition of titanium nitride using a CO₂ laser as the heat source is being examined. In order to understand the kinetics of the process, the relationship between the processing parameters and the resulting deposit thickness is being studied. Initial findings indicate that the laser chemical vapor deposition (LCVD) is two or three orders of magnitude faster than the conventional CVD process. Associated transport phenomena for LCVD is modeled to understand the kinetics of the process. LIF, multiwavelength pyrometry spectroscopic techniques are employed for on-line monitoring of temperature and chemistry of the process. Kinetics of the reaction are determined by LIF for a certain range of reactant mixtures.

The goal is to develop a science base for producing nanocrystalline intermetallics such as Nb-aluminides, known for their limited ductility, using the laser ablation technique. The specific objectives are to develop a fundamental understanding of the rate of generation of nanocrystalline niobium aluminides by laser ablation with a view to improving upon the rates available from the gas condensation technique; on-line measurement of the nucleation rate of niobium aluminide in the vapor phase; a correlation between laser process parameters with yield and properties; and to study the diffusivity of substitutional elements from group V metals for structure manipulation.

The main objective of this research is to investigate energy coupling of high-power, short-pulse lasers with solids and plasmas, and specifically to investigate using lasers for near diffraction-limited micromachining and forming. The ultimate goal is to understand and control laser beams irradiating solids to create high-intensity filaments on the order of microns. Achieving this would increase the spatial resolution and depth of field of current laser-machining and joining processes, but the greatest impact could be on the fabrication of micron-sized structures for 3-D micromachines.

Mass removed during high-power laser ablation of solids not only depends on laser energy, pulse duration and wavelength, and material and surface properties, but is also influenced by the laser spot size. The mass removal rate per unit area exponentially changes with laser energy when the spot size drops below a threshold value. For this project, the mechanisms involved with mass removal rates are investigated as a function of laser and target geometries and laser energy. Understanding laser energy coupling with solids at small spot sizes will help achieve high spatial resolution for laser sampling, machining, and joining of solids.

A comprehensive system of mathematical models is being developed for the continuous steel slab casting process. Separate models simulate fluid flow, heat flow, mass transfer, solidification, shrinkage, and stress generation within the liquid, solidifying steel shell and mold. The models are being verified with industrial measurements and applied to improve the process by better understanding how defects arise.

Clogging in Continuous Casting Tundish Nozzles
Clogging of tundish nozzles caused by deposition of alumina inclusions limits the productivity and adversely affects the quality of continuous casting. This project seeks to quantify, understand, and predict the formation of these clogs, particularly when they occur in the tundish well and require changing the tundish. Mathematical models are being developed to predict flow in the nozzle and clogging, including the effects of nozzle heat loss, argon bubble motion, and turbulence. The results are being evaluated in conjunction with plant measurements and metallographic analysis of clogged nozzles from various casters.

Fluid Flow, Heat Transfer, and Inclusion Movement in the Liquid Pool
Three-dimensional finite-difference models are being applied to predict turbulent fluid flow in the liquid pool contained by the solidifying shell. The models include two-phase flow effects of argon gas injection, dissipation of superheat, and the movement of inclusion particles. The effect of transient motion of the top ``free'' surface is currently being incorporated mathematically and investigated. The results are needed to understand and prevent flow and inclusion-related defects.

Mass Transfer in the Liquid Pool
Intermixing in the liquid pool has been found to be significant and produces rejected steel during a grade change. A fast and accurate mathematical model has been developed to simulate mixing in the tundish and the mold and accurately predict the composition profiles that develop in the solidified slabs and blooms. The model has been calibrated with 3-D model calculations and water model measurements and verified with measurements at several plants. It is being further refined to run on personal computers at the steel plants to allow efficient grade classification, minimizing both customer complaints and yield losses. The effect of casting variables are now being investigated with the model to optimize grade changes for any casting operation.

Behavior of Top Surface Flux Layers
A 3-D finite-element model of heat transfer and fluid flow has been developed of the top surface of the continuous caster to predict the thickness and behavior of the powder and flux layers. Mold powder melts to form a protective flux layer on the top surface of the steel, which helps to trap impurities and inclusions. Results indicate that the thickness of the beneficial liquid layer is highly nonuniform. Detrimental solid inclusions caused by surface turbulence are most likely in the thinnest regions, whose locations have been successfully predicted by the model. Mold flow conditions, which are affected by the inlet nozzle geometry, are found to control this layer and are being investigated.

