Description

Moving results between a MIDAS GEN NX model, Excel, and a separate calculation program is repetitive, hard to version, and easy to break by hand. This first part covers the MIDAS GEN NX API as a direct connection between the model and Python code: what the four HTTP request types do, how a script authenticates with the MAPI key issued by the running MIDAS GEN NX session, and what the setup needs — Python with the midas_gen library, an IDE, and a virtual environment holding NumPy, Polars, Matplotlib, scikit-learn, and SteelPy for AISC section properties. The complete Python script for the example is published as a repository and linked by QR code on the slides, so the session can be followed line by line and the code reused directly.

The example builds a 12 m span warehouse roof truss in an empty model without drawing anything in the interface. Geometry, A36 material, four section groups (double-angle chords, single-angle verticals and diagonals), elements and supports are all created from the script, and member end releases are written as a raw JSON POST request because that function is not yet exposed in the library. Loads follow the same route: dead, roof live and four wind cases generated in code, with roof Cp values interpolated from NSCP tables, then load combinations and envelopes.

The script does not stop at a finished model. It triggers the run, reads beam forces back, computes steel quantity through SteelPy, and evaluates DCR against tension yielding and rupture, buckling, and slenderness, resizing sections by A_i = A_(i-1) × DCR^0.6 until DCR approaches 1.0 and total steel weight changes by less than 2% between iterations.