About STAC

Project Information Center

Hydrogen Education

Previous Solicitations

News

Presentations

Contact

Home

STAC

Project Information Center

STATEMENT OF WORK

Project Title:

Preventing Solidification Defects in Large Superalloy Castings Used in Advance Electric Power Systems

Contractor:

West Virginia University Research Corp. on behalf of West Virginia University
West Virginia University Mechanical & Aerospace Engineering
PO Box 6106
Morgantown, WV 26506
(304) 293-3111 x2324

Program Area:

Materials Sciences

Partners:

GE Energy
Special Metals Corp.
Pennsylvania State University

Project Description:

The objectives of WVU's project is to study macrosegregation in superalloy remelting processes.  The ultimate goal is to develop a predictive technology that can be applied commercially to prevent solidification defects for large superalloy castings used in advanced electric power systems.

Management Plan

The unique nature of STAC requires that projects be supported by multiple State entities, and to the extent necessary any other entity. As indicated in the STAC Agreement, it is the Contractor’s responsibility to coordinate the execution of work under the Contract, incorporated by reference hereto. Contractor, in conjunction with the other State entities, and to the extent necessary any other entity, shall conduct the project in accordance with the Management Plan –  described below.

Administration

Project management will be objective-oriented with an emphasis on successfully completing proposed tasks on a timely manner.  Direct interactions and communication between project team members through conference calls and emails are customary.  West Virginia University (WVU) will visit Special Metals Corp. (SMC) for there will be at least two conference calls per year to review the progress and decide the research direction for the next half year.  As the PI institution, WVU will visit GE Energy at least once per year and SMC twice per year.  Based on prior collaboration experiences, the research team expects to make significant progress on one of the most critical challenges in the development of large superalloy casting for power generation industries.

Task 1: Solidification Modeling

1.a. Thermodynamics database and partition coefficients

Thermodynamic modeling (LiuZi-Kui/PSU)

The goal of this subtask is to obtain the partition coefficient of the model systems Ni-Cr-Nb, Ni-Cr-Fe, Ni-Cr-Ti, Ni-Cr-Fe-Nb, and Ni-base superalloys 706 & 718, through first-principle calculation and thermodynamic modeling.

First-principles calculations, based on density functional theory, require only knowledge of the atomic species and crystal structure, and hence are predictive in nature.  The first-principles methods to be used here include the full-potential linearized augmented plane wave (FLAPW) method (the "benchmark" for accuracy in density-functional-based methods), and the highly efficient Vienna ab-initio Simulation Package (VASP).  These methods yield quantities related to the electronic structure and total energy of a given system, and can be used to accurately predict phase stabilities of compounds at 0 K.  By combining first-principles techniques with statistical mechanics methods such as the cluster expansion/Monte Carlo approach, one can explore, without any fitting parameters, thermodynamic phenomena such as phase transformation temperatures, phase diagrams, and short-range order.  Coupling with frozen phonon or linear response techniques opens the possibility for exploring finite-temperature vibrational effects.  Furthermore, these approaches are applicable to any phases of a given alloy system, not only the equilibrium ones.  Hence, first-principles techniques can provide a method to obtain properties of metastable phases, which are often crucial to microstructure evolutions before equilibrium states are reached, but can be difficult to isolate and study experimentally.  In recent years, we have been actively working on the above topics.  Furthermore, we are using the Alloy-Theoretic Automated Toolkit (ATAT) to develop efficient routes to predict the enthalpy and entropy of mixing for solid solutions phases through the special quasirandom structures (SQS).

As engineering materials are typically multicomponent, it is not computationally tractable, for the foreseeable future, to use first-principles calculations alone to determine the total free energy directly for multicomponent systems, at least not with accuracy comparable to experimental phase diagram measurements.  On the other hand, semi-empirical methods based on  the CALPHAD approach have been very successful in predicting the phase equilibria of multicomponent commercial alloys.  CALPHAD uses a wealth of thermodynamic information, some taken from first-principles calculations, to model multicomponent systems.  In the CALPHAD approach, a large number of thermochemical and phase equilibrium data are used to extract parameters describing the alloy energetics, which are then used in calculations of thermodynamic properties, phase equilibria, phase diagrams, and phase transformations through the minimization of free energy and calculation of thermodynamic driving forces.  The approach produces reliable phase diagrams and stability maps for complicated multicomponent commercial alloys.  CALPHAD modeling begins with the evaluation of descriptions of unary and binary systems.  By combining the constitutive binary systems and ternary experimental data, ternary interactions and Gibbs energy of ternary phases are obtained.  Materials databases thus developed cover the whole composition and temperature ranges, including experimentally uninvestigated regions.  In this approach, properties of individual phases are modeled, and the model parameters are collected in databases.  The modeling of the Gibbs energy of individual phases and the coupling of phase diagram and thermochemistry are the keys to developing unambiguous thermodynamic descriptions of multi-component materials with sound fundamentals and predictive power because the two sets of data are deduced from the Gibbs energy of individual phases under given constraints.  Our first-principles calculations provide critically needed thermochemical data, which are difficult to obtain experimentally.