Model of Inclusion Agglomeration and Removal
Removing inclusions from the molten steel is important to prevent surface defects in the steel products. To clarify the mechanism of inclusion removal from molten steel, a model of inclusion collision is being developed to simulate the evolution of the inclusion size distribution. The model includes the Saffman model for inclusion collision in turbulent flow and Stokes law for inclusion migration. The results should clarify the relationship between the dissipation rate of turbulent energy and the deoxidation rate. In conjunction with other models, this work will help explain how to better remove inclusions from ladle, tundish, and mold.

Inclusion Motion in the Mold and ``Pencil Pipe'' Defects
Large inclusion clusters containing alumina and/or mold flux components and associated with argon bubbles lead to occasional but severe surface defects, including ``pencil pipe'' defects in ultralow-carbon steel slabs. Several models are being used to investigate the mechanisms causing this problem and to suggest ideas for improvement of existing operating practice and SEN design. Detailed models of argon bubble and inclusion motion through the fluid steel are being used to study the inclusion build-up rate on individual bubbles. The 3-D finite-difference flow model is then used to obtain the velocity field in the mold cavity and to trace particle paths and entrapment positions for different casting conditions.

Slab Shape Prediction
The final shape of a continuous cast slab is not always uniform, or even rectangular. Of particular concern is the great variation in final slab width of certain grades, which occurs during slow-downs in operation. This creates production and logistic problems due to uncertain tonnage, in addition to quality concerns. Models are being developed to simulate heat flow, creep bulging, and associated distortion throughout the secondary cooling zone, in order to understand and quantify the mechanism for slab width variations. Model predictions are being compared with data collected from plant trials. The calibrated models eventually will be applied as a tool to predict and control slab shape in the plant under the range of casting conditions encountered.

Thermal Stress Model of Solidifying Shell
A coupled, two-dimensional, transient finite-element model has been developed to predict temperature, shrinkage, and stress development in both horizontal and vertical sections through the solidifying shell as it moves down through the caster. The model includes the effects of the volume change during phase transformation, ferrostatic pressure, the generalized plane strain stress state, the constraining influence of the mold, creep plasticity, and the dynamic effect of solidification shrinkage on heat transfer across the interfacial gap between the mold and the shell. Efforts are currently underway to apply the models to simulate and understand phenomena occurring during the early stages of solidification.

Macrosegregation and Interdendritic Crack Formation
Crack defects are caused by a combination of stress and metallurgical embrittlement, due in large part to segregation. One-dimensional mathematical models are being developed to investigate the coupled phenomena of segregation, heat flow, and stress generation during solidification, focusing on the continuous casting of steel. The models will feature transient calculation of the composition, volume, and flow of the interdendritic liquid, coupled with the CON1D and CON2D models developed previously. They will be applied to investigate the effect of segregation on the solid shell resistance and bulging of continuously cast steel slabs. The models are important to provide fundamental understanding of the complex mechanisms behind surface cracks and internal hot tears.

This project focuses on applying design sensitivity and optimization techniques to the design of both products and processes for automotive castings. In particular, we consider the design of casting gating systems for improved casting quality at reduced cost and design methods for dimensional control of cast parts.

Some processes for molding composite parts require a preform: a flat fabric of the reinforcing fibers is formed into the shape of the part. The fabric should not wrinkle or tear during preforming. We are developing computer-aided design tools for preforms. A mapping is established between the initial flat fabric and the curved 3-D preform shape. This mapping is optimized to eliminate wrinkling and tearing as much as possible. The results show whether the shape can be formed and where problems will occur, and provide the data needed to predict other properties of the finished part.

When a fiber-reinforced plastic is injected into a mold, the mold filling flow can cause preferential orientation of the fibers. This alters the mechanical properties of the part and causes many parts to warp. In the past we developed mathematical models for fiber orientation and implemented those models in flow simulations of injection and compression molding. Careful comparison with experiments showed that the governing equation for fiber orientation was not as accurate as desired, due to a closure approximation. We have developed a new class of closure approximations, called orthotropic closures. Using this, we are developing simulations with much better accuracy.

Integrated circuits are processed on wafers sliced from single crystals of silicon. The demand for better crystals has led to the use of magnetic fields to control heat and mass transfer during the growth of crystals by the Czochralski. With models for arbitrary, nonuniform axisymmetric magnetic fields, the magnetic field can be tailored to achieve optimal crystal properties.

Many optoelectronic devices require indium-phosphide (InP) crystals with small defect densities. Most crystal-growth processes involve large temperature gradients, and the associated thermal stresses produce large defect densities in the weak InP crystals. The liquid-encapsulated Kyropoulos process involves crystal growth with very small temperature gradients, so that the InP crystals have very small defect densities. A magnetic field is needed to stabilize the melt motion and to eliminate turbulent temperature fluctuations. Analytical and numerical models are being developed to guide process optimization. The purpose of the modeling effort is to complement an experimental program being conducted at an air force laboratory.