SAP2000 Automation

Python tooling to drive SAP2000 on Windows through the COM API.

What it does:

• open an existing .sdb model
• run analysis
• run steel design (default code: AISC360-16)
• export analysis log / design summary ratios to Excel
• optimize member sections with a genetic algorithm to reduce structural weight

Requires Windows + a licensed SAP2000 installation. The COM bridge will not work on macOS/Linux.

Setup

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
pip install -e .

Optional:

$env:SAP2000_EXE = "C:\Program Files\Computers and Structures\SAP2000 24\SAP2000.exe"

Put your .sdb files under models/ (gitignored).

Usage

Analysis / design

sap2000-auto analyze models\modelFixed.sdb -o outputs
sap2000-auto design models\modelFixed.sdb -o outputs --code AISC360-16

Genetic weight optimization

1. Copy configs/optimize.example.json
2. Set section_pool to section property names that already exist in the model
3. Set member_groups (frames that must share one section).
   If you leave member_groups empty, groups are auto-built from the current section assignments.
4. Run:

sap2000-auto optimize models\modelFixed.sdb -c configs\optimize.example.json -o outputs --save-model models\modelFixed_opt.sdb

Fitness is:

• minimize total steel weight
• if max design ratio > ratio_limit, apply a large penalty

Outputs:

• outputs/ga_history.csv
• outputs/ga_best_assignment.json

Python API

from sap2000_automation.config import AppConfig
from sap2000_automation.optimize import load_optimize_config, run_weight_optimization

config = AppConfig.from_env(model_path=r"models\modelFixed.sdb", output_dir="outputs")
problem = load_optimize_config("configs/optimize.example.json")
result = run_weight_optimization(config, problem)
print(result.best_sections, result.total_weight, result.max_ratio)

Layout

src/sap2000_automation/
  client.py
  analysis.py
  design.py
  genetic.py
  optimize.py
  cli.py
configs/optimize.example.json
models/
outputs/
notebooks/
tests/

Notes

• Section names in section_pool must already exist as frame properties in SAP2000.
• Keep model/analysis binaries out of git.
• GA unit tests run without SAP2000; full optimize needs a local SAP install.

License

MIT — see LICENSE.