FSAE EV Endurance Simulation

Endurance simulation and optimization for UConn Formula SAE Electric. Predicts lap times, energy consumption, and competition points to find optimal vehicle configuration and driver strategy.

Cars

	CT-16EV (2025)	CT-17EV (2026)
Pack	110S4P Molicel P45B	100S4P Molicel P50B
Mass (with driver)	288 kg	~279 kg
Motor/Inverter	Shared	Shared
Controls	Shared	Shared

Architecture

Vehicle Config (YAML)  →  Simulation Engine  →  Results  →  Dashboard
                              ↑
               Track (from GPS telemetry)
               Driver Strategy (swappable)

Simulation method: Quasi-static point-mass, calibrated against real AiM telemetry from the 2025 Michigan endurance event. For each track segment, resolves speed from force balance and driver strategy, steps battery state, enforces BMS limits.

Modules:

Module	Purpose
fsae_sim.vehicle	Vehicle, powertrain, and battery configuration
fsae_sim.track	Track representation from GPS segments
fsae_sim.driver	Driver strategy / control policy (swappable)
fsae_sim.sim	Simulation engine
fsae_sim.scoring	FSAE endurance + efficiency point formulas
fsae_sim.optimization	Parameter sweep runner
fsae_sim.analysis	Post-processing metrics, Pareto computation
fsae_sim.data	Telemetry and simulation data loaders
dashboard	Dash web app for viewing results

Quick Start

With Docker

docker compose -f docker/docker-compose.yaml up
# Browser → http://localhost:3000

Without Docker

pip install -e ".[dev]"
python -m dashboard
# Browser → http://localhost:3000

Run tests

pytest -v

Project Structure

├── src/fsae_sim/          # Simulation Python package
│   ├── vehicle/           # Vehicle, powertrain, battery models
│   ├── track/             # Track representation
│   ├── driver/            # Driver strategy
│   ├── sim/               # Simulation engine
│   ├── scoring/           # FSAE scoring formulas
│   ├── optimization/      # Parameter sweeps
│   ├── analysis/          # Metrics and post-processing
│   └── data/              # Data loaders
├── dashboard/             # Dash web app (port 3000)
│   └── pages/             # Dashboard pages
├── configs/               # Vehicle config YAML files
├── Real-Car-Data-And-Stats/  # Telemetry and battery data
├── results/               # Simulation outputs (gitignored)
├── tests/                 # pytest test suite
└── docker/                # Dockerfile and compose

Data

• AiM telemetry: 2025 Endurance Data.csv — 20Hz, ~37k samples, ~100 channels from Michigan endurance
• Voltt battery sim: Cell and pack level CSVs — voltage, SOC, current, temperature, heat generation for 110S4P P45B
• BMS tune: Endurance Tune2.txt — discharge limits, SOC taper, inverter/motor parameters

Roadmap

Phase 1 — Foundation ✅ (Done)

Repository scaffold, Docker, dashboard skeleton, vehicle configs, data loaders

Phase 2 — Core Simulation (Nearly Done)

Battery model, powertrain model, vehicle dynamics with 4-wheel Pacejka tire model, driver model (CalibratedStrategy, zone-based), simulation engine validated against real telemetry (~2% energy error, 8/8 metrics pass). Remaining: driver model finalization and quality/accuracy validation checks.

Phase 3 — Optimization & Comparison (Next)

Swappable strategies, parameter sweeps, car comparison, Pareto frontier, dashboard buildout

Phase 4 — Scoring & Decision Support (Future)

FSAE scoring model, field estimation, points maximization, final decision dashboard