This is a Streamlit web application that simulates and optimizes the operation of a battery energy storage system (BESS) based on day-ahead electricity market prices. Access the app here.
You manage a battery connected to the electricity grid. Given historical day-ahead market prices (e.g., from SMARD.de as shown in the screenshot below), the app helps determine the optimal charge/discharge schedule over the year to maximize profit — while considering grid fees, round-trip efficiency, and battery degradation costs.
- 📁 Upload your own SMARD CSV with hourly electricity prices
- ⚙️ Customize battery parameters (capacity, power, efficiency, costs)
- 🧠 Run a rolling optimization using
cvxpy(free and open-source convex optimization library) - 📈 Visualize battery charge/discharge behavior and SoC (State of Charge)
- 💾 Export results as a downloadable CSV file.
- 💡 Built with Streamlit for simplicity and interactivity
Customize how your battery behaves using intuitive sliders and input fields:
- Set the battery’s energy capacity, max power, and round-trip efficiency
- Adjust grid fee and degradation cost
- Load your own price data and hit Optimize to simulate realistic operations
Once the optimization runs, the app displays:
- A time series chart of charge, discharge, and state of charge (SoC)
- Summary of total profit
- Option to download results as CSV
git clone https://github.com/NoCh-Git/battery-optimizer-app.git
cd battery-optimizer-apppython -m venv .venv
source .venv/bin/activate # On Windows: .venv\Scripts\activatepip install -r requirements.txtstreamlit run app.pyOnce running, if it didn't happen automatically, open http://localhost:8501 in your browser.
Upload a CSV file downloaded from SMARD.de with the following options:
- File type: CSV
- Category: Market Data → Day-ahead Prices
- Region: Germany/Luxembourg [or any other region you're interested in]
- The CSV should have a
;delimiter (as SMARD exports by default).
battery_optimization_app/
├── app.py ← Main Streamlit app
├── optimization.py ← CVXPY-based optimization model
├── utils.py ← CSV loading and plotting utilities
├── requirements.txt ← Python dependencies
├── images/ ← Images used in README.md
└── data/ ← Optional inputs to be kept locally (ignored by git).
This app simulates realistic battery behavior by aligning with the actual publication schedule of electricity market prices in Germany.
- 📅 Day-ahead prices are published by EPEX SPOT SE (via SMARD.de) every day around 12:45–13:00 CET.
- 🕛 These prices cover the next calendar day from 00:00 to 23:00.
- 📤 Bids are submitted by 12:00 CET, and market results become available about 45–60 minutes later.
-
Between 00:00 and ~13:00, the system can only "see" prices up to 23:00 on the same day — meaning the actual forecast horizon shrinks as the day progresses (from 24h at midnight down to ~10h before the new prices are published).
-
After ~13:00, the system gains visibility into the next day → enabling a longer optimization window of ~11h + 24h = ~35h.
The rolling optimizer in this app adapts accordingly, switching between §13h and 35h lookahead depending on the current hour of operation — just like a real energy storage operator would.
To maintain realism, the last 12 hours of the dataset are excluded from optimization. This ensures that our app's operational decisions are based on more realistic and informed price visibility.
This is a hands-on project for a battery optimization app. Feel free to contribute, improve, and expand upon it!
- How the inclusion of photovoltaic (PV) generation would change the approach and results.
- How the non-shiftable load would change the approach and results.
Narges Chinichian
Feel free to reach out for questions, ideas, or collaboration!
MIT License