Digital Twin – Guide for Students

The research topic focuses on developing a cloud-based digital twin architecture for real-time monitoring, forecasting, and optimization of physical systems. This architecture integrates physical simulators with real-world data streams, collects sensor data from IoT devices, processes large volumes of time-series data, and performs machine learning pipelines on cloud platforms, implementing closed-loop control systems for autonomous decision-making. Applications span diverse domains including renewable energy systems (solar farms, wind turbines, virtual power plants), smart city infrastructure (traffic flow optimization, energy grid management), and robotics.

This website is a reference guide to the tools and technologies for building a cloud-based digital twin supported by a solid MLOps infrastructure. Required knowledge includes Python programming, core ML libraries (pandas, numpy, scikit-learn), basic data preprocessing, model training, evaluation metrics, and Git fundamentals. You’re expected to quickly learn Docker, Databases (TimescaleDB), Cloud Deployment (AWS), MLOps (MLflow), and Streaming (Kafka). For a concrete example, explore the Virtual Power Plant (VPP) project.

Core Infrastructure & Cloud Services

Container Orchestration & Deployment

Cloud Platforms

Specific AWS Services

Data Streaming, Brokers & Databases

Real-time Data Processing

Message Queues

Databases

Machine Learning & MLOps

ML Frameworks & Libraries

MLOps Platforms

Backend & Frontend Development

API Frameworks (Backend)

Data Processing

Web Frameworks (Frontend)

IoT & Hardware Integration

Microcontrollers & SBCs

IoT Communication Protocols

IoT Libraries (Python)

Simulation, Optimization & Monitoring

Optimization & Programming

Energy & Physical System Simulation

Monitoring & Observability

Visualization

3D Visualization & DT Platforms

Recommended Learning Path

  1. Start with Infrastructure: Docker, Docker Compose, basic cloud services.
  2. Data Streaming: Kafka fundamentals and real-time data processing.
  3. IoT Integration: Raspberry Pi, sensors, MQTT protocol.
  4. Database Design: TimescaleDB for time-series data.
  5. ML Pipeline: MLflow for experiment tracking and model deployment.
  6. API Development: FastAPI for backend services.
  7. Frontend Visualization: React + Chart.js for real-time dashboards.
  8. Optimization: PuLP for decision-making algorithms.
  9. Monitoring: Prometheus + Grafana for system observability.