🎯 Wind Energy - Agentic AI Framework For Maintenance & Monitoring of Windmill

---

Overview

A multi-agent AI system designed specifically for Windmill Energy Domain that combines computer vision and natural language processing to provide comprehensive maintenance and monitoring solutions. The system processes visual inspections through AI-powered damage detection and handles technical queries through intelligent conversational agents.

---

System Architecture

---

Agentic Workflow

1. 🤖 Agent 1: Windmill Damage Detection - Computer vision agent using YOLOv8 model to detect structural damages in windmill components
2. 🤖 Agent 2: Turbine Blade Damage Detection - Computer vision agent using YOLOv8 model to detect damages in turbine blades and components
3. 🤖 Agent 3: General QnA Agent - Bedrock Agent + OpenSearch KB for general windmill-related queries
4. 🤖 Agent 4: Component QnA Agent - Bedrock Agent + OpenSearch KB for windmill component-specific information
5. 🤖 Agent 5: Maintenance QnA Agent - Bedrock Agent + OpenSearch KB for maintenance procedures and guidance

---

Technical Stack

Component	Technology Used
Frontend	Streamlit
LLM Models	Claude 3 Sonnet (Bedrock)
Computer Vision	YOLOv8 Models (SageMaker)
Vector Database	OpenSearch
Storage	Amazon S3
Model Deployment	Amazon SageMaker
Logging	Amazon CloudWatch

---

Run The Code

- Clone the repo.
- Create a py env.
- Install the requirements in the env.
- streamlit run app.py.
- Note: Only the UI will render, you wont get any output because we have dis-integrated the aws components due to cost.

---

🎬 Demo Videos

Solution.mp4

2_Demo.mp4