Writing

Working notes on control, inverters, and the gap between a paper's results section and a simulation that actually runs.

• 23 Jul 2026 What a grid-forming inverter is actually doing

  The distinction between grid-following and grid-forming is not about hardware. It's about which quantity you treat as given.

• 20 Jul 2026 The affine lift that fixed my Koopman model

  A single constant observable took multi-step prediction error from 16.8% to 2.25%. Here's why that isn't a trick.