Integrated Computational Materials Engineering

Process-Structure-Property-Performance

At the heart of any ICME framework are the relationships between chemistry, process, structure, and properties, which dictate how a component will perform in a given application. This kind of structural hierarchy stems from the philosophies and writings of Smith [2] and Cohen [3]. Figure 1 below has been adapted from Olson [4].

The key links in this paradigm are process-structures-properties-performance. To fully predict how a material will perform in a given component, it is critical to understand the “cause and effect” relationships like how the composition and processing route influence microstructure, and how microstructure will influence material properties. Linking microstructural models to micro and macroscopic process simulations allows for process variability to be included in the component definition and could also be used in an inverse way to design materials for specific performance requirements and to predict location specific properties.

To describe these relationships, both models (simulation tools) and materials data are required. Since ICME spans the entire process and manufacturing chain, a variety of models and software tools that describe multiple length scales (microscale, mesoscale, macroscale and up to the component level) are needed.

Materials Data for ICME

Software employed in an ICME framework typically requires some materials data. Common sources of data include handbooks, but these are usually limited to only the most common alloys, are rarely temperature dependent, and do not capture composition variation or contain data for novel/experimental materials. This limits the accuracy of the simulations.

To assess the sensitivity to temperature-dependent values, Smith et al [5] compared the effect of using a constant handbook value on the specific heat capacity for 316L, with temperature dependent CALPHAD generated data. They concluded that it resulted in an over 500K difference in peak melt pool temperature and two times difference in the size of the melt pool. Figure 2 illustrates a comparison of the apparent heat capacity calculated using Thermo-Calc with a constant handbook value found from the literature.

Thermo-Calc also offers more than just look-up-tables of data. Many properties can be calculated under different conditions that reflect differences in processing. For example, properties that may vary with cooling rates or through non-isothermal heat treatments, and so on.

A list of these properties that can be predicted using Thermo-Calc can be found here:

Integration (Workflows, Frameworks, Interoperability)

The bridging of inputs and outputs between models and codes are what makes the “I” in ICME. In our experience there is no one single solution to developing an ICME framework and each implementation is unique, depending on the tools that are being used and the goals of the simulation framework.

At Thermo-Calc Software, we develop software tools that are fundamental to ICME frameworks and methodologies. Using Python™ as a ‘binder’ language coupled with the TC-Python API, Thermo-Calc offers a fully customizable foundation for ICME frameworks, calculating chemistry and process specific thermophysical properties that can be fed into other ICME tools like finite element codes, property models, or even machine learning models. Outputs like thermal history can be fed back into Thermo-Calc to make a prediction about the microstructure of the material.

To get started implementing an ICME framework, taking a systems level approach can be very useful to help an engineering team decide what variables are most important, and what properties to optimize [4]. These can be represented through a material systems design chart, which is a first-order representation of the full material system that helps to explicitly depict the microstructural subsystems that are controlling the properties of interest and the substages of processing that are governing the evolution of each subsystem. Such systems design charts are usually made for each specific application, but as an example to illustrate the concept, a more generalized materials system design chart is shown in Figure 3.

ICME for Processes

Simulating integrated processes (i.e. casting, additive, hot rolling) from start to finish requires linking simulation tools across length scales. An ICME framework for processes should account not only for process variables, but chemistry variables as well, to fully capture microstructure and predict mechanical properties. One example of such a framework is detailed in a paper by Nomoto et al. They used a combination of CALPHAD, finite element modeling, and phase field modeling to simulate the microstructure and properties during casting and subsequent hot rolling of a duplex stainless steel. More details can be found in the paper:

Another example can be found in a paper by Luo et al. They used the temperature history calculated by finite-element models as the input data for DICTRA solidification simulations, to predict location-specific microstructures in the real industrial casting processes:

In another example, Wang and Xiong used the systems design approach to develop an ICME framework to optimize an HSLA-115 chemistry for the additive manufacturing process. They used a collection of CALPHAD based models to optimize for hot/cold cracking resistance, low temperature toughness, while maintaining a yield strength of over 115 ksi: 

ICME for Components

For high value add components, such as turbine blades, an ICME framework can accelerate the entire design and manufacturing process.  Methodologies such as the accelerated insertion of materials (AIM) methodology [6] can reduce the number of tests required to forecast the 1% property minimums on a probability density curve. Bolcavage et al. have published a thorough review of ICME applications to aerospace components and have detailed the impact they have had in accelerating component design. More details can be found in the paper:

References

1. National Research Council (2008). Integrated Computational Materials Engineering: A Transformational Discipline for Improved Competitiveness and National Security.
   https://doi.org/10.17226/12199 
2. Smith, Cyril Stanley; Goodstein, David (1984). A Search for Structure. Selected Essays on Science, Art and History. American Journal of Physics, 52, pp 94-95.
   https://doi.org/10.1119/1.13843
3. Cohen, Morris (1976). Unknowables in the essence of materials science and engineering. Materials Science and Engineering, 25, pp 3-4.
   https://doi.org/10.1016/0025-5416(76)90043-4
4. Olson, G.B. (1997). Computational Design of Hierarchically Structured Materials. Science, Vol. 277, Iss 5330, pp 1237-1242.
   https://doi.org/10.1126/science.277.5330.1237
5. Smith, Jacob; Xiong, Wei; Yan, Wentao; Lin, Stephen; Cheng, Puikei; Kafka, Orion L; Wagner, Gregory J.; Cao, Jian; Liu, Wing Kam (2016). Linking process, structure, property, and performance for metal-based additive manufacturing: computational approaches with experimental support. Computational Mechanics, 57, pp 583-610.
   https://doi.org/10.1007/s00466-015-1240-4
6. National Research Council (2004). Accelerating Technology Transition: Bridging the Valley of Death for Materials and Processes in Defense Systems.
   https://doi.org/10.17226/11108