Domestic Water Pipe Sizing: The Fixture Unit Method Explained

Why domestic water sizing isn’t the same problem as hydronic sizing

A chilled-water pipe serves a coil with a known, steady design flow — calculate it once from a cooling load, pick a velocity limit, size the pipe. Domestic water piping serves a population of fixtures that are never all in use at once. A 40-unit apartment building has far more total fixtures than any single riser will ever see running simultaneously. Size the main by adding up every fixture’s full flow rate and you get a wildly oversized, unnecessarily expensive pipe.

The fixture unit method converts a fixture count into a statistically realistic peak demand, based on decades of empirical data on how plumbing fixtures actually get used. It traces back to Dr. Roy Hunter’s research for the National Bureau of Standards in the 1940s, and it’s still the basis for the sizing tables in both the IPC (Appendix E) and the Uniform Plumbing Code (UPC), with fixture-unit approaches referenced in various forms by other national plumbing codes worldwide.

Two related but different IPC tables come up here, and they’re easy to conflate: Table 604.3 gives the minimum flow rate and pressure required at each individual fixture outlet — what that one fixture needs to work properly. Table E103.3(2) gives each fixture’s Water Supply Fixture Unit value — how much it contributes to the probable peak demand on the pipe serving it. One is a per-fixture performance requirement; the other is the system-sizing input you actually total up.

If you came here looking to size chilled-water or hydronic piping instead, that’s a different problem with a different method — see our Pipe Size Calculator, which sizes hydronic pipe directly from a known flow rate and design velocity.

What a Water Supply Fixture Unit actually is

A Water Supply Fixture Unit (WSFU) is a dimensionless load value assigned to a fixture type, sized so fixtures of different flow rates and usage patterns can be added together on a common scale. It is not a flow rate. A fixture with 2 WSFU doesn’t mean “2 gpm” — it means “this fixture contributes this much probable demand,” with the odds of that fixture being in use at any given moment already built into the number.

2012 IPC Table E103.3(2). For fixtures not listed, assign loads by comparison to a listed fixture with similar flow. Hot and cold branch loads are each three-quarters of the fixture’s total load, rounded to the nearest tenth — the Total column is the fixture’s true combined value, not the sum of the Cold and Hot columns.

The sizing method, step by step

This follows IPC Appendix E’s Segmented Loss Method (Section E103.3), which is fundamentally a pressure-budget exercise: everything the system loses in pressure — elevation, meter, backflow preventer, taps, fittings, and pipe friction — has to add up to no more than what’s actually available at the source, with fixture demand determining how much friction loss each pipe segment can afford.

1. Build a complete fixture schedule. List every fixture on the branch, riser, or main — nothing skipped. Hose bibbs and irrigation points are the most commonly forgotten, and on a commercial site they add up.
2. Assign the correct WSFU value to each fixture from Table E103.3(2), using the private or public row as appropriate, and the correct flush-tank or flushometer-valve row for water closets and urinals.
3. Total the WSFU separately for cold, hot, and combined — a hot water riser is sized from the hot-branch total, not the combined total. Continuous-flow items like hose bibbs and lawn sprinklers aren’t run through the fixture-unit conversion at all; their gpm demand is added directly to the converted fixture demand afterward.
4. Convert total WSFU to peak demand (gpm) using Table E103.3(3), selecting the flush-tank or flushometer-valve side depending on which type dominates. Don’t shortcut this with a straight-line ratio — the relationship is deliberately non-linear: a small increase produces a comparatively large jump in demand at low WSFU counts, while the curve flattens at high counts as the odds of everything running at once keep dropping.
5. Build the pressure budget. Starting from the minimum pressure available at the main, subtract the highest pressure required at the fixture (Table 604.3), static elevation gain (0.433 psi per foot of rise), meter loss, tap loss (Table E103.3(4)), and any special device losses. What’s left is available for pipe and fitting friction.
6. Select a pipe size whose friction loss — at the segment’s peak gpm, over its actual length plus equivalent fitting length — fits inside that budget. If it doesn’t fit, move up a size and recheck.

Velocity isn’t a separate pass/fail check written into this method as a standalone numeric limit — but it isn’t ignored either. It’s implicit in the friction figures (a pipe run at unreasonably high velocity shows correspondingly high friction loss, which the pressure budget will reject on its own), and most engineers still track it separately against industry velocity guidance as a noise/erosion/water-hammer sanity check — see further down.

