Syntaxed-IEGMS (Intelligent Energy Grid Management System)

Quantexera Hackathon 2026 · Real-time grid monitoring + ML prediction + optimization dashboard + AI Agent

Team members:

• Prashant Goundadkar
• Hussain Shaikh
• Abhishek A Pillai
• Jagadeesh Kadlimatti
• Yashsai Dessai

---

Architecture

┌─────────────────────┐     MQTT (broker.hivemq.com)      ┌─────────────────────┐
│   IoT Simulator     │  ──────────────────────────────►  │   FastAPI Backend   │
│  simulator.py       │   topic: iegms/syntaxed/grid/data  │   main.py           │
│  (publishes every   │                                    │   mqtt_client.py    │
│   2 seconds)        │ ◄───────────────────────────────   │   ml_model.py       │
│                     │   topic: .../time_sync             │   optimizer.py      │
└─────────────────────┘                                    │   agent_orchestrator│
                                                           └──────────┬──────────┘
                                                                      │ REST API
                                                                      │ (HTTP polling)
                                                           ┌──────────▼──────────┐
                                                           │   Frontend (3D)     │
                                                           │   index.html        │
                                                           │   app.js / style.css│
                                                           │   Three.js          │
                                                           └─────────────────────┘

Folder Structure

project-root/
├── backend/
│   ├── main.py                   ← FastAPI server (with AI agent endpoints)
│   ├── agent_orchestrator.py     ← AI Grid Agent (Ollama/LangChain, optional)
│   ├── mqtt_client.py            ← MQTT subscriber + in-memory store
│   ├── ml_model.py               ← Linear interpolation load predictor
│   ├── optimizer.py              ← Priority load shedding + battery logic
│   ├── test_agent.py             ← AI agent test suite
│   ├── example_client.py         ← Example API client
│   ├── requirements.txt          ← Dependencies
│   └── .env.example              ← Configuration template
├── simulator/
│   └── simulator.py              ← IoT load data publisher (MQTT)
├── frontend/
│   ├── index.html                ← 3D Simulation UI
│   ├── app.js                    ← Three.js scene + API polling
│   └── style.css                 ← Dark-mode glassmorphism design
└── README.md

---

How to Run

1. Install Python Dependencies

cd backend
pip install -r requirements.txt

2. Start the Backend (Terminal 1)

cd backend
python -m uvicorn main:app --host 0.0.0.0 --port 8000 --reload

The API will be available at http://localhost:8000 Interactive docs: http://localhost:8000/docs

3. Start the IoT Simulator (Terminal 2)

cd simulator
python simulator.py

The simulator publishes a new grid reading every 2 seconds via MQTT.

4. Open the Frontend

cd frontend
python -m http.server 3000
# Visit http://localhost:3000

The dashboard auto-refreshes every 2 seconds.

---

##AI Agent (Optional)

An Ollama + LangChain-powered AI agent is included for intelligent grid management decisions.

Setup

1. Install optional dependencies:

pip install langchain langchain-community ollama

2. Run Ollama on a machine (locally or remote):

ollama serve
ollama pull gemma3:1b

3. Configure the endpoint:

# backend/.env
OLLAMA_ENDPOINT=http://localhost:11434
OLLAMA_MODEL=mistral

AI Agent Endpoints

Method	Endpoint	Description
POST	/agent/configure	Configure Ollama endpoint/model
GET	/agent/config	Get agent configuration
POST	/agent/decide	Request AI grid decision
GET	/agent/status	Get agent status & analysis

---

🔌 Grid API Endpoints

Method	Endpoint	Description
GET	/health	Health check
GET	/data	Last 20 MQTT grid readings
GET	/optimize	Optimization decision + battery/shed status
POST	/set_time	Jump simulation to a specific hour
GET	/docs	Interactive Swagger UI

---

Optimization Logic

Condition	Status	Action
Load ≤ 900 MW	✅ Normal	Charge battery with excess
Load > 900 MW (battery helps)	⚠️ Warning	Discharge battery, alert
Load > 900 MW (still deficit)	🚨 Critical	Shed by priority (Residential → School)

Priority: Hospital > School > Industry > Residential

---

Demo Script (For Judges)

1. Open the 3D dashboard — live voxel city with real-time telemetry
2. Drag the Time Slider — watch buildings light up/dim as day turns to night
3. Power Plants — two plants feed the grid battery via blue supply lines
4. Orchestrator Log — AI takes live shedding decisions every 2 seconds
5. Battery Meter — charges when excess grid power, discharges on peak demand
6. Status badge — NORMAL → WARNING → CRITICAL as load increases

---

🛠️ Tech Stack

Layer	Technology
Backend	Python 3.11+, FastAPI, Uvicorn
ML	Custom time-interpolation model
AI Agent	LangChain + Ollama (optional)
IoT Transport	MQTT via paho-mqtt
MQTT Broker	broker.hivemq.com (public)
Frontend 3D	HTML5, Three.js r128
Styling	Vanilla CSS (dark glassmorphism)
Storage	In-memory deque (no DB required)