Skip to content
This repository was archived by the owner on Mar 7, 2026. It is now read-only.

README.md

Latest commit

 

History

History
156 lines (120 loc) · 5.98 KB

README.md

File metadata and controls

156 lines (120 loc) · 5.98 KB
Raw

Grid Balancing Swarm

Autonomous energy trading using Agent OS

"Watch 100 DER agents negotiate grid stability in real-time."

🎬 Demo Video

Grid Balancing Demo

Script (60 seconds):

[0:00] "Grid operator announces price spike at 6 PM."
[0:10] [Dashboard: 100 DER agents activate]
[0:20] [Agents negotiating: Solar-01 bids 50kW @ $0.15]
[0:30] [Consensus forming: 15 agents reach agreement]
[0:40] [Grid frequency: 60.02 Hz - stable]
[0:50] "100 agents. 30 second negotiation. Zero policy violations."

🚀 Quick Start (One Command)

cd examples/grid-balancing
cp .env.example .env
docker-compose up

# Wait 30 seconds, then open:
# → http://localhost:8082  (Demo UI)
# → http://localhost:3002  (Grafana Dashboard - admin/admin)
# → http://localhost:16688 (Jaeger Traces)

📊 Live Dashboard

┌─────────────────────────────────────────┐
│ Grid Balancing - DER Coordination       │
├─────────────────────────────────────────┤
│ DERs Active:             100            │
│ Grid Load:               450 MW         │
│ Grid Frequency:          60.02 Hz       │
│ Negotiations/sec:        1,247          │
│ Consensus Rate:          97.3%          │
│ Policy Violations:       0              │
└─────────────────────────────────────────┘

Overview

This demo simulates a distributed energy grid with 100 Distributed Energy Resources (DERs):

  • Solar panels
  • Home batteries
  • Electric vehicles

When the grid operator broadcasts a price signal, agents autonomously negotiate to balance supply and demand.

Architecture

┌─────────────────────────────────────────────────────────────────────┐
│                     GRID OPERATOR                                   │
│                  "Price spike at 6 PM"                              │
└──────────────────────────┬──────────────────────────────────────────┘
                           │ AMB (Agent Message Bus)
                           ▼
┌─────────────────────────────────────────────────────────────────────┐
│                    100 DER AGENTS                                   │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐               │
│  │ Solar-01 │ │Battery-15│ │  EV-42   │ │ Solar-99 │  ...          │
│  │ forecast │ │  trader  │ │ dispatch │ │ forecast │               │
│  └────┬─────┘ └────┬─────┘ └────┬─────┘ └────┬─────┘               │
│       │            │            │            │                      │
│       └────────────┴─────┬──────┴────────────┘                      │
│                          │                                          │
│              ┌───────────▼───────────┐                              │
│              │   IATP Policy Check   │                              │
│              │   (Signed Contracts)  │                              │
│              └───────────────────────┘                              │
└─────────────────────────────────────────────────────────────────────┘

Agent Types

1. Forecast Agent

  • Predicts solar output using weather data
  • Publishes forecasts to AMB topic: grid/forecast

2. Trader Agent

  • Listens for grid operator price signals
  • Bids battery discharge capacity
  • Uses IATP to sign binding contracts

3. Dispatch Agent (Mute Agent)

  • Only acts when IATP-signed contract received
  • Controls actual battery discharge
  • Returns NULL if contract invalid

Key Features

Agent Message Bus (AMB)

  • 1,000+ messages/second throughput
  • Priority lanes for emergency signals
  • Backpressure to prevent cascade failures

Inter-Agent Trust Protocol (IATP)

  • Agents verify each other's signatures
  • No action without signed contract
  • Tamper-proof audit trail

Policy Enforcement

  • Max discharge limits enforced at kernel level
  • IPC Pipes: trader | policy_check("max_discharge") | dispatch
  • Shadow Mode for testing without real dispatch

Quick Start

# Run the demo
docker-compose up

# Or run locally
pip install -e .
python demo.py

# Run with 100 agents
python demo.py --agents 100

# Run with price spike simulation
python demo.py --scenario price_spike

Demo Scenarios

Scenario 1: Price Spike

Grid operator broadcasts high price signal. Agents compete to sell stored energy.

Scenario 2: Solar Surplus

Too much solar generation. Agents coordinate to store excess.

Scenario 3: Emergency

Grid frequency drops. Agents respond in <100ms with emergency discharge.

Metrics

Metric Value
Agents 100
Negotiations/minute 1,000+
Average latency 15ms
Policy violations 0
Grid stabilization time <30 seconds

License

MIT