Try it: WSFU and peak demand calculator

Worked example

Consider a small commercial restroom branch serving:

• 2 × public lavatories, faucet (2.0 WSFU each = 4.0)
• 2 × public water closets, flushometer valve (10.0 WSFU each = 20.0)
• 1 × public urinal, ¾ in. flushometer valve (5.0 WSFU)

Total: 29.0 WSFU, all on the cold branch (flushometer fixtures carry no separate hot-water load in Table E103.3(2)).

This is a flushometer-dominated system, so the total is read against the flushometer side of Table E103.3(3). The table doesn’t list every whole-number load, so you interpolate between the bracketing rows — at 25 WSFU the table gives 38.0 gpm, and at 30 WSFU it gives 42.0 gpm, so 29 WSFU lands at approximately 41 gpm.

Compare that to simply adding each fixture’s own rated flow directly — three flushometer valves alone can each individually flow well over 25 gpm, which would put a naive sum north of 70–80 gpm. Sizing on that basis would push the branch one or two pipe sizes larger than the demand table actually supports, for a peak condition that essentially never occurs. That gap is the entire point of the fixture-unit method — and it’s also why the pressure budget and friction check are what actually pick the final pipe size, not the 41 gpm figure alone. The same 41 gpm could land on different final pipe sizes on two projects with different available pressure, elevation change, or developed length.

Common mistakes

A note on pipe material and velocity

The 2012 IPC’s own sizing method doesn’t state a single blanket velocity limit as a numeric pass/fail rule the way it does for WSFU values — velocity is folded into the friction-loss figures used in the final sizing step. That said, most engineers still track velocity separately as a practical sanity check against noise, erosion, and water hammer, using widely-cited industry figures:

• Copper, steel, and stainless steel: commonly held to around 8 fps on cold water, dropping to roughly 5 fps (or lower, for hot water above ~140°F) — the tighter hot-water limit exists because hot water accelerates erosion-corrosion in metallic pipe.
• PEX and CPVC: commonly held to around 5 fps for both hot and cold — partly because plastic pipe of the same nominal size typically has a smaller actual internal diameter than copper, so the same flow moves faster through it.

Treat these as industry-standard practice, not a quoted IPC numeric limit — confirm against your pipe manufacturer’s data and any velocity guidance your specific code edition or local amendment does state. Codes outside the US (BS EN 806 and similar) use different reference velocities entirely, so don’t assume these figures transfer to a project running under a different code.

Standards referenced

• International Plumbing Code (IPC), 2012 edition — Section 604 (Table 604.3 fixture flow/pressure, Table 604.5 minimum supply pipe sizes) and Appendix E, Section E103.3 Segmented Loss Method (Table E103.3(2) fixture unit values, Table E103.3(3) demand conversion, Table E103.3(4) tap/tee loss)
• Uniform Plumbing Code (UPC) — equivalent fixture-unit and demand-conversion provisions, with different table numbers and some different values
• Local/regional plumbing codes and amendments — always govern over generic references when they conflict, and later IPC editions may carry revised values

Every fixture-unit value, demand figure, and table reference above was read directly from a 2012 IPC copy rather than a secondary source. If your project is permitted under a different edition or a different code entirely, treat this as a worked illustration of the method rather than a source of design values — pull the actual numbers from your adopted code before using them on real work.

Frequently asked questions

Is a Water Supply Fixture Unit the same as a Drainage Fixture Unit (DFU)?

No. WSFU values describe supply-side demand; DFU values describe drainage-side load. They use different tables and different fixture values — don’t mix them.

Can I just add up every fixture’s rated flow instead of using WSFU?

You can, but it will oversize the pipe. Simple summation assumes every fixture runs at once, which the fixture-unit method exists specifically to avoid.

Why does a flushometer toilet have such a higher WSFU than a tank toilet?

A flushometer valve draws its full flow directly from the supply main for the duration of the flush, producing a much higher instantaneous demand than a tank, which refills relatively slowly after each use.

Does low-flow fixture technology change these numbers?

Modern low-flow fixtures use less water than the fixtures the original tables were built around, and some jurisdictions and alternative methods account for this more directly. Whether an alternative method is permitted in place of the standard WSFU approach is a jurisdiction- and code-edition-specific question — check with your local authority having jurisdiction.

Does IPC actually require a specific maximum velocity?

Not as a standalone numeric rule in the 2012 edition’s Segmented Loss Method — velocity is implicit in the friction-loss figures used to size the pipe. Check whether your specific code edition or local amendment states one explicitly.

Do I need to size hot and cold branches separately?

Yes. Each branch is sized from its own WSFU total (three-quarters of the fixture’s total value per branch), because the hot and cold sides can each peak independently.