F1 2026 Energy Deployment Optimization — Monza

Research-grade energy deployment optimization model for the 2026 Formula 1 power unit regulations, focused on minimizing lap time at Monza.

Objective

Determine the optimal electrical energy deployment strategy that minimizes lap time while respecting:

• Battery SOC limits
• Power deployment/harvest rate limits
• 2026 regulatory constraints

Quick Start

# Install dependencies
pip install -r requirements.txt

# Run single lap optimization
python -m simulations.single_lap

# Run with validation
python -m simulations.single_lap --validate --initial-soc 0.8

# Run sensitivity analysis
python -m analysis.sensitivity

# Generate visualization plots
python -m analysis.visualizations

Repository Structure

f1-2026-energy-optimization/
├── data/
│   ├── config.py           # Central configuration (all assumptions)
│   └── monza_track.py      # Monza track model with discretization
├── models/
│   ├── vehicle_model.py    # Longitudinal vehicle dynamics
│   ├── energy_model.py     # Power unit energy model (ICE + MGU-K)
│   └── lap_time_model.py   # Segment-level lap time computation
├── optimization/
│   └── dynamic_programming.py  # DP solver (backward induction)
├── simulations/
│   └── single_lap.py       # Main simulation driver
├── analysis/
│   ├── sensitivity.py      # Parameter sensitivity sweeps
│   └── visualizations.py   # Publication-style matplotlib plots
├── results/                # Generated outputs
├── requirements.txt
└── README.md

Regulatory Assumptions (2026 Simplified)

Parameter	Value	Notes
MGU-H	Removed	Per 2026 regulations
MGU-K Peak Power	350 kW	~50% of total power
ICE Peak Power	400 kW	Simplified constant output
Battery Capacity	4.0 MJ	Parameterized
Max Deploy Rate	350 kW	Same as MGU-K peak
Max Harvest Rate	350 kW	During braking only
SOC Bounds	[0%, 100%]	Strictly enforced
Deploy Efficiency	95%	Battery → Wheels
Harvest Efficiency	90%	Braking → Battery

All assumptions are defined in data/config.py and can be modified.

Mathematical Formulation

State Space

• Segment index: i ∈ {0, 1, ..., N-1} (spatial)
• SOC: Discretized into M states {0, Δsoc, 2Δsoc, ..., 1}

Action Space

• Deploy fraction: a ∈ {0.0, 0.1, ..., 1.0} (fraction of max MGU-K power)

State Transition

SOC_{i+1} = SOC_i + H_i - E_i

where:
  E_i = deploy_power × Δt / battery_capacity  (deployment)
  H_i = harvest_power × Δt × η_harvest / cap  (harvesting, braking only)

Objective

minimize Σ Δt_i

where:
  Δt_i = Δs_i / v_i(P_total)
  P_total = P_ICE + a × P_MGU-K × η_deploy

Algorithm

Bellman Equation (Backward Induction):

V*(i, soc) = min_a { cost(i, soc, a) + V*(i+1, soc') }

Track Model — Monza

• Total length: 5,793 m
• Discretization: 25 m segments (~232 segments)
• Segment classification:
  • Straight: High-speed, full deployment favorable
  • Braking: Deceleration zone, harvest favorable
  • Corner: Grip-limited, minimal deployment
  • Acceleration: Corner exit, partial deployment

Key zones (approximate distances):

• T1 Chicane: 750-920 m
• Curva Grande: 1,250-1,450 m
• Lesmo 1-2: 2,580-3,150 m
• Ascari: 4,050-4,350 m
• Parabolica: 5,200-5,650 m

Vehicle Model

Longitudinal dynamics:

m·a = F_drive - F_drag - F_roll - F_grade

where:
  F_drag = 0.5 × ρ × Cd × A × v²
  F_roll = Crr × m × g
  F_grade = m × g × sin(θ)

Key parameters:

• Mass: 798 kg
• Cd: 1.0
• Frontal Area: 1.5 m²

Outputs

Console Output

=== Dynamic Programming Solver ===
Segments: 232
SOC states: 51
Actions: 11
...
Optimal Lap Time: 1:XX.XXX
Improvement: X.XXX s (X.XX%)

Saved Files (in results/)

• optimal_deployment.npy — Optimal deploy fraction per segment
• soc_trajectory.npy — SOC evolution over lap
• speed_trajectory.npy — Speed at each segment
• optimization_summary.txt — Text summary

Plots

• deployment_profile.png — Deployment vs distance
• soc_trajectory.png — SOC vs distance
• speed_comparison.png — Optimized vs baseline speed
• sensitivity_analysis.png — Parameter sweep results

Extension Ideas

1. Multi-lap optimization — Account for initial/final SOC constraints
2. Race strategy — Optimize across full race distance
3. MPC formulation — Receding horizon control
4. Uncertainty — Stochastic DP for traffic/weather
5. Multi-track — Generalize to other circuits

Non-Goals (Explicitly Out of Scope)

• Web UI / Streamlit dashboards
• Real-time telemetry integration
• DRS / active aero modeling
• Tire degradation
• Fuel consumption

Dependencies

• Python 3.8+
• NumPy ≥ 1.24.0
• SciPy ≥ 1.10.0
• Matplotlib ≥ 3.7.0

License

Internal R&D project. Not for public distribution.

---

This project is designed to read like internal motorsport performance engineering tooling, prioritizing accuracy and clarity over polish.