Optimizing Gigascale BESS Arbitrage
With the connection of the 7.8GWh Saudi Arabia BESS and the shifting de-rating factors in the Polish capacity market, energy storage operators face a complex optimization problem. This challenge requires building a production-grade pipeline to forecast energy price volatility and optimize battery dispatch strategies while accounting for physical de-rating and cycle-induced degradation. You will leverage Metaflow for workflow orchestration and Gemini 2.5 Flash to synthesize market sentiment from real-time news sources via Browserless, feeding these qualitative insights into a quantitative optimization framework.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a hybrid forecasting and optimization pipeline using Metaflow, Gemini 2.5 Flash, and Pyomo to maximize revenue for a utility-scale battery storage system under dynamic regulatory constraints.
How work is evaluated
The project will be evaluated based on the net profit achieved in a simulated market environment, the accuracy of the price forecasts, and the robustness of the optimization constraints.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Formulate and solve a mixed-integer linear programming (MILP) model for battery dispatch optimization.
Develop a LLM-driven sentiment analysis module to adjust price forecasts based on regulatory news.
Implement a robust data pipeline for multi-market energy price ingestion and feature engineering.
Evaluate the impact of de-rating factors on long-term project IRR and state-of-health (SoH) preservation.
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
The project will be evaluated based on the net profit achieved in a simulated market environment, the accuracy of the price forecasts, and the robustness of the optimization constraints.
Challenge input
CSV file with hourly price and demand data.
Agent execution
No agent tool is listed for this challenge.
Evaluated output
CSV file with hourly charge/discharge commands and resulting SoC.
- Battery SoC must never exceed 100% or drop below 0%.
- Maximum discharge power must respect the input de-rating factor.
- Profitability Index target: 0.85
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Battery SoC must never exceed 100% or drop below 0%.
- Maximum discharge power must respect the input de-rating factor.
Metrics
- Profitability Index · Target 0.85
- Forecasting RMSE · Target 5
Workflow tasks
- Task 1
Input: CSV file with hourly price and demand data.
Output: CSV file with hourly charge/discharge commands and resulting SoC.
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.