InDeS Platform
What Is InDeS?
Multi-Scale Inverse Design Software (InDeS) is AIM EFRC’s flagship AI-powered platform for the discovery and optimization of composite materials and manufacturing processes. InDeS starts with a product’s target performance and computes backward, identifying optimal material architectures and process parameters within energy, economy, and environmental (E³) constraints.
1. AI-Powered Inverse Design
InDeS leverages invertible neural networks (INNs) to map product performance requirements to the optimal combination of materials and manufacturing processes. These models enable bidirectional prediction, meaning users can not only simulate how a material will perform but also generate designs to meet specific target properties. This approach allows for rapid, on-demand exploration of composite architectures, reducing reliance on trial-and-error experimentation and significantly accelerating the design cycle.
2. Multi-Scale Modeling
The platform spans a wide range of modeling scales, from atomic-level simulations to full part-scale predictions. It integrates coarse-grained molecular dynamics (CGMD), particle-based simulations like dissipative particle dynamics (DPD), and continuum-based models such as finite element analysis. This multi-scale framework allows InDeS to capture material behavior across hierarchical length scales, enabling accurate predictions of performance based on microstructure, chemistry, and processing conditions.
3. E³-Constrained Decision Making
InDeS is designed to optimize composite design under energy, economy, and environment (E³) constraints. It incorporates Life Cycle Assessment (LCA) tools directly into the design workflow to account for factors such as energy consumption, cost, carbon footprint, and end-of-life impact. This ensures that every design generated by the platform is not only technically sound but also aligned with sustainability and economic goals.
4. Closed-Loop Innovation Pipeline
The platform integrates AIM EFRC’s three research thrusts—simulation (Thrust I), AI modeling (Thrust II), and experimental validation (Thrust III)—into a unified, closed-loop system. Physical data from AIM’s test beds (e.g., thermoforming, 3D printing, injection molding) are used to continually validate and improve model accuracy. This feedback loop strengthens the reliability of predictions and supports continuous refinement of materials and process designs in real-world conditions.​
Real-World Validation
InDeS is being actively validated across AIM’s experimental test beds:
-
Thermoforming
-
3D Printing
-
Injection Molding
-
Ultrasonic Welding
Each test bed serves as a proving ground where AI-generated designs are translated into physical components and evaluated against real-world performance metrics.
1. Reduced Warpage in Injection-Molded Parts
InDeS is being used to address key manufacturing challenges, such as minimizing warpage in injection-molded composite parts. By integrating AI models with high-fidelity simulation data, the platform allows engineers to predict and control deformation behavior, resulting in more dimensionally stable components and reduced post-processing needs.
2. Optimized Composite Formulations Under E³ Constraints
The platform supports the design of composite formulations that are not only high-performing but also energy-efficient, cost-effective, and environmentally sustainable. By embedding E³ constraints into the optimization process, InDeS helps industries meet both performance and sustainability targets without compromise.
3. Data-Efficient Alternatives to Trial-and-Error Methods
InDeS replaces traditional, labor-intensive trial-and-error methods with intelligent, AI-guided workflows. This significantly reduces material waste, testing time, and development costs, while increasing the speed and precision of innovation across sectors like aerospace, automotive, and consumer products.
4. Industry Collaboration for Product Prototyping and Scale-Up
The platform is already being used in collaborative projects with industry partners to design, prototype, and refine composite components. Its predictive capabilities are helping companies de-risk new product development and accelerate the path from concept to manufacturing scale-up.
What’s Coming Next?
1. Unified Interface with Standardized Inputs and Outputs
A streamlined version of InDeS is currently under development, offering users a consolidated interface for seamless access to its suite of tools. This version will feature standardized input/output formats, making it easier to integrate into existing digital workflows and collaborate across teams.
2. Embedded Training, Documentation, and Tutorials
To ensure accessibility and ease of use, the upcoming version will include built-in training modules, user documentation, and interactive tutorials. These resources will support both technical users and new adopters, accelerating platform onboarding and application.
3. Cloud-Deployable Modules Hosted on GitHub and CCIT Portals
The platform is being engineered for deployment through cloud infrastructure, enabling scalable, remote access. Toolkits will be available on GitHub and institutional portals such as CCIT, ensuring broad accessibility for research teams, industry collaborators, and educators.
4. Public Demonstration-Ready Version by Early 2026
A fully operational, public-facing version of the InDeS platform is planned for release by early 2026. This milestone will mark the transition from internal development to broader dissemination, allowing stakeholders across sectors to directly engage with AIM’s AI-driven composite design infrastructure.
Interested in Collaborating?
Want to explore how InDeS can support your research or product development?
Contact Dr. Srikanth Pilla for demo access, partnerships, or training opportunities at spilla [at] udel [dot] edu