⚡ AlphaESS Charging Optimizer

Dynamic battery arbitrage against Czech day-ahead electricity prices

Quick Start • Lambda Deploy • How It Works • Configuration

---

Automatically charge your AlphaESS battery when electricity is cheap and discharge when expensive. Uses 15-minute price slots from OTE (Czech day-ahead market) to maximize savings through smart arbitrage cycles.

┌─────────────────────────────────────────────────────────────────┐
│  💰 CHEAP (Valley)          💸 EXPENSIVE (Peak)                 │
│  ════════════════          ═══════════════════                  │
│  03:00-06:00 @ 45€         17:00-20:00 @ 180€                   │
│  ↓ CHARGE ↓                ↓ DISCHARGE ↓                        │
│  Grid → Battery            Battery → Home                       │
│                                                                 │
│  Spread: 135 €/MWh  ✨                                          │
└─────────────────────────────────────────────────────────────────┘

✨ Features

Feature	Description
🔍 Dynamic Detection	Auto-detects valleys & peaks from daily price patterns
🔄 Arbitrage Cycles	Pairs charge windows with discharge windows for max spread
🔋 Battery-Aware	Sizes windows based on actual SOC and capacity
📊 15-min Granularity	Uses OTE's 96 daily price slots for precision
☁️ Serverless Ready	Deploy to AWS Lambda or run locally

🚀 Quick Start

Prerequisites

• Python 3.12+
• uv (fast Python package manager)
• AlphaESS Open API credentials

Installation

# Clone
git clone https://github.com/michaelkrasa/AlphaESS-charging-optimizer.git
cd AlphaESS-charging-optimizer

# Install dependencies
uv sync

Configuration

Create a .env file with your credentials:

cp .env.example .env
# Edit .env with your values

# Required - AlphaESS API
APP_ID=your_alpha_ess_app_id
APP_SECRET=your_alpha_ess_app_secret
SERIAL_NUMBER=your_system_serial

Tune behavior in config.yaml:

charge_to_full: 3        # Hours to charge 0→100%
price_multiplier: 1.2    # Valley/peak threshold vs daily mean
min_soc: 10              # Don't discharge below this %
max_soc: 100             # Charge target %

Run

# Default: Single optimization for today (run at midnight)
uv run optimizer.py

# Dry run for a specific day (no API changes)
uv run optimizer.py --date 15

---

☁️ AWS Lambda Deployment

Run as a serverless function - no server required, pay only for execution time.

Quick Deploy

# 1. Configure (edit .env with AWS settings)
cp .env.example .env

# 2. Deploy to AWS
./deploy-lambda.sh

The script will:

• ✅ Build Docker image (arm64 for Graviton)
• ✅ Push to Amazon ECR
• ✅ Update Lambda function
• ✅ Configure environment variables

Lambda Configuration

Setting	Value
Architecture	arm64 (Graviton)
Timeout	30 seconds
Memory	256 MB
Trigger	EventBridge @ 00:00 UTC daily

Lambda Execution

Lambda runs once at 00:00 daily and optimizes for the current day. No configuration needed - just schedule it via EventBridge.

Schedule with EventBridge

# Daily at 00:00 UTC (01:01 CET)
aws events put-rule \
  --name "ess-daily-optimization" \
  --schedule-expression "cron(1 0 * * ? *)"

---

🧠 How It Works

┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│  Fetch 96    │────▶│  Detect      │────▶│  Build       │
│  Price Slots │     │  Valleys &   │     │  Arbitrage   │
│  (15-min)    │     │  Peaks       │     │  Cycles      │
└──────────────┘     └──────────────┘     └──────────────┘
                                                 │
                                                 ▼
┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│  Done! ✅    │◀────│  Program     │◀────│  Size to     │
│              │     │  AlphaESS    │     │  Battery     │
│              │     │  API         │     │  SOC         │
└──────────────┘     └──────────────┘     └──────────────┘

Price Analysis

1. Smooth prices with moving average to reduce noise
2. Calculate daily mean price
3. Detect valleys: price < mean / price_multiplier
4. Detect peaks: price > mean × price_multiplier
5. Find mid-day dips between peaks for extra opportunities

Arbitrage Matching

• Each valley pairs with the next sequential peak
• Discharge windows extend to cover all profitable hours
• Up to 2 cycles per day (AlphaESS API limitation)

Battery Intelligence

• Reads actual SOC from device
• Pulls capacity (gross × usable %)
• Sizes charge windows to actual need
• Accounts for consumption between windows

---

⚙️ Configuration

Environment Variables (.env)

Variable	Required	Description
APP_ID	✅	AlphaESS API app ID
APP_SECRET	✅	AlphaESS API secret
SERIAL_NUMBER	✅	Your ESS serial number
AWS_ACCOUNT_ID	Lambda	AWS account for ECR
ECR_REPO	Lambda	ECR repository name

Optimization Settings (config.yaml)

Setting	Default	Description
charge_to_full	3	Hours to charge 0→100%
price_multiplier	1.2	Threshold factor vs daily mean
min_soc	10	Minimum discharge SOC %
max_soc	100	Target charge SOC %
avg_day_load_kw	1.8	Avg household load for SOC estimation
min_window_slots	2	Minimum window size (×15 min)
smoothing_window	2	Price smoothing window (×15 min)

---

📁 Project Structure

├── optimizer.py        # Main optimizer orchestration
├── models.py           # Data models (PriceWindow, ArbitrageCycle, etc.)
├── price_analyzer.py   # Price analysis and valley/peak detection
├── battery_manager.py  # Battery state calculations
├── ess_client.py       # AlphaESS API client
└── price_cache.py      # Price caching logic
├── config.py           # Configuration loader
├── config.yaml         # Optimization settings
├── lambda_handler.py   # AWS Lambda entry point
├── Dockerfile          # Lambda container (arm64)
├── deploy-lambda.sh    # One-command AWS deployment
├── .env.example        # Environment template
└── test_ess.py         # Test suite

---

🧪 Testing

uv run pytest test_ess.py -v

---

⏰ Automation

Cron (Linux/macOS)

# Run daily at 00:00 (midnight + 1 minute)
1 0 * * * cd /path/to/AlphaESS-charging-optimizer && uv run optimizer.py

AWS Lambda + EventBridge

See Lambda Deployment section above.

---

📝 Notes

• Target market: Czech OTE day-ahead prices (15-min granularity)
• API limitation: Max 2 charge + 2 discharge windows per day
• Schedule: Run at 00:00 daily to optimize for that day (prices are published the day before)

---

📄 License

MIT

---

Happy arbitrage! ⚡🔋