Use cases of Heaven Platforms

It’s great playing around with case studies on Heaven 🇦🇺

We ran a simulation focused on solving one of the biggest headaches in Australia’s National Electricity Market (NEM): wind curtailment in Victoria (VIC1).

The Objective

Set up a utility-scale wind farm co-located with a grid-scale battery (BESS) and test how AI dispatch can bypass physical transmission bottlenecks to capture spilled revenue.

How It Played Out

• Text-to-SCADA Asset Synthesis: Dropped in a 300 MW onshore wind farm paired with a 150 MW / 300 MWh battery connected to a 220kV line using plain English prompts.

• Grid Physics & Transmission Constraints: Modeled local high-voltage line limits with PyPSA (aligned with the NEM Integrated System Plan) alongside live weather telemetry from Open-Meteo.

• AI Optimization: Heaven’s MARL (Multi-Agent Reinforcement Learning) optimizer charged the BESS during forced curtailment hours and dumped power back into the grid when spot prices peaked.

• The Bottom Line: Saved the spilled energy to deliver a +42.9% revenue uplift 📈

Under the Hood (Data & Logic)

• Market Data: 5-minute spot pricing and FCAS data pulled straight from AEMO.

• Weather & Physics: Open-Meteo API wind vector curves paired with PyPSA network topology.

• Smart Modeling Assumptions: Uses a 24-hour perfect foresight lookahead to calculate maximum theoretical revenue limits, factors in market cannibalization co-optimization for large assets (>50MW), and accounts for static battery cycle degradation.

Super cool being able to deploy your own generators and operate straight against regional spot prices in real-time.

Check out the platform here: https://heavenplatforms.space/ 🚀

Would love to hear your thoughts or feedback from any energy, climate-tech, or quant folks in the community!