Focused child case · power electronics · verified 28 July 2026

Make the DC-link assumption visible before selecting hardware.

Two inspectable relationships estimate voltage ripple for an ideal triangular capacitor-current waveform and minimum capacitance for a defined constant-power energy pulse.

Engineering question

What capacitance follows from the stated waveform or energy window?

These are first-pass checks for exposing current, frequency, energy, bus-voltage, and allowable-drop assumptions before switching, lifetime, thermal, tolerance, and component studies.

DC-link system map with source, capacitor, converter, load, ideal ripple waveform, voltage window, and sizing equation

Explanatory architecture

Source, capacitor, converter, and load define the boundary; waveform and voltage-window assumptions determine which relationship applies.

Open full-size figure

Energy-pulse relationship

Cmin = 2E / (V₁² − V₂²)

A transparent check—not a component selection.

The companion ripple relationship assumes a zero-mean triangular capacitor-current waveform: ΔVpp = Ipp / (8 f C).

Parameter Unit Physical Domain Constraint & Invariant
V₁ (Initial bus voltage) V V₁ > 0 Must exceed minimum voltage: V₁ > V₂
V₂ (Minimum post-pulse voltage) V 0 < V₂ < V₁ V₂ = V₁ − ΔV; ΔV < V₁ required
E (Pulse energy) J E > 0 E = P · Δt for constant power P and duration Δt
Cmin (Minimum capacitance) F Cmin > 0 Invalid windows (ΔV ≥ V₁ or non-finite inputs) rejected
Worked numerical example (agrees with reference figure and CSV):

For a 20 kW pulse of duration 20 ms on an 800 V bus with an allowed voltage drop of 80 V (V₂ = 720 V):

  • Pulse energy: E = 20,000 W × 0.020 s = 400 J
  • Voltage window difference: V₁² − V₂² = 800² − 720² = 640,000 − 518,400 = 121,600 V²
  • Minimum capacitance: Cmin = (2 × 400 J) / 121,600 V² = 800 / 121,600 F ≈ 6.58 mF (0.00658 F)

Verified result

Inputs, units, and sensitivities stay on the figure.

Validation and evidence

Three automated checks cover the released DC-link relationships.

Triangular ripple

The unit test verifies the implemented Ipp / (8 f C) relationship.

Energy pulse

A valid constant-power pulse produces positive capacitance from the exact capacitor-energy change.

Invalid voltage window

The implementation rejects an allowed drop that equals or exceeds the initial bus voltage.

Parity boundary

MATLAB equivalents are present. The automated six-test repository count comes from the Python suite across thermal and DC-link families.

Reproduction instructions

Run the model separately from the retained website figure.

1. Run the model (battery-power-models repository)

Clone the pinned model repository to execute the analytical functions and unit tests:

git clone https://github.com/mohammadrezwankhan/battery-power-models.git
cd battery-power-models
git checkout 7d9f61f1ebdca73f9facce5129f79805b1747ed0
python src/dc_link.py
python -m unittest discover -s tests -v

Prerequisites: Python 3.10+ (standard library only). Optional MATLAB parity script matlab/dc_link_ripple.m is provided for environments with a licensed MATLAB runtime.

2. Rebuilding the website figure is unavailable to public readers

The retained SVG, PNG, and CSV are downloadable. Rebuilding this site figure depends on a generator in the private portfolio repository; it has no established public license and is unavailable anonymously.

The public battery-power-models Python source remains runnable independently. Retaining these generated outputs does not mean the private website generator is publicly reproducible.

3. Inspect retained results

Inspect or download the retained dataset directly: Download dc-link-sizing.csv.

Engineering boundary

Replace idealized inputs before making a component decision.

Parent and related context

Use the relationships to expose assumptions, not to skip detailed design evidence.Return to parent overviewOpen pinned repository

Evidence summary

Reviewed 2026-11-12; evidence records retain their own scope and dates.

Portfolio source record
6 tests · source-reported · 7d9f61f · 7d9f61f1ebdca73f9facce5129f79805b1747ed0
Current release
Unknown from the reviewed evidence.
Current verification
Successful hosted run · 2026-11-10 · target 31cad57f6224a6460e07a68c74df7c219549be61
Repository head at 2026-10-17 inventory
682d43c769e82c0fe0c8d6ed2a2934c9fad6be35 · inventory snapshot only