Automated Power-Electronics Simulation

Context: BMW Group · Electric Mobility & Charging
Period: 2026 · Simulation automation
Focus: Engineering Software
Tools / methods: Python · SPICE · parametric analysis

Engineering question

Single-case simulations are difficult to compare across switching conditions, models, and circuit variants. Repeatability becomes especially important when extracting waveform metrics sensitive to numerical settings or transient boundaries.

My work

  • Worked on parameterized circuit simulation and engineering post-processing tools.
  • Used structured sweeps and waveform-derived metrics to compare transient behavior across operating conditions.
  • Investigated approaches to traceable model inputs, numerical reproducibility, and simulation-to-test workflows.

Technical approach

A reusable architecture separates configuration, simulation orchestration, signal extraction, metric calculation, and results validation. Examples of useful outputs include switching energy, peak overshoot, slew-rate estimates, and ringing descriptors—provided each uses documented numerical conventions.

Engineering considerations

Transient resampling, sign conventions, solver tolerances, integration boundaries, and baseline comparisons affect apparent trends. Optimization outputs are meaningful only when their physical constraints and uncertainty are explicitly tracked.

Evidence and maturity

Engineering study and automation work; no representation that an independently validated production tool has been released.

Public scope

No internal source code, optimization datasets, proprietary circuit models, or test campaigns are shared.

Shivam Vashishtha
Shivam Vashishtha
Engineer

Electrical Engineering and Computer Science student working across power, AI, energy systems, and emerging technologies.