Our Research
At ICARI, we believe that the most impactful innovations emerge at the intersection of disciplines.
Our research is organized into three complementary pillars, with foundational strength in computational engineering and information technology. By integrating these domains, we create a unified research environment where complex problems can be addressed from multiple angles simultaneously — from fundamental algorithm development to applied system simulation.
Computational Engineering and Simulation
Computational
Engineering and Simulation
Our engineering research is conducted entirely through advanced computational modelling and simulation, creating virtual prototypes and digital twins of complex physical systems. We focus on developing high-fidelity in-silico environments to predict performance, optimize design, and de-risk innovation before any physical implementation.
Key Research Areas
Computational Materials and Multiphysics Simulation
- Finite Element Analysis (FEA): Structural integrity assessment, stress-strain analysis, and failure prediction for mechanical components, aerospace structures, and biomedical implants.
- Computational Fluid Dynamics (CFD): Simulating fluid flow, heat transfer, and chemical reactions in energy systems, turbomachinery, microfluidics, and environmental dispersion.
- Multiphysics Coupling: Integrating thermal, electrical, magnetic, and mechanical phenomena to model complex interactions in batteries, fuel cells, sensors, and actuators.
- Molecular Dynamics (MD): Understanding material behaviour at the atomic scale for designing novel polymers, nanocomposites, and crystalline materials.
Digital Twin and Systems Modelling
- Industrial Process Digital Twins: Simulating manufacturing lines, chemical plants, and power generation facilities to optimize throughput, predict maintenance needs, and reduce energy consumption.
- Energy System Modelling: Creating dynamic models of electrical grids, renewable energy farms, and energy storage systems to support the transition to sustainable infrastructure.
- Infrastructure Digital Twins: Developing models of bridges, tunnels, pipelines, and urban environments for structural health monitoring, risk assessment, and lifecycle management.
Computational Design and Optimization
- Generative Design: Algorithm-driven generation of optimized geometries based on performance requirements, manufacturing constraints, and material properties.
- Topology Optimization: Determining optimal material distribution within a design space to achieve maximum stiffness, minimum weight, or other target objectives.
- Surrogate Modelling and Optimization: Building computationally efficient approximations of expensive simulations to enable rapid exploration of design spaces using genetic algorithms and Bayesian optimization.
Multiscale and Multidisciplinary Simulation
- Coupled Microscale-Macroscale Models: Linking molecular dynamics simulations with continuum-level FEA to predict material failure mechanisms originating at the nanoscale.
- Fluid-Structure Interaction (FSI): Simulating interaction between deformable structures and surrounding fluid flows for aerospace, biomedical, and civil engineering applications.
Information Technology and Artificial Intelligence
This division develops the core computational architectures, algorithms, and data frameworks that power the centre’s digital research ecosystem. We specialize in creating scalable, secure, and intelligent systems for data synthesis, knowledge extraction, and autonomous decision-making.
Key Research Areas
Foundations of Distributed and Intelligent Systems
- Edge Computing Paradigms: Architectures that process data closer to its source (IoT sensors, industrial equipment) to reduce latency, bandwidth requirements, and cloud dependency.
- Federated Learning Frameworks: Systems enabling collaborative model training across decentralized datasets without data leaving its original location—critical for privacy-sensitive applications.
- Secure Multi-Party Computation (SMPC): Cryptographic protocols allowing multiple parties to jointly compute functions over private inputs while maintaining confidentiality.
- High-Performance Computing (HPC): Designing and optimizing algorithms for parallel execution on clusters, cloud infrastructure, and GPU-accelerated systems.
AI for Simulation, Analysis, and Decision Support
- AI Agents and Reinforcement Learning: Autonomous agents that interact with simulations and real-world systems to learn optimal policies for industrial process control, energy grid management, and design space exploration.
- Natural Language Processing (NLP) and Knowledge Graphs: Systems that ingest, organize, and synthesize information from millions of scientific documents to accelerate literature-based discovery.
- Computer Vision for Industrial Applications: Deep learning for defect detection in manufacturing, structural health monitoring, analysis of scientific imagery, and autonomous systems.
- Explainable AI (XAI): Methods to make AI model decisions interpretable and transparent, enabling trust and adoption in regulated domains.
Predictive Analytics and Complex Systems Informatics
- Graph Neural Networks (GNNs): Neural network architectures for graph-structured data applied to molecular property prediction, materials discovery, network analysis, and interconnected systems modelling.
- Time Series Analysis and Forecasting: Advanced methods for predicting behaviour of dynamic systems—equipment degradation, energy demand, environmental conditions, and financial markets.
- Anomaly Detection and Fault Diagnosis: Algorithms identifying unusual patterns or incipient failures in streaming data from sensors, industrial equipment, and network monitoring systems.
- Causal Machine Learning: Moving beyond correlation to understand cause-effect relationships using causal discovery, double/debiased machine learning, and counterfactual inference.
Software Engineering and Research Infrastructure
- Research Software Engineering: Applying software engineering best practices (version control, testing, documentation, continuous integration) to ensure reliable, reusable research code.
- Data Management and Curation: Systems for managing research data throughout its lifecycle, including secure storage, versioning, metadata annotation, and FAIR principles compliance.
- API Development and System Integration: Building interfaces enabling integration of research tools into external systems for technology transfer and real-world adoption.
Digital Public Health and Computational Epidemiology
Complementing our core strengths in engineering and information technology, we apply computational methods to understand health systems, predict outcomes, and inform evidence-based policy.
Key Research Areas
- Health Systems Informatics and Modelling: Agent-based modelling and discrete-event simulation of healthcare delivery to test care pathways, resource allocation strategies, and digital health interventions.
- Computational Epidemiology: Disease transmission models (mechanistic and statistical) for outbreak forecasting and intervention evaluation.
- Data Science for Health Equity: Causal inference models and spatial analytics applied to administrative and social determinants of health data to identify disparities and evaluate policy impacts.
- Risk Stratification and Predictive Health Analytics: Machine learning on integrated datasets to predict individual and population health trajectories.
The Integrative Advantage
The unique strength of ICARI lies in the seamless integration of our computational domains into a unified research pipeline. A typical project might unfold as follows:
- Computational engineers create a digital twin of an industrial facility or urban environment
- IT and AI researchers integrate real-time sensor data and develop machine learning models for predictive analytics or autonomous control
- Data scientists apply causal inference to understand how system parameters affect outcomes, informing operational decisions or policy recommendations
This end-to-end capability — from fundamental algorithm development to applied decision support — defines our approach to creating impactful, evidence-based innovation in a virtual space.
Collaborate with Us
ICARI actively seeks partnerships with industry, government, and academic institutions. We offer:
- Collaborative Research Projects: Joint development and execution of research programs aligned with partner needs
- Technology Transfer: Licensing of software, algorithms, and methodologies developed at the centre
- Consulting and Advisory Services: Expert guidance on computational modelling, AI strategy, and data infrastructure
- Training and Professional Development: Courses and workshops in computational methods, AI, and research software engineering
