MAE professor receives federal grants for model-based research

Posted: July 19, 2021

Ohio State professor Soheil Soghrati recently received two federal grants for his research totaling $1 million. Soghrati is an associate professor of Mechanical and Aerospace Engineering with a joint appointment in the Department of Materials Science and Engineering. He is the director of Automated Computational Mechanics Laboratory (ACML), where he and his research team focus on developing novel computational techniques and artificial intelligence (AI) algorithms for modeling synthetic and biomaterials.


Soghrati’s first grant comes from the Air Force Office of Scientific Research (AFOSR) for his work titled, “Mesh generation and AI-enhanced algorithms for modeling complex materials systems.” The goals of this project will be to allow a parallel mesh generation algorithm named CISAMR the ability for automated modeling of 3D crack growth problems. The research also aims to develop a cost-efficient, deep learning based framework to simulate mechanical behaviors of massive structures with complex microstructures.

According to Soghrati, the project will expand on capabilities developed during a recent AFSOR project, where Soghrati’s research group developed and automated computational framework relying on CISAMR algorithms for modeling heterogenous materials.

Besides modeling materials systems that are of interest to the U.S. Air Force, CISAMR has extensively been used in other projects in ACML for modeling a variety of composite materials used for aerospace and automotive applications,” Soghrati said.

The ACML has also licensed the CISAMR code for use by companies such as Ford and Owens Corning.

Examples of material modeling
Examples of CISAMR modeling capabilities

In this project, Soghrati’s team is looking to expand CISAMR to address two types of challenging problems in the development of next-generation composite materials and light-weight metals used by the Air Force.

First, the ACML will add new capabilities to CISAMR, to automate the simulation of complex 3D crack growth problems. The new meshing capabilities developed by ACML for modeling facture problems also have the ability to be integrated with commercial FE software, and delivered to the Air Force Research Laboratory to facilitate related research efforts.

The project’s second phase involves developing a new AI-enhanced computational framework. This will combine advanced deep learning algorithms and domain decomposition techniques for simulating the mechanical behavior of materials and structures.

Soghrati said this modeling capability could reduce computational costs associated with massive simulations, and also find application in other areas.

This method will have several other applications besides the field of computational materials engineering such as augmented reality and digital manufacturing, where AI can be used as a surrogate to analytical or computational methods to simulate the mechanical behavior of a system in real time,” Soghrati said.

CISAMR model
CISAMR modeling of the growth of multiple interacting cracks

The second grant Soghrati received is on a collaborative project with North Carolina State University (NCSU). The project, “Optimization of self-healing fiber reinforced polymer composites via convolutional neural networks,” received funding from the Strategic Environmental Research and Development Program (SERDP), a Department of Defense (DoD) sub-program.

The project provides a perfect collaborative environment between an experimental research team (NCSU) and a computational research group (ACML) to achieve ambitious objectives of this project,” Soghrati said.

The NCSU team has already made significant advances in the manufacturing and testing such self-repaired composites. But achieving an improved design capable of complete restoration of mechanical properties under perpetual healing cycles would not be feasible without advanced AI-enhanced modeling capabilities.

Soghrati’s research team will lead this computational design effort using advanced AI algorithms, which involves optimizing the shape and thickness of a 3D printed self-healing thermoplastic polymer embedded in woven composite microstructure to achieve this goal.

The research project aims to develop a sustainable self-healing composite system capable of complete restoration in interlaminar fracture resistance, without compromising mechanical properties.

“Internal delamination damage in fiber-reinforced composites is difficult to detect and nearly impossible to repair by conventional methods,” Soghrati said. “To date, this failure mechanism remains one of the most significant factors limiting the reliability and leads to wasteful design/replacement of composites in lightweight structures.”

The team is taking a collaborative approach by merging the fields of polymer mechanics and chemistry, emergent manufacturing, advanced computing, and deep learning. 

Soghrati said the work being done has the potential to benefit the Department of Defense, and the environment by eliminating the need for costly inspection, reducing overall maintenance, and enhancing structural fiber-composite reliability and performance.

The proposed developments could bring advancement in defense sectors including aerospace, civil, and naval. The developments also have the potential to improve life protection systems such as body, vehicle, and structure armor, by allowing them to endure higher loadings and increased numbers of impact events.

“The resulting self-healing, structural composites will increase safety, resilience, and reliability,” said Soghrati. “This breakthrough technology will also enhance tactical capabilities for competitive advantage in DoD operations while helping to preserve the as-built environment.”  

Categories: FacultyResearch