Research

Composites, concrete, soils, rocks, bone, and plants can be considered multiphase hierarchical structures, whose multiphase composition evolves over multiple length scales, with heterogeneities ranging from micrometers to centimeters. These structures can be designed by combining various functional properties at different length scales to achieve favorable mechanical properties at the macroscale. This can open opportunities in the fabrication of bioinspired materials, civil infrastructure applications, multiscale additive manufacturing, and biomechanical applications. 

Our goal is to address the key theoretical and computational limitations of current approaches in this direction.

Bamboo - an example of multiphase hierarchical structure (Wegst et. al, Nature  Materials, 2015)

Cantilever Benchmark: Evolution of macroscale density and eq. plastic strains with optimal material configuration at meso and microscales

Predicting self-adapting mechanisms through concurrent material and structure optimization concepts

The idea of concurrent material and structure optimization fits perfectly well with the fact that biological systems adapt their “form” (or shape/structure) against the dynamic external environment and improve the “microstructure architecture,” fulfilling the local needs imposed by physiological, phylogenetic, and reproductive constraints. We applied our framework to simulate simple optimum microstructure configurations in hierarchically graded bamboo stems, and the optimum design identified by our framework corresponds to material configurations at different scales that are observed in nature.

One of my long-term research goals is to establish a connection between computational growth and remodeling with this optimization framework. The overall structure of our framework is suitable for building a growth and adaption framework in a material and structure optimization paradigm. An apparent advantage is the possible seamless connection of the macroscale constitutive and growth evolution equations with the underlying microscales, which can allow us to reflect on the multiscale origin of adaption in biological materials.

Collaborators:  Dr. Heuschele (US Dept. of Agriculture), Prof. Kevin Smith (Agronomy and Plant Genetics), Prof. Guala (St. Anthony Falls Lab, Civil Engineering), Prof. Annor (Food Sci. & Nutrition), and Prof. Fok (Minnesota Dental Research Center for Biomat. and Biomech.)

An essential prerequisite for the applications involving hierarchical structures such as biomechanical tailoring of plant materials is to accurately relate mechanical behavior to compositional and morphological properties across different length scales.

From a multiscale analysis viewpoint, the cost for resolving hierarchical scales computationally, for instance through computational homogenization, increases exponentially with each additional scale for multiphase hierarchical materials. Moreover, biological materials often exhibit a random microstructure with complex geometric characteristics of their constituents, and only partial statistical information such as the volume fraction, the shape of constituents, and their interaction with other constituents can be derived from the data.  Due to these roadblocks, current computational approaches remain out of reach for practical engineering problems involving multiphase hierarchical structures. 

Experimental characterization and micromechanical modeling of plant stem materials

The analytical framework of continuum micromechanics provides a rigorous foundation for deriving homogenized estimates of a microheterogeneous representative volume element (RVE)  utilizing its statistical description. In this work, we developed a multiscale modeling approach within the continuum micromechanics framework to predict the macroscale stiffness and strength of multiphase hierarchical materials focusing on a broad class of plant materials.  

The hierarchical organization of plants is statistically characterized with the help of microimaging (micro-CT, scanning electron microscopy, light microscopy, transmission electron microscopy) and chemical analysis data. This statistical information provides all the data. Our micromechanics approach purely relies on this statistical information for the estimation of associated macroscale stiffness and strength properties. We validate our model against the four-point bending experiments and demonstrate that the micromechanics model provides excellent accuracy without any further phenomenological calibration. For more information, follow [Mech. Mater.], [Biomech. Model. Mechanobiol.]. 

Experimental characterization and micromechanical modeling with  validation for plant stem materials

Multiscale modeling driven in silico design of crops

The current yield loss potential of cereals due to lodging in storms is documented as 30-90%, which significantly affects the cereal crop production (Berry et al.,  Adv. in Agro., 2004). In the context of the foreseen challenges of global food shortage, plant geneticists are currently focusing on breeding crop varieties with improved lodging resistance. Conventional breeding methods are based on the identification of traits that strongly correlate with lodging, mainly through visual inspections of a large number of genetic lines in the field. 

In the case of cereals, however, the correlations between measurable traits and lodging are often unreliable and poor. Reasons include the complex anatomy and morphology of crops at multiple length scales, the complicated interaction of the traits that contribute to lodging, and environmental factors. Our micromechanical model provides an opportunity to perform "in-silico" experiments or computer simulations of lodging behavior and quantitatively relates the plant traits at different scales with the lodging behavior. 

We presented a unique collaboration among disciplines for plant science, modeling and simulations, and experimental fluid dynamics in a broader context of breeding lodging resilient wheat and oat. We established that the integration of our multiscale material model-based simulations revealed the multiscale origin of failure mechanisms leading to lodging that would not have been possible with a purely experimental approach. The gained insight from these simulations will help in devising breeding strategies for lodging-resistant cereals reducing time and labor costs. This work was supported by the MnDRIVE fellowship [report]. 

Historical Context: Norman Borlaug (Nobel Peace Prize, 1970; Padma Vibhushan, 2006), a University of Minnesota alumnus, was an instrumental figure behind the green revolution in the 1960s in developing economies such as Mexico and India. He developed semi-dwarfed lodging resistant wheat varieties that led to a major boost in wheat production and improved food security in these nations. I grew up in the region that was beneficiary of his work, which in combination with a very interesting and relatable foreseen application of computational mechanics led my interest in this project. 

Fracture mechanics inspired "analysis-suitable" segmentation

Fully automatic patient-specific image-to-simulation (Im2Sim) framework

Towards our goal of patient-specific simulations, we have developed a prototype image-to-simulation (Im2Sim) software workflow that can be easily integrated with current CT/Xray postprocessing computer infrastructure and does not require highly trained expert supervision.  

In the next steps, we will evaluate the predictive accuracy of osteoporosis predictions based on pre-diagnosed clinical data sets and work towards establishing simulation-based clinical protocols in consultation with medical experts.