Splitter silencer pressure drop calculator
Enter the duct size, splitter count/thickness/length and airflow for a rectangular splitter (baffle) silencer, and get the gap velocity, Reynolds number and total pressure loss through it.
What this calculates, and why not VDI 2081
VDI 2081 publishes an empirical pressure-loss correlation specifically for splitter silencers, with coefficients by splitter thickness. We looked into implementing it directly, but without the source page to verify the exact term structure against, a reconstruction from a garbled formula came out roughly 10–50× too high against realistic silencer pressure drops — a sign the term placement was still wrong, not just a rounding issue. Rather than risk shipping a real engineering tool with a hidden error of that size, this version uses a general model built entirely from formulas already verified elsewhere on this site.
The model
Air entering the silencer contracts from the full duct area into the gaps between splitters, runs through the gaps as a straight rectangular duct, then expands back to the full duct area on the way out. Each stage uses a formula already checked independently — Darcy-Weisbach + Colebrook-White against Moody-chart reference points, and the Borda-Carnot contraction/enlargement coefficients against basic momentum conservation:
Contraction (duct → gaps): Kc = 0.5 × (1 − Agap/Aduct)
Friction (through the gaps): Δpf = f × (L/Dh) × ρVgap²/2
Expansion (gaps → duct): Ke = (1 − Agap/Aduct)²
Δptotal = KcρVgap²/2 + Δpf + KeρVgap²/2
The gap width is calculated automatically from duct width, splitter count and splitter thickness — for N splitters evenly spaced with a gap on each side, s = (W − N×dk) / (N+1) — or you can override it directly.
This is a general approximation, not the VDI 2081 correlation. It doesn't capture VDI 2081's empirical corrections for splitter nose/tail shape or the specific test-derived behavior at different splitter thicknesses. Treat results as directional. Insertion loss (dB attenuation by octave band) and self-generated noise aren't included in this version — both need a real, citable acoustic model (the kind of source this site uses for its other acoustic calculator), which isn't available yet.