Thermodynamic modeling will initially focus on the model systems Ni-Cr-Nb, Ni-Cr-Fe, Ni-Cr-Ti, and Ni-Cr-Fe-Nb.  In our recent work supported by NASA, we worked on the Ni-Al-Mo-Ta system.  Integrating with the first-principles calculations, we developed a new phase diagram for the Ni-Mo system.  We also have a project on Ni-Al-Cr-Pt supported by CSIR-Manufacturing and Materials Technology in South Africa.  Therefore, by combining thermodynamic modeling available in the literature and from our related projects, we will extrapolate them to IN706 and IN718.

Experimental investigation on partition coefficients (Liu/WVU):

As mentioned the "literature review" section, because of the difficulty in conducting experiments, there is little experimental data on partition coefficients available currently.  During recent years, WVU has successfully developed a modified DTA quenching testing + square-mesh systematic point count metallographic technique to measure the solute redistribution profiles in Ni-base superalloys.  The interdendritic liquid composition, equilibrium partition coefficient, and dendrite core composition of several commercial available superalloys and Ni-Cr-Nb, Ni-Cr-Fe, Ni-Cr-Ti, and Ni-Cr-Fe-Nb model alloys have been measured.

In the proposed project, the square-meshing technique and modified DTA interrupt quenching will be employed to study the solidification characteristics of IN706 and IN718, and several model Nb-, or Ti- content model alloys.  The solidification characteristics include partition coefficient, liquid composition, core dendrite composition, and liquid fraction as the function of temperature.  The experimental outcome will be used to calibrate thermodynamic calculations.

1.b.  Interaction between minor elements and major elements

Thermodynamic modeling (LiuZi-Kui/PSU)

With the low concentration of minor elements, we will use supercells in first-principles calculations to study their interactions with major elements.  We will first carry out binary calculations between minor elements and major elements.  Then we will use binary SQS between Ni and major elements to study the interaction and behavior of minor elements in binary solutions.

1.c. Kinetic effect on partition coefficients during solidification

Thermodynamic + Diffusion modeling (LiuZi-Kui/PSU)

In this subtask, we will use the Dictra code to simulate the solidification with diffusion in both liquid and solid phases considered. Dictra code takes thermodynamic properties and phases directly from the thermodynamic database developed in the task listed as 1.1 and 1.2.  It was used in our recent research on liquidation in Al-Cu alloys.

Experimental study (Liu/WVU):

During the past years, WVU has developed an interrupt quenching method to study the kinetic effect on solidification by modifying a DTA machine.  In this subtask, the interrupt quenching tests with various cooling rate will be conducted.  The cooling rate effect on partition coefficient and other solidification characters including undercooling, liquid fraction as the function of temperature and eutectic fraction, will be investigated.

1.d. Setup of solidification modeling (Chang/WVU, Yang/SMC)

During the past several years, under the sponsorship of SMC, WVU systematically studied the freckle formation in superalloy ingots.  Several freckle criteria in the literature were examined under experimental data of the upward directional solidification of Pb-Sn-Sb systems.  It was found that the freckle criterion given by Flemings gives the best description of the freckle tendency.  Furthermore, WVU developed a Rayleigh number from Flemings' criterion, which ahs the following formula:

Ra=∆ρgΠ _g 1 

     νf_L   R

Where ∆ρ is the density, g is the gravity, Π is the permeability, ν is the liquid viscosity, and fL is the liquid fraction, and R is the crystal growth speed.  The permeability Π was calculated by Poirier:

Π = 3.75 x 10^-4f_L²d₁²(m²)

where d₁ is the primary dendrite arm spacing (PDAS).  It was found that the critical Rayleigh number R_er is unity.

Furthermore, WVU applied this Rayleigh number, along with other thermodynamic approach to model the solidification behaviors of Ni-base superalloys.  The model was successfully employed to predict the ingot size capability of the industry melting shops, which is limited mainly by freckle defects.

It was clear that all the parameters determining Rayleigh number are related to partition coefficients of alloying elements.  In the past, WVU used the Thermocalc software and corresponding database to get the values of partition coefficients.  In this subtask of the proposed project, the partition coefficients will be obtained from the subtasks 1.1. to 1.3 and substituted into the WVU model.  The new model with improved partition coefficients will be employed to simulate the solidification of superalloys and predict macrosegregation behavior.  The simulation results will be calibrated by tasks 2 and 3.

Task 2: Directional Solidification Verification

The goal of this task is to experimentally study the effects of composition and kinetic parameters during solidification on the formation of macrosegregation defects, especially freckles.  A directional solidification (DS) furnace will be built and employed in the investigations.

2.a. Verification for model alloys (Liu/WVU)

Based on commonly used commercial superalloys, a series of simplified alloyed compositions with different eutectic phases, such as Laves phase, δ-Ni₃Nb, or η-Ni₃Ti, will be prepared to verify and modify phase diagram database and solidification modeling.

2.b. Effect of processing parameters (Liu/WVU)

Direction solidification tests will be conducted to study the influence of solidification conditions, including cooling rate, thermal gradient, isotherm speed, solidification range, and local solidification time, on the freckle & center segregation formation.

Task 3: Verification for Commercial Alloys (WVU/SMC/GE Energy)

The calibrated solidification model will be applied to analyze the industrial data provided by SMC.  Composition and processing effects will be characterized so that a comprehensive database is available in usable format. An alloy index of macrosegregation tendency can be determined for complex alloy compositions.  The guideline for composition control and processing optimization will be obtained from the model and applied in the industrial VAR/ESR practice.

Project Tasks, Status, and Deliverables

Last Updated: 10/24/06