Skip to content
2nd Place

Pravāh

Electric grid operators are overwhelmed with the data needed to manage complex grid operations, including weather, state-space estimation, network congestion, bidirectional electricity flow, DER coordination, and asset f

Code Video Cerebral Valley

Core trick

Five agents each optimize a separate objective inside a power-flow solver, with contingency checks that fall back to conservative actions.

Tags

agentsscience

Description

Electric grid operators are overwhelmed with the data needed to manage complex grid operations, including weather, state-space estimation, network congestion, bidirectional electricity flow, DER coordination, and asset failures. We developed an agentic grid orchestration framework that enables operators to utilise GPT-5 with multimodal data inputs within a power-flow solver, incorporating fail-safe contingencies. We built five agents - each with their own objective function, such as asset health monitoring, reliability, safety, power purchase cost reduction, etc., to operate an optimal system under physical constraints. We struggled to fine-tune the agents, so actions defaulted to conservative choices, and we were unable to connect the refined front end to our backend fully.
pravah-hq/pravah-open-ai-hackathon★ 00 forksPythonMITCreated Aug 9, 2025 · last push Aug 10, 2025

README excerpt

# OpenDSS-Sim: Production-Ready Grid Simulator

A **config-driven** electric distribution grid simulator with FastAPI streaming, safety validation, event engine, and plug-and-play tools for OpenAI Agents SDK integration. Built for the Pravah OpenAI Hackathon.

## 🚀 Features

- **Config-Driven**: Single `config.yaml` controls everything - no code changes needed
- **IEEE-123 Ready**: Pre-configured with realistic 123-bus distribution feeder
- **Real OpenDSS Integration**: Uses OpenDSSDirect.py with mock fallback for testing
- **FastAPI Streaming**: WebSocket data broadcast + REST API for agent control
- **Beautiful UI**: Self-hosted test console at `http://localhost:8787/ui/` to drive endpoints
- **Safety Validation**: Voltage/current limit checking with violation reporting
- **Event Engine**: Time-based events (heatwaves, equipment faults, load scaling)
- **Weather/Market Data**: Meteostat live generation or pre-loaded Parquet files
- **MPC Scheduling**: Battery energy storage optimal control with price signals
- **Agent Tools**: JSON schemas and client utilities for OpenAI Agents SDK
- **Production Ready**: Comprehensive tests, logging, error handling, cross-platform

## 🎯 Architecture

```
┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Agent Tools   │    │   FastAPI API    │    │  Simulation     │
│                 │    │                  │    │     Core        │
│ • Client        │◄──►│ • /simulate      │◄──►│ • OpenDSS       │
│ • Tool Specs    │

